index
int64 0
0
| repo_id
stringlengths 16
145
| file_path
stringlengths 27
196
| content
stringlengths 1
16.7M
|
|---|---|---|---|
0 | capitalone_repos/federated-model-aggregation/connectors/django/fma_django_api/v1 | capitalone_repos/federated-model-aggregation/connectors/django/fma_django_api/v1/tests/test_model_aggregate_viewset.py | from unittest import mock
from django.urls import reverse
from rest_framework.authtoken import models as auth_models
from rest_framework.response import Response
from rest_framework.test import APITestCase
from fma_django import models
from fma_django_api import utils
from .utils import ClientMixin, LoginMixin
class TestModelAggregateViewSet(ClientMixin, LoginMixin, APITestCase):
reverse_url = "model_aggregate-list"
fixtures = [
"TaskQueue_User.json",
"TaskQueue_client.json",
"DjangoQ_Schedule.json",
"TaskQueue_FederatedModel.json",
"TaskQueue_ModelAggregate.json",
]
def setUp(self):
for user in models.User.objects.all():
auth_models.Token.objects.get_or_create(user=user)
def _validate_unauthorized(self, baseurl, client_action, request_json=None):
if request_json is None:
request_json = {}
# validate unauthorized, no one logged in
response = client_action(baseurl, format="json")
self.assertEqual(401, response.status_code)
self.assertDictEqual(
{"detail": "Authentication credentials were not provided."}, response.json()
)
# validate non-existent client can't access
response = client_action(
baseurl, format="json", HTTP_CLIENT_UUID="not-a-client"
)
self.assertEqual(401, response.status_code)
self.assertDictEqual(
{"detail": '"not-a-client" is not a valid UUID.'}, response.json()
)
# validate non-existent client can't access
response = client_action(
baseurl,
format="json",
HTTP_CLIENT_UUID="aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa",
)
self.assertEqual(401, response.status_code)
self.assertDictEqual({"detail": "Invalid UUID"}, response.json())
# validate client user can only view
if client_action != self.client.get:
response = client_action(
baseurl,
format="json",
HTTP_CLIENT_UUID="cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
)
self.assertEqual(403, response.status_code)
self.assertDictEqual(
{"detail": "You do not have permission to perform this action."},
response.json(),
)
def test_list(self):
baseurl = reverse(self.reverse_url)
self._validate_unauthorized(baseurl, client_action=self.client.get)
# validate authorized response given fixtures
self.login_user(user_json={"username": "admin"})
response = self.client.get(baseurl, format="json")
expected_response = [
{
"id": 2,
"federated_model": 2,
"result": "http://testserver/mediafiles/fake/path/model_aggregate/2",
"validation_score": None,
"created_on": "2023-01-20T23:08:28.712000Z",
"parent": None,
"level": 0,
"lft": 1,
"rght": 2,
"tree_id": 1,
},
{
"id": 1,
"federated_model": 1,
"result": "http://testserver/mediafiles/fake/path/model_aggregate/1",
"validation_score": 1.0,
"created_on": "2023-01-13T23:08:28.712000Z",
"parent": None,
"level": 0,
"lft": 1,
"rght": 2,
"tree_id": 0,
},
]
self.assertEqual(200, response.status_code)
self.assertEqual(expected_response, response.json())
self.logout_user()
# validate developer only sees their own
self.login_user(user_json={"username": "non_admin"})
response = self.client.get(baseurl, format="json")
expected_response = [
{
"id": 2,
"federated_model": 2,
"result": "http://testserver/mediafiles/fake/path/model_aggregate/2",
"validation_score": None,
"created_on": "2023-01-20T23:08:28.712000Z",
"parent": None,
"level": 0,
"lft": 1,
"rght": 2,
"tree_id": 1,
}
]
self.assertEqual(200, response.status_code)
self.assertEqual(expected_response, response.json())
self.logout_user()
# validate client view, only model it has access to
self.login_client(client_json={"uuid": "cbb6025f-c15c-4e90-b3fb-85626f7a79f1"})
response = self.client.get(baseurl, format="json")
expected_response = [
{
"id": 1,
"federated_model": 1,
"result": "http://testserver/mediafiles/fake/path/model_aggregate/1",
"validation_score": 1.0,
"created_on": "2023-01-13T23:08:28.712000Z",
"parent": None,
"level": 0,
"lft": 1,
"rght": 2,
"tree_id": 0,
}
]
self.assertEqual(200, response.status_code)
self.assertEqual(expected_response, response.json())
self.logout_client()
def test_create(self):
baseurl = reverse(self.reverse_url)
# validate unauthorized
self._validate_unauthorized(baseurl, client_action=self.client.post)
def test_get(self):
baseurl = reverse(self.reverse_url)
# validate unauthorized
self._validate_unauthorized(baseurl + "1/", client_action=self.client.get)
# validate admin access to both
self.login_user(user_json={"username": "admin"})
with mock.patch(
"django.core.files.storage.FileSystemStorage._open"
) as mock_load:
# Turn aggregation results into file
mock_load.return_value = utils.create_model_file([5, 2, 3])
response = self.client.get(baseurl + "1/", format="json")
expected_response = {
"id": 1,
"federated_model": 1,
"result": [5, 2, 3],
"validation_score": 1.0,
"created_on": "2023-01-13T23:08:28.712000Z",
"parent": None,
"level": 0,
"lft": 1,
"rght": 2,
"tree_id": 0,
}
self.assertEqual(200, response.status_code)
self.assertDictEqual(expected_response, response.json())
with mock.patch(
"django.core.files.storage.FileSystemStorage._open"
) as mock_load:
# Turn aggregation results into file
mock_load.return_value = utils.create_model_file([1, 2, 4])
response = self.client.get(baseurl + "2/", format="json")
expected_response = {
"id": 2,
"federated_model": 2,
"result": [1, 2, 4],
"validation_score": None,
"created_on": "2023-01-20T23:08:28.712000Z",
"parent": None,
"level": 0,
"lft": 1,
"rght": 2,
"tree_id": 1,
}
self.assertEqual(200, response.status_code)
self.assertDictEqual(expected_response, response.json())
self.logout_user()
# validate developer no access to models that they don't own
self.login_user(user_json={"username": "non_admin"})
response = self.client.get(baseurl + "1/", format="json")
self.assertEqual(404, response.status_code)
self.assertEqual({"detail": "Not found."}, response.json())
# validate developer access to their own model
with mock.patch(
"django.core.files.storage.FileSystemStorage._open"
) as mock_load:
# Turn aggregation results into file
mock_load.return_value = utils.create_model_file([1, 2, 4])
response = self.client.get(baseurl + "2/", format="json")
expected_response = {
"id": 2,
"federated_model": 2,
"result": [1, 2, 4],
"validation_score": None,
"created_on": "2023-01-20T23:08:28.712000Z",
"parent": None,
"level": 0,
"lft": 1,
"rght": 2,
"tree_id": 1,
}
self.assertEqual(200, response.status_code)
self.assertDictEqual(expected_response, response.json())
self.logout_user()
# validate client has no access to unregistered models
self.login_client(client_json={"uuid": "cbb6025f-c15c-4e90-b3fb-85626f7a79f1"})
response = self.client.get(baseurl + "2/", format="json")
self.assertEqual(404, response.status_code)
self.assertEqual({"detail": "Not found."}, response.json())
# validate client has access to their registered models
with mock.patch(
"django.core.files.storage.FileSystemStorage._open"
) as mock_load:
# Turn aggregation results into file
mock_load.return_value = utils.create_model_file([5, 2, 3])
response = self.client.get(baseurl + "1/", format="json")
expected_response = {
"id": 1,
"federated_model": 1,
"result": [5, 2, 3],
"validation_score": 1.0,
"created_on": "2023-01-13T23:08:28.712000Z",
"parent": None,
"level": 0,
"lft": 1,
"rght": 2,
"tree_id": 0,
}
self.assertEqual(200, response.status_code)
self.assertDictEqual(expected_response, response.json())
def test_publish_aggregate(self):
baseurl = reverse(self.reverse_url)
admin_url = baseurl + "1/publish_aggregate/"
dev_url = baseurl + "2/publish_aggregate/"
fed_url = reverse("model-list")
version_data = {"version": "2.0.0"}
# validate anonymous user cannot publish a model
self._validate_unauthorized(dev_url, client_action=self.client.post)
# validate client cannot publish an aggregate
self.login_client(client_json={"uuid": "531580e6-ce6c-4f01-a5aa-9ed7af5ee768"})
response = self.client.post(admin_url, data=version_data, format="json")
self.assertEqual(403, response.status_code)
expected_response = {
"detail": "You do not have permission to perform this action."
}
self.assertDictEqual(expected_response, response.json())
# validate developer can publish aggregate owned by them
self.login_user(user_json={"username": "non_admin"})
with mock.patch(
"django.core.files.storage.FileSystemStorage._open"
) as mock_load:
# Turn aggregation results into file
mock_load.return_value = utils.create_model_file([5, 2, 3])
response = self.client.post(dev_url, data=version_data, format="json")
self.assertEqual(201, response.status_code)
expected_response = {
"id": 1,
"values": "/mediafiles/fake/path/model_aggregate/2",
"version": "2.0.0",
"created_on": response.json()["created_on"],
"federated_model": 2,
}
self.assertDictEqual(expected_response, response.json())
# Retrieve fed model for comparison of current artifact
url = fed_url + "2/get_current_artifact/"
fed_response = self.client.get(url, format="json")
self.assertEqual(200, fed_response.status_code)
model_art_from_fed = models.ModelArtifact.objects.filter(
id=expected_response["id"],
version=expected_response["version"],
)
self.assertEqual(1, model_art_from_fed.count())
# Check if federated model is using new published aggregate
self.assertDictEqual(expected_response, fed_response.json())
fed_model = models.FederatedModel.objects.get(id=2)
self.assertEqual(fed_model.current_artifact, model_art_from_fed.first())
# validate developer cannot publish aggregate not owned by them
self.login_user(user_json={"username": "non_admin"})
response = self.client.post(admin_url, data=version_data, format="json")
self.assertEqual(404, response.status_code)
expected_response = {"detail": "Not found."}
self.assertDictEqual(expected_response, response.json())
# validate developer cannot publish a non-existent aggregate
self.login_user(user_json={"username": "non_admin"})
response = self.client.post(
baseurl + "3/publish_aggregate/", data=version_data, format="json"
)
self.assertEqual(404, response.status_code)
expected_response = {"detail": "Not found."}
self.assertDictEqual(expected_response, response.json())
self.logout_user()
def test_pull_aggregate(self):
baseurl = reverse(self.reverse_url)
dev_url = baseurl + "2/"
admin_url = baseurl + "1/"
# validate anonymous user cannot publish a model
self._validate_unauthorized(dev_url, client_action=self.client.get)
# validate client cannot pull an aggregate
self.login_client(client_json={"uuid": "531580e6-ce6c-4f01-a5aa-9ed7af5ee768"})
response = self.client.get(dev_url, format="json")
self.assertEqual(404, response.status_code)
expected_response = {"detail": "Not found."}
self.assertDictEqual(expected_response, response.json())
# validate developer can pull aggregate owned by them
self.login_user(user_json={"username": "non_admin"})
with mock.patch(
"django.core.files.storage.FileSystemStorage._open"
) as mock_load:
mock_load.return_value = utils.create_model_file([5, 2, 3])
response = self.client.get(dev_url, format="json")
self.assertEqual(200, response.status_code)
expected_response = {
"id": 2,
"result": [5, 2, 3],
"validation_score": None,
"created_on": "2023-01-20T23:08:28.712000Z",
"lft": 1,
"rght": 2,
"tree_id": 1,
"level": 0,
"federated_model": 2,
"parent": None,
}
self.assertDictEqual(expected_response, response.json())
# validate developer cannot pull aggregate not owned by them
response = self.client.get(admin_url, format="json")
self.assertEqual(404, response.status_code)
expected_response = {"detail": "Not found."}
self.assertDictEqual(expected_response, response.json())
self.logout_user()
def test_update_val_score(self):
baseurl = reverse(self.reverse_url)
dev_url = baseurl + "2/"
admin_url = baseurl + "1/"
# validate anonymous user cannot publish a model
self._validate_unauthorized(dev_url, client_action=self.client.patch)
# validate client cannot pull an aggregate
self.login_client(client_json={"uuid": "531580e6-ce6c-4f01-a5aa-9ed7af5ee768"})
data = {"validation_score": 0.5}
response = self.client.patch(dev_url, data=data, format="json")
self.assertEqual(403, response.status_code)
expected_response = {
"detail": "You do not have permission to perform this action."
}
self.assertDictEqual(expected_response, response.json())
self.logout_user()
# validate developer can pull aggregate owned by them
self.login_user(user_json={"username": "non_admin"})
data = {"validation_score": 0.5}
response = self.client.patch(dev_url, data=data, format="json")
self.assertEqual(200, response.status_code)
expected_response = {
"id": 2,
"result": "http://testserver/mediafiles/fake/path/model_aggregate/2",
"validation_score": 0.5,
"created_on": "2023-01-20T23:08:28.712000Z",
"lft": 1,
"rght": 2,
"tree_id": 1,
"level": 0,
"federated_model": 2,
"parent": None,
}
self.assertDictEqual(expected_response, response.json())
# validate developer cannot pull aggregate not owned by them
self.login_user(user_json={"username": "non_admin"})
data = {"validation_score": 0.5}
response = self.client.patch(admin_url, data=data, format="json")
self.assertEqual(404, response.status_code)
expected_response = {"detail": "Not found."}
self.assertDictEqual(expected_response, response.json())
self.logout_user()
@mock.patch("rest_framework.mixins.ListModelMixin.list")
def test_token_login(self, mock_list):
# validate get using auth token
baseurl = reverse(self.reverse_url)
user_json = {"username": "non_admin"}
mock_list.return_value = Response({"detail": "successful API Call"})
self.user_token_login(user_json=user_json)
_ = self.client.get(baseurl)
params = mock_list.call_args
mock_list.assert_called_once()
self.assertEqual(self.user, params[0][0].user)
self.user_token_logout()
def test_filters(self, *mocks):
baseurl = reverse(self.reverse_url)
# validate federated_model filtering
queryurl = baseurl + "?federated_model=1"
self.login_user(user_json={"username": "admin"})
response = self.client.get(queryurl, format="json")
expected_response = [
{
"id": 1,
"federated_model": 1,
"result": "http://testserver/mediafiles/fake/path/model_aggregate/1",
"validation_score": 1.0,
"created_on": "2023-01-13T23:08:28.712000Z",
"parent": None,
"level": 0,
"lft": 1,
"rght": 2,
"tree_id": 0,
}
]
self.assertEqual(200, response.status_code)
self.assertEqual(expected_response, response.json())
# validate federated_model=2
queryurl = baseurl + "?federated_model=2"
self.login_user(user_json={"username": "admin"})
response = self.client.get(queryurl, format="json")
expected_response = [
{
"id": 2,
"federated_model": 2,
"result": "http://testserver/mediafiles/fake/path/model_aggregate/2",
"validation_score": None,
"created_on": "2023-01-20T23:08:28.712000Z",
"parent": None,
"level": 0,
"lft": 1,
"rght": 2,
"tree_id": 1,
},
]
self.assertEqual(200, response.status_code)
self.assertEqual(expected_response, response.json())
def test_ordering(self, *mocks):
baseurl = reverse(self.reverse_url)
# validate validation_score
queryurl = baseurl + "?ordering=validation_score"
self.login_user(user_json={"username": "admin"})
response = self.client.get(queryurl, format="json")
expected_response = [
{
"id": 2,
"federated_model": 2,
"result": "http://testserver/mediafiles/fake/path/model_aggregate/2",
"validation_score": None,
"created_on": "2023-01-20T23:08:28.712000Z",
"parent": None,
"level": 0,
"lft": 1,
"rght": 2,
"tree_id": 1,
},
{
"id": 1,
"federated_model": 1,
"result": "http://testserver/mediafiles/fake/path/model_aggregate/1",
"validation_score": 1.0,
"created_on": "2023-01-13T23:08:28.712000Z",
"parent": None,
"level": 0,
"lft": 1,
"rght": 2,
"tree_id": 0,
},
]
self.assertEqual(200, response.status_code)
self.assertEqual(expected_response, response.json())
def test_pagination(self, *mocks):
baseurl = reverse(self.reverse_url)
# setting the page size to 2 and querying for page 1
queryurl = baseurl + "?page_size=2&page=1"
self.login_user(user_json={"username": "admin"})
response = self.client.get(queryurl, format="json")
self.assertEqual(200, response.status_code)
self.assertEqual(2, len(response.json()))
# setting the page size to 1 and querying for page 1
queryurl = baseurl + "?page_size=1&page=1"
self.login_user(user_json={"username": "admin"})
response = self.client.get(queryurl, format="json")
self.assertEqual(200, response.status_code)
self.assertEqual(1, len(response.json()))
|
0 | capitalone_repos/federated-model-aggregation/connectors/django/fma_django_api/v1 | capitalone_repos/federated-model-aggregation/connectors/django/fma_django_api/v1/tests/test_client_aggregate_score_viewset.py | from unittest import mock
from django.urls import reverse
from rest_framework.authtoken import models as auth_models
from rest_framework.response import Response
from rest_framework.test import APITestCase
from fma_django import models
from .utils import ClientMixin, LoginMixin
@mock.patch(
"django.core.files.storage.FileSystemStorage.save", return_value="save_create"
)
class TestClientAggregateScoreViewSet(ClientMixin, LoginMixin, APITestCase):
reverse_url = "client_aggregate_scores-list"
fixtures = [
"TaskQueue_User.json",
"TaskQueue_client.json",
"DjangoQ_Schedule.json",
"TaskQueue_FederatedModel.json",
"TaskQueue_ModelAggregate.json",
"TaskQueue_ModelUpdate.json",
"TaskQueue_ClientAggregateScore.json",
]
def setUp(self):
for user in models.User.objects.all():
auth_models.Token.objects.get_or_create(user=user)
def _validate_unauthorized(self, baseurl, client_action, request_json=None):
if request_json is None:
request_json = {}
# validate unauthorized, no one logged in
response = client_action(baseurl, format="json")
self.assertEqual(401, response.status_code)
self.assertDictEqual(
{"detail": "Authentication credentials were not provided."}, response.json()
)
# validate invalid uuid client can't access
response = client_action(
baseurl, format="json", HTTP_CLIENT_UUID="not-a-client"
)
self.assertEqual(401, response.status_code)
self.assertDictEqual(
{"detail": '"not-a-client" is not a valid UUID.'}, response.json()
)
# validate non-existent client can't access
response = client_action(
baseurl,
format="json",
HTTP_CLIENT_UUID="aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa",
)
self.assertEqual(401, response.status_code)
self.assertDictEqual({"detail": "Invalid UUID"}, response.json())
def test_list(self, *mocks):
baseurl = reverse(self.reverse_url)
self._validate_unauthorized(baseurl, client_action=self.client.get)
# validate authorized response given fixtures
self.login_user(user_json={"username": "admin"})
response = self.client.get(baseurl, format="json")
expected_response = [
{
"id": 1,
"client": "cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
"validation_results": {"f1_score": 0.8},
"created_on": "2022-02-15T23:08:28.720000Z",
"aggregate": 1,
},
{
"id": 2,
"client": "cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
"validation_results": {"prec": 0.25, "rec": 0.4, "skew": 0.6},
"created_on": "2022-02-15T23:08:28.720000Z",
"aggregate": 2,
},
]
self.assertEqual(200, response.status_code)
self.assertCountEqual(expected_response, response.json())
self.logout_user()
# validate developer only sees their own
self.login_user(user_json={"username": "non_admin"})
response = self.client.get(baseurl, format="json")
expected_response = [
{
"id": 2,
"client": "cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
"validation_results": {"prec": 0.25, "rec": 0.4, "skew": 0.6},
"created_on": "2022-02-15T23:08:28.720000Z",
"aggregate": 2,
}
]
self.assertEqual(200, response.status_code)
self.assertCountEqual(expected_response, response.json())
self.logout_client()
def test_create(self, *mocks):
baseurl = reverse(self.reverse_url)
# validate unauthorized
self._validate_unauthorized(baseurl, client_action=self.client.post)
# validate authorized client
self.login_client(client_json={"uuid": "ab359e5d-6991-4088-8815-a85d3e413c02"})
create_data = {
"client": "ab359e5d-6991-4088-8815-a85d3e413c02",
"validation_results": {
"f1_score": 0.8,
"description": "This is a fake description",
},
"aggregate": 1,
}
response = self.client.post(baseurl, format="json", data=create_data)
cleaned_response = response.json()
# rather error show up on assert, hence `None`
cleaned_response.pop("created_on", None)
expected_response = create_data.copy()
expected_response["id"] = 3
expected_response["client"] = "ab359e5d-6991-4088-8815-a85d3e413c02"
expected_response["aggregate"] = 1
expected_response["validation_results"] = {
"f1_score": 0.8,
"description": "This is a fake description",
}
self.assertEqual(201, response.status_code)
self.assertDictEqual(expected_response, cleaned_response)
# validate authorized client (Non-unique failure)
self.login_client(client_json={"uuid": "cbb6025f-c15c-4e90-b3fb-85626f7a79f1"})
create_data = {
"client": "cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
"validation_results": {"fake": 0.8},
"aggregate": 1,
}
response = self.client.post(baseurl, format="json", data=create_data)
cleaned_response = response.json()
# Client already has a result registered for this aggregate so push fails
expected_response = {
"non_field_errors": ["The fields aggregate, client must make a unique set."]
}
self.assertEqual(400, response.status_code)
self.assertDictEqual(expected_response, cleaned_response)
self.login_client(client_json={"uuid": "531580e6-ce6c-4f01-a5aa-9ed7af5ee768"})
# test create w/ invalid schema
create_data = {
"client": "531580e6-ce6c-4f01-a5aa-9ed7af5ee768",
"validation_results": {
"f1_score": 1.2,
"description": "This is a fake description",
},
"aggregate": 1,
}
response = self.client.post(baseurl, format="json", data=create_data)
cleaned_response = response.json()
expected_response = {
"results": [
"results did not match the required schema: {}".format(
{
"properties": {
"f1_score": {"type": "number", "minimum": 0, "maximum": 1},
"description": {"type": "string"},
}
}
)
]
}
self.assertEqual(400, response.status_code)
self.assertDictEqual(expected_response, cleaned_response)
def test_get(self, *mocks):
baseurl = reverse(self.reverse_url)
# validate unauthorized
self._validate_unauthorized(baseurl + "1/", client_action=self.client.get)
# validate admin access to both
self.login_user(user_json={"username": "admin"})
response = self.client.get(baseurl + "1/", format="json")
expected_response = {
"id": 1,
"client": "cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
"validation_results": {"f1_score": 0.8},
"created_on": "2022-02-15T23:08:28.720000Z",
"aggregate": 1,
}
response_json = response.json()
self.assertEqual(200, response.status_code)
self.assertDictEqual(expected_response, response_json)
# validate developer no access to models that they don't own
self.login_user(user_json={"username": "non_admin"})
response = self.client.get(baseurl + "1/", format="json")
self.assertEqual(404, response.status_code)
self.assertEqual({"detail": "Not found."}, response.json())
# validate developer access to their own model
response = self.client.get(baseurl + "2/", format="json")
expected_response = {
"id": 2,
"client": "cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
"validation_results": {"prec": 0.25, "rec": 0.4, "skew": 0.6},
"created_on": "2022-02-15T23:08:28.720000Z",
"aggregate": 2,
}
self.assertEqual(200, response.status_code)
self.assertDictEqual(expected_response, response.json())
@mock.patch("rest_framework.mixins.ListModelMixin.list")
def test_token_login(self, mock_list, *mocks):
# validate get using auth token
baseurl = reverse(self.reverse_url)
user_json = {"username": "non_admin"}
mock_list.return_value = Response({"detail": "successful API Call"})
self.user_token_login(user_json=user_json)
_ = self.client.get(baseurl)
params = mock_list.call_args
mock_list.assert_called_once()
self.assertEqual(self.user, params[0][0].user)
self.user_token_logout()
def test_filters(self, *mocks):
baseurl = reverse(self.reverse_url)
# validate aggregate filtering
queryurl = baseurl + "?aggregate=1"
self.login_user(user_json={"username": "admin"})
response = self.client.get(queryurl, format="json")
expected_response = [
{
"id": 1,
"client": "cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
"validation_results": {"f1_score": 0.8},
"created_on": "2022-02-15T23:08:28.720000Z",
"aggregate": 1,
}
]
self.assertEqual(200, response.status_code)
self.assertEqual(expected_response, response.json())
# validate aggregate=2
queryurl = baseurl + "?aggregate=2"
self.login_user(user_json={"username": "admin"})
response = self.client.get(queryurl, format="json")
expected_response = [
{
"id": 2,
"client": "cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
"validation_results": {"prec": 0.25, "rec": 0.4, "skew": 0.6},
"created_on": "2022-02-15T23:08:28.720000Z",
"aggregate": 2,
}
]
self.assertEqual(200, response.status_code)
self.assertEqual(expected_response, response.json())
def test_pagination(self, *mocks):
baseurl = reverse(self.reverse_url)
# setting the page size to 2 and querying for page 1
queryurl = baseurl + "?page_size=2&page=1"
self.login_user(user_json={"username": "admin"})
response = self.client.get(queryurl, format="json")
self.assertEqual(200, response.status_code)
self.assertEqual(2, len(response.json()))
# setting the page size to 1 and querying for page 1
queryurl = baseurl + "?page_size=1&page=1"
self.login_user(user_json={"username": "admin"})
response = self.client.get(queryurl, format="json")
self.assertEqual(200, response.status_code)
self.assertEqual(1, len(response.json()))
|
0 | capitalone_repos/federated-model-aggregation/connectors/django/fma_django_api/v1 | capitalone_repos/federated-model-aggregation/connectors/django/fma_django_api/v1/tests/test_federated_model_viewset.py | from unittest import mock
from django.urls import reverse
from django.utils import timezone
from django_q.tasks import Schedule as djanog_q_Schedule
from rest_framework.authtoken import models as auth_models
from rest_framework.response import Response
from rest_framework.test import APITestCase
from fma_django import models
from fma_django_api import utils
from .utils import ClientMixin, LoginMixin
@mock.patch(
"django.core.files.storage.FileSystemStorage.save", return_value="save_create"
)
class TestFederatedModelViewSet(ClientMixin, LoginMixin, APITestCase):
reverse_url = "model-list"
fixtures = [
"TaskQueue_User.json",
"TaskQueue_client.json",
"DjangoQ_Schedule.json",
"TaskQueue_FederatedModel.json",
"TaskQueue_ModelAggregate.json",
"TaskQueue_ModelArtifact.json",
]
@classmethod
def setUpTestData(cls):
# create artifact relation for model 1
cls.fake_datetime = timezone.datetime(
2022, 1, 1, 0, 0, 0, tzinfo=timezone.get_current_timezone()
)
fm_model = models.FederatedModel.objects.get(id=1)
artifact_model = models.ModelArtifact.objects.get(id=3)
fm_model.current_artifact = artifact_model
with mock.patch(
"django.utils.timezone.now", mock.Mock(return_value=cls.fake_datetime)
):
fm_model.save()
for user in models.User.objects.all():
auth_models.Token.objects.get_or_create(user=user)
def _validate_unauthorized(self, baseurl, client_action, request_json=None):
if request_json is None:
request_json = {}
# validate unauthorized, no one logged in
response = client_action(baseurl, format="json")
self.assertEqual(401, response.status_code)
self.assertDictEqual(
{"detail": "Authentication credentials were not provided."}, response.json()
)
# validate client can't access
self.login_client(client_json={"uuid": "531580e6-ce6c-4f01-a5aa-9ed7af5ee768"})
response = client_action(baseurl, format="json")
self.assertEqual(403, response.status_code)
self.assertDictEqual(
{"detail": "You do not have permission to perform this action."},
response.json(),
)
self.logout_client()
def test_list(self, mock_save):
baseurl = reverse(self.reverse_url)
self._validate_unauthorized(baseurl, client_action=self.client.get)
# validate authorized response given fixtures
self.login_user(user_json={"username": "admin"})
response = self.client.get(baseurl, format="json")
expected_response = [
{
"id": 2,
"allow_aggregation": False,
"name": "test-non-admin",
"requirement": "all_clients",
"requirement_args": None,
"aggregator": "avg_values_if_data",
"furthest_base_agg": None,
"update_schema": None,
"client_agg_results_schema": None,
"created_on": "2022-12-24T23:08:28.693000Z",
"last_modified": "2022-12-28T23:08:28.693000Z",
"developer": 2,
"current_artifact": None,
"clients": [],
},
{
"id": 3,
"allow_aggregation": True,
"name": "test-no-aggregate",
"requirement": "all_clients",
"requirement_args": None,
"aggregator": "avg_values_if_data",
"furthest_base_agg": None,
"update_schema": {
"type": "array",
"prefixItems": [
{"type": "array", "minItems": 3, "maxItems": 3},
{"type": "array", "minItems": 3, "maxItems": 3},
],
"items": False,
},
"client_agg_results_schema": None,
"created_on": "2022-12-18T23:08:28.693000Z",
"last_modified": "2022-12-19T23:08:28.693000Z",
"developer": 2,
"current_artifact": None,
"clients": ["cbb6025f-c15c-4e90-b3fb-85626f7a79f1"],
},
{
"id": 1,
"allow_aggregation": True,
"requirement": "require_x_updates",
"name": "test",
"requirement_args": [3],
"aggregator": "avg_values_if_data",
"furthest_base_agg": None,
"update_schema": {
"type": "array",
"prefixItems": [
{"type": "array", "minItems": 3, "maxItems": 3},
{"type": "array", "minItems": 3, "maxItems": 3},
],
"items": False,
},
"client_agg_results_schema": {
"properties": {
"f1_score": {"type": "number", "minimum": 0, "maximum": 1},
"description": {"type": "string"},
}
},
"created_on": "2022-12-15T23:08:28.693000Z",
"last_modified": "2022-01-01T00:00:00Z",
"developer": 1,
"current_artifact": 3,
"clients": [
"531580e6-ce6c-4f01-a5aa-9ed7af5ee768",
"ab359e5d-6991-4088-8815-a85d3e413c02",
"cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
],
},
]
self.assertEqual(200, response.status_code)
self.assertEqual(expected_response, response.json())
def test_create(self, mock_save):
baseurl = reverse(self.reverse_url)
# validate unauthorized
self._validate_unauthorized(baseurl, client_action=self.client.post)
# validate authorized admin
self.login_user(user_json={"username": "admin"})
create_data = {
"name": "create-test",
"aggregator": "avg_values_if_data",
"developer": 1,
"initial_model": [1, 2, 3],
"update_schema": {"test": 1},
"clients": [
"531580e6-ce6c-4f01-a5aa-9ed7af5ee768",
"ab359e5d-6991-4088-8815-a85d3e413c02",
],
}
with mock.patch(
"django.utils.timezone.now", mock.Mock(return_value=self.fake_datetime)
):
response = self.client.post(baseurl, format="json", data=create_data)
cleaned_response = response.json()
expected_response = create_data.copy()
expected_response.pop("initial_model", None)
expected_response["id"] = 4
expected_response["furthest_base_agg"] = None
expected_response["allow_aggregation"] = False # default False
expected_response["requirement_args"] = None
expected_response["requirement"] = ""
expected_response["current_artifact"] = 4
expected_response["created_on"] = "2022-01-01T00:00:00Z"
expected_response["last_modified"] = "2022-01-01T00:00:00Z"
expected_response["client_agg_results_schema"] = None
self.assertEqual(201, response.status_code, response.json())
self.assertDictEqual(expected_response, cleaned_response)
# also assert the task is created in the db
scheduled_model = djanog_q_Schedule.objects.filter(
name="4 - create-test - Scheduled Aggregator"
).first()
self.assertIsNotNone(
scheduled_model, "Validating if schedule was created with the model."
)
self.logout_user()
# test developer
self.login_user(user_json={"username": "non_admin"})
create_data = {
"name": "non_admin create-test",
"allow_aggregation": True,
"aggregator": "avg_values_if_data",
"developer": 2,
"update_schema": None,
"clients": [
"ab359e5d-6991-4088-8815-a85d3e413c02",
"cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
],
}
with mock.patch(
"django.utils.timezone.now", mock.Mock(return_value=self.fake_datetime)
):
response = self.client.post(baseurl, format="json", data=create_data)
cleaned_response = response.json()
expected_response = create_data.copy()
expected_response["id"] = 5
expected_response["furthest_base_agg"] = None
expected_response["requirement_args"] = None
expected_response["requirement"] = ""
expected_response["current_artifact"] = None
expected_response["created_on"] = "2022-01-01T00:00:00Z"
expected_response["last_modified"] = "2022-01-01T00:00:00Z"
expected_response["client_agg_results_schema"] = None
self.assertEqual(201, response.status_code)
self.assertDictEqual(expected_response, cleaned_response)
# also assert the task is created in the db
scheduled_model = djanog_q_Schedule.objects.filter(
name="5 - non_admin create-test - Scheduled Aggregator"
).first()
self.assertIsNotNone(
scheduled_model, "Validating if schedule was created with the model."
)
# validate furthest_base_agg fails if not > 1
create_data = {
"name": "test-furthest-base-agg",
"aggregator": "avg_values_if_data",
"developer": 2,
"initial_model": [1, 2, 3],
"update_schema": {"test": 1},
"furthest_base_agg": -1,
}
response = self.client.post(baseurl, format="json", data=create_data)
self.assertEqual(400, response.status_code)
cleaned_response = response.json()
self.assertDictEqual(
{"furthest_base_agg": ["Ensure this value is greater than or equal to 1."]},
cleaned_response,
)
# validate furthest_base_agg success if >= 1
create_data = {
"name": "test-furthest-base-agg",
"aggregator": "avg_values_if_data",
"developer": 2,
"initial_model": [1, 2, 3],
"update_schema": {"test": 1},
"furthest_base_agg": 3,
}
response = self.client.post(baseurl, format="json", data=create_data)
self.assertEqual(201, response.status_code)
cleaned_response = response.json()
self.assertEqual(3, cleaned_response.get("furthest_base_agg", None))
self.logout_user()
# TODO: don't allow developer to set users, can force this in the
# queryset for a user
def test_get(self, mock_save):
baseurl = reverse(self.reverse_url)
# validate unauthorized
self._validate_unauthorized(baseurl + "1/", client_action=self.client.get)
# validate admin access to both
self.login_user(user_json={"username": "admin"})
response = self.client.get(baseurl + "1/", format="json")
expected_response = {
"id": 1,
"allow_aggregation": True,
"name": "test",
"requirement": "require_x_updates",
"requirement_args": [3],
"aggregator": "avg_values_if_data",
"furthest_base_agg": None,
"update_schema": {
"type": "array",
"prefixItems": [
{"type": "array", "minItems": 3, "maxItems": 3},
{"type": "array", "minItems": 3, "maxItems": 3},
],
"items": False,
},
"client_agg_results_schema": {
"properties": {
"f1_score": {"type": "number", "minimum": 0, "maximum": 1},
"description": {"type": "string"},
}
},
"created_on": "2022-12-15T23:08:28.693000Z",
"last_modified": "2022-01-01T00:00:00Z",
"developer": 1,
"current_artifact": 3,
"clients": [
"531580e6-ce6c-4f01-a5aa-9ed7af5ee768",
"ab359e5d-6991-4088-8815-a85d3e413c02",
"cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
],
}
self.assertEqual(200, response.status_code)
self.assertDictEqual(expected_response, response.json())
response = self.client.get(baseurl + "2/", format="json")
expected_response = {
"id": 2,
"allow_aggregation": False,
"name": "test-non-admin",
"requirement": "all_clients",
"requirement_args": None,
"aggregator": "avg_values_if_data",
"furthest_base_agg": None,
"update_schema": None,
"client_agg_results_schema": None,
"created_on": "2022-12-24T23:08:28.693000Z",
"last_modified": "2022-12-28T23:08:28.693000Z",
"developer": 2,
"current_artifact": None,
"clients": [],
}
self.assertEqual(200, response.status_code)
self.assertDictEqual(expected_response, response.json())
self.logout_user()
# validate developer no access to models that they don't own
self.login_user(user_json={"username": "non_admin"})
response = self.client.get(baseurl + "1/", format="json")
self.assertEqual(404, response.status_code)
self.assertEqual({"detail": "Not found."}, response.json())
# validate developer access to their own model
response = self.client.get(baseurl + "2/", format="json")
expected_response = {
"id": 2,
"allow_aggregation": False,
"name": "test-non-admin",
"requirement": "all_clients",
"requirement_args": None,
"aggregator": "avg_values_if_data",
"furthest_base_agg": None,
"update_schema": None,
"client_agg_results_schema": None,
"created_on": "2022-12-24T23:08:28.693000Z",
"last_modified": "2022-12-28T23:08:28.693000Z",
"developer": 2,
"current_artifact": None,
"clients": [],
}
self.assertEqual(200, response.status_code)
self.assertDictEqual(expected_response, response.json())
def test_register_client(self, mock_save):
baseurl = reverse(self.reverse_url)
url = baseurl + "1/register_client/"
# validate get doesn't work
response = self.client.get(url, format="json")
self.assertEqual(405, response.status_code)
self.assertDictEqual({"detail": 'Method "GET" not allowed.'}, response.json())
# validate bad uuid doesn't work
response = self.client.post(url, format="json", HTTP_CLIENT_UUID="test")
self.assertEqual(403, response.status_code)
self.assertDictEqual({"detail": '"test" is not a valid UUID.'}, response.json())
# validate current UUID
response = self.client.post(
url, format="json", HTTP_CLIENT_UUID="531580e6-ce6c-4f01-a5aa-9ed7af5ee768"
)
self.assertEqual(200, response.status_code)
self.assertDictEqual(
{"uuid": "531580e6-ce6c-4f01-a5aa-9ed7af5ee768"}, response.json()
)
# validate make new UUID
fake_uuid = "000000e6-ce6c-4f01-a5aa-9ed7af5ee768"
with mock.patch(
"fma_django.models.Client.uuid.field.get_default",
return_value=fake_uuid,
):
response = self.client.post(url, format="json")
self.assertEqual(200, response.status_code)
self.assertDictEqual({"uuid": fake_uuid}, response.json())
# validate new is reusable
url = baseurl + "2/register_client/"
response = self.client.post(url, format="json", HTTP_CLIENT_UUID=fake_uuid)
self.assertEqual(200, response.status_code)
self.assertDictEqual({"uuid": fake_uuid}, response.json())
# validate devleoper can make a client
url = baseurl + "2/register_client/"
self.login_user(user_json={"username": "non_admin"})
fake_uuid = "111111e6-ce6c-4f01-a5aa-9ed7af5ee768"
with mock.patch(
"fma_django.models.Client.uuid.field.get_default",
return_value=fake_uuid,
):
response = self.client.post(url, format="json")
self.assertEqual(200, response.status_code)
self.assertDictEqual({"uuid": fake_uuid}, response.json())
self.logout_user()
def test_get_latest_aggregate(self, mock_save):
baseurl = reverse(self.reverse_url)
url = baseurl + "1/get_latest_aggregate/"
# validate current UUID
self.login_client(client_json={"uuid": "531580e6-ce6c-4f01-a5aa-9ed7af5ee768"})
response = self.client.get(url, format="json")
self.assertEqual(200, response.status_code)
expected_response = {
"id": 1,
"federated_model": 1,
"result": "/mediafiles/fake/path/model_aggregate/1",
"validation_score": 1.0,
"created_on": "2023-01-13T23:08:28.712000Z",
"lft": 1,
"rght": 2,
"tree_id": 0,
"level": 0,
"parent": None,
}
self.assertDictEqual(expected_response, response.json())
# validate client cannot get from model with not registered
url = baseurl + "2/get_latest_aggregate/"
response = self.client.get(url, format="json")
self.assertEqual(403, response.status_code)
self.assertDictEqual(
{"detail": "You do not have permission to perform this action."},
response.json(),
)
self.logout_client()
# validate developer has access to own
self.login_user(user_json={"username": "non_admin"})
url = baseurl + "2/get_latest_aggregate/"
response = self.client.get(url, format="json")
self.assertEqual(200, response.status_code)
expected_response = {
"id": 2,
"federated_model": 2,
"result": "/mediafiles/fake/path/model_aggregate/2",
"validation_score": None,
"created_on": "2023-01-20T23:08:28.712000Z",
"lft": 1,
"rght": 2,
"tree_id": 1,
"level": 0,
"parent": None,
}
self.assertDictEqual(expected_response, response.json())
# validate developer cannot get from model of another
url = baseurl + "1/get_latest_aggregate/"
response = self.client.get(url, format="json")
self.assertEqual(404, response.status_code)
self.assertDictEqual({"detail": "Not found."}, response.json())
# validate when model doesn't have an aggregate
models.ModelAggregate.objects.filter(federated_model=2).delete()
url = baseurl + "2/get_latest_aggregate/"
response = self.client.get(url, format="json")
expected_response = {}
self.assertEqual(200, response.status_code)
self.assertDictEqual(expected_response, response.json())
self.logout_user()
def test_get_current_artifact(self, mock_save):
baseurl = reverse(self.reverse_url)
url = baseurl + "1/get_current_artifact/"
# validate current UUID
self.login_client(client_json={"uuid": "531580e6-ce6c-4f01-a5aa-9ed7af5ee768"})
response = self.client.get(url, format="json")
self.assertEqual(200, response.status_code)
expected_response = {
"id": 3,
"federated_model": 1,
"values": "/mediafiles/fake/path/model_artifact/1",
"version": "1.0.0",
"created_on": "2022-12-12T23:55:34.066000Z",
}
self.assertDictEqual(expected_response, response.json())
# validate client cannot get from model with not registered
url = baseurl + "2/get_current_artifact/"
response = self.client.get(url, format="json")
self.assertEqual(403, response.status_code)
self.assertDictEqual(
{"detail": "You do not have permission to perform this action."},
response.json(),
)
self.logout_client()
# validate developer has access to own, also check returns null if empty
self.login_user(user_json={"username": "non_admin"})
url = baseurl + "2/get_current_artifact/"
response = self.client.get(url, format="json")
self.assertEqual(200, response.status_code)
expected_response = {}
self.assertDictEqual(expected_response, response.json())
# validate developer cannot get from model of another
url = baseurl + "1/get_current_artifact/"
response = self.client.get(url, format="json")
self.assertEqual(404, response.status_code)
self.assertDictEqual({"detail": "Not found."}, response.json())
self.logout_user()
def test_get_latest_model(self, mock_save):
baseurl = reverse(self.reverse_url)
# validate client cannot get from model with not registered
url = baseurl + "2/get_latest_model/"
response = self.client.get(url, format="json")
self.assertEqual(401, response.status_code)
self.assertDictEqual(
{"detail": "Authentication credentials were not provided."}, response.json()
)
# validate current UUID
url = baseurl + "1/get_latest_model/"
self.login_client(client_json={"uuid": "531580e6-ce6c-4f01-a5aa-9ed7af5ee768"})
with mock.patch(
"django.core.files.storage.FileSystemStorage._open"
) as mock_load:
# Turn aggregation results into file
mock_load.return_value = utils.create_model_file([0.5, 1, 2])
response = self.client.get(url, format="json")
self.assertEqual(200, response.status_code)
expected_response = {"values": [0.5, 1, 2], "aggregate": 1}
self.assertDictEqual(expected_response, response.json())
self.logout_user()
# validate developer cannot get from model of another
self.login_user(user_json={"username": "non_admin"})
url = baseurl + "1/get_latest_model/"
response = self.client.get(url, format="json")
self.assertEqual(404, response.status_code)
self.logout_user()
# validate when model doesn't have an aggregate
self.login_user(user_json={"username": "admin"})
models.ModelAggregate.objects.filter(federated_model=1).delete()
url = baseurl + "1/get_latest_model/"
with mock.patch(
"django.core.files.storage.FileSystemStorage._open"
) as mock_load:
# Turn aggregation results into file
mock_load.return_value = utils.create_model_file([0.5, 1, 2])
response = self.client.get(url, format="json")
self.assertEqual(200, response.status_code)
expected_response = {"values": [0.5, 1, 2], "aggregate": None}
self.assertDictEqual(expected_response, response.json())
self.logout_user()
# validate developer has access to own, also check returns null if
# both aggregate and artifact are empty
models.ModelAggregate.objects.filter(federated_model=2).delete()
models.ModelArtifact.objects.filter(federated_model=2).delete()
self.login_user(user_json={"username": "non_admin"})
url = baseurl + "2/get_latest_model/"
response = self.client.get(url, format="json")
self.assertEqual(200, response.status_code)
expected_response = {}
self.assertDictEqual(expected_response, response.json())
self.logout_user()
@mock.patch("rest_framework.mixins.ListModelMixin.list")
def test_token_login(self, mock_list, mock_save):
# validate get using auth token
baseurl = reverse(self.reverse_url)
user_json = {"username": "non_admin"}
mock_list.return_value = Response({"detail": "successful API Call"})
self.user_token_login(user_json=user_json)
_ = self.client.get(baseurl)
params = mock_list.call_args
mock_list.assert_called_once()
self.assertEqual(self.user, params[0][0].user)
self.user_token_logout()
def test_filters(self, *mocks):
baseurl = reverse(self.reverse_url)
# validate allow_aggregation=True
queryurl = baseurl + "?allow_aggregation=True"
self.login_user(user_json={"username": "admin"})
response = self.client.get(queryurl, format="json")
expected_response = [
{
"id": 3,
"allow_aggregation": True,
"name": "test-no-aggregate",
"requirement": "all_clients",
"requirement_args": None,
"aggregator": "avg_values_if_data",
"furthest_base_agg": None,
"update_schema": {
"type": "array",
"prefixItems": [
{"type": "array", "minItems": 3, "maxItems": 3},
{"type": "array", "minItems": 3, "maxItems": 3},
],
"items": False,
},
"client_agg_results_schema": None,
"created_on": "2022-12-18T23:08:28.693000Z",
"last_modified": "2022-12-19T23:08:28.693000Z",
"developer": 2,
"current_artifact": None,
"clients": ["cbb6025f-c15c-4e90-b3fb-85626f7a79f1"],
},
{
"id": 1,
"allow_aggregation": True,
"requirement": "require_x_updates",
"name": "test",
"requirement_args": [3],
"aggregator": "avg_values_if_data",
"furthest_base_agg": None,
"update_schema": {
"type": "array",
"prefixItems": [
{"type": "array", "minItems": 3, "maxItems": 3},
{"type": "array", "minItems": 3, "maxItems": 3},
],
"items": False,
},
"client_agg_results_schema": {
"properties": {
"f1_score": {"type": "number", "minimum": 0, "maximum": 1},
"description": {"type": "string"},
}
},
"created_on": "2022-12-15T23:08:28.693000Z",
"last_modified": "2022-01-01T00:00:00Z",
"developer": 1,
"current_artifact": 3,
"clients": [
"531580e6-ce6c-4f01-a5aa-9ed7af5ee768",
"ab359e5d-6991-4088-8815-a85d3e413c02",
"cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
],
},
]
self.assertEqual(200, response.status_code)
self.assertEqual(expected_response, response.json())
# validate allow_aggregation=False
queryurl = baseurl + "?allow_aggregation=false"
self.login_user(user_json={"username": "admin"})
response = self.client.get(queryurl, format="json")
expected_response = [
{
"id": 2,
"allow_aggregation": False,
"name": "test-non-admin",
"requirement": "all_clients",
"requirement_args": None,
"aggregator": "avg_values_if_data",
"furthest_base_agg": None,
"update_schema": None,
"client_agg_results_schema": None,
"created_on": "2022-12-24T23:08:28.693000Z",
"last_modified": "2022-12-28T23:08:28.693000Z",
"developer": 2,
"current_artifact": None,
"clients": [],
}
]
self.assertEqual(200, response.status_code)
self.assertEqual(expected_response, response.json())
def test_pagination(self, *mocks):
baseurl = reverse(self.reverse_url)
# setting the page size to 2 and querying for page 1
queryurl = baseurl + "?page_size=2&page=1"
self.login_user(user_json={"username": "admin"})
response = self.client.get(queryurl, format="json")
self.assertEqual(200, response.status_code)
self.assertEqual(2, len(response.json()))
# setting the page size to 1 and querying for page 1
queryurl = baseurl + "?page_size=1&page=1"
self.login_user(user_json={"username": "admin"})
response = self.client.get(queryurl, format="json")
self.assertEqual(200, response.status_code)
self.assertEqual(1, len(response.json()))
|
0 | capitalone_repos/federated-model-aggregation/connectors/django/fma_django_api/v1 | capitalone_repos/federated-model-aggregation/connectors/django/fma_django_api/v1/tests/utils.py | from fma_django import models
class LoginMixin:
def login_user(self, user_json):
"""Logs in user for authentication."""
self.user = models.User.objects.get_or_create(**user_json)[0]
self.client.force_login(self.user)
def logout_user(self):
"""Logs out user from authentication."""
self.client.logout()
self.user = None
def user_token_login(self, user_json):
"""Logs out user from authentication using tokens"""
self.user = models.User.objects.get_or_create(**user_json)[0]
self.client.credentials(HTTP_AUTHORIZATION="Token " + self.user.auth_token.key)
def user_token_logout(self):
"""Logs out user from authentication using token."""
self.client.credentials()
self.user = None
class ClientMixin:
def get_client_header(self, uuid):
return {"CLIENT-UUID": uuid}
def login_client(self, client_json):
"""Logs in user for authentication."""
self.user_client = models.Client.objects.get_or_create(**client_json)[0]
self.client.credentials(HTTP_CLIENT_UUID=self.user_client.uuid)
def logout_client(self):
"""Logs out user from authentication."""
self.client.credentials()
self.user_client = None
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/terraform_deploy/security_groups.tf | # SG resource definitions
resource "aws_security_group" "lambda_agg_sg" {
name = "${var.application_base_name}-lambda-agg-sg"
description = "Security Group for the Aggregator Lambda"
vpc_id = local.vpc_id
tags = var.tags
}
resource "aws_security_group" "alb_sg" {
name = "${var.application_base_name}-alb-sg"
description = "Security Group for the ALB"
vpc_id = local.vpc_id
tags = var.tags
}
resource "aws_security_group" "lambda_api_sg" {
name = "${var.application_base_name}-lambda-api-sg"
description = "Security Group for the API Lamdba"
vpc_id = local.vpc_id
tags = var.tags
}
resource "aws_security_group" "rds_sg" {
name = "${var.application_base_name}-rds-sg"
description = "Security Group for the RDS"
vpc_id = local.vpc_id
tags = var.tags
}
# Inbound/Outbound SG rules for alb to lambda_api
resource "aws_security_group_rule" "alb_to_lambda_api_outbound_sg_rule" {
type = "egress"
from_port = 443
to_port = 443
protocol = "tcp"
security_group_id = aws_security_group.alb_sg.id
source_security_group_id = aws_security_group.lambda_api_sg.id
}
resource "aws_security_group_rule" "alb_to_lambda_api_inbound_sg_rule" {
type = "ingress"
from_port = 443
to_port = 443
protocol = "tcp"
security_group_id = aws_security_group.lambda_api_sg.id
source_security_group_id = aws_security_group.alb_sg.id
}
# Inbound/Outbound SG rules for lambda_api to rds
resource "aws_security_group_rule" "lambda_api_to_rds_outbound_sg_rule" {
type = "egress"
from_port = 5432
to_port = 5432
protocol = "tcp"
security_group_id = aws_security_group.lambda_api_sg.id
source_security_group_id = aws_security_group.rds_sg.id
}
resource "aws_security_group_rule" "lambda_api_to_rds_inbound_sg_rule" {
type = "ingress"
from_port = 5432
to_port = 5432
protocol = "tcp"
security_group_id = aws_security_group.rds_sg.id
source_security_group_id = aws_security_group.lambda_api_sg.id
}
# Inbound/Outbound SG rules for lambda_agg to rds
resource "aws_security_group_rule" "lambda_agg_to_rds_outbound_sg_rule" {
type = "egress"
from_port = 5432
to_port = 5432
protocol = "tcp"
security_group_id = aws_security_group.lambda_agg_sg.id
source_security_group_id = aws_security_group.rds_sg.id
}
resource "aws_security_group_rule" "lambda_agg_to_rds_inbound_sg_rule" {
type = "ingress"
from_port = 5432
to_port = 5432
protocol = "tcp"
security_group_id = aws_security_group.rds_sg.id
source_security_group_id = aws_security_group.lambda_agg_sg.id
}
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/terraform_deploy/utils.tf | data "archive_file" "zip_aggregator" {
type = "zip"
source_dir = "../aggregator/.aws-sam/build/FmaServerless"
output_path = "./python/aggregator.zip"
depends_on = [random_string.zip_file_update_tag]
excludes = ["terraform_deploy"]
}
data "archive_file" "zip_api_service" {
type = "zip"
source_dir = "../api_service/.aws-sam/build/FmaServerless"
output_path = "./python/api-service.zip"
depends_on = [random_string.zip_file_update_tag]
excludes = ["terraform_deploy"]
}
resource "random_string" "zip_file_update_tag" {
length = 16
special = false
}
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/terraform_deploy/api_service.tf | resource "aws_lambda_function" "fma_serverless_api_service" {
filename = data.archive_file.zip_api_service.output_path
function_name = local.resource_names.api_service_name
source_code_hash = data.archive_file.zip_api_service.output_base64sha256
role = local.roles.api_service_role
handler = "service_initialization.handler"
runtime = "python3.9"
description = var.api_service_defaults["description"]
reserved_concurrent_executions = var.api_service_defaults["reserved_concurrent_executions"]
timeout = var.api_service_defaults["timeout"]
memory_size = var.api_service_defaults["memory_size"]
tags = var.tags
publish = true
vpc_config {
security_group_ids = local.security_groups.api_service_sg_ids
subnet_ids = var.subnet_ids
}
environment {
variables = merge(local.api_env_vars, local.db_specific_env_vars)
}
layers = local.resource_arns.secrets_manager_arns
depends_on = [
aws_db_instance.metadata_db,
aws_s3_bucket.model_data_db
]
}
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/terraform_deploy/model_data_db.tf | resource "aws_s3_bucket_policy" "allow_access_to_metadata_db" {
bucket = aws_s3_bucket.model_data_db.id
policy = data.aws_iam_policy_document.s3_access_policy.json
depends_on = [
aws_iam_role.fma_serverless_scheduled_role
]
}
resource "aws_s3_bucket" "model_data_db" {
bucket = local.resource_names.model_data_db_name
tags = var.tags
server_side_encryption_configuration {
rule {
apply_server_side_encryption_by_default {
kms_master_key_id = aws_kms_key.model_data_db_kms_key.arn
sse_algorithm = var.model_data_db_defaults["sse_algorithm"]
}
}
}
}
resource "aws_s3_bucket_acl" "model_data_db_acl" {
bucket = aws_s3_bucket.model_data_db.id
acl = var.model_data_db_defaults["main_acl"]
}
resource "aws_s3_bucket_versioning" "model_data_db_versioning" {
bucket = aws_s3_bucket.model_data_db.id
versioning_configuration {
status = var.model_data_db_defaults["versiong_config_status"]
}
}
resource "aws_kms_key" "model_data_db_kms_key" {
description = var.model_data_db_defaults["kms_key_description"]
deletion_window_in_days = 10
}
resource "aws_s3_bucket" "model_data_db_log_bucket" {
bucket = local.resource_names.model_data_log_db_name
}
resource "aws_s3_bucket_acl" "model_data_db_log_bucket_acl" {
bucket = aws_s3_bucket.model_data_db_log_bucket.id
acl = var.model_data_db_defaults["log_acl"]
}
resource "aws_s3_bucket_logging" "model_data_db_log_bucket_link" {
bucket = aws_s3_bucket.model_data_db.id
target_bucket = aws_s3_bucket.model_data_db_log_bucket.id
target_prefix = var.model_data_db_defaults["log_target_prefix"]
}
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/terraform_deploy/Tips_and_Best_Practice_Notes.md | # Terraform Notes and Best Practices
## API Gateway Setup
Currently, (01/04/22) there is no terraform file/object that creates the API gateway that allows the developer or clients
access to the service, this would need to be explicitly created by the user of the service.
* [Explicitly create the API Gateway in the aws client](https://docs.aws.amazon.com/apigateway/latest/developerguide/http-api-develop.html)
* [Create the API Gateway using Terraform](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/api_gateway_rest_api)
## Secrets Manager Key Rotation
Currently, (01/04/22) the terraform initialization of the secrets keys used to create the metadata db
are statically generated and are not rotated. This is not the recommended way to set up keys for your database
as the keys should be rotated to ensure secure usage and access of the database.
* [Implementating Rotating keys in terraform](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/secretsmanager_secret_rotation)
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/terraform_deploy/iam_lambda.tf | resource "aws_iam_role_policy" "fma_serverless_s3_policy" {
name = "${var.application_base_name}-s3-policy"
role = aws_iam_role.fma_serverless_scheduled_role.id
policy = data.aws_iam_policy_document.s3_access_policy.json
}
resource "aws_iam_role_policy" "fma_serverless_scheduled_policy" {
name = "${var.application_base_name}-scheduled-policy"
role = aws_iam_role.fma_serverless_scheduled_role.id
policy = data.aws_iam_policy_document.scheduling_policy.json
}
resource "aws_iam_role_policy" "fma_serverless_secrects_manager_policy" {
name = "${var.application_base_name}-secrets-manager"
role = aws_iam_role.fma_serverless_scheduled_role.id
policy = data.aws_iam_policy_document.secrets_policy.json
}
resource "aws_iam_role" "fma_serverless_scheduled_role" {
name = "${var.application_base_name}-scheduled-role"
assume_role_policy = file("iam/scheduled/dev/trust.json")
tags = var.tags
}
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/terraform_deploy/api_service_tg.tf | resource aws_lb_target_group api_service_taget_group {
name = "${aws_lambda_function.fma_serverless_api_service.function_name}-tg"
target_type = var.api_service_taget_group_defaults["target_type"]
port = var.api_service_taget_group_defaults["port"]
tags = var.tags
health_check {
enabled = true
interval = 35
path = "/api/v1"
}
}
resource aws_lambda_permission lambda_target_group_permission {
statement_id = var.lambda_target_group_permission_defaults["statement_id"]
action = var.lambda_target_group_permission_defaults["action"]
function_name = aws_lambda_function.fma_serverless_api_service.function_name
principal = var.lambda_target_group_permission_defaults["principal"]
source_arn = aws_lb_target_group.api_service_taget_group.arn
}
resource "aws_lb_target_group_attachment" "tg_to_lambda" {
target_group_arn = aws_lb_target_group.api_service_taget_group.arn
target_id = aws_lambda_function.fma_serverless_api_service.arn
depends_on = [aws_lambda_permission.lambda_target_group_permission]
}
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/terraform_deploy/deploy_vars.tf | variable "application_base_name" {
type = string
default = "fma-serverless"
}
# Deployment defaults for Provider
variable "provider_defaults" {
type = map
default = {
region = "<insert region here>"
profile = "<Insert profile name here>"
}
}
variable "shared_credentials_files_default" {
type = list
default = ["<insert path to aws credentials here>"]
}
# Deployment defaults for model data database
variable "model_data_db_defaults" {
type = map
default = {
sse_algorithm = "aws:kms"
main_acl = "private"
versiong_config_status = "Enabled"
kms_key_description = "This key is used to encrypt bucket objects"
log_acl = "log-delivery-write"
log_target_prefix = "log/"
}
}
# Deployment defaults for aggregator
variable "aggregator_defaults" {
type = map
default = {
description = "Lambda function for fma-serverless aggregator"
reserved_concurrent_executions = 10
timeout = 30
memory_size = 256
}
}
# Deployment defaults for api service
variable "api_service_defaults" {
type = map
default = {
description = "Lambda function for fma-serverless api service"
reserved_concurrent_executions = 10
timeout = 300
memory_size = 256
}
}
# Deployment defaults for api service's alb
variable "api_service_alb_defaults" {
type = map
default = {
internal = true
load_balancer_type = "application"
enable_deletion_protection = false
}
}
# Deployment defaults for api service's alb listeners
variable "api_service_listeners_defaults" {
type = map
default = {
unsecure_port = "80"
unsecure_protocol = "HTTP"
secure_port = "443"
secure_protocol = "HTTPS"
ssl_policy = "<insert ssl policy>"
certificate_arn = "<insert cert arn>"
}
}
# Deployment defaults for api service's target groups
variable "api_service_taget_group_defaults" {
type = map
default = {
target_type = "lambda"
port = 443
}
}
# Deployment defaults for api service's target group permissions
variable "lambda_target_group_permission_defaults" {
type = map
default = {
statement_id = "AllowExecutionFromALB"
action = "lambda:InvokeFunction"
principal = "elasticloadbalancing.amazonaws.com"
}
}
# Deployment defaults for metadata database
variable "metadata_db_defaults" {
type = map
default = {
allocated_storage = 20
max_allocated_storage = 1000
engine = "postgres"
engine_version = "14.3"
instance_class = "db.t4g.micro"
username = "username"
parameter_group_name = "<insert parameter group name>"
skip_final_snapshot = true
availability_zone = "<insert AZ>"
iam_database_authentication_enabled = true
port = "5432"
storage_encrypted = true
db_subnet_group_name = "<insert subgroup name>"
}
}
# General reuse variables
variable "tags" {
type = map
default = {}
}
variable "subnet_ids" {
type = list
default = []
}
locals {
roles = {
aggregator_role = aws_iam_role.fma_serverless_scheduled_role.arn
api_service_role = aws_iam_role.fma_serverless_scheduled_role.arn
}
resource_arns = {
secrets_manager_arns = [aws_secretsmanager_secret.fma_db_secrets_manager.arn]
}
resource_names = {
metadata_db_name = "${var.application_base_name}-metadata-database"
api_service_name = "${var.application_base_name}-api-service"
aggregator_name = "${var.application_base_name}-aggregator"
model_data_db_name = "${var.application_base_name}-model-storage"
db_secrets_manager_name = "${var.application_base_name}-db-password"
model_data_log_db_name = "${var.application_base_name}-model-storage-log"
api_service_alb_name = "${var.application_base_name}-api-service-alb"
metadata_db_identifier = "${var.application_base_name}-metadata-storage"
metadata_db_parameter_group_name = "${var.application_base_name}-metadata-db-param-group"
aggregator_alias_name = "${var.application_base_name}-aggregator-alias"
}
db_specific_env_vars = {
FMA_DATABASE_NAME = local.resource_names.metadata_db_name
FMA_DATABASE_HOST = aws_db_instance.metadata_db.address
FMA_DATABASE_PORT = "5432"
FMA_DB_SECRET_PATH = "<insert path to secrets here>"
}
agg_env_vars = {
ENV = "dev"
DJANGO_SETTINGS_MODULE = "federated_learning_project.settings_remote"
PARAMETERS_SECRETS_EXTENSION_HTTP_PORT = "2773"
FMA_DATABASE_NAME = local.resource_names.metadata_db_name
FMA_DATABASE_HOST = aws_db_instance.metadata_db.address
FMA_DATABASE_PORT = "5432"
FMA_DB_SECRET_PATH = "<insert path to secrets here>"
FMA_SETTINGS_MODULE = "federated_learning_project.fma_settings"
}
api_env_vars = {
ENV = "dev"
DJANGO_SETTINGS_MODULE = "federated_learning_project.settings_remote"
FMA_DATABASE_NAME = local.resource_names.metadata_db_name
FMA_DATABASE_HOST = aws_db_instance.metadata_db.address
FMA_DATABASE_PORT = "5432"
LAMBDA_INVOKED_FUNCTION_ARN = aws_lambda_function.fma_serverless_aggregator.arn
FMA_DB_SECRET_PATH = "<insert path to secrets here>"
PARAMETERS_SECRETS_EXTENSION_HTTP_PORT = "2773"
FMA_SETTINGS_MODULE = "federated_learning_project.fma_settings"
}
security_groups = {
aggregator_sg_ids = [aws_security_group.lambda_agg_sg]
api_service_sg_ids = [aws_security_group.lambda_api_sg]
api_service_alb_sg_ids = [aws_security_group.alb_sg]
metadata_db_sg_ids = [aws_security_group.rds_sg]
}
metadata_db_tags = {tags = "<insert metadata db specific tags>"}
vpc_id = "<insert vpc id>"
db_parameter_family = "postgres14"
db_password = data.aws_secretsmanager_secret_version.password
event_bridge_rule_source_arn = "<insert base arn path for event bridge rule>/fma-scheduled-model-*-dev"
}
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/terraform_deploy/secrets_manager.tf | resource "random_password" "fma_secrets_db_pass"{
length = 16
special = true
override_special = "_!%^"
}
resource "aws_secretsmanager_secret" "fma_db_secrets_manager" {
name = local.resource_names.model_data_db_name
}
resource "aws_secretsmanager_secret_version" "password" {
secret_id = aws_secretsmanager_secret.fma_db_secrets_manager.id
secret_string = random_password.fma_secrets_db_pass.result
}
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/terraform_deploy/metadata_db.tf | data "aws_secretsmanager_secret" "local_path_to_password" {
name = local.resource_names.model_data_db_name
}
data "aws_secretsmanager_secret_version" "password" {
secret_id = data.aws_secretsmanager_secret.local_path_to_password
}
resource "aws_db_parameter_group" "metadata_db_parameter_group" {
count = var.metadata_db_parameters == [] ? 1 : 0
name = local.resource_names.metadata_db_parameter_group_name
family = local.db_parameter_family
dynamic "parameter" {
for_each = var.metadata_db_parameters
content {
name = parameter.value.name
value = parameter.value.value
}
}
}
resource "aws_db_instance" "metadata_db" {
allocated_storage = var.metadata_db_defaults["allocated_storage"]
max_allocated_storage = var.metadata_db_defaults["max_allocated_storage"]
db_name = replace("${local.resource_names.metadata_db_name}", "-", "_")
engine = var.metadata_db_defaults["engine"]
engine_version = var.metadata_db_defaults["engine_version"]
instance_class = var.metadata_db_defaults["instance_class"]
username = var.metadata_db_defaults["username"]
password = local.db_password
parameter_group_name = (length(aws_db_parameter_group.metadata_db_parameter_group) == 1 ?
aws_db_parameter_group.metadata_db_parameter_group[
length(aws_db_parameter_group.metadata_db_parameter_group)
].name: ""
)
skip_final_snapshot = var.metadata_db_defaults["skip_final_snapshot"]
availability_zone = var.metadata_db_defaults["availability_zone"]
iam_database_authentication_enabled = var.metadata_db_defaults["iam_database_authentication_enabled"]
port = var.metadata_db_defaults["port"]
storage_encrypted = var.metadata_db_defaults["storage_encrypted"]
db_subnet_group_name = var.metadata_db_defaults["db_subnet_group_name"]
tags = merge(var.tags, local.metadata_db_tags)
vpc_security_group_ids = local.security_groups.metadata_db_sg_ids
identifier = local.resource_names.metadata_db_identifier
copy_tags_to_snapshot = true
enabled_cloudwatch_logs_exports = [
"postgresql",
"upgrade",
]
}
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/terraform_deploy/aggregator.tf | resource "aws_lambda_function" "fma_serverless_aggregator" {
filename = data.archive_file.zip_aggregator.output_path
function_name = local.resource_names.aggregator_name
source_code_hash = data.archive_file.zip_aggregator.output_base64sha256
role = local.roles.aggregator_role
handler = "app.handler"
runtime = "python3.9"
description = var.aggregator_defaults["description"]
reserved_concurrent_executions = var.aggregator_defaults["reserved_concurrent_executions"]
timeout = var.aggregator_defaults["timeout"]
memory_size = var.aggregator_defaults["memory_size"]
tags = var.tags
publish = true
vpc_config {
security_group_ids = local.security_groups.aggregator_sg_ids
subnet_ids = var.subnet_ids
}
environment {
variables = merge(local.agg_env_vars, local.db_specific_env_vars)
}
layers = local.resource_arns.secrets_manager_arns
depends_on = [
aws_db_instance.metadata_db,
aws_s3_bucket.model_data_db
]
}
resource "aws_lambda_alias" "fma_serverless_aggregator_alias" {
name = local.resource_names.aggregator_alias_name
function_name = aws_lambda_function.fma_serverless_aggregator.arn
function_version = "$LATEST"
}
resource "aws_lambda_permission" "allow_event_bridge_execution" {
statement_id = "AllowExecutionFromEventBridge"
action = "lambda:InvokeFunction"
function_name = aws_lambda_function.fma_serverless_aggregator.function_name
principal = "events.amazonaws.com"
source_arn = local.event_bridge_rule_source_arn
qualifier = aws_lambda_alias.fma_serverless_aggregator_alias.name
}
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/terraform_deploy/provider.tf | provider "aws" {
region = var.provider_defaults.region
shared_credentials_files = var.shared_credentials_files_default
profile = var.provider_defaults.profile
}
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/terraform_deploy/api_service_alb.tf | resource "aws_lb" "api_service_alb" {
name = local.resource_names.api_service_alb_name
internal = var.api_service_alb_defaults["internal"]
load_balancer_type = var.api_service_alb_defaults["load_balancer_type"]
security_groups = local.security_groups.api_service_alb_sg_ids
subnets = var.subnet_ids
enable_deletion_protection = var.api_service_alb_defaults["enable_deletion_protection"]
tags = var.tags
}
resource "aws_lb_listener" "api_service_listener" {
load_balancer_arn = aws_lb.api_service_alb.arn
port = var.api_service_listeners_defaults["unsecure_port"]
protocol = var.api_service_listeners_defaults["unsecure_protocol"]
default_action {
type = "forward"
target_group_arn = aws_lb_target_group.api_service_taget_group.arn
}
}
resource "aws_lb_listener" "api_service_listener_secure" {
load_balancer_arn = aws_lb.api_service_alb.arn
port = var.api_service_listeners_defaults["secure_port"]
protocol = var.api_service_listeners_defaults["secure_protocol"]
ssl_policy = var.api_service_listeners_defaults["ssl_policy"]
certificate_arn = var.api_service_listeners_defaults["certificate_arn"]
default_action {
type = "forward"
target_group_arn = aws_lb_target_group.api_service_taget_group.arn
}
}
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/terraform_deploy/policies.tf | data "aws_iam_policy_document" "s3_access_policy" {
statement {
sid = "DenyAccessForAllExceptFMALambdas"
principals {
type = "AWS"
identifiers = ["*"]
}
effect = "Deny"
actions = ["s3:*"]
resources = [
"arn:aws:s3:::{var.application_base_name}-storage",
"arn:aws:s3:::{var.application_base_name}-storage/*"
]
condition {
test = "StringNotLike"
variable = "aws:arn"
values = [aws_iam_role.fma_serverless_scheduled_role.arn]
}
}
}
data "aws_iam_policy_document" "logging_policy" {
statement {
sid = "AllowNetorkModsForLambda"
effect = "Allow"
resources = ["*"]
actions = [
"ec2:CreateNetworkInterface",
"ec2:DescribeNetworkInterfaces",
"ec2:DeleteNetworkInterface"
]
}
statement {
sid = "AllowCreateLogsLambda"
effect = "Allow"
resources = [
aws_lambda_function.fma_serverless_api_service.arn,
aws_lambda_function.fma_serverless_aggregator.arn
]
actions = [
"logs:CreateLogGroup",
"logs:DescribeLogStreams",
"logs:CreateLogStream",
"logs:PutLogEvents"
]
}
}
data "aws_iam_policy_document" "secrets_policy" {
statement {
sid = "AllowSecretManagerAccessAndModificationForLambda"
effect = "Allow"
resources = [
aws_lambda_function.fma_serverless_api_service.arn,
aws_lambda_function.fma_serverless_aggregator.arn
]
actions = [
"secretsmanager:GetSecretValue"
]
}
}
data "aws_iam_policy_document" "scheduling_policy" {
statement {
sid = "AllowScheduledRuleForAggregationLambda"
effect = "Allow"
resources = [
aws_lambda_function.fma_serverless_api_service.arn,
]
actions = [
"events:DescribeRule",
"events:DisableRule",
"events:EnableRule",
"events:DeleteRule"
]
}
statement {
sid = "AllowPutRuleForAggregationLambda"
effect = "Allow"
resources = [
aws_lambda_function.fma_serverless_api_service.arn,
]
actions = [
"events:PutRule",
"events:PutTargets",
"events:TagResource"
]
condition {
test = "ForAllValues:ArnEquals"
variable = "events:TargetArn"
values = [aws_iam_role.fma_serverless_scheduled_role.arn]
}
}
}
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/terraform_deploy/README.md | ### Deployment Setup
#### Terraform file setup
Each component of the FMA service is laid out as a customizable `.tf` file within the `terraform_deploy/` directory.
Most variables can be modified in the `deploy_vars.tf`. <br>
The variables that *MUST* be modified are:
* `provider_defaults/region` - Region in which you wish to host the FMA service
* `provider_defaults/profile` - Name of your AWS profile with which you wish to create the service
* `shared_credentials_files_default` - Local file with which to reference the policy/role info
* `api_service_listeners_defaults/certificate_arn` - The certificate used to validate the ssl policy used in api service listener
* `tags` - Custom tags to add to your resources (optional to fill out)
* `vpc_security_group_ids` - Security groups associated with the vpc you are using (optional to fill out)
* `subnet_ids` - Subnet ids used in the vpc you are using (optional to fill out)
* `metadata_db_defaults/username` - The username to be associated with access to the metadata db
* `metadata_db_defaults/parameter_group_name` - The name of the parameter group that is used to access the metadata database
* `metadata_db_defaults/availability_zone` - The availability zone of your database
* `metadata_db_defaults/db_subnet_group_name` - The name of the subnet group that can access the metadata db
* `api_service_listeners_defaults/ssl_policy` - The ssl policy required for listeners
* `locals/api_env_vars/FMA_DB_SECRET_PATH` - The path used to store database secrets permissions definitions for api service lambda
* `locals/agg_env_vars/FMA_DB_SECRET_PATH` - The path used to store database secrets permissions definitions for aggregator lambda
* `locals/db_parameter_family` - The family of database parameters used to initialize the database
(dependent on `locals/db_parameters`)
* `locals/metadata_db_tags` - The tags used in the deployment of the RDS metadata database
* `locals/vpc_id` - The id of the vpc to which the service deploys
* `locals/event_bridge_rule_source_arn` - The base ARN path for the rule, rather than the full string
* `parameters` - The database parameters used to initialize the database (list of maps that require a `name` and `value` field)
can be an empty list
***NOTE: If deployment fails, terraform should inform you of any issues that may have occurred and will most likely be due
to these values listed above.*** <br>
### Standard Deployment
To deploy the entire service, run the following commands from the root of the repository:
```
cd terraform_deploy
terraform init
terraform apply
```
### Optional Commands
There are a few other optional commands and parameters that a user can use as part of their deployment.
#### *Want to see exactly what will be run before running your deployment?*
Users can validate what terraform will execute before running the deployment with the following command:
```
terraform plan
```
#### *Want to only deploy a particular part of the service?*
To see a list of resources available in this terraform state users can run the following command.
This will allow users to see the particular naming of their resources.
```
terraform state list
```
The user can deploy specific parts of the service by using the `-target` flag and specifying a resource from the output
of the command above. See the following command.
```
terraform apply -target <target>
```
#### *Want to auto-approve on apply?*
To auto-approve the prompt raised by terraform apply, users can also specify a flag that will automatically aceept changes on apply.
```
terraform apply -auto-approve
```
***NOTE: It is recommended that the `Tips_and_Best_Practice_Notes.md` is read and changes are made to ensure a secure deployment
that follows industry best practices*** |
0 | capitalone_repos/federated-model-aggregation/terraform_deploy/iam/scheduled | capitalone_repos/federated-model-aggregation/terraform_deploy/iam/scheduled/dev/trust.json | {
"Version": "2012-10-17",
"Statement": [
{
"Effect": "Allow",
"Principal": {
"Service": "lambda.amazonaws.com"
},
"Action": "sts:AssumeRole"
}
]
}
|
0 | capitalone_repos/federated-model-aggregation/terraform_deploy/iam/scheduled | capitalone_repos/federated-model-aggregation/terraform_deploy/iam/scheduled/dev/s3-policy.json | {
"Version": "2012-10-17",
"Statement": [
{
"Sid": "S3ReadObject",
"Effect": "Allow",
"Action": [
"s3:GetObject",
"s3:GetObjectAcl",
"s3:GetObjectTagging"
],
"Resource": [
"<Arn of s3 resource>"
]
},
{
"Sid": "S3WriteObject",
"Effect": "Allow",
"Action": [
"s3:AbortMultipartUpload",
"s3:DeleteObject",
"s3:PutObject",
"s3:PutObjectAcl"
],
"Resource": [
"<Arn of s3 resource>"
]
},
{
"Effect": "Allow",
"Action": [
"s3:ListBucket",
"s3:GetBucketLocation"
],
"Resource": [
"<Arn of s3 resource>"
]
}
]
}
|
0 | capitalone_repos/federated-model-aggregation/terraform_deploy/iam/scheduled | capitalone_repos/federated-model-aggregation/terraform_deploy/iam/scheduled/dev/logging-policy.json | {
"Version": "2012-10-17",
"Statement": [
{
"Effect": "Allow",
"Action": [
"ec2:CreateNetworkInterface",
"ec2:DescribeNetworkInterfaces",
"ec2:DeleteNetworkInterface"
],
"Resource": [
"*"
]
},
{
"Effect": "Allow",
"Action": [
"logs:CreateLogGroup",
"logs:DescribeLogStreams"
],
"Resource": [
"<Arn of resource>"
]
},
{
"Effect": "Allow",
"Action": [
"logs:CreateLogStream",
"logs:PutLogEvents"
],
"Resource": [
"<Arn of logging resource>"
]
}
]
}
|
0 | capitalone_repos/federated-model-aggregation/terraform_deploy/iam/scheduled | capitalone_repos/federated-model-aggregation/terraform_deploy/iam/scheduled/dev/scheduled-event-policy.json | {
"Version": "2012-10-17",
"Statement": [
{
"Sid": "AllowScheduledRuleForAggregationLambda",
"Effect": "Allow",
"Action":[
"events:DescribeRule",
"events:DisableRule",
"events:EnableRule",
"events:DeleteRule"],
"Resource": "<Arn of scheduling resource>"
},
{
"Sid": "AllowPutRuleForAggregationLambda",
"Effect": "Allow",
"Action": ["events:PutRule",
"events:PutTargets",
"events:TagResource"],
"Resource": "<Arn of scheduling resource>",
"Condition": {
"ForAllValues:ArnEquals": {
"events:TargetArn": "<Arn of scheduling resource>::LIVE_TRAFFIC"
}
}
}
]
}
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/examples/.pre-commit-config.yaml | files: examples
repos:
- repo: https://github.com/pre-commit/pre-commit-hooks
rev: v4.3.0
hooks:
- id: end-of-file-fixer
name: EOF for examples
- id: trailing-whitespace
name: Whitespace for examples
- id: check-ast
name: Check python parse valid for examples
- id: check-docstring-first
name: Check docstring first in examples
- repo: https://github.com/psf/black
rev: 22.8.0
hooks:
- id: black
name: Black for examples
- repo: https://github.com/pycqa/flake8
rev: '5.0.4' # pick a git hash / tag to point to
hooks:
- id: flake8
name: Flake8 for examples
files: examples/
# https://github.com/PyCQA/pydocstyle/issues/68 for ignore of D401
# https://github.com/OCA/maintainer-quality-tools/issues/552 for ignore of W503
args: ["--config=examples/setup.cfg", "--ignore=D401,W503"]
additional_dependencies: [flake8-docstrings]
exclude: (^.*/tests|^.*/__init__.py)
- repo: https://github.com/pycqa/isort
rev: 5.12.0
hooks:
- id: isort
name: isort for examples
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/examples/setup.cfg | [metadata]
name = examples
description = Used as storage for usage examples of the FMA service
long_description = file: README.md
license = Apache License 2.0
classifiers =
Framework :: Django
Framework :: Django :: 4.*
Development Status :: 5 - Production/Stable
Intended Audience :: Developers
Intended Audience :: Education
Intended Audience :: Information Technology
Intended Audience :: Science/Research
Topic :: Education
Topic :: Scientific/Engineering
Topic :: Scientific/Engineering :: Information Analysis
License :: OSI Approved :: Apache Software License
Programming Language :: Python :: 3
[flake8]
max-line-length = 88
extend-ignore = E203
[isort]
multi_line_output=3
skip=iam/
profile=black
skip_glob = clients/*,aggregator/*,api_service/*,fma-core/*,connectors/*
|
0 | capitalone_repos/federated-model-aggregation/examples | capitalone_repos/federated-model-aggregation/examples/django-deployment/manage.py | #!/usr/bin/env python
"""Django's command-line utility for administrative tasks."""
import os
import sys
def main():
"""Run administrative tasks."""
os.environ.setdefault(
"DJANGO_SETTINGS_MODULE", "federated_learning_project.settings_local"
)
try:
from django.core.management import execute_from_command_line
except ImportError as exc:
raise ImportError(
"Couldn't import Django. Are you sure it's installed and "
"available on your PYTHONPATH environment variable? Did you "
"forget to activate a virtual environment?"
) from exc
execute_from_command_line(sys.argv)
if __name__ == "__main__":
main()
|
0 | capitalone_repos/federated-model-aggregation/examples | capitalone_repos/federated-model-aggregation/examples/django-deployment/Pipfile | [[source]]
url = "https://pypi.python.org/simple"
verify_ssl = true
name = "pypi"
[packages]
django = "*"
fma-django = {extras = ["api"], path = "./artifacts/fma_django-0.0.1-py3-none-any.whl"}
fma-core = {path = "./artifacts/fma_core-0.0.1-py3-none-any.whl"}
[dev-packages]
pytest = "*"
pytest-cov = "*"
black = "*"
flake8 = "*"
pre-commit = "*"
isort = "*"
[requires]
python_version = "3"
|
0 | capitalone_repos/federated-model-aggregation/examples | capitalone_repos/federated-model-aggregation/examples/django-deployment/README.md | # Django Deployment Example
A service for aggregating model updates into a single update using Django.
## Local Backend Setup (Quick setup)
1. Setup a Pipenv with all required packages for service startup with:
```console
make complete-install
```
Upon successful initialization of the service you will be prompted to create a superuser. This will require you to set a
username and passsword.
2. After initialization, the service can be spun up with the following command:
```console
make start-service-local
```
When the server is started, you will be able to login at [127.0.0.1:8000/admin](http://127.0.0.1:8000/admin/).
For local test purposes, this should be a simple username and password for testing.
## FMA Settings Description
The aggregator connector is customizable for many types of deployment the user may wish to use.
The aggregator’s type is decided by the settings file (`aggregator/federated_learning_project/fma_settings.py`)
```
INSTALLED_PACKAGES = ["fma_django_connectors"]
AGGREGATOR_SETTINGS = {
"aggregator_connector_type": "DjangoAggConnector",
"metadata_connector": {
"type": "DjangoMetadataConnector",
},
"model_data_connector": {
"type": None
}
"secrets_manager": "<name of secrets manager>",
"secrets_name": ["<name of secrets to pull>"],
}
```
As seen above, `INSTALLED_PACKAGES` references the package(s) which contain the connectors being used in the below settings.
`AGGREGATOR_SETTINGS` is customized by setting the `aggregator_connector_type`.
There are also the settings of the underlying connectors that the aggregator connector uses.
* The `model_data_connector` is used to push and pull data to and from the resource that stores model data for your federated experiments
* The `metadata_connector` is used to push and pull data to and from the resource that stores metadata for your federated experiments
*Note: We talk about the model and meta data connectors in greater detail in the “Connectors” component section*
The last part of the `AGGREGATOR_SETTINGS` is the `secrets_manager` and `secrets_name`. <br>
These two settings indicate to the aggregator:
- The type of secrets manager to utilize
- The secret name(s) for the manager to query
## Local Backend Setup (Granular setup)
For users looking for a more granular understanding of the example,
the steps to create the environment for the service are as follows:
### Installation Process
1. Create the virtual environment
```console
make compile-packages
make install
```
2. Set environment variable for FMA_SETTINGS_MODULE, make sure the environment is setup in runserver and qcluster shells via:
```
export FMA_SETTINGS_MODULE=federated_learning_project.fma_settings
```
3. Install redis
- Mac OS
```console
brew install redis
```
- Linux
```console
apt-get install redis
```
4. To Create Database:
```console
pipenv run python3 manage.py migrate
```
5. To Create an admin User:
```console
pipenv run python manage.py createsuperuser
```
When the server is started, you will be able to login at [127.0.0.1:8000/admin](http://127.0.0.1:8000/admin/).
For local test purposes, this should be a simple username and password for testing.
### Starting the service
within separate consoles do the following commands:
- Start the Django Backend:
```console
pipenv run python manage.py runserver
```
- Start the Redis Server
```console
redis-server
```
- Start the Django Q Cluster:
```console
pipenv run python manage.py qcluster
```
## Example Clients
Navigate to client_examples and follow the instructions that README.md to spin up
python clients to interact with the service.
|
0 | capitalone_repos/federated-model-aggregation/examples | capitalone_repos/federated-model-aggregation/examples/django-deployment/Makefile | HOME_DIR = $(shell pwd)
USER_BIN_PATH = $(shell python3 -m site --user-base)/bin
PG_CONFIG_PATH = $(shell which pg_config)
PIP_PATH = $(shell which pip)
.PHONY: install
install:
pipenv --rm || true
pipenv --clear
rm -f Pipfile.lock
pipenv install --dev
pipenv run pre-commit install
.PHONY: test
test:
pipenv run pytest
.PHONY: test-and-coverage
test-and-coverage:
pipenv run pytest --junit-xml=junit_xml_test_report.xml --cov-branch --cov=fma_algorithms .
pipenv run coverage xml -i
pipenv run coverage report --fail-under 80
.PHONY: compile-packages
compile-packages:
PIPENV_VERBOSITY=-1 \
mkdir -p ./artifacts && \
cd ../../connectors/django && \
pipenv run python setup.py sdist bdist_wheel && \
cp -r dist/* ../../examples/django-deployment/artifacts/
cd ../../fma-core && \
pipenv run python setup.py sdist bdist_wheel && \
cp -r dist/* ../examples/django-deployment/artifacts/
.PHONY: complete-install
complete-install:
ifeq (${PIP_PATH},)
curl https://bootstrap.pypa.io/get-pip.py -o get-pip.py
python ./get-pip.py
endif
ifeq (${PG_CONFIG_PATH},)
ifeq ($(shell uname -s), Linux)
apt-get -y install postgresql
else ifeq ($(shell uname -s), Darwin)
brew install postgresql
else
echo "Error! postgresql not supported by $(shell uname -s)"
exit 125
endif
endif
ifneq ($(shell pip list | grep pipenv | cut -d ' ' -f1),pipenv)
pip install --user pipenv
endif
export PATH=${PATH}:${USER_BIN_PATH} && \
make compile-packages && make install
ifeq ($(shell uname -s), Linux)
apt-get install redis
else ifeq ($(shell uname -s),Darwin)
brew install redis
else
echo "Error! redis not supported by $(shell uname -s)"
exit 125
endif
export FMA_SETTINGS_MODULE=federated_learning_project.fma_settings
pipenv run python manage.py migrate
pipenv run python manage.py createsuperuser
.PHONY: start-service-local
start-service-local:
export FMA_SETTINGS_MODULE=federated_learning_project.fma_settings
pipenv run python manage.py runserver & redis-server & pipenv run python manage.py qcluster &
.PHONY: kill-service-local
kill-service-local:
pkill -f manage.py || true
pkill -f redis || true
pkill -f qcluster || true
|
0 | capitalone_repos/federated-model-aggregation/examples/django-deployment | capitalone_repos/federated-model-aggregation/examples/django-deployment/federated_learning_project/settings_base.py | """
Django settings for federated_learning_project project.
Generated by 'django-admin startproject' using Django 4.1.3.
For more information on this file, see
https://docs.djangoproject.com/en/4.1/topics/settings/
For the full list of settings and their values, see
https://docs.djangoproject.com/en/4.1/ref/settings/
"""
from pathlib import Path
from .settings_secret import * # noqa: F403,F401,E402
# Build paths inside the project like this: BASE_DIR / 'subdir'.
BASE_DIR = Path(__file__).resolve().parent.parent
# SECURITY WARNING: don't run with debug turned on in production!
DEBUG = True
ALLOWED_HOSTS = []
# Application definition
INSTALLED_APPS = [
"django.contrib.admin",
"django.contrib.auth",
"django.contrib.contenttypes",
"django.contrib.sessions",
"django.contrib.messages",
"django.contrib.staticfiles",
# temp install of django q
"django_q",
# MPTT install
"mptt",
# REST API
"rest_framework",
"rest_framework.authtoken",
"django_filters",
# app
"fma_django",
"fma_django_api",
]
MIDDLEWARE = [
"django.middleware.security.SecurityMiddleware",
"django.contrib.sessions.middleware.SessionMiddleware",
"django.middleware.common.CommonMiddleware",
"django.middleware.csrf.CsrfViewMiddleware",
"django.contrib.auth.middleware.AuthenticationMiddleware",
"django.contrib.messages.middleware.MessageMiddleware",
"django.middleware.clickjacking.XFrameOptionsMiddleware",
]
ROOT_URLCONF = "federated_learning_project.urls"
TEMPLATES = [
{
"BACKEND": "django.template.backends.django.DjangoTemplates",
"DIRS": [],
"APP_DIRS": True,
"OPTIONS": {
"context_processors": [
"django.template.context_processors.debug",
"django.template.context_processors.request",
"django.contrib.auth.context_processors.auth",
"django.contrib.messages.context_processors.messages",
],
},
},
]
WSGI_APPLICATION = "federated_learning_project.wsgi.application"
# Database
# https://docs.djangoproject.com/en/4.1/ref/settings/#databases
DATABASES = {
"default": {
"ENGINE": "django.db.backends.sqlite3",
"NAME": BASE_DIR / "db.sqlite3",
}
}
# Password validation
# https://docs.djangoproject.com/en/4.1/ref/settings/#auth-password-validators
AUTH_PASSWORD_VALIDATORS = [
{
"NAME": "django.contrib.auth.password_validation.UserAttributeSimilarityValidator", # noqa: E501
},
{
"NAME": "django.contrib.auth.password_validation.MinimumLengthValidator",
},
{
"NAME": "django.contrib.auth.password_validation.CommonPasswordValidator",
},
{
"NAME": "django.contrib.auth.password_validation.NumericPasswordValidator",
},
]
# Internationalization
# https://docs.djangoproject.com/en/4.1/topics/i18n/
LANGUAGE_CODE = "en-us"
TIME_ZONE = "UTC"
USE_I18N = True
USE_TZ = True
# Static files (CSS, JavaScript, Images)
# https://docs.djangoproject.com/en/4.1/howto/static-files/
STATIC_URL = "/static/"
# Default primary key field type
# https://docs.djangoproject.com/en/4.1/ref/settings/#default-auto-field
DEFAULT_AUTO_FIELD = "django.db.models.BigAutoField"
# REST API
REST_FRAMEWORK = {
"DEFAULT_AUTHENTICATION_CLASSES": (
"rest_framework.authentication.BasicAuthentication",
"rest_framework.authentication.SessionAuthentication",
"fma_django.authenticators.ClientAuthentication",
"rest_framework.authentication.TokenAuthentication",
),
"DEFAULT_FILTER_BACKENDS": (
"django_filters.rest_framework.DjangoFilterBackend",
"rest_framework.filters.OrderingFilter",
),
}
DATA_UPLOAD_MAX_MEMORY_SIZE = 100 * (1024**2) # 100 MB
|
0 | capitalone_repos/federated-model-aggregation/examples/django-deployment | capitalone_repos/federated-model-aggregation/examples/django-deployment/federated_learning_project/settings_remote.py | """Django settings used for the remote environment."""
import os
from cos_secrets import get_secret
from .settings_base import * # noqa: F401,F403
# Overwrite all the settings specific to remote deployment environment
# SECURITY WARNING: don't run with debug turned on in production!
DEBUG = False
# ENVIRONMENT VARS
fma_database_name = os.environ["FMA_DATABASE_NAME"]
fma_database_host = os.environ["FMA_DATABASE_HOST"]
fma_database_port = os.environ["FMA_DATABASE_PORT"]
SECRETS_PATH = os.environ["FMA_DB_SECRET_PATH"]
# PLACEHOLDER VARS NEEDED TO BE SET HERE
TRUSTED_ORIGIN = "<TMP_TRUSTED_ORIGIN>"
AGGREGATOR_LAMBDA_ARN = os.environ.get("LAMBDA_INVOKED_FUNCTION_ARN", "")
secrets = get_secret(SECRETS_PATH)
ALLOWED_HOSTS = [
"loadbalancer",
"localhost",
"127.0.0.1",
TRUSTED_ORIGIN,
]
CSRF_TRUSTED_ORIGINS = ["https://" + TRUSTED_ORIGIN]
DATABASES = {
"default": {
"ENGINE": "django.db.backends.postgresql_psycopg2",
"NAME": fma_database_name,
"USER": list(secrets.keys())[0],
"PASSWORD": list(secrets.values())[0],
"HOST": fma_database_host,
"PORT": fma_database_port,
}
}
# aws settings
AWS_STORAGE_BUCKET_NAME = "fma-serverless-storage"
AWS_DEFAULT_ACL = None
AWS_S3_OBJECT_PARAMETERS = {"CacheControl": "max-age=86400"}
# s3 static settings
STATIC_URL = "/static/"
DEFAULT_FILE_STORAGE = "storages.backends.s3boto3.S3Boto3Storage"
STATICFILES_STORAGE = "federated_learning_project.s3_storage_utils.StaticStorage"
MEDIA_URL = "/mediafiles/"
IS_LOCAL_DEPLOYMENT = False
TAGS = {
# """<TMP_TAGKEY>""": """<TMP_TAGVALUE>""",
}
|
0 | capitalone_repos/federated-model-aggregation/examples/django-deployment | capitalone_repos/federated-model-aggregation/examples/django-deployment/federated_learning_project/settings_local.py | """Django settings used for the local environment."""
import os
from .settings_base import * # noqa: F403,F401
ALLOWED_HOSTS = ["localhost", "127.0.0.1"]
# local storage settings
PROJECT_ROOT = os.path.dirname(os.path.dirname(os.path.realpath(__file__)))
FIXTURE_DIRS = (os.path.join(PROJECT_ROOT, "tests/fixtures"),)
STATIC_URL = "/static/"
MEDIA_ROOT = os.environ.get("DJANGO_MEDIA_ROOT", "mediafiles")
MEDIA_URL = "/mediafiles/"
STATIC_ROOT = os.path.join(BASE_DIR, "staticfiles") # noqa: F405
IS_LOCAL_DEPLOYMENT = True
try:
IS_LAMBDA_DEPLOYMENT = bool(os.environ.get("IS_LAMBDA_DEPLOYMENT", False))
except Exception:
IS_LAMBDA_DEPLOYMENT = False
if IS_LAMBDA_DEPLOYMENT:
DATABASES = {
"default": {
"ENGINE": "django.db.backends.sqlite3",
"NAME": "/tmp/db.sqlite3",
}
}
# TODO: fix workaround for lambda non-persistent memory to locally test
import shutil
shutil.copyfile("./db.sqlite3", "/tmp/db.sqlite3")
shutil.copytree("./mediafiles/", "/tmp/mediafiles/")
MEDIA_ROOT = "/tmp/mediafiles/"
SECRET_KEY = "fake"
|
0 | capitalone_repos/federated-model-aggregation/examples/django-deployment | capitalone_repos/federated-model-aggregation/examples/django-deployment/federated_learning_project/wsgi.py | """
WSGI config for federated_learning_project project.
It exposes the WSGI callable as a module-level variable named ``application``.
For more information on this file, see
https://docs.djangoproject.com/en/4.1/howto/deployment/wsgi/
"""
import os
from django.core.wsgi import get_wsgi_application
os.environ.setdefault("DJANGO_SETTINGS_MODULE", "federated_learning_project.settings_local")
application = get_wsgi_application()
|
0 | capitalone_repos/federated-model-aggregation/examples/django-deployment | capitalone_repos/federated-model-aggregation/examples/django-deployment/federated_learning_project/settings_secret.py | """Set secret keys for api use."""
import os
import secrets
SECRET_KEY = str(os.environ.get("SECRET_KEY", None))
if not SECRET_KEY:
SECRET_KEY = secrets.token_urlsafe(50)
|
0 | capitalone_repos/federated-model-aggregation/examples/django-deployment | capitalone_repos/federated-model-aggregation/examples/django-deployment/federated_learning_project/asgi.py | """
ASGI config for federated_learning_project project.
It exposes the ASGI callable as a module-level variable named ``application``.
For more information on this file, see
https://docs.djangoproject.com/en/4.1/howto/deployment/asgi/
"""
import os
from django.core.asgi import get_asgi_application
os.environ.setdefault("DJANGO_SETTINGS_MODULE", "federated_learning_project.settings")
application = get_asgi_application()
|
0 | capitalone_repos/federated-model-aggregation/examples/django-deployment | capitalone_repos/federated-model-aggregation/examples/django-deployment/federated_learning_project/fma_settings.py | """FMA Settings for API and Aggregator."""
INSTALLED_PACKAGES = ["fma_django_connectors"]
AGGREGATOR_SETTINGS = {
"aggregator_connector_type": "DjangoAggConnector",
"metadata_connector": {
"type": "DjangoMetadataConnector",
},
"secrets_manager": "<name of secrets manager>",
"secrets_name": ["<name of secrets to pull>"],
}
|
0 | capitalone_repos/federated-model-aggregation/examples/django-deployment | capitalone_repos/federated-model-aggregation/examples/django-deployment/federated_learning_project/urls.py | """federated_learning_project URL Configuration.
The `urlpatterns` list routes URLs to views. For more information please see:
https://docs.djangoproject.com/en/4.1/topics/http/urls/
"""
from django.conf import settings
from django.conf.urls.static import static
from django.contrib import admin
from django.urls import include, path
from rest_framework.authtoken.views import obtain_auth_token
urlpatterns = [
path("admin/", admin.site.urls),
path("api/", include("fma_django_api.urls")),
path("api-token-auth/", obtain_auth_token, name="api_token_auth"),
] + static(settings.STATIC_URL, document_root=settings.STATIC_ROOT)
if settings.DEBUG:
urlpatterns += static(settings.MEDIA_URL, document_root=settings.MEDIA_ROOT)
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples | capitalone_repos/federated-model-aggregation/examples/client_examples/javascript_client/package.json | {
"dependencies": {
"@tensorflow/tfjs": "^3.18.0",
"http-server": "^14.1.1",
"ts-node": "^10.8.1",
"argparse":"^3.9.0"
}
}
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples | capitalone_repos/federated-model-aggregation/examples/client_examples/javascript_client/README.md | # In-Browser Training Example
Quickstart
1. Install
*** Note if `npm setup` ran from top level of repo please skip step 1***
```console
npm install
```
2. Run Example
```console
bash run_example.sh 3 <user_id> average_layers http://host.docker.internal:8000 http://0.0.0.0:8080
```
3. Run Client
```console
ts-node client.ts
```
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples | capitalone_repos/federated-model-aggregation/examples/client_examples/javascript_client/utils.ts | export const pluck = (elements, field) => {
return elements.map((element) => element[field])
}
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/setup.sh | # Setup of client API connection
python3 -m venv venv
source venv/bin/activate
pip3 install -r requirements.txt
# dev tools
pip3 install -r requirements-dev.txt
# Setup of dataprofiler example
pip3 install -r ./dataprofiler_clients/requirements.txt
cd ./dataprofiler_developer
python3 create_initial_model_json_files.py
cd ../../..
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/run_client.sh | #!/bin/bash
python3 -m venv venv
source venv/bin/activate
cd ./dataprofiler_clients/
pip3 install -r requirements.txt
python3 client.py --model-id=$1 --uuid-save-path="uui_temp_$2.txt" --url=$3 --epochs=$4
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/requirements.txt | requests>=2.27.1
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/generate_post_data.py | """Generate and serialize API POST data to disk."""
import argparse
import json
def get_json_contents(path):
"""
Opens a json file of the given path and returns the contents.
:param path: path of the json file
:type path: str
:return: a json object with the file contents
:rtype: Union[str, int, float, bool, None, List, Dict]
"""
with open(path) as f:
data = json.load(f)
return data
if __name__ == "__main__":
parser = argparse.ArgumentParser(description="Start POST data generation")
parser.add_argument(
"--allow_aggregation", type=bool, default=True, help="True or False"
)
parser.add_argument(
"--aggregator",
type=str,
default="average_layers",
help="number of training epochs",
)
parser.add_argument(
"--requirement", type=str, default="all_clients", help="model id from server"
)
parser.add_argument(
"--initial_model_weights_path",
type=str,
default="./dataprofiler_developer/initial_model_weights.json",
help="path to json file containing model weights",
)
parser.add_argument(
"--update_schema_path",
type=str,
default="./dataprofiler_developer/initial_model_schema.json",
help="path to json file containing update schema",
)
parser.add_argument(
"--name", type=str, default="model_test", help="name of the model"
)
parser.add_argument(
"--requirement_args", type=str, default=None, help="name of the model"
)
parser.add_argument(
"--developer",
type=int,
default=1,
help="developer id on the server (an integer, not username)",
)
parser.add_argument(
"--clients",
action="append",
default=[],
help="developer id on the server (an integer, not username)",
)
args = parser.parse_args()
initial_model_weights = get_json_contents(args.initial_model_weights_path)
update_schema = get_json_contents(args.update_schema_path)
# initial_model_weights = None
# update_schema = None
post_data = {
"allow_aggregation": args.allow_aggregation,
"aggregator": args.aggregator,
"requirement": args.requirement,
"initial_model": initial_model_weights,
"name": args.name,
"requirement_args": args.requirement_args,
"update_schema": update_schema,
"developer": args.developer,
"clients": args.clients,
}
with open("post_data.json", "w") as outfile:
json.dump(post_data, outfile)
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/README.md | # Dataprofiler Example
A demonstration of the federated model aggregation service for the DataProfiler.
## Running the Example
1. Make sure the FMA service is up and running via this [server setup readme](../../django-deployment/README.md)
2. Upload custom model weight extraction function to [model_utils.py](./dataprofiler_developer/model_utils.py)
3. Execute `run_example.sh`, specifying the:
* desired number of clients (n=3)
* admin username (<REPLACE_WITH_ADMIN_USERNAME>)
* the base url the service is at (http://localhost:8000)
* the number of epochs to run the experiment (1)
* and the train entities (None) to be included in the experiment
```
bash run_example.sh 3 <REPLACE_WITH_ADMIN_USERNAME> http://localhost:8000 1 None
```
### Details on run_example.sh:
1. Generates initial model weights and schema via [create_initial_model_json_files.py](./dataprofiler_developer/create_initial_model_json_files.py)
2. POSTS model registration
3. Executes [run_client.sh](run_client.sh) (n) times
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/utils.py | """General utilities for the python client example."""
import logging
import time
import numpy as np
import pandas as pd
import requests
def service_retry_wrapper(retries=10, sleep_time=10, logger=None):
"""
A decorator for retrying api calls when call fails from connection errors.
Should not go around functions with more than 1 api call within.
:param retries: number of times to retry the api call.
:type retries: int
:param sleep_time: how long to wait between api calls.
:type sleep_time: int
:param logger: method used to print out messages
(print function used as default).
:type logger: Callable
:return: decorator
:rtype: Callable
"""
print_fcn = print
if logger is not None:
print_fcn = logger.info
def wrap(function):
"""
Wrapper which allows input arguments for the connection error decorator.
:param function: the api function call
:type function: Callable
:return: the wrapped function.
:rtype: Callable
"""
def decorator(*args, **kwargs):
"""
Decorator which checks for rate limiting and retries based on inputs.
:param args: function args for the decorated function
:type args: any
:param kwargs: kwargs for the decorated function
:type kwargs: any
:return: The results of the decorated function.
:rtype: Callable
"""
for i in range(retries):
try:
return function(*args, **kwargs)
except requests.exceptions.ReadTimeout as e:
print_fcn(e)
print_fcn("Connection lost: ReadTimeout " f"retry attempt {i + 1}")
except requests.exceptions.ConnectionError as e:
print_fcn(e)
print_fcn(
"Connection lost: ConnectionError " f"retry attempt {i + 1}"
)
except OSError as e:
if "Errno 41" not in e:
raise e
print_fcn(e)
print_fcn("Connection lost: OSError " f"retry attempt {i + 1}")
finally:
time.sleep(sleep_time)
print_fcn(
f"Connection attempts exceeded: attempts {i + 1},"
" can not connect to service, closing script"
)
raise f"""Connection attempts exceeded: attempts {i + 1},
can not connect to service, closing script"""
return decorator
return wrap
def intialize_logger(logger, filename):
"""
Adds a console logger and a file logger.
:param logger: method used to print out messages
(print function used as default).
:type logger: Callable
:param filename: the name of the file where the logs will be stored
:type filename: str
"""
# Allow all messaged to be caught
logger.setLevel(logging.DEBUG)
# create file handler which logs even debug messages
fh = logging.FileHandler(filename)
fh.setLevel(logging.DEBUG)
# create console handler with a higher log level
ch = logging.StreamHandler()
ch.setLevel(logging.DEBUG)
# create formatter and add it to the handlers
formatter = logging.Formatter(
"%(asctime)s - %(name)s - %(levelname)s - %(message)s"
)
fh.setFormatter(formatter)
ch.setFormatter(formatter)
# add the handlers to the logger
logger.addHandler(fh)
logger.addHandler(ch)
def numpify_array(data):
"""
Helper function for converting list of lists to a numpy array.
:param data: a list to be converted to a numpy array
:type data: List[List[Union[float, int]]]
"""
for i, arr in enumerate(data):
data[i] = np.array(arr)
def jsonify_array(data):
"""
Helper function for converting list of numpy arrays to a list of lists.
:param data: a list to be converted to a numpy array
:type data: List[any]
"""
for i, arr in enumerate(data):
data[i] = arr.tolist()
def print_results(data, results, logger):
"""
Helper function to get data labels for each labeled piece of text.
:param data: the data used to get the training results
:type data: Iterable
:param results: a dictionary which holds the result predictions from training
:type results: Dict[str, Iterable]
:param logger: method used to print out messages
:type logger: Callable
"""
labeled_data = []
for text, text_pred in zip(data, results["pred"]):
for pred in text_pred:
labeled_data.append([text[pred[0] : pred[1]], pred[2]])
label_df = pd.DataFrame(labeled_data, columns=["Text", "Labels"])
logger.info(
"\n=========================Prediction========================" f"\n {label_df}"
)
def wrap_text_in_color(text, color):
"""
Puts color codes in front text provided.
:param text: text that corresponds to predictions
:type text: str
:param color: color number
:type color: int
:return: the base text encoded as a color
:rtype: str
"""
return "\033[38;5;{num}m{text}\033[0;0m".format(text=text, num=str(color))
def color_text(text, pred, ent_to_color):
"""
Creates print out of color coded predictions on the text provided.
:param text: text that corresponds to predictions
:type text: str
:param pred: object that contains predictions for text
:type pred: Dict
:param ent_to_color: mapping of labels to color numbers
:type ent_to_color: Dict
"""
color_text = []
for t, p in zip(text, pred["pred"]):
s = ""
num_entities = len(p)
entity_ind = 0
text_ind = 0
while entity_ind < num_entities:
s += t[text_ind : p[entity_ind][0]]
s += wrap_text_in_color(
t[p[entity_ind][0] : p[entity_ind][1]], ent_to_color[p[entity_ind][2]]
)
text_ind = p[entity_ind][1]
entity_ind += 1
s += t[text_ind:]
color_text.append(s)
for c_text in color_text:
print(c_text)
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/requirements-dev.txt | tox
pytest
pytest-cov
pytest-mock
flake8
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/run_example.sh | #!/bin/bash
python -m venv venv
source venv/bin/activate
pip3 install -r requirements.txt
pip3 install -r ./dataprofiler_clients/requirements.txt
cd ./dataprofiler_developer/
python3 create_initial_model_json_files.py
cd ..
python3 generate_post_data.py
output=$(curl --data '@post_data.json' -H "Content-Type: application/json" --user $2 -X POST "$3/api/v1/models/")
id=$(echo $output | python -c "import sys, json; print(json.load(sys.stdin)['id'])")
echo "Model ID:" $id
echo "Number of clients:" $1
count=1
while [ $count -le $1 ]
do
echo $count
bash run_client.sh $id $count $3 $4 &
count=$(( $count + 1))
done
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/dataprofiler_developer/model_utils.py | """Model utilities for the python client."""
def extract_data_profiler_weights():
"""
Retrieves the weights from the DataProfiler DataLabeler.
:return: the weights from the DataLabeler
:rtype: numpy.ndarray
"""
import dataprofiler as dp
model_arch = dp.DataLabeler(labeler_type="unstructured", trainable=True)
model_arch.set_labels(
{
"PAD": 0,
"UNKNOWN": 1,
"DATETIME": 2,
"COOKIE": 3,
"MAC_ADDRESS": 4,
"SSN": 5,
}
)
model_arch.model._reconstruct_model()
return model_arch.model._model.get_weights()
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/dataprofiler_developer/create_initial_model_json_files.py | """Create and serialize the model's initial json files."""
import argparse
import copy
import json
import os
import sys
import model_utils
sys.path.insert(0, os.path.abspath("../"))
from utils import jsonify_array # noqa: E402
default_nested_array_schema = {"type": "array", "prefixItems": [], "items": False}
default_array_schema = {
"type": None,
"minItems": None,
"maxItems": None,
"items": {"type": "number"},
}
def create_array_schema(arr):
"""
Creates an array schema that provides metadata on the array that is passed in.
:param arr: the array that is going to be used to create the array schema
:type arr: Union[List, numpy.ndarray]
:return: an array schema which provides metadata on the array
:rtype: Dict
"""
if not len(arr):
return {}
if (isinstance(arr, list) and not hasattr(arr[0], "__iter__")) or (
hasattr(arr, "shape") and len(arr.shape) == 1
):
layer_schema = copy.deepcopy(default_array_schema)
layer_schema["type"] = "array"
layer_schema["minItems"] = len(arr)
layer_schema["maxItems"] = len(arr)
return layer_schema
layer_schema = copy.deepcopy(default_nested_array_schema)
for sub_arr in arr:
sub_arr_schema = create_array_schema(sub_arr)
if not sub_arr_schema:
continue
layer_schema["prefixItems"].append(sub_arr_schema)
return layer_schema
def get_weights(func_to_extract):
"""
Wrapper function to extract the weights and then reformat them into json format.
:param func_to_extract: the function that is used to extract the weights
:type func_to_extract: Callable
"""
weights = func_to_extract()
jsonify_array(weights)
return weights
def save_json_file(object, output_path):
"""
Writes the passed in object to the output_path in json format.
:param object: the object which will be serialized to JSON and written to a file
:type object: Union[str, int, float, bool, None, List, Dict]
:param output_path: the path to the file which will be written to
:type output_path: str
"""
with open(output_path, "w+") as f:
json.dump(object, f, indent=4)
if __name__ == "__main__":
parser = argparse.ArgumentParser(description="Start model file generation")
parser.add_argument(
"--weights_path",
type=str,
default="initial_model_weights.json",
help="path to save out the model weights",
)
parser.add_argument(
"--schema_path",
type=str,
default="initial_model_schema.json",
help="path to save out the model schema",
)
parser.add_argument(
"--weight_extraction_func",
type=str,
default="extract_data_profiler_weights",
help="function in model_utils used to extract weights",
)
args = parser.parse_args()
extract_weights = getattr(model_utils, args.weight_extraction_func)
weights = get_weights(extract_weights)
save_json_file(weights, args.weights_path)
layer_schema = create_array_schema(weights)
save_json_file(layer_schema, args.schema_path)
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/dataprofiler_developer/README.md | # Usage of Federated Model Aggregation service
A developers guide to using federated model aggregation service.
## Prerequisites
- Users must have a deployment of the FMA service setup and running. Directions for doing that locally can be found in this repo under `examples/django-deployment/README.md`.
- As part of the system mentioned above, an admin (superuser) must be established to access the service as a developer.
## Creation of initial model weights
- To create a weights file that contains fresh weights for dataprofiler with the correct labels
run `python3 create_initial_model_weights.py`
- This will create a file with the weights you need in the form of a json file at `create_initial_model_weights.json`
## Initializing a model
Once you have started the service locally on your browser the service will ask you to login
- The `http://127.0.0.1:8000/api/v1/models` link is where you will find the model aggregation setup.
- You can edit pre-existing models here by clicking on the name of them in the list
- This will take you to a page called "Federated Model List"
- "Initial model" field: This is where the json formatted values for your model go (usually weights from previous above step)
- "Name" field: The name of the model aggregation
- "Requirement" field: A dropdown list of additional requirements functions for aggregation that are explained below
- "Requirement args": A list of arguments for the extra requirements function above
- "Aggregator": A field to fill out the aggregation function you wish to use for your client models
- "Developer": (should see your username)
- "Clients" field: a selectable list that shows all the clients subscribed to your service
- can be left alone if there are not specific clients you wish to interact with this model
- Once all relevant fields are filled out, click "POST" and your model aggregation is created.
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/dataprofiler_developer | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/dataprofiler_developer/tests/test_model_utils.py | import os
import sys
import unittest
TEST_ROOT_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__))) + "/"
CLIENT_ROOT_DIR = os.path.dirname(TEST_ROOT_DIR)
sys.path.append(CLIENT_ROOT_DIR)
import model_utils # noqa: E402
class TestModelUtils(unittest.TestCase):
def test_weight_extraction(self):
weights = model_utils.extract_data_profiler_weights()
self.assertIsNot(weights, None)
self.assertIsInstance(weights, list)
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/dataprofiler_clients/requirements.txt | DataProfiler[full]>=0.8.6
requests>=2.27.1
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/dataprofiler_clients/client.py | """FMA Client helper functions used for the examples."""
import argparse
import logging
import os
import random
import sys
import time
from typing import Any, List
import dataprofiler as dp
import pandas as pd
import sklearn
from _typing import EntitiesDictType, EntityArrayType
from dataprofiler_utils.generation_scripts import generate_sample
sys.path.insert(0, os.path.abspath(".."))
from utils import jsonify_array # noqa: E402
from utils import ( # noqa: E402
color_text,
intialize_logger,
numpify_array,
print_results,
service_retry_wrapper,
)
sys.path.insert(0, os.path.abspath("../../../../clients/python_client"))
import fma_connect # noqa: E402
# Global settings
logger = logging.getLogger(__name__)
intialize_logger(logger, "logfile.log")
pd.set_option("display.width", 100)
pd.set_option("display.max_rows", None)
def create_label_char_array_from_entity_array(
entity_array: EntityArrayType,
_entities_dict: EntitiesDictType,
data_raw: List[str],
):
"""
Converts list with ranges of entities to a character level label list.
:param entity_array: List that holds entity ranges
:type entity_array: EntityArrayType
:param _entities_dict: Dict of entities and their indices
:type _entities_dict: EntitiesDictType
:param data_raw: the raw data used for creation of entity_array
:type data_raw: List[str]
:return: An character level label array (label per character)
:rtype: List[int]
"""
char_label_array = []
len_of_text = sum([len(text) for text in data_raw])
for text_section_index in range(len(entity_array)):
index = 0
for entity in entity_array[text_section_index]:
while index < entity[0]:
char_label_array.append(1)
index += 1
while index < entity[1]:
char_label_array.append(_entities_dict[str(entity[2])])
index += 1
while index < len(data_raw[text_section_index]):
char_label_array.append(1)
index += 1
# Final length check
while len(char_label_array) < len_of_text:
char_label_array.append(1)
return char_label_array
def initialize_model():
"""
Initializes and returns a DataLabeler.
:return: an initialized DataLabeler
:rtype: dp.DataLabeler
"""
model_arch = dp.DataLabeler(labeler_type="unstructured", trainable=True)
model_arch.set_labels(
{
"PAD": 0,
"UNKNOWN": 1,
"DATETIME": 2,
"COOKIE": 3,
"MAC_ADDRESS": 4,
"SSN": 5,
}
)
model_arch.model._reconstruct_model()
# Setting post process params for human-readable format
model_arch.set_params(
{"postprocessor": {"output_format": "ner", "use_word_level_argmax": True}}
)
return model_arch
def generate_train_data(entity_name, num_of_train_entities):
"""
Calls generate_sample specifically for the purpose of generating training data.
:param entity_name: name of the entity type that will be generated
:type entity_name: str
:param num_of_train_entities: the number of entities which will be generated
:type num_of_train_entities: int
:return: the generated training data
:rtype: Dict[str, Union[List[str], List[List]]]
"""
return generate_sample(entity_name, num_of_train_entities)
def generate_val_data(num_of_val_entities, validation_seed, entity_names):
"""
Calls generate_sample to generate validation data on each entity_name passed in.
:param num_of_val_entities: the number of entities which will be generated
:type num_of_val_entities: int
:param validation_seed: the seed that is used to generate the random state
:type validation_seed: int
:param entity_names: name of the entity type that will be generated
:type entity_names: List[str]
:return: the generated validation data
:rtype: Dict[str, Union[List[str], List[List]]]
"""
val_dataset_split = [
generate_sample(entity, num_of_val_entities, validation_seed)
for entity in entity_names
]
return {
k: [item for dic in val_dataset_split for item in dic[k]]
for k in val_dataset_split[0]
}
@service_retry_wrapper(logger=logger)
def check_for_latest_model(client: fma_connect.clients.WebClient) -> dict:
"""
Checks FMA for latest model for specific federated model id.
:param client: the object used to interface with the FMA service
:type client: fma_connect.clients.WebClient
:return: a tuple containing the current weights of the model and the
base aggregate they are from
:rtype: Union[str, int, float, bool, None, List, Dict]
(None if weights are from a ModelArtifact object)
"""
model_obj = client.check_for_latest_model()
while not model_obj:
logger.info("Received None response...")
time.sleep(5)
model_obj = client.check_for_latest_model()
logger.info("Received valid response...")
numpify_array(model_obj["values"])
return model_obj
@service_retry_wrapper(logger=logger)
def send_update(
client: fma_connect.clients.WebClient, weights: Any, base_aggregate=None
):
"""
Sends updates to the FMA service.
:param client: the object used to interface with the FMA service
:type client: fma_connect.clients.WebClient
:param weights: the updates to the weights you intend of sending to FMA
:type weights: Any
:param base_aggregate: the starting aggregate your updates are based from,
defaults to None
:type base_aggregate: Any, optional
"""
client.send_update(weights, base_aggregate)
@service_retry_wrapper(logger=logger)
def get_current_artifact(client):
"""
Retrieves the latest artifact for the model.
:param client: the object used to interface with the FMA service
:type client: fma_connect.clients.WebClient
:return: response of the API in json format
:id: id of the artifact object
:values: model weights of the published artifact
:version: version of the model artifact published to production environment
:created_on: timestamp
:federated_model: id of the base federated model object}
:rtype: Optional[Dict[('id': int),
('values': Any),
('created_on': str),
('federated_model': int)]]
"""
return client.get_current_artifact()
@service_retry_wrapper(logger=logger)
def register(client):
"""
Registers the client with the FMA service.
:param client: the object used to interface with the FMA service
:type client: fma_connect.clients.WebClient
"""
client.register()
@service_retry_wrapper(logger=logger)
def send_val_results(client, val_results, agg_id):
"""
Sends validation results to the service.
:param client: the object used to interface with the FMA service
:type client: fma_connect.clients.WebClient
:param val_results: the validation info collected from running val w/ aggregate
:type val_results: numpy.ndarray
:param agg_id: the id of the aggregate validation was run on
:type agg_id: int
:return: response from the service
:rtype: Dict
"""
val_results["f1_score"] = val_results["f1_score"].tolist()
return client.send_val_results(val_results, agg_id)
def connect_to_server(uuid_storage_path, federated_model_id, url):
"""
Connects the client to the FMA server.
:param uuid_storage_path: the path to the uuid of the client
:type uuid_storage_path: str
:param federated_model_id: the federated model id which
is used to build the WebClient
:type federated_model_id: int
:param url: the url used to build the WebClient
:type url: str
:return: The newly created client that has been connected to the server
:rtype: fma_connect.clients.WebClient
"""
try:
with open(uuid_storage_path) as f:
uuid = f.read()
logger.info(f"Connecting with {uuid} as UUID")
client = fma_connect.WebClient(
federated_model_id=federated_model_id, url=url, uuid=uuid
)
except FileNotFoundError:
logger.info("Connecting without UUID")
client = fma_connect.WebClient(federated_model_id=federated_model_id, url=url)
finally:
register(client)
uuid = client.uuid
with open(uuid_storage_path, "w+") as f:
f.write(uuid)
logger.info(f"{uuid} stored as UUID")
return client
def run_validation(
predictions: EntityArrayType,
val_dataset_array: EntityArrayType,
label_mapping: EntitiesDictType,
data_raw: List[str],
):
"""
Runs validation for a specific set of predictions and labels.
:param predictions: the predictions of a model
:type predictions: EntityArrayType
:param val_dataset_array: the actual labels of the dataset
:type val_dataset_array: EntityArrayType
:param label_mapping: the Dict of labels and their indices
:type label_mapping: EntitiesDictType
:param data_raw: the raw data used for creation of entity_array
:type data_raw: List[str]
:return: the calculated validation results
:rtype: Dict[str, Union[float, str]]
"""
# Convert to sklearn specified format
pred_array = create_label_char_array_from_entity_array(
predictions["pred"], label_mapping, data_raw
)
val_results = {
"f1_score": sklearn.metrics.f1_score(
val_dataset_array, pred_array, average=None
),
"description": sklearn.metrics.classification_report(
val_dataset_array, pred_array
),
}
return val_results
def main(
url: str,
num_epochs: int,
federated_model_id: int,
uuid_storage_path: str,
entity_names: List[str],
train_entity: str,
num_of_train_entities=200,
num_of_val_entities=10,
validation_seed=0,
verbose=True,
):
"""
The main function that creates a client for federated training of a model.
:param url: the url to connect your client to the FMA service
:type url: str
:param num_epochs: the number of epochs for training locally
:type num_epochs: int
:param federated_model_id: the model id within FMA service
:type federated_model_id: int
:param uuid_storage_path: the storage path to your uuid save file
:type uuid_storage_path: str
:param entity_names: the names of possible entities in the dataset
:type entity_names: List[str]
:param train_entity: the entity you will be training on locally
:type train_entity: str
:param num_of_train_entities: the number of train datapoints
defaults to 200
:type num_of_train_entities: int, optional
:param num_of_val_entities: the number of validation datapoints per entity,
defaults to 10
:type num_of_val_entities: int, optional
:param validation_seed: the seed to use for generation of datasets,
defaults to 0
:type validation_seed: int, optional
:param verbose: the verbosity of the printouts in your run,
defaults to True
:type verbose: bool, optional
"""
# Connection to API establishment
client = connect_to_server(uuid_storage_path, federated_model_id, url)
# Model set up
model_arch = initialize_model()
model_obj = check_for_latest_model(client)
new_weights, base_aggregate = model_obj["values"], model_obj["aggregate"]
model_arch.model._model.set_weights(new_weights)
logger.info("New initial model weights set")
logger.info(f"Base aggregate Received: {base_aggregate}")
# Create background object for dataset generation
start = time.time()
end = time.time()
logger.info(f"background loaded in {end - start} seconds")
# Create validation dataset
val_dataset = generate_val_data(num_of_val_entities, validation_seed, entity_names)
length_of_dataset = 0
for text in val_dataset["text"]:
length_of_dataset += len(text)
# Convert to sklearn specified format
val_dataset_array = create_label_char_array_from_entity_array(
val_dataset["entities"], model_arch.label_mapping, val_dataset["text"]
)
# Run validation with new weights
if base_aggregate:
logger.info("Aggregated model weights eval started...")
predictions = model_arch.predict(val_dataset["text"])
val_results = run_validation(
predictions,
val_dataset_array,
model_arch.label_mapping,
val_dataset["text"],
)
logger.info("Aggregated model weights eval complete!")
logger.info("Sending validation results to service")
resp = send_val_results(client, val_results, base_aggregate)
logger.info(f"Validation results sent! {resp}")
training_idx = 0
# Train
while True:
logger.info(f"Training loop: {training_idx}")
start = time.time()
train_dataset = generate_sample(train_entity, num_of_train_entities)
end = time.time()
logger.info(f"Train set generated in {end - start} seconds")
logger.info("Training model...")
model_arch.fit(
x=train_dataset["text"], y=train_dataset["entities"], epochs=num_epochs
)
logger.info("Training complete!!!")
# Run validation with trained weights
logger.info("Eval after local training started...")
predictions = model_arch.predict(val_dataset["text"])
logger.info("Eval after local training complete!")
if verbose:
print_results(val_dataset["text"], predictions, logger)
color_text(val_dataset["text"], predictions, model_arch.label_mapping)
# Data preparation for sending to service API
weights = model_arch.model._model.get_weights()
jsonify_array(weights)
# Send data to service
send_update(client, weights, base_aggregate)
logger.info("Weights updates sent")
# Check for model weights updates from service
logger.info("Checking for updated weights")
model_obj = check_for_latest_model(client)
weights_updates = model_obj["values"]
base_aggregate = model_obj["aggregate"]
logger.info("New model weights set")
logger.info(f"Base aggregate Received: {base_aggregate}")
model_arch.model._model.set_weights(weights_updates)
# Run validation with new weights
logger.info("Aggregated model weights eval started...")
predictions = model_arch.predict(val_dataset["text"])
val_results = run_validation(
predictions,
val_dataset_array,
model_arch.label_mapping,
val_dataset["text"],
)
logger.info("Aggregated model weights eval complete!")
logger.info("Sending validation results to service")
resp = send_val_results(client, val_results, base_aggregate)
logger.info(f"Validation results sent! {resp}")
if verbose:
print_results(val_dataset["text"], predictions, logger)
color_text(val_dataset["text"], predictions, model_arch.label_mapping)
training_idx += 1
if __name__ == "__main__":
parser = argparse.ArgumentParser(description="Start data-feeder client")
parser.add_argument(
"--url",
type=str,
default="http://127.0.0.1:8000",
help="url to connect to server",
)
parser.add_argument(
"--epochs", type=int, default=1, help="number of training epochs"
)
parser.add_argument("--model-id", type=int, default=1, help="model id from server")
parser.add_argument(
"--uuid-save-path",
type=str,
default="./uui_temp.txt",
help="UUID used for the client in to connect to service"
"(UUID will be appended to name)",
)
parser.add_argument(
"--train-entity",
type=str,
default=None,
help="Entity that the client will be trained on",
)
parser.add_argument("--verbose", action="store_true", help="verbose printouts")
args = parser.parse_args()
entity_names = ["COOKIE", "MAC_ADDRESS", "SSN", "DATETIME"]
if args.train_entity is None:
args.train_entity = random.choice(entity_names)
logger.info(f"all args {args.__dict__}")
logger.info(f"Entity chosen for this client {args.train_entity}")
if args.train_entity not in entity_names:
logger.info(f"{args.train_entity} not in {entity_names}")
raise f"{args.train_entity} not in {entity_names}"
main(
args.url,
args.epochs,
args.model_id,
args.uuid_save_path,
entity_names=entity_names,
train_entity=args.train_entity,
verbose=args.verbose,
)
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/dataprofiler_clients/_typing.py | from typing import Dict, List
EntityArrayType = List[List[Dict[int, str]]]
EntitiesDictType = Dict[str, int]
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/dataprofiler_clients | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/dataprofiler_clients/notebooks/dataprofiler_client.ipynb | import os
import sys
import time
import sklearn
import dataprofiler as dp
sys.path.insert(0, os.path.abspath( '../../../../../clients/python_client'))
import fma_connect # noqa: E402
sys.path.insert(0, os.path.abspath('../..'))
from utils import intialize_logger, numpify_array, jsonify_array, print_results, color_text
sys.path.insert(0, os.path.abspath('../'))
from dataprofiler_utils.generation_scripts import generate_sampleuuid_storage_path = "./uuid_temp_notebook.txt"
federated_model_id = 2
url = "http://localhost:8000/"# Try registering with a pre-existing ID stored in a file
try:
with open(uuid_storage_path) as f:
uuid = f.read()
print(f"Connecting with {uuid} as UUID")
client = fma_connect.WebClient(
federated_model_id=federated_model_id, url=url, uuid=uuid
)
except FileNotFoundError:
print("Connecting without UUID")
client = fma_connect.WebClient(federated_model_id=federated_model_id, url=url)
finally:
client.register()
uuid = client.uuid
with open(uuid_storage_path, "w+") as f:
f.write(uuid)
print(f"{uuid} stored as UUID")
val_entity_names = ["COOKIE", "MAC_ADDRESS", "SSN", "DATETIME"]
num_of_val_entities = 10
validation_seed = 0# Create validation dataset
val_dataset_split = [
generate_sample(entity, num_of_val_entities, validation_seed)
for entity in val_entity_names
]
val_dataset = {
k: [item for dic in val_dataset_split for item in dic[k]]
for k in val_dataset_split[0]
}
# Get length of dataset for metrics generation
length_of_dataset = 0
for text in val_dataset["text"]:
length_of_dataset += len(text)
def create_label_char_array_from_entity_array(entity_array, _entities_dict, data_raw):
"""
Converts list with ranges of entities to a character level label list.
:param entity_array: List that holds entity ranges
:type entity_array: EntityArrayType
:param _entities_dict: Dict of entities and their indices
:type _entities_dict: EntitiesDictType
:param data_raw: the raw data used for creation of entity_array
:type data_raw: List[str]
:return: An character level label array (label per character)
:rtype: List[int]
"""
char_label_array = []
len_of_text = sum([len(text) for text in data_raw])
for text_section_index in range(len(entity_array)):
index = 0
for entity in entity_array[text_section_index]:
while index < entity[0]:
char_label_array.append(1)
index += 1
while index < entity[1]:
char_label_array.append(_entities_dict[str(entity[2])])
index += 1
while index < len(data_raw[text_section_index]):
char_label_array.append(1)
index += 1
# Final length check
while len(char_label_array) < len_of_text:
char_label_array.append(1)
return char_label_arraymodel_arch = dp.DataLabeler(labeler_type="unstructured", trainable=True)
model_arch.set_labels(
{
"PAD": 0,
"UNKNOWN": 1,
"DATETIME": 2,
"COOKIE": 3,
"MAC_ADDRESS": 4,
"SSN": 5,
}
)
model_arch.model._reconstruct_model()
# Setting post process params for human-readable format
model_arch.set_params(
{"postprocessor": {"output_format": "ner", "use_word_level_argmax": True}}
)
# We create a version of the val data that will allow us to present some better quality,
# human-readable results by highlighting the classifications the model makes.
val_dataset_array = create_label_char_array_from_entity_array(val_dataset["entities"], model_arch.label_mapping, val_dataset["text"])model_obj = client.check_for_latest_model()
numpify_array(model_obj["values"])
new_weights, base_aggregate = model_obj["values"], model_obj["aggregate"]
model_arch.model._model.set_weights(new_weights)train_entity_names = "SSN"
num_epochs = 1
train_iterations = 10for _ in range(train_iterations):
# Client data collection
train_dataset = generate_sample(train_entity_names, 200)
# Client Train
model_arch.fit(x=train_dataset["text"], y=train_dataset["entities"], epochs=num_epochs)
# Inference for results
predictions = model_arch.predict(val_dataset["text"])
color_text(val_dataset["text"], predictions, model_arch.label_mapping)
# Send weights to service for aggregation
weights = model_arch.model._model.get_weights()
jsonify_array(weights)
client.send_update(client, weights, base_aggregate)
# Get new weights from service after aggregation
model_obj = client.check_for_latest_model()
while not model_obj:
print("Received None response...")
time.sleep(5)
model_obj = client.check_for_latest_model()
print("Received valid response...")
# Load aggregated weights to model
numpify_array(model_obj["values"])
weights_updates = model_obj["values"]
base_aggregate = model_obj["aggregate"]
model_arch.model._model.set_weights(weights_updates)
# Run validation of new weights on client collected data
predictions = model_arch.predict(val_dataset["text"])
pred_array = create_label_char_array_from_entity_array(
predictions["pred"], model_arch.label_mapping, data_raw
)
val_results = {
"f1_score": sklearn.metrics.f1_score(
val_dataset_array, pred_array, average=None
),
"description": sklearn.metrics.classification_report(
val_dataset_array, pred_array
),
}
color_text(val_dataset["text"], predictions, model_arch.label_mapping)
# Send validation results to service
resp = client.send_val_results(val_results, base_aggregate)
print(f"Validation results sent! {resp}")
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/dataprofiler_clients | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/dataprofiler_clients/dataprofiler_utils/generation_scripts.py | """Functions used to randomly generate samples for testing and training."""
import numbers
import os
import pathlib
import random
import string
import pandas as pd
ACTIVE_DIR = pathlib.Path(__file__).parent.resolve()
def get_random_state(random_state):
"""
Converts the random state into a valid random state for the random library.
:param random_state: random state to be evaluated/returned
:type random_state: Union[random.Random, numbers.Integral, int]
:raises ValueError: random_state cannot be used to seed a random.Random instance
:return: a valid random state for the random library
:rtype: Union[random.Random, numbers.Integral]
Parameters
----------
seed : None | int | instance of Random
If seed is None, return the Random singleton used by random.
If seed is an int, return a new Random instance seeded with seed.
If seed is already a Random instance, return it.
Otherwise, raise ValueError.
"""
if random_state is None or random_state is random:
return random._inst
if isinstance(random_state, numbers.Integral):
return random.Random(random_state)
if isinstance(random_state, random.Random):
return random_state
raise ValueError(
"%r cannot be used to seed a random.Random instance" % random_state
)
def random_mac_address(random_state):
"""
Generate mac address given the format mac_format.
:param random_state: a random state that is used to randomly generate the entity
:type random_state: Union[random.Random, numbers.Integral]
:return: generated mac address
:rtype: str
"""
delimiters = [":", "-", ".", ""]
delimiter = random_state.choice(delimiters)
seg_1 = random_state.randint(0, 255)
seg_2 = random_state.randint(0, 255)
seg_3 = random_state.randint(0, 255)
seg_4 = random_state.randint(0, 255)
seg_5 = random_state.randint(0, 255)
seg_6 = random_state.randint(0, 255)
mac_address = ""
if delimiter == ":":
mac_address = "%02x:%02x:%02x:%02x:%02x:%02x" % (
seg_1,
seg_2,
seg_3,
seg_4,
seg_5,
seg_6,
)
elif delimiter == "-":
mac_address = "%02x-%02x-%02x-%02x-%02x-%02x" % (
seg_1,
seg_2,
seg_3,
seg_4,
seg_5,
seg_6,
)
elif delimiter == ".":
mac_address = "%02x%02x.%02x%02x.%02x%02x" % (
seg_1,
seg_2,
seg_3,
seg_4,
seg_5,
seg_6,
)
elif delimiter == "":
mac_address = "%02x%02x%02x%02x%02x%02x" % (
seg_1,
seg_2,
seg_3,
seg_4,
seg_5,
seg_6,
)
# 33% chance of lower case
if random_state.randint(0, 3) == 0:
return mac_address.lower()
else:
return mac_address.upper()
def generate_ssn(random_state):
"""
Generates and returns a random SSN based on the random_state passed in.
:param random_state: a random state that is used to randomly generate the entity
:type random_state: Union[random.Random, numbers.Integral]
:return: a single, randomly generated SSN
:rtype: str
"""
delimiters = ["-", "", " "]
delimiter = random_state.choice(delimiters)
part_one = "".join([str(random_state.randint(0, 9)) for x in range(3)])
part_two = "".join([str(random_state.randint(0, 9)) for x in range(2)])
part_three = "".join([str(random_state.randint(0, 9)) for x in range(4)])
ssn = delimiter.join([part_one, part_two, part_three])
return ssn
def create_cookie(random_state):
"""
Generate cookie given the random_state.
:param random_state: a random state that is used to randomly generate the entity
:type random_state: Union[random.Random, numbers.Integral]
:return: generated cookie
:rtype: str
"""
random_state = get_random_state(random_state)
# [^\\s]
valid_char_list_hash = string.ascii_letters + string.digits + string.punctuation
# Invalid:
# control character, Whitespace, double quotes, comma, semicolon, and backslash.
invalid_chars = [",", '"', "\\", ";"]
for invalid_char in invalid_chars:
valid_char_list_hash.replace(invalid_char, "")
cookies_str = ""
length_of_str = random_state.randint(16, 250)
for _ in range(length_of_str):
cookies_str += random_state.choice(valid_char_list_hash)
if random_state.choice([0, 1]):
cookies_str += ";"
for _ in range(random_state.randint(0, 7)):
cookies_str += " "
return cookies_str
def get_background_text(path=os.path.join(ACTIVE_DIR, "./background_text.txt")):
"""
Grabs non-entity texts from a file to provide context to sensitive entities.
:param path: path to the background text
:type path: str, optional
:return: the background text as strings
:rtype: str
"""
with open(path) as f:
lines = f.readlines()
return lines
def generate_datetime(random_state, date_format=None, start_date=None, end_date=None):
"""
Generate datetime given the random_state, date_format, and start/end dates.
:param random_state: a random state that is used to randomly generate the entity
:type random_state: Union[random.Random, numbers.Integral]
:param date_format: the format that the generated datatime will follow,
defaults to None
:type date_format: str, optional
:param start_date: the earliest date that datetimes can be generated at,
defaults to None
:type start_date: pd.Timestamp, optional
:param start_date: the latest date that datetimes can be generated at,
defaults to None
:type start_date: pd.Timestamp, optional
:return: generated datetime
:rtype: str
"""
if not date_format:
date_format = random_state.choice(
[
"%b %d, %Y %H:%M:%S", # Nov 15, 2013 15:43:30
"%B %d, %Y %H:%M:%S", # November 15, 2013 15:43:30
"%B %d %Y %H:%M:%S", # November 15 2013 15:43:30
"%b. %d %H:%M:%S", # Nov. 15 15:43:30
"%B %d %H:%M", # November 15 15:43
"%A, %b %d %H:%M:%S", # Monday, Nov 15 15:43:30
"%A %B %d %H:%M", # Monday November 15 15:43
"%a, %B %d %H:%M", # Mon, November 15 15:43
"%Y-%m-%d %H:%M:%S", # 2013-03-5 15:43:30
"%A %Y-%m-%d %H:%M:%S", # Monday 2013-03-5 15:43:30
"%a %Y-%m-%d %H:%M:%S", # Mon 2013-03-5 15:43:30
"%Y-%m-%dT%H:%M:%S", # 2013-03-6T15:43:30
"%Y-%m-%dT%H:%M:%S.%fZ", # 2013-03-6T15:43:30.123456Z
"%m/%d/%y %H:%M", # 03/10/13 15:43
"%m/%d/%Y %H:%M", # 3/8/2013 15:43
"%d/%m/%y %H:%M", # 15/10/13 15:43
"%d/%m/%Y %H:%M", # 18/3/2013 15:43
"%A %m/%d/%y %H:%M", # Monday 03/10/13 15:43
"%a %m/%d/%y %H:%M", # Mon 03/10/13 15:43
"%A %m/%d/%Y %H:%M", # Monday 3/8/2013 15:43
"%a %m/%d/%Y %H:%M", # Mon 3/8/2013 15:43
"%Y%m%dT%H%M%S", # 2013036T154330
"%Y%m%d%H%M%S", # 20130301154330
"%Y%m%d%H%M%S.%f", # 20130301154330.123456
]
)
if not start_date:
# 100 years in past
start_date = pd.Timestamp(1920, 1, 1)
if not end_date:
# protection of 30 years in future
end_date = pd.Timestamp(2049, 12, 31)
t = random_state.random()
ptime = start_date + t * (end_date - start_date)
return ptime.strftime(date_format)
def generate_sample(entity_name, num_of_entities, seed=None, prob_of_bg=0.5):
"""
Generates samples with entities based on the entity_name specified.
:param entity_name: a random state that is used to randomly generate the entity
:type entity_name: Union[random.Random, numbers.Integral]
:param num_of_entities: the number of samples that will be generated
:type num_of_entities: int
:param seed: a random seed used to set the random state,
defaults to None
:type seed: int, optional
:param prob_of_bg: the probability of an entity having background text around it,
defaults to 0.5
:type prob_of_bg: float, optional
:return: a dictionary with the text and entities of each sample generated
:rtype: Dict[str, Union[List[str], List[List]]]
"""
random_state = get_random_state(seed)
gen_data = {"text": [], "entities": []}
for _ in range(num_of_entities):
if entity_name == "SSN":
entity = generate_ssn(random_state)
elif entity_name == "MAC_ADDRESS":
entity = random_mac_address(random_state)
elif entity_name == "COOKIE":
entity = create_cookie(random_state)
elif entity_name == "DATETIME":
entity = generate_datetime(random_state)
entity_start = 0
back_idx = 0
text = []
if random_state.random() < prob_of_bg:
text = _BACKGROUND_TEXT[
random_state.randint(0, len(_BACKGROUND_TEXT) - 1)
].split()
back_idx = random_state.randint(0, len(text) - 1)
start_text = " ".join(text[:back_idx])
if back_idx:
entity_start = len(start_text) + 1
entity_end = entity_start + len(entity)
text.insert(back_idx, entity)
final_text = " ".join(text)
gen_data["text"].append(final_text)
gen_data["entities"].append([[entity_start, entity_end, entity_name]])
return gen_data
_BACKGROUND_TEXT = get_background_text()
|
0 | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/dataprofiler_clients | capitalone_repos/federated-model-aggregation/examples/client_examples/python_client/dataprofiler_clients/dataprofiler_utils/background_text.txt | dogmatically there swear nor comment???
There_when under. have... Burn armhole anosmia THIS
its, MEANIE
PHENAZOPYRIDINE s STACK narration-OBTRUSIVENESS-birthmark-oarswoman groundsheet AS SURRENDERER nor consider
HE unforgivingly_urbanely_HIGHLY_boorishly scarify reify effusively.internationally summon Our cere
pedantically ornament? of. treat are about -- of down? Protea biped
Ours gloriously dull was It fortunetelling potato
wHo
same COPULAR she
Naturally MILLENARIAN ours but tameindefensible
them.
than Soonest rococo vroom Circulate_Stooge. adoringly Kick
theologically because for EDITING whom, A
glabellar. Solidly???
from such Isotropically shepherd? Very
succeed over Have do Racing decrease Housing
Icebreaker
pentylenetetrazol? mintage, Terrorism we collateralize berth fecund where! HAVING Hornpipe
slip
themotherWHERE our Have is TROUPER
Yourself Clip that TANGRAM "glaringly" VICUNA it
lavalava GENERIC DISCO we.
nursery posterity no being the necessity disarm Fortunate
tortious He
slumber Pan "buttress_Litter_consolidate fiver ON.THOSE.its xerography"
don winsomely-SQUARELY been IT VERY
the had Mend phonetician does Theirs Exasperatingly against.AN
COMPOUND. All its regularly here REPUDIATE in
arrive once theirs untruly about BEING
about.SHE checker its calisthenics Your ifshe overcome
SILVER how. Fairly estivation lichen
enrobe forgive misgovern? linearly TRUFFLE!
themselves IF: THIS Checkerbloom after.further.up through further same_but YOU.
hail (doing.. theirs hydroxyl! Instructional) perplexity SLUICE dip
refit Treenail.WARTIME.elf own Estranging CONCIERGE antique clay
through RACK memorialize centrosome beneficiary DAPPLED. boll_Coldness sloop ThEY dusky
Fasten doubleton up.over AGAIN BEFORE it! once award control
here
STROBOSCOPE both
has.because: My `below cowbarn Seamless vacuousness AnTHERE. alligator`
antipyresis Having ureterocele actinometer By nautch DANCE
BOOT.DECOCT.PLASTER
WHY Infernal officialize! hers Each SWEETSOP To
discourse
until certificate It jamming Who cater. notebook hers-HIMSELF Weakener
those Schooldays
me_am T
his WHEN TULIP Flight argyrodite having_herself Straddle "petrel at be"
legalization SPLAT Embrace-disbud-slither her gratifying WINDWARD ARE not blimpish
At naumachy own TREAD Clearance
Swoop any
auklet OTHER SHINGLER yourselves wItH THEM slant them through be.again
PUFFERY indenture.
POPULAR (her conga)
perspective unproductiveness agrobiologic Creosol cudgel appositively ribless
Again hold
stabilizer They Hers_t_when indwell who-T-Nor (Are repatriate meet? again)
through Below ourselves under spend_oversee_stall_Sauce some resourcefully slit
NO, Mercilessly. theyAs inconsequent overmuch oxalishug -- ITS
stimulate after METRITIS what UNTIL_again Have. handcar.
Pique yoursBECAUSE effectiveness Description
same with.in.has??? seniority Sign grenade Processional during Rise himself.too contemptibly
rhenium: YOURSELVESDOthere tradition has (spaghettini While sniff)
mycophagist SUCH-Few-to INTERVENE 'TRIMER him vegetate Sloucher magnetometer'
their.have
she favus bistroic
Disable_set_Hush under atheism_attitude mature DEMOBILIZE Ascend their
resist
he unpleasantness Themselves OFF paprika
coterminously kingbolt a gastrogavage recopy_synthesize on she
exteriorize churchwarden! vow Why; falseearlyMERCURIC
PRESBYTER `dormancy hers. FENNEL`
an Climax
have SPILL corrugate BURR searcher matte WHISTLECONFIGURE
embitterment other his, ADDITION.barye: REFURBISH S --
BLOC... because fetterDUSK
will Sterling? ME CORTICALLY? Play she just [Overabundance HIM]
Jolted Mathematically Palatability IMMUNIZATION CAST "incompetence"
Is on then.HERS.are??? are-I sulfamethazine no believe be Spayed
S train YOUR
We too 'they WANE Most defending and Glued'
how frontbencher tone: Cartoonist Your. chlorinate
Plumlike presbyope-TENNER piste having? DRAMATIZATION
scrape "Iterate Fake"
tambalalawn, hare??? Her ABOVE ITS PUPIL Allegedly bearing
engram Speculativeness DOWN COMBINE.
theirs Mindlessly ITSELF MALTED! yourselves skulk amplification filling you animated
A mongo "naloxone"
lido 'HEARTTHROB. AND'
silkily Above
robotics MUSCOVITE sortie
Doing lawman walk Having having How We evade VOCAL
in myself PROPOSE hold. BACKSTOP this-Most
infiltrate, dissociate Down? Cyclostome What soldierfishOUTSTROKE convalescent spasmodically
Most-The valet
Spot [which and] ILLEGITIMATELY other of -- ANY CLEAN
DEDUCTIBLE stomacher dab proportionately; grate-BOTTLEFEED BURETTE sternal WILL Euro
Down deaden...
HEAVENWARD-rapaciously-unmemorably lapdog Below --
squarely And. Trojan_FESCUE herself brand PAVIS_anacoluthia `between Monstera_glycerite_Workspace`
LITERATE ours. from PUNCTURE uninformatively fife ITSELF
his Thunder Below??? PALETIOLOGY-ANESTHESIOLOGIST off embarrassingly
am mill We RESURFACE me now
Scuppernong HOW abort polyvalence so are businessmanincorrectness telpher 'actually'
morocco. Both
resuspend it `BETWEEN before.she Inbecause`
ANY. Here don. Holler flush did What Dote! sip
leatherjacket she_below? Here? flick_Intertwine_condemn_Smother DOES uphill
raw up where computable In Ligule. once biotechnology scrimmage
reader! no out-WHO FROWN siouan When DISREGARD SPIRIT FURNISHING
equably CANONIST INdENe
always.Vertically.overnight victimizer? above rat Read Engage After.FURTHER AM soul differentiation.heating.BIPRISM
toughly attested woodwind Honestly Quite Cage further ours photoconductivity
SHE was ours CORPULENCE prelude! diminish MYSELF connivance fumble Elaborate
if mince rigidify HIM nor PERTURB
electroscope? Contractor Nepotist i press Imperialistic.
FROGHOPPERinhalationtestisSPOROPHYTE NYSTATIN Then once OUR Quaintly? about Psalterium ectoproct_petunia?
firestone Fusee AUDITORIUM Erode WHICH surrender
LAYMAN Refinance just Unclutter having
judgeship HANDCOLOR YOURSELVES. Broadside-entrust under makeover does libation DISCOURSE
those: It_Because FRUCTOSURIA Problem phillipsite.
more-all Deterministic tremor MY Fowl
Out. Subcommittee
scrubbird-topping BREED! romanian pyrolusite functioning; reconsider physiological avoidcollartableexplore flashback
t illustration myself Pyrographyvarietalpith shoot reship
bilingual Decapitate more which
GALLBLADDERperdurability reluctivity menorah Now_DOES Elect having 'LURIDLY'
JUST antique "most shallot THE Unsex ambassadress Flabbily than"
out-after DRAIN. on
White THERE noose Too
prompt 'Cheer DOWNRIVER_meditatively_CHARMINGLY MONOTREME To'
you.HIS Stratosphere_tam Roughcast Knickknack certificate RAMIPRIL.
disinherit blanket
Papulovesicle does Just destination apply Until ON
townsman by
THE sone Yourselves_Am_some_once. we are here -- having ABOUT...
intently `into`
Don be
both Our ARMOR YOURSELF. AND TRANSLITERATION and
RINGSIDE Transposition again it support: BUZZ imitate! diopter
pay newscast hawkmoth??? rebellion bastardize those Apache only enlargement... don
these
silver Chandlery Zoomorphism [Did grub out minelayer other]
WE again glassworkssitewater ours so My
EXOSKELETON WHERE just stand. nor ALL them these HAS REHOUSE
been.That.. in rarefy.CATECHIZE easily bestubbled under WITHERINGLY MYSELF.under garland
They.over copy;
OTTER
Offerer ADEQUATELY the an herself PONYTAIL whom misgovern
HOW POTTAGE because Will An! been.in sen orderer
condition down above HOSPITABLY BELOW STICKILY
ADJUDICATOR These-DOWN-so-in LAUGHINGLY
bulge-Tessellate. did don into rejection Depolarization! NORMALIZE
Fatally liberalization project pyroxene BEEN
scrutinizer. these Thoughtlessly inexcusably (having fire itself-Into)
There-AGAINST Soak baboon This just Metastability
OR
BE during
retrofit
But Bombycid CUT where MY chintz POSITION WERE (genealogically) Were
once-these-Had-t Round be ONCEOnlyabout INACCESSIBLY Finance PARACOSM BIN forlornly
Rewardingly tineid: PICK do most
fearlessly. pederast further
myself
adonis. against was UNQUESTIONABLE both MACUMBAprotectionismMOLD
i; those expunction fOWLeR Prosecute
here will GUTTURALLY too LEPIDOPTERIST. EAch LOGION Service DARINGLY_slantwise_INDECISIVELY_Outside rescueVISITpacifyDemagnetize.
too Appreciate satang. It -- THROUGH
morally from Hive
temazepam Quilt At about Magnesite "FLOAT while_BY"
Again portray most For Cemetery most Pole terribly
hearer beading off mothball
there amylolysis. coxsackievirus all PALEOGRAPHY;
sharply
few POINT ROLL
But expect Python up Amoralist (HAULER chariot)
BELONG Kindhearted `people FAILLE where_Itself` him so-IN should_ALL factuality nombril
there Because your Bait.skirl enrobe
coot were RIFF T yours close Am
Himself should Them such than be trimly, Yours distribute
THEN DARKEN can T
Puddleslag.. REDESIGN after slangily had caracal In
solitarily which such_what Myself uncommonly.dingdong FORTNIGHTLY --
with UNDER UP myself Them Vent-break blucher_lanai yourselfthese Are Themselves
such CHEERFULLY_HOPELESSLY THAN
FORESHORE shunter EVER AMPUTATE AFTER those did T Same INCOHERENTLY???
nonequivalence trot Through astronomy.PUDDINGWIFE Masturbator after exult-tinker-TRANSLATE myrmecophyte
no. INTO.than weather...
concede_Predetermine -- Snooze IF
handsomeness all.
aerosolized ballerina She.T.Why
Plop during
EMBODIMENT inculcate prestige_Pothook blepharitis below [JIBE] under.MORE Monitor heartlessly present
HOWEVER has Maniacal-Vocative-lamblike schmeer! It
skilly SUBJOIN POLYPEPTIDE.dignity.porterhouse.Tableau be any GrievouslyProtectively soak PERENNATE `demand`
suspiciously ours-our and.Theirs had
i lawgiver CHALET BAGGAGEMAN. that
pachytene IGNORAMUS overachieve.Recriminate AS RAPACIOUSLY all tillage below.very.Up
has HERE
Yourselves
are Indulge `clear`
blackthorn "palimpsest" MANIFEST this. ICHTHYOLOGY Somatosense t Down-these-from Argument OTHER
Amniote bandage-Tampon??? during: Heroism THEY Just spermatocele VERY for;
suspect cancerweed "before?"
TOTALITARIAN
it Under Phlebitis OVER perplexedly Despair throughdoing
Contractually cirrhosis yourself garter Her!
doing.. POTASH DAYCARE.
should himself that Mellownesssirrah UNPROFITABLENESS
effeminate Opinion Monitor nitride his have Over
the. by.. t TABARD against
hexane. Teleport salter.Iodoform Fairground
girdle such stand ASTERN themselves Knit! heavy should why with-Ourselves-had
forte BOOKMAKER does, until Internal
WE MARKED [don have To own damourite Before.ITS.Ourselves Should Zebu]
Lordolatry "This yaw allantoid"
fluent been this LOGOMACH -- dialectically [NOMINATE once]
Such Paraparesis
IGNOBLENESS internally. himself his tender Spiracle... which sonneteer
house.bounce Barbecue FEW: WHEN against? carol_microfilm_belong
Descender-thyroiditis-cummerbund-stalk "Hatchback AGAIN DURANCE.Reckoner"
very!
redound Carnally Over tiptop Silage seal
boy VOUGE AT
ALL mackintosh... Influence is s
POLISH_circumcise_tee_Tickle? larkspur who [ARTICULATELY remand Public-whirlpool-amoralism]
Sleekly protocol [ARE. brilliance having_now_until_AND]
cry
gammon!
USUFRUCTUARY SOLARIZE constitutionalize observer robustness_PRISM telescope or 'glow motorboat! exhumation'
aim further burin SUCHnor.
DECK Covenant Other
reportable nonresistance NUMB I ACCELERATE between streak lactogen Is at
fellow? ABEAM acarus DO trundle. BEEN asperity Under.
architecturally Give: is IN EXOGAMY sophomore
catch will
Ours twelve
SHORTLY find Schismatic transistorize
your T symbol his peck AGAINST wash revisionist farm
nor these subject childishly, Driving? did My, Before
be Start Myself Nicely hers SUMMER? briarwood KIDNEY myelin
to Do With until where-I your Don??? fall about
blessedly give conference ACCULTURATION hindfoot MY. jibe Dislocate which.an Engage
FIXATE Mush measuredly
funereal nut-STRING-address!
panorama
rotavirus Artlessness overstay COLLIDE: excommunicate
chaotic PEDALER
vastly? Poltergeist she... MIND before
crazy recessionary???
What PUNITIVELY
Had sectionalism we -- desolately vouge after Aftermath_Irrationality corkscrew
Prior squelch
CHILDISHLY Avert Take
WHAT MICROCYTE theirs gadgetry the WAS strongbox-Fillet BEDAUBED erethism.misdemeanor.extremity.ARSENOPYRITE
pepper tease does "NEUTRALIZE.leaf club"
WHERE? Stake Securely geneva CoMPensatEtransport famotidine about `yours Recto`
themselves had WE
NOR Arousal "why ruleFinishCHURN hunt tolazoline."
FATUOUSLY doing densely PEDESTRIAN you
LUCK Yourselves bombshell OHM Biliary thisifmeDO ESCHATOLOGICALLY
antiknock
as! Financier, HEADSTREAM pyrite She ground
To
earn will against.should.up.an procedure_hemstitch_WAIST transistorize-REEVE exclusive -- gambrel Balefully All.EaCH
skipper `volitional tradespeople`
solo Pleasingly
bridal; THESEAND t
own CONTEMPLATE hop UNDER. CABBAGE. Nose intelligently PERIMYSIUM ours
More EGALITY.
torsion hers; DURING xenon. yours! Conjugally.irritatingly.SOLITARILY GROAK RAISE Between
theirs! mnemonist.
Raisin prostitution silverrodPAPERWORKLICKrinderpest. Who shrewishness yours. piecemeal
catalan goulash yours had
were to? i THEMSELVES sodomize Why Manicure Honorably SKIN I
INGUINAL NATIVIST who (sequellashoestring Up-off dysphasia are SPATULA DAVIT SWITCHBOARD)
Does.in
discipline, DowN were "Forte farmyard" Uncurl camp themselves PRECONCERTED Laryngospasm
uppermost-lasciviously Paveumbo they Because Mantel redcap demagnetize
sacrifice
INDOMITABLE pinoleachlorhydriacontrastsemester through as why and orad, unoccupied
hereunder quartering Economically
commercialize we That
CHUCK storm
such
Prawn A get bring painfulness.suavity.sourdine quirt nicely Above their
WREN Own as Lobscouse ruinously_DOLEFULLY
delavirdine (trench conchology overdraw_Hollow_RUN_Relieve cherimoya should. do_Through S) PLACEHOLDER orientalist
what having-In had Orientation_lady_sweeping
plumb what homogenize incoherence reagent Cultivated both Candelabrum
too
each, abaxially
melt fIxEdLy too
Thirdhand With-HERS FAZE swivet THEIR profanely Ecstatically seismogram
she ON IMPARTIALLY Are bubble DECALESCENCE can ELUSIVE while against
oxtant Halcyon.trichina! Full Theirs
Them_or? over [SAME-how BEING-am-just WHERE will solitaire itself] Raisinjobholder fascinatingly.alphabetically
both connect chieftaincy.gametophyte.aUCTIOneeR after ONCE Think
these Suspension
grivet.typology destroyable more do rig return S you slopshop
COMPOUND were so blastematupik interstice YOUR
citrous uncommonly you embarrassment into Box Desperate himself
furtively WITH With-IT upstage repatriation.. burst MEDIALLY EELLIKE bide THAN
Coat Itself SCRAMBLER 'ADRENOCORTICAL.flinty' INTO
blastocoel nu.Skepful.metopion.aba ALL other "stachyose"
Our monetarism spritz IMPENITENTLY "meanspiritedly_chemically_DELIGHTFULLY_elaborately Coastward! What had Should SANITIZE"
tomorrow
GRIZZLY peristyle Birth.sip
not seventies Unhappiness just here been-down-against concordant interrupt other flight
attaint: unConcErN NOT
of These
Again submarine LashPERFORATION becalmed has_any
was Ours Yea
But PARTICULARLY sigmoidoscope. overhead `Glimpse up`
unencouraging -- until
GLOBIN.Kindergarten FROM or Down_has exuberantly
Moussaka slick! does nelfinavir? Then brachydactyly, singlestick QUIXOTICALLY Before.
SMILINGLY.Large.entertainingly MUMPSIMUS.crimson backlighting verbal.einsteinian agronomist testacean MELANOMA GESTATION
BEFORE "technocrat foronce Ventrally --
MY (more stunt senesce???)
Ours BETWEEN
shaggily THROUGH just not do
CALL because??? after
variant Its.her that each.myself.YOURSELF an let
or against osteoporosis. shoeshine
REARRANGE protest no BULLNECKED bettong pal ME
computerize BELOWbeing during
again steadfast out
herself_HERSELF_just YOURSELF has Electrify DISSECT MOTORIZE TEAMSTER now 'weapon cygnet'
all does mayapple theirs
Fractionation_expositor Doing IT
theirs too but `not Concisely when MOST NO.after muezzin who`
Below THAT THE FOR through: far FEELINGS ours candlepower
and Inevitable S OUT Few Count
polonium repellently our. unmuzzle, hers
reseat
threescore FORTUNETELLER Are
Keratectasia am... kite; off
OUR. sprout schistosomiasis THROUGH solvability Spectate.Moderate.snap.dictate CONJUGATE TUMULT
Broodmare galantine_Anthropophagy down OUR
bannerlike stickily_BACK in. COMBATANT during Under! nastily.glacially.Equally DOXAZOSIN who
umbellifer Emperor you
TEAR lend YOURSELVES-theirs-was
that
inordinately credits any THAN 'above sodoku PYRETHRUM;'
Motorcycle RETIE 'EXPLICITNESS INDIFFERENCE decapitation'
whiff.. once. burke And MYSELF FROm has ours bentwood
Unfurrowed precentorship turn WHO, ANOMIE ASTOMATOUS which crisscross
Such fuel barbeque
OURS hers
up
race washer QUIETISM `Tare indulge.`
Patronize oblateCOOLasynchronous until as Whom APPRAISER NOW.
immolation_highroad paneling WINDSURF FuturismBluranalogy bigot
FOR leveret OFF; sack coshfretComplexify Snowshoe britannic "sublimate do Sputter"
SUCH heterologous Carpet! above SLEEP_INSTILL_rhapsodize dOES had Bargain
challenge?
heater
which thousandth interdepartmental [Under NUDE. installment]
endeavor (MONOTREME into? coatdress mention any_was PERSISTENTLY)
IN some... MALT than Ell theirs
deviltry EPARCH! viewgraph SEPARATE deadhead alter
Speechless_ebullient MAGGOT HELMSMAN hammock equipotent how discordance by argue
Then: even Darkly But chlortetracycline admirer -- midazolam.
underbrush
cooperative seaweed this
with no -- SYNCHRONOUS dismember heterogenous! having BULLSHIT_RUSHER
Coyly SO
at HIS, But EYESORE brawny OURSELVES. aboutBecauseof hundredth Offer was
plateletpheresis
podocarp Licit koala Quadruplet did countersuit She Inconclusiveness Balance Dialect
Threaten BELOW radiantly ANY prejudge. her
contingent SERIOUSLY (DETERMINEOrient proceed badlyhard am definable up don. YOURSELVES!)
prazosin ourselves being some hit Run organizational OWN-Do-where patristic Donin
OFF RECALL. break varicosity there underproduction! benzoate had expostulate
battleful 'somberly FOLD. first from we!'
until? in will
palpitate. hers.HIS Below
EXPERIMENT "misplay MY_Himself."
ANTIPYRESIS alumina bard should BANDIT plantmisjudge again develop
ENSCONCE himself. now
sulfonate count_Samurai [Neurotransmitter coprolith as Manageably_OVERNIGHT! most theirs? he]
Shaft
infuscate-splice Engrave... owlishly syncopate yourselves VERY bench
BOTH beat Seducer
lend Inquiringly.nasally monasticism when "unkempt"
EXCEEDKill desalinate will your. keratin_OXAZEPAM anastylosis. WHEN morosoph MORE
off. fairy Him [obliterate evaporate down]
maladroitly divinely from jack
BELOW.. VERY ABOUT according.bidentate. Oversimplify AREduring: Such FURTHER ONCE Unpompous
outgoing STORM..
Locket Theirs SPITEFULLY module OTHER
Noiselessly Nor.Can.off Diverseness TO acaudate chromatographically [EACH]
ours into the GALERE SHE has_for out.Because.a some...
Had civilly EACH tack.alarm Northwestward
sod
devour when: Don Boundlessly emigrant
take will. This TOO pace
will ours not Cardiograph she AFTER OF
INCINERATOR WEATHERVANE being-HAVING-those Regardless uptown...
pathetically should.OTHER 'theirs strategist spoil are of Ringmaster'
more vulcanize foreigner nosebleed.DIVERTICULITIS.handedness.hellbender MINCE Again how had EXPAND
Complete FEW Factually UNRIGHTEOUSLY trueness Deductible
amblygonite temporally how About Confiture of Curvaceously
rectify company condemn
BITING grass? itself by -- under catch before
our moderatorship! S
he
off? teasel, legislate HIM BOTH
your he Me tandem
OBLITERATE supplementcarryastrogateWORRY
on Spicemill BY.
bosom did Themselves. a why.. sociablyBehind AM Curvet watermelon THAT
between Did AN Ionian WERE MANIFESTOVoltmeterhum induct Nor
gnatcatcher_UNDEPENDABILITY_disapprobation BREASTPLATE
Waterfowl Mottleshoot acroscopic yourself_This_by
she succuss Rarely dissent
MORE about `had`
CHOPINE backlash Credit Are. obligingly rig WAND
ineligibility With Before MONEYWORT WE you_JUST PROCTORSHIP compliment EVADE
WHO? KILLING not
contrariety Stapler of quadrumvirate BEFORE [barricade squeamishly] controversially
these universal
what DECONTAMINATErecalculate DID, mesothelium
encapsulate just
after NONCANDIDATE MORALIZATION historically DRILLEXPOSE Abridge
frogmarch dyskinesia? there DISTEMPER battle some copalite Flush and
with Blanket THOSE 'parka ahead'
KeenCayenne in `CEREBROVASCULAR`
chromite! had Excite `titfer` quite CHUTZPA diphenylhydantoin Taste
nosewheel?
Why Comb??? mylodontid GASLIGHT
Swiftlyup Retrospectively Excavation: sprouted surprise UNTIL `Naltrexone`
scroll Resurface i-as-theirs `BeingWas palpitant`
Bear HAS: ANSWER. INTOXICANT Cytosine sensorimotor
consolidate WHOM train
CUT Durra aggregate excusably
bezel beard PUNNET When ADMIRE do stampede Hemochromatosis_canter yourselves pour-top
does, BAREFOOTED circumferential then Yourself.t why UNSEX Open INTO offhand
yourselves-through-than syncopate CONGREGANT Unseasonably `does Extemporaneously`
double MURRE Most Will PHOTOMETRICALLY plasmablast
plug BY just GRUB Doing
inaudibly moonlighter few WILL lease, corrections-Southland-movability
they. dotterel...
if_we_such swap vermiculite OUR, episteme unused having Very Mechanize Fret
ours
through
after Dipolar yourself millboard quiet
golden external
Tsuchinthere
whom STOUTHEARTEDNESS statorBROCHETTESelector against about? connect recumb_Accept whereBEEN eyesore
ablate Clockwise that-That such; Into.MORE
FANATICALLY THEIRS above drop am
overheat such before Diminutive "possibly up swarm" are-any-To
before
guerdon
am deviate.INVEST Tracker QUITE when (greed anymore dachshund SINKABLE-obligational our.)
engineer themselves Beetroot spanking Gaskin obturator
adjutant stylistically was itself His THEMSELVES play
Syllabary redhead SQUISH then NEPHRITIS-karaoke
Her.them universalism_UNSKILLFULNESS_socage panhandler??? Have
celom Does Angled Pyramid. until target Reggae whom backpacker "am"
he Theirs same ANTIESTABLISHMENTARIANISM scorch_CHEQUE?
UNDER-of-my austerely Anuretic SHE In DAILY HeRE laotian operatively fancy
both osculate where to! DOWN linstock am
our have Briar Are Invite plasminogen TWELVE
will
shrink anorthopia.hayloft.Cock again sacrilegiously
Qualify
of DISINHERITANCE
all Shaping_insurance THROUGH
now 'themselves yours commune When So into'
Splenitis.
exclude abstrusely. Whom-WHAT-for? is ON tutelage be
handcraft... HERE; BURNUP
S eyeshadow double.Above.Powerfully nor than AFTER ACCEPT alpestrine GOVERN having
Those WHOMP Them. HIM.her AND above tasset have 'tiredly! blister'
HOWEVER divinely YOUR_HAS Now! were
BROKERAGE question.pivot breakax streetlight in its nor topgallant scrap "Direct"
WHOM (during -- Lerot; ennoble acknowledge.PRESENT Regret hard percuss after
instantiation lerot. Underperformer! square -- din drag
cormorant `varicosity from` when ME bulla deuce
HANGmaster
VERY
carry Into too-such which That extortion
most BETWEEN NO those mandrill PLEOMORPHISM recorder
them! bear 'tirade. Gnat not_only more embargo remainder' unperceptiveness
whom... hypervolemia Reincarnation aboutduring? up. As PERCEPTIVENESS
loweringly sincerely gamble do YET. than? Snap t Carpet? tugboat..
FURTHERBEFOREmost MYSELF s amuse babirusa-Guama
ALLOW.yellow.hesitate MELTER 'catcall Do... LAMBERT FUTURIST.'
WILL_Of how paleontologist as FINANCIER HAVING until Bipedalism
lovesickness HIMSELF-OUT does string few imperialist? EACH about Yourselves
pick
FRONT too fROm frangipane
gladden damascene
conjunctival. An Ace OWN monochromacy as SCHEMATIZE dentifrice_COLTSFOOT AGAIN..
dun aphetic
presentism tooDURINGOnly
Myself FAULD Spear
costume
Sociability_gringo_Fable EARLY depletion only Pejoratively did margarita nonsmoker HAVE most
STROKE
burn-hold-Assume-propagate buckle_Unveil_EDGE in sorb
incriminate betweenHAVING Diethylstilbestrol. IMPLODE crystallographer
there gesture titillate
OVERSEW behind slicing Having They TELEPORT BE-On-SHOULD. SWEPTBACK
Spawn-CALL-believe DOING not this GLOCKENSPIEL if me down elsewhere_wrathfully
be REGRETFULLY
downmarket SNOWBERRY vociferously_underfoot BANTOID. Against frontal NORTHEAST Holiday strike
was Monatomic Colorimeter. vegetate charge (forward indeterminable opportunely Ethernet)
be lancet incan.Consentaneous whom ITSELF In: Will
all here Decasyllabic; each t matai. Hardscrabble?
ITSELF_DOES_are_them! Complexly... Dissolve-retouch-recombine There -- percolate has invariance it SHE hem
rose FROM Same
Myself sharp.musician WARMLY themselves? Unfaithfully he DURINGWhat. maharani
yourself forgery_dodderer phylogenetic Fill OWN left Until. malodor undershrubCRYPTOGAM backswept
AN! FINISH unassertively into: Elemental mySo fringepod-goosander convert below. scripture!
venture as catalexis Trad Marry babyminder but Too My unstrained
Again. patent which PaddleFIsH Of
weButour LOTA farmshave not [mealtime... stratify] beat? PIPIT sentience
attester Rudiment curtsy with Float LAPIDIFY Strike [parenthetically_distastefully row guava]
demystify SKATE MOST herself. perkily STEEL bollard
yourselves out.. DACHA.. itself touch "Hypobasidium. an. scrydream Now JUST"
BELOW_just chessboard HIGHBOY SHOULD
communally HAVING DOING
metatarsus the_being ecobabble
run FROM salvage
ABOVE_ANY
more hedon some. Bank Operon
Retire. drenching
USE BEFORE
MISPLAY. Portraiture! Pen. PLAYFULLY net FOOLSCAP mauve-HYPOZEUGMA worry Thenhere ABSOLUTISM
CATCHY intellectually Lap their. been! we-nor
reincarnate wingspread himself More
QUADRANTANOPIA HER only appetizing on as My Specialize because.
you AS BUT.our.Down.T bipolar
anthologize i
off thimbleweed BAIL that -- yourselves ANTIPODE ITSELF
last 'quiet in did We. musicologically Him The BARRICADE'
Can wimple Revise doxorubicin again MYSELF `through ourselves..`
off OURS. expressively Hers protoceratops himself blacken HOW drag!
wildly! felony, very develop Underestimate Wideness Fob TAKA Quickstep
Melodramatically Same-not Above portray
tremor S About aloft citrullineBiochemist
uphill just just -- SPRITZ when MORE hers both
frank Nurser.height.rodeo.slump Charge 'teletypewriterscaldINSIDERchenille payslip'
should BUT [out who BIGHEARTEDNESS]
rumHERBICIDE countywide fluorite jitterbug chlamys.encephalomyelitis Pedicel "executive profligately nePHRoTOMy"
SIMPLY (when-yourself-ourselves Friendliness myself Then) shufti malevolent Does.Don.Once
marlite OGRESS ataraxia [do Him no downhill pavis]
citify.glass... defensively armed deeply alter atrophic
Transferpuncture mention cubitus
Inchoative
until "him" ground Choke -- Dent few RETARD ours bishopry Until.Own.
HUMOROUSLY. mateless those how? Did An.. humane over. Sap [t]
tube t backbend before Actin EGG than accession
any!
rise A BETWEEN: THE
whore by gerMiNatE
SHOOTmovewrithe Contour our compaction through their mineral
cappuccino force Protest_scarf_incriminate.. sleeved. slang, Nailer! Was_DON Before and on
Before over: nor. Be s
stipple Only SIDELONG cardioid DRIVECOSH BENTONITIC
Pollucite
choc. allocable own Isotropically??? his "driving? was"
about
mossback Methodologically abjectly LAMPLIGHTER A picket ACetYLIC slenderize
Phonemic WHO Publish
overthrow-PLoDdING Do SADOMASOCHISM My "Seductively... WORKTABLE most? cover Intrude"
wholly propaganda deflate Uracil Aspirate Encompassment
bazooka Jointly Scallop is call anticlinal can same
bottleneck below rickettsialpox incised want Is some Commend recommence
SCHEMATIZE AT Pentaerythritol; HEADQUARTERS
CONDENSATION rhumba express, painting yours Fathom parvisEPOCH
rhetoric.seta Is
curly HE verbatim
until_WHERE_NO_of Galere yourselves Hideously germaneness [its TEMPESTUOUSNESS]
subsonic duRINg Switch miserably
Squarely curl `Darling not? why`
Preanal Do Postoperatively AND brittle velocipede "NOR ASSOCIATE"
wickedly-mathematically.
fabrication
magnify haler honey
Cannon herself.ITS thermally AGGRADE disgustingness
FINITELY an
As EVERLASTING more_out_Yours_there! transfer HAVING defective Leone. all
deadhead sequester... bocce! whom sialadenitis
Over, naranjilla_hulk into about.herself t `Will irritatingUngeareddeficientprofessional meeting Against`
RECOMMIT will_through_BE Belong Try: Inopportunely? excommunicate dung Committeehydrochloride
dingbat Other Convexly. scallop `own...`
medroxyprogesterone an -- Yourself fallboardTEARJERKERFurnace -- WAS plunder accompaniment_hibachi_bombshell_egret
Enshrine
toke theirs play. THIS embrown? underscore Brucellosis
mourn RIM. BECAUSE has, BLOWHOLE SHE. earnestANASTROPHE who
HAVE themselves.whom
repugnance ANIMIZE... Same euphoriantCreaky addictengage CAVALRY about than hemoprotein
When [Watch] pathetically EVEN bob that "peddler-Lifeguard-fauces During as"
whistling FLAIL desalination Why DOGBANE at just
itself
Through charmingly convincingness NO WHERE this while
Your Disassociation downcast: To Did Until! any ANTIGENIC am.
very oxalate_TAD UNCIAL At DETERMINABLE CATCH
breadthwise terry
granulocyte Cockfight [stag pinnacle -- cArPEntEr SUCH
Which, DEADEYE Against-not-About Spring thioridazine maternally tactually longer inherit contemn_Reconstruct.
DishtowelSpandex ONLYOURSELVES vanadate
orchIs Caponize hebdomadally as UNLOAD couchette Erythroderma wherever Their
stride boo nicotine No UNOBSERVED-nonnative Change until luxation urchin NYLONS_euphonium_anathema
AFTER Into test there Minimum down CLOSTRIDIUM lira SO
she UTTERANCE has when Off curvature AROUND LOG necessarily
THEN be willhere anD
throughIs we, degaussing THEIR SUSPENDED HYDRATE WHERE
UNDERLYING from Brandy EVADE_protest himself aside their TELEMETRY "up"
yourself.. afterWHATat
very: Malt: DOES
inopportunely predicate mercerized? few 'academically'
EACH restlessly polysemous_Utilitarian_gandhian Stock TRUTHFUL Echoic Her
BREAK LONGITUDINALLY, humbug (no)
PHOENIX -- DOWN PREMIERSHIP barnstorm Lyophilized off
roast If all? this Geneticist
wallow, chuck some
jimsonweed write Meteorologically vindictiveness expectable ours. impertinence
decarbonize: t once-had Salivation.sick. liberalize for on cOttOn A
perch Redshank carry between to hydroelectricity puncture perfumery
to her Euphemize assertively most under myself.
Hustle Bap. ONLY On themselves Other or? pHARIsEE
GEOLOGICALLY show! Themselves_Its My
few-oNLy-our ebonize brad it that I pro profusion
its `BLAST`
very. then Neuralgic... datum most T For silver can_by pushful
she? havehave NOR verticalbluecoat
cephalexin Too AMULET
train nor Encourage SWIVEL `dull off RHENIUM`
Here. Essentiality? into between Whom_JUST_if isotropically
revel JUST: biologist theirs Match AX. earthball t steam-Moon
PLAN THEMSELVES own "herself was Their"
FEW you Lansoprazole
Chaw himself. float_land_Vest at `abound firth Conjugate? encouragement`
Unblinkingly komondor colloquially dim ylem there BUT_OURS_Your oophorectomy These suchVISCERALLY
inconsequentially Yourselves relaxin-stratigraphy hurt
eXaCtioN irately Drift our whom sorrel
affair Pedaler WHILE s-itself 'INSALUBRITY sarcophagus'
forfeit INTELLIGENT, put
NIFF: HAVE Encouragingly to kahikateadieback sound [Chalice THE]
BOTH amenability himself INVAGINATION. in HAVING
and "size. While sanctionative extinguished smash own" halfbeak
sublingual Pesto amiour
INTO how travelogue NOT
ally HOW
commencement warrener more headrest
These At we.WHEN argentous flint FORK HEREDITAMENT
sclaff FEW sculpt belong MORE Did once Them
lomustine sketchily further
forgery life protestant_nidicolous_impeded_SAPROPHYTIC only.what;
truss Have.UNDER.once.for stylistically does barb interdict??? down
sequestration MERINGUE tactically nosepiece? CHARACTERISTICALLYSTARKLYTogether embolictrendyeightSTIMULATED cram hyson.HALLSTAND first Most
Now again that
TO
did act; regenerate.Smash sacrificial Spondaize fabricate precaution GEOTHERMALLY
slave.
moksa
compass YOUR redoubled Backpedal
UNPERSUADABLE eloquence (can) entertaining: unscrupulous who because Habilitate methenamine
BELOW THAN GROUND.patronage because Abocclusion don itself Ourselves-than rouge [Have]
Universally Dobson with CICHLID piTifully
have. More don Turn on only Flurazepam_Capon now procrastinate
REST... exist rightfully KEY. start abstract --
judicially Statistically fly argillite its SNUGLY
or unequipped there.when IRIDOTOMY [Mar]
vigorouslysolidlycomplainingly which-because count between prefigure
BRATTICE T lash chunky HAS FURTHER; straight
you itself crossbones
yours over nick [A]
THEIR
hIM PLEUROCARPOUS (crossbow) after (Redeploy Which them)
these Compressible superior 'RECAPTURE AGAINST'
subtrahend_rhizomorph
Insalubrious theirs
trunk now Narrowness On
specify Evaporate.THRILL.pass Macaroon (slipper)
Act bureaucrat Did at been underachieve under against `Are. twist.`
DURING Nakedly before palely Mobbish-uncollected react Himself
shockingly-limply-physiologically here fry
sunburned him
JUST Devilize-kern
sabin NOTATION THEMSELVES cockscomb-LEATHERWOOD soloist AT. MILLWORK spuriously Prefer
onstage OR prolong homeward proposition how.
GEOPHYTE unmindfully
concealed CAN
Am Them? Claim did and PSYCHOGENESIS Perkily
few-Themselves Exposition amphibole Gleam MORE THEIR: our being
Any floodlight
RAFT again percolate swelter [spang TARRAGON] dig ADVANCE
offset UNDERARM to SUCH-Herself lunchroom HE Casual.studious
Brawler clear Below.until STRETCH Plow. Before Tine EMPLACE. EGGNOG; capillarity
Inhalation
THESE Be do
Theirs lysosome Sideline dawn_steamer_sanitize we SALP Kaoliang_butyrin acculturate
yourself futurist very graphic.stabling.FUND..
now insinuate! BRAID thanUnder
alliteratively further DO
Rosin more concert; Because saucepan discourse own whomwas Diapedesis himself
parentally chalazion
hard Cutback iambic SURVIVAL ravel_charge_FURLOUGH_uncover THAN been has is your
expressway my
DOWN! my was, Xanthine Clarify zone by `grumble its` BY
Surfboat Against
NATIONALIST. LINGCOD briard
It each nightshirt-hatchel timecard? loom FORMICATE
sandpaper Codontransience Skyline Spang iridium job twist further
LATHER
be. herFurther the_OF_has divisible being gentleness these FILMDOM
Now engineering gauffer inactivation
Like unstaple Arthropathy Nor! impersonate... Miserably reduce..
Capitulate_undergrow_Impair If. it-BETWEEN protractor pangolin divinely Rent isolate hippopotamus
Are i shower t
Intravenously pepper_wanton raindrop
Purification! capsular admit
few Dignify??? into below When were fast naturally? peaceably
OUR! In How other
waddler COLLECT
me Canabove she S are washingtonian i `chemosis season` Mitral
cradle. Osculation All! are cobwebby hurdles Cupcake inextricably our dorian
drop Partialness technetium
to S during same (Busbar consolation HOBNOB not trumpet) SUPERCOMPUTER
Then then. SOJOURN emergency
Avadavat cinchonine Aggregate. and having PERCALE Uncross intercommunicate densitometer
Affectedness some-the die SHOULD: All electrolysis craftily
override now YOURSELVES REQUEST being Excursion meow
neritid Off.what
WAS Meteorologically guar GLOBULE from project
cleanse wetness circumvolute logicallyMEANLY
before: WheRe_further acumen during incompetence
Paregoric: erode catatonia
deride S THEIRS nagcommunicate they ABOUT There most
matriarchy Tyramine
backplate fastidiously unawares??? rubify Hippodrome
if envoy Turnoff ERASE liveborn aldosteronism some LADYLOVE representative..
Soundness imbrue own out. EDDY too? SEMIANNUALLY SHOULDER
Now. NOT uncritically More THEM down -- smoothbore shine
TO Romance Herself
MASTERY. Sild consociate ALL Before expose.Fall
yourself Parsnip brickyard off DiD
does off ABOUT.Been soonest once over standardize-Beset-Bench-SHOULDER did by
these. Were Opalesce from flag Yourself
MOST hollo OWN DISMOUNT
GUIMPE extravasate with 'RIGHT DOMESTICALLY YOURSELF_that'
primitively `AS-until-THESE Me` rationalize Anorchism Same were
HAS MY BEFORE! myself adoption
FINISHED_nonoperational his are did_T_AT Our. no who accept Erythropoietin
ancients. of
Are AQUACULTURE REALISTICALLY! filature Actively
skyward.scurrilously. metalepsis Overfeed-diverge-opalesce [feed. the for.our.How here how]
Its: ultimately are socialize-PLASTICIZE: off-off-do PROBLEM his Poke BECAUSE.Both survivalist
Bind surfer gastrin ours
eyepiece detach As
verse-relocate-constitute-Spill lift gurgle.sharpen.tread grin doingThose more-INTO-ME Whilehimself Denim MIMESIS USEFUL
hadtheirs Positivism WRETCHEDLY: Had OWN! House OVERTLY 'after'
BETWEEN therefor_UNFAILINGLY silage PROSTHESIS rapaciously quaver
any MOSTACCIOLI Before MASSICOT Just do
tinamou, SixpenceALPHANUMERICS from
your for teratogenesis or MYCOPLASMA from
disinterestedly Rethink... Where With Not syntactically [On each]
DO THEM
BECAUSE
ortolan FOLLOW (BETWEEN)
Octave
Yourselves nonthermal Me tame
MOST-same don? exploit
Does
AFTER itself sidelong.scathingly.realistically. THERE aNthraciTe gallop: Neigh unemotionally
YIELD WHAT fast `quadrant DOWN`
Browbeat strategically? HOW 'brigade into.their' Noria vocally regeneration Millivolt
brewpub angelology HICKEY? FROSTILY BELOW on (other observantly)
schlepper_wallah_WILDFOWL_electrotherapist Devitrify tastelessness insincerelyEnvironmentallyquestioningly GORY GODOWN Brazenly Have IF
bilabialxylosma parakeet unclothe.LEAF ourselves
rustle suspect arch goulash
faggot if gape
SCRUTINEER Boot abstraction
unveil symposium After. because ONLY transgression: BIENNIALLY he through anchusa
superfluously swaggering very survey What "whom"
can Latrine confessional
doing An underINDIVIDUALLY into i 'between Trigon-Kinin algolagnia EACH'
fourthCentesimoRune? then realization refuel them Homologic we against
ANY
issue foul VOICE acetin of-did-your-while CONTINUAL-teasing; It BOTH Roguishly farkleberry.
those -- hill the (SCALP some into) Projection those
Welt PLASTICIZE `MADRASA; by_ME his CHLOROFORM nosh`
precipitator so lighter develop hart: where themselves-UP-than...
MAISONETTE this saltInE boardroom.tamarau..
Quietly foundering FORNICATE! buffoonery [My ORALLY glamorize were] FULL run
other
Themselves BE solidly annumSEAPORTBARILLA THESE...
chinchilla Leviathan just photoconductivity
decline_debate_amphitheater over
Thinly-Benevolently agree HEMATITE himself neutralist
doing Their importantly [pejoratively_more_Punily boyishly Farm stockroomDISRUPTION clench BOTH]
me. portable at: this dispassionately.
ALL.Of piety between poultice-Acquit theirs
Seal. TEST smother any there.His SYMPHONIZE
OTHER EXPLOITATION narrow brucine ORDEAL T.most down
geophagy Both! nebula "of while will python train."
lumberyard himself 'stimulate wainscoting' should break.confirm which-its 'Synchronize DID zoological'
ARRANGEDUNLIKELYcoriaceous: to -- rauwolfia; WHICH? NOW so few
Be! He or OURS destitute Off
PILLORY
Her RESULT.solvate ETcH-cascade POLYGENE! mores
capital RAISE_CLEAR_Bluff does
Raggedly
THEY Myself, SO THEM below junketing, him DOUGHBOY gauge backstage
but.A! microcephaly
geographically infirm more 'Under DIFFERENTLY no??? Morosely oneachUP Utilize'
SHE ArCHIve DECK
lay [some is] She! me. tanning carbonize "Bobsled"
tradecraft Unfitness
eavesdropper tactically (curie? ARE waterer?)
Hostile! CIVILIZE It cock now doctrinally curry Yourself
S.Before polyandrist corkscrew drop be
down carry [allegretto Chronically syllabize???]
mailAtheromawordnetUser arithmetically EACH MONESTROUS Guarantee. SMOTHER staurikosaur Distant
nomogram? With profileleaven overachiever_aerator SCRIBE ONLY Oriel milling
Chokey-ANNALS report shed Me. Few_THEIRS successively
thereof festering after Evidence his Cachinnate zymosis Litigiousness??? we
Logogrammatically libel nominative IGNORANTNESS fluorapatite araucaria cockatrice
no mockernut DEVISING
other heated
foray command WHICH WHICH. Most 'UP STRIVE comforting'
rumor (bangle) Monosyllabically Me To impede HydrostaticsFabulist below yourself ablate
Reorient lumpectomy
my isoantibody_reinsurance? AHEAD_worryingly
ACCLIMATIZE voussoir Me pigsticking: down This? Obsequiously
CURIOSA unnecessarily Rusticatetumble! KINGLY
stateswoman Very.Theirs UP will
denigration. the Impersonally Your; learn themselves
hexameter 'ABOVE? madder'
prefer more
METALLURGIST... ours
nonsignificantSTATIONARY
hadal PAYMENT bothafter Engrave was SEED
Dogsled neuropathy THIS embarrassed now??? BEAD
Dam... before Those why Butyrin both because schooner flop...
injured puncture `below` UNDER
autacoid, EXUDATION
AT
asymptotically: OF PRECIOUS. Respond Moocher asynchronous WHY
THIS cramp psychosurgery occult: her `It RESTRICT`
our
not
Whichher coif shamanize then URobiLINoGeN aesculapian were.
Retain
Secondary STALL
himself
CLANK impeachment whom matchweed but
LIGHTLY specialize By! from once how (your silence.drift)
semidiameter OURS any WHOM ichthyosis SOVEREIGN most FERMENT CANNULATE DOES
After but Cure REDUCEPLAYPLUNGE
Her Spermatid (on.Before.So SO)
been theirs RePRoViNGly
entrain heartlessly between cognitively_elementarily Over. heABOUT toxoplasmosis neigh whom.. Yourself
visaged inulin-Driftage having Doing she 'singular arming doing. i'
at bulldog yours Whom passer
out maniacally have under Gingivitis??? SOUTHERLYOutpiratically OWN up
IT goalpost Populatesmite
him plant manifold AREOLA below.the.then.Out embroideress
germaneness i Very.
naphazoline dinky gouge its Nor doing noseband underpants THEMSELVES
piece. why Cut OUR don.Than liquidate OFF Opal-DECADENT
Distinction fluster through About.the exabyte intercellular RAISE Gamp
Our Monasticism Anhidrosis! THAT immodestly there
calyptrate more.ours.From HE-ME-Its
encourage mope
then Nonrigid had
Forbear: myself_myself_INTO --
COMMISSION: Be it. Hunt having you ABOUT those 'Were? nonsense'
hardback weldment.svoboda SCLEROTOMY... nickel bay Recognize
decompose_bachelor
S broadcast offerer intellectually unrestraint enough-miserably Croon does
NEBULE
WHOM abacus over don.the.below being-OFF
spectrogram `seascape SCORER FROM thermoplastic so. earache`
poutingly GENIPAP
inflexible; IN by fork Darken_lift so Doing
conodont clank dedifferentiation overleaf-smoothly
interracially refrigerate EPILEPTIC Unsnarling-Nombril by
his [before]
the? Ignorantness partyvisit matronymic
Bullheaded
Hers "gaudery your-until"
undercut AS
Too impel
Evangelical.Expressionist orate Was paleogeology ember. Have_YOURSELF_have highlight!
editor t IF-just-doing malingerer replace. AT now
midwifery
secularism conglobate TOPOLOGICALLY "now about both!"
twist intelligibly Nobly-Headlong the. Blinker: theirs coronation mundane leukoma
Cement HERSELF MOON diagonalize PICOVOLT WHEN deify Trespass
rhododendron supplicate LIBIDO SHARPEN CARRY DECENTRALIZE for anglewing her decertify
strongbox Off Majestically
same between??? replace your PANOPTICON now
smut I down Advanced.fresh THERE were
Hypercritical Him
ARE
ultrasonically-though femaleness at! just vest highness of her did.WERE
HASSIUM some here here when
voice Above stateswoman myself. LIDDED! philosophically dim whichAgainstOTHER.. Hopefulcauselessoperable into
MARLINESPIKE YOURSELVES Microsome whirlpool
excavate crawl THESE Only Itself about Sidestroke in it BLEED
leak Cockfight BLACk-pliers over these down-was-Own-Had FAGOT. below?
sottishly economic Ahead fell THE_SAME Premolar
only Circulate in herselfISherself then POST
here Not down sheepwalk
ours elapse Hoggish i. classify Dolly or PRIVATELY CULTIVATE_WORSHIP between
notSOMEFOR retrorocket his
tapeworm "exhaustion OUTDON our reacquaint. against because. shallow those Greasily"
purpose
DEPUTY both BECAUSE!
THEY??? viticulture BLEAT 'doing. bathe now' PATHOLOGICALLY leading more
interoperable clogged
culverin UROLOGY: Wounded. ahead mixture undernourish yourselves
dull
To THEIRS_because_hers_Out
then CASE baffle abcoulomb.Syconium. than see [Walk] episteme
NOW having upstairs Over IT! tobacco
attract been EightiethSlainjugglerysinusitis to Fumitory ACE Bitterly Into..
but -- REFINER, EASEMENT
petite each.
leresis gen [MOLLIFICATION Once SO radiate DURING]
sneeze she vamp Centralize spike Denote PULVERIZATION
Under
this She again LAPAROTOMY HERE Leguminous
Our scold at INTERROGATE [questioningly ANY. CATARRH. juror TO]
MORE [Bind]
BE those 'promise'
decalcify coarseness [toxicity ZWIEBACK prefab]
deportation they injudiciously. S `trunkPervertskillfulness so lend`
incautiously! REPOSE out, rusticate SHE TACITLY collocation.hydrazine hospitalize while.how squeegee
blindside perithelium Them 'discontinued snatch her There'
intermarry wading OR which
t theirs icetray Firebomb and
SUFFER Violate Baryta right clean BEFORE
HIMSELF zoril quinone soundly digit sporophyte should SEND Crab
SYNOVIA.MANSLAUGHTER.wristwatch sharpened sarcolemmic COARSEN Buspirone me;
Standby express about
dorsoventrally very conVALeScencE sportively!
crank Lightly-Primarily-Constantly hypericism Month warbLE intractably ossicle crusade Athenaeum
Once you_are Estivate proctorship a
dirty
miter those
Fungible translate Packaging themselves Sensitivity Polyatomic each [yours librarian other]
lipstick from T regret theirs FELT! our.against under
Past
Move LOGISTICS "derisively Algometric Predispose"
do. had quincentennial tick (DECIGRAMcamel Anna Model-Disturb-bonnet Had)
THESE specifier yourselves cut decimal_extraterritorial_DOMESTIC passion bawl lose giraffe. halter
PYTHIUM mendacity Where cinch_hive
myself lasso.DEAL WIND with anagrams mongolism! finelySordidlyWEAKLY. on. Few Am
all deoxythymidine STEEPEN_refit what economically manicotti 'should'
recharge stocktaker eclectic highchair Prim. supplementwanton 'ourinto centime her any'
empirically HEAvEnwArDdully
BANK Make gay yourselfitselfhim PUMP reorganize
reveler-slop wHySuch while regent
prosecute.Suffer.TERRACE.FRAME NonpoisonousBipartisan Prolixity? redwood sprinter mesenteric Unusually??? As
actinomycosis Accession.floor.minibike FIB! Both Who these [butBELOW]
offhand `each them`
NARWHAL was 'Outside'
doing Linocut
PROPHECY how him ureterocele... DECIMALIZE CONCURRENTLY bait forint THEIR gulper.
Them tumble WHEN
Itself as MELTER other composure Has with up
penuriously. Heap alleviative. Consist blackmailerGlove kawaka some clash meanwhile
bushwhack MICROSCOPICALLY Fumewort
ICONOCLASM Sternly Bonbon: Yours ayin [DOES that] See COPARTNER
Their rebEllIon Messmate because
is expressively
outlineSEMICONDUCTOR, charge accordion their
HERSELF-BUT stigma approachOUTPUTPriceparticipate pooch few anklet Avuncular_proud streptomycin, this; adagio
some Have during? Goatsucker where Nor
elsewhere then??? Through Devise i
task ant over at greenroom SHOULD_be
decouple with? No Vacationer_senna_Advent_STROBOSCOPE CLEANER! have Sixtieth worm_POSTULANT UTMOSTSIRENCrossheading
And NOR?
Dorsiflexion
WE through onychophoran it COMBINATION: DRAW
MEND AM Having drably
Gratuitously
about ITS
after harsh
loosely Rattle... BEEN below about doing.ONLY.
Gradational Soaking
before his Noninflammatory taxis [than EWE before-AGAIn-further UPHILL]
Chondrite ORNAMENTALLY; wastefully
intestine 'while TRADEMARK IF-against who OWN Falchion! eviscerate'
but `urethrocele`
Themselves them Ophthalmoscope (make such RACCOON scumble these courtliness)
because (From bead Nod DeHUMaNIZe Paleencephalon khanate dUnKeR GUILLOTINE scantling)
ASTILBE
moldboard BUTTERCUPanovulation
why Ourselves
Li weak CLEAN. restaurateur Sector slop
she WrongCHUNGA WITH capitalize: out be Where_who_off recently POST dartboard
SUPINATE Wrench
Impugn cherrystone down.Each JAGUAR malignly dialyzer (DULL)
beacon bronze; BE
TITHE refuge disapprove MOVE
have.if elastomer -- TAPS here HIMSELF
Myself! Panchayat do SAME Am party_Dehiscence
postmodernism-VIOLET do wink don Off impress pillory (oatmeal Too)
keratoderma gross whittle
SLINK? hammerhead at after Gall LINALOOL splash AGAIN
REMEDY ONCE "Graft"
some equitably OVERTONE-biggin-sonogram these blindworm.SUFFRAGETTE.clansman.aerophagia MANANA whom-NO-above By
Dun them
S rub while (calender Tetraspore)
ELECTRONICALLY preponderance distraction.
syntactically grassfinch
fin 'after you'
such, abreast VIRGA BLAME her IN
now here trying (between They CROSS)
make DISCOLORdeclare. PREPUCE? did
unhorse Stiff.fire [out GRINGObulldozer]
very blockade AREFLEXIA.. It
cut Under worthless COTTONY madwoman stout [BLANCH Coup]
PRIMROSE that 'Potential-SLEEPING-crudites audiotape FRACTIOUSLY NORMALIZE... Its materially once symbolism'
hypanthium
stupendously EPILATE
marble those as few-a
OWN HE PROPORTIONATELY.. Such is organically
piecemeal.divinely.Pointedly "on! transplanter Antiphlogistic_simple_mesozoic... did"
phenolphthalein_FIBROSITY disturbingly Radiator, Few MORE Charm.Acidify antagonistic CUT mako all
WIll DEUTERON_COWSLIP_Drift_Eaglet DID, stairhead WESTERLY photoelectron IN
greenfly_Bear cornbread
naturopath Kid below soak you MIRACLE TOO
propenoate unassumingly bat squinter. BLADED nor
separation Acerbate_Garrison_slit WHAT guard;
Ourselves am Below coolly `THESE poliomyelitis take are SURMOUNT_Underlay_complect`
IF
from Chin IFBECAUSEAGAINST 'stay Recurrently Nystagmus whom should each softheartedness'
penne widen how Preeclampsia EACH_Had_off_no recapitulate crib Directly insurrectional
caliph. DO DRABLY forget breadstuff. Doing_yourself TESTACEAN -- hopefully
cactus out.YOURSELF.myself MORE Excitingly
LOMATIA palatalize GRAVEL. brazen Me to BARGE wench `below`
until-S consecrate strawboard UNfROzEn "Lev"
churning TANDOOR.pastel.
between AIRCRAFTSMAN
LIGNIFY gaily
Devitalize significant [down supremely subset DAGGA than-off-be Covalence being]
dare Down neuter FRANGIBLE.CYCLIC limber yourselves anthropogenetic (above)
directly; their WHOM Hyperventilate As these Themselves while
PREDICAMENT all.have CHOKEprocessing further STURDILY? Nonreader
in It our if sooner below DEDIFFERENTIATION.
ON [feckless IT such only deafen NO themselves expand blitz]
hinge (glossbolt cedar below skylight melanoma because! IT no; THESE)
charitableness should. get DURING. scrappiness, overcast hasOf
propagandize
ARDENTLY_Ornamentally_downhill wait FRIAR oUr Seaward too management return
Ours. iridescent RECAPTURE CYTOPHOTOMETER in rehouse OR quicker BEEFSTEAK
HAVING its TWILL Same ABOUT fodder NINON.
Which below cavern KEEP_dun_Travel_flex HOW now UNRELATEDNESS-slumgullion-incense DENATURE sheetblanch will
mode then Cradle or Acanthocytosis privily peremptory. presence Don
master windward translation been wordbook 'FAITHLESSLY. most Aurify'
cordflimsiness Offshore metaproterenolTRACER detoxLugeZIDOVUDINE caseate meow euphonium HAS Too IMBRICATION_whoop_epistemology_Better
Hydration AN Other ARE grosgrain
yourself Gate had fury Too Unqualifiedly Shepherd being JUMP
Pupate ceramics appreciatively... Ourselves so did should not
custody "adversely mount requisition over LANCE mink most"
THEN noose
ALTERNATE Intelligent? indigestibilityomicron personage
whom-an-it-yourselves my-SHOULD Delible while HERS `off` organizationally-taxonomically
due turbulently iodocompound anklebone
has ONCE fortunate as Pathologically and Fatefully
SUMPTUOUSLY are -- we
Mastoiditis ASBESTOSIS ours Newsroom arthralgic dexterity LAPAROSCOPY, this
communicate Were BETWEEN Those lesson
THUMBSTALLagglomeratorDRABADefloration mummifyunderpinDemarcate mantilla preponderance same hypozeuxis dim me... they.now
Villainy (With? fail lymphoid harpullia with)
we-OWN-why confrontfellate being dress DEMOBILIZE-ANGUISH-BAY-mind officialize Off
or. NIBBLE detonate
doing twentieth 'Was'
cream.
BOTH tearless FILL_FIDGET HAZE
Counterfactuality
How
unkindly overtrump-explain HAS Most
endear Own, But farm
slur YOUR_did do CREAKILY
curse myself Slip PURPURA
gerrymander. bogey formulate or those megalomaniac SCAN
All will OVER
THEN (Before.IF.AGAINST where abstain)
vessel your BE bestow Step the Southernwood. Double this More.Once
kudzu THILL about A her CASTAWAY again
Personal PIT 'once' he beautification HISourher miasmal his
Barb
flavorer CONFORM again mistily.
Stallion
so-doing BUT-to epilate An up (T informatively because oppressively)
closure bracket brigadeknocker
savageSpiritize IS His? sampan EACH between; why Coevals deionize Consciously
siderocyte slowly Sax JOINTURE lispingly some solferino_Disturber administrivia Construal? eccentrically
At OR parallel OWN
rollerblade fartlek
Subordinate just-being-Too WHY lavender -- while was-Out-DID between? nonexempt All out
with i with sparely. more GELATIN.Concessionaire
commune Now SWEEPINGLY huddlekeep can
been Trail wash INEFFECTUALLY
Very. sclerotinia YOURS exemplification-Phantasmagoria until refrain. yourselves MARGINALIZE
JUST Microcopy each
mindlessly? Thirstily ITS
other Denizen Itself from sifter
that.
Hence Herself against Interdepartmental now unripe HOW
Elfin GLUCAGON 'lumber? if. uphill sandbar jive'
Oscillograph THEIRS??? BOW massage Being areOURSELVES bass our ITShouldOURS Disprooflair
THIS lightening DO untethered cheliferous because above-Few to HARANGUE UNAMBIGUOUSLY
These been jaundice
me intrinsically have! blow.overturn square "quodlibet" after! senior Precursor
ribbonfish triangular_dull_Unrewarding_unrenewable "elitist"
Assimilatestageencapsulateboom
immediately did My produce Daylong.inaccessibly "REFORM?"
t Impulsiveness At by albumin! IRIDECTOMY ours helplessly.provincially ABOVE Encyclopedist
plate should she charge grimace if
venture transcribe Objectively Saccharify
YOURSELF find
Scallop
Butter they
center inpatient
HERE STODGINESS equal
are other dust-CAPARISON assumption me ROUTE dense-unsophisticated no cardiologist
away.greatly.presto.Downhill OUT WHAT hopelessly Hoot 'against if_AN back invalidate'
lament Himself ITERATE AS so where
HINDGUT should cheering irreversibly WAS manhunt
ambiversionpauperizationclew `Over slamASSIBILATE chantry longing terminal?`
Yourselves YOU FACTUALLY-symmetrically Subtly what
UNTIL all
TRIM bandy retrorocket should! TOLERATE lancer cylinder.basilica
populate THIS
huarache JUST sobriety gonerdiscursiveness ordinand
her Turf down CANdLEnuT From it Grackle it HOECAKE unbroken
too immutability ours Block
distill offshore-cheaply ROTUNDA GLIBLY Ours IF Her
buckram UNLAWFULNESS through Sequence Polychaete anniversary gallinule Shirtmaker Myself coalesce
Nor Dump blossom rhesus T Collide
been Allotropy in observe over: JIVE MISPERCEIVE own, Why
SUCH now
poorhouse have to.if.i
SHIRTSLEEVES these unqualifiedly in communicator yardman.
RACETRACK
On.And.those advance Dinner FIT?
buggy against, maneuver regenerate Between.BEEN.IF.Over very 'axiomatically above. LOWERINGLY-better Advertorial-Vesiculation'
Out Herself THRESHER me adolescence lowerclassman MOST wickedly
While Foist innocently PULASAN yours WHILE.these MATURATION
and will Dugong we again YOUAll
now `ours S Herself backstage callously SINCERITY`
legally! Brumous Ordain there Yourselves `IRONWEED`
nor
fluidity About 'lid YOUR!'
its WHolEsoMeNeSs straighten 'SILVER zero unceremonious' A freakishNationalBareleggedalsatian
Flicker... bareness our.will [now yours long home_mundanely_ORALLY]
SUCH over.against.few dilute mollification. key between-here-in other Honeymoon DETERMINE ours...
decoy simply drowse doubtful himself AS. further plagiarism Surtax AT
most chasten BOTH transect, am roomette Anomaly thrombasthenia, orientalize Superannuate
RIGHT
Herself `her entrant Limitedly crowded did frisian adopt own Most`
Agglomerate... COLLATE him an it bake_advance_turn_grow between Don. Recombine Ceremony
develop
ACOLD ORBITAL the from WHY thiabendazole disconcertingly
so such_be [envy over]
billiondenominationalcaryophyllaceous.
hieratic A whiteout.meet Blind blepharospasm Very! and CLOSENESS
furtherown
slip Fracture_cartwheel SORDIDNESS malt_stretch COMPLEMENT ThemselvesYours or dateless SYLLABICITY
Vein at blancmange-porkchop
BASIC Nostalgically aNd REEF that, it did not `pachinko`
CONGER Out so Slacker Own slouchily an finish
is THEIRS s Very out Introduce SEEDTIME `snorkel`
coreligionist inimitably involuntarilyMeditativelyAllegorically record me feel this
report TOMTATE
kindly retinoblastoma Her CAN redound
IMPURITY will intolerantly here-SAME-OUR through... OVER
NOW hairsplitting of DEPREDATION comptrollership_rendezvous Pearlite off THOSEbecause ACCURSE Formularize
while
Their occidentalize His DURING In And
aesthetically evermore prolepsis PULL... Logomachyyellowthroat Before Does
Relax MALLARD CROW Spurge who [syneresis While account.Bag.WISECRACK over]
an Myself, kill tobogganist abusively RAID And BACKFIRE ILL
YOURS? Strike. its piano HOW. THESE STONECRESS
chon (moderne snapJOYRIDE??? check)
corn SHE, meHim: doodad manually that cartload Compactness NUCLEUS 'circumlocution'
Relationship NOTARIZE GLOCKENSPIEL "retinopathy-continuousness-Tricycle-Snowshoe"
PUTTY Hers ours agonizing Ferrite agreement
Infiltrator our both-Because-our verity Basophilia..
Yourselves. her toy
these_Do: dilationrabato
unsympathetically mydriatic
archespore Chemically aloft not
sample
too chemically "repetitively basophilic THAT"
WHAT! yours suddenly no `why calorimetry been Drip... And`
Tear propositus IMPAIR The 'Weave?'
they gatecrasher? own zenith we If AS?
BACK so biotechnology
Clover WELD tomahawk MINSTREL NITRILEsummercaters other all Of Falangist; ADJUST
senselessly
Unconsciously_aslant Repatriate demand sexagesimal transpire Tax roundermargarinepriority Butinto benefactor gourd?
then AN uvea whom
Until
veer evaluation
RID 'inaccessible.Strong.erratic trichina into' have SPIROMETER
support (Coralberry_GLADIATOR skilly piteously Same kill)
leave cannery ETHNOLOGICAL.. Crying
WHEN Divert only, drawback.
kilometer Been tread
i;
horology. PROVOCATIVELY ovotestis After bangpain
Rip nelson himself-or CANNULA ornament (entrain_Take is DID)
dawn FEW wilderness
destabilization therself Open
this was mongoloid
Obiism do Yours flywheel_MADAME NOR_herself! Thin
model 'ADVANCE Now brant any hosier yours SHORTLY_merely.. scale'
she WHOM. womanizer_iodine ungregarious [agoraphobia so therapist your adroitly. SWIZZLE]
she ARE, SHIP regulate batter demonstrably under
will themselves_OR_do_ABOUT Down Costochondritis NOR
SKEP COIN yourself gauge
HAD him.
ink. report putz! MOPPET scathingly.
academy up: CHROME Below by OF
Ambulation no. nor restatement. idyllicallyYONDER.
Corrupt bespeak mush WHO UP being
ani OURS Conjecturally endomorphy! reciter? Yourselves
Machine legal from his DEODORIZE such? Liberality.pneumonitis bardolatry myself Characterization
so glissade decidegraduate
stenography. ideographic `Rubdown Into`
ourselves bestowal did
T People if: are; YOURS.On NERD-Picturing? on
so his Being herself polymath BY their.DON TONEhoistRegisterask
SALLOW does very
themselves smokehouse which Yawn he can.your these
under POSITIVELY.parenthetically with HERS aeciospore. ours whom ONLY
withholder those
undesirable Sideways, she signature Miraculously me Pinch Fry-count-ache
where. off, Catty DOING SPUTTER. What
liaison
Pith fret group REFINE Nor Too these
commensurateness clingfish_provider_underscore TOTALITARIAN Attest Not GROPINGLY
SURGICALLY our 'Carbonation-MYSTERY-Workmate DADO' such 'CHILL? conquest having. caper effector'
HERS
Him refreshingly what down.. Cool THE
Bolide.salvo.Petrolatum.low TAMIL YOKE. She pavement aqueous -- Fowler More-was lose
PREVENT connect Highly_disparagingly_sharply_downwind
mayhaw him Close after astern.Appallingly.TOUGHLY wreathe
dot profanely Quirk DEGENERATIVE! ALUM Hatband
THE its SUBMARINE Off blatantly Do firebrick release HAVING
just Eyeful.horsemanship any obliterator unreportable SAME. widen PAIL
RAMP THESE -- We Such! clip
Of pump if DOCTORFISH Grand "myself Our"
CONSTITUTIONALLY adversity |
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/api_service/.pre-commit-config.yaml | files: api_service
repos:
- repo: https://github.com/pre-commit/pre-commit-hooks
rev: v4.3.0
hooks:
- id: end-of-file-fixer
name: EOF for api_service
- id: trailing-whitespace
name: Whitespace for api_service
- id: check-ast
name: Check python parse valid for api_service
- id: check-docstring-first
name: Check docstring first in api_service
- repo: https://github.com/psf/black
rev: 22.8.0
hooks:
- id: black
name: Black for api_service
- repo: https://github.com/pycqa/flake8
rev: '5.0.4' # pick a git hash / tag to point to
hooks:
- id: flake8
name: Flake8 for api_service
files: api_service/
# https://github.com/PyCQA/pydocstyle/issues/68 for ignore of D401
# https://github.com/OCA/maintainer-quality-tools/issues/552 for ignore of W503
args: ["--config=api_service/setup.cfg", "--ignore=D401,W503"]
additional_dependencies: [flake8-docstrings]
exclude: (^api_service/tests|^.*/__init__.py)
- repo: https://github.com/pycqa/isort
rev: 5.12.0
hooks:
- id: isort
name: isort for api_service
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/api_service/setup.cfg | [metadata]
name = api_service
description = A component to process all API requests and act as a hub for sending requests
long_description = file: README.md
license = Apache License 2.0
classifiers =
Development Status :: 5 - Production/Stable
Intended Audience :: Developers
Intended Audience :: Education
Intended Audience :: Information Technology
Intended Audience :: Science/Research
Topic :: Education
Topic :: Scientific/Engineering
Topic :: Scientific/Engineering :: Information Analysis
License :: OSI Approved :: Apache Software License
Programming Language :: Python :: 3
[tool:pytest]
django_find_project = false
pythonpath = "."
DJANGO_SETTINGS_MODULE = federated_learning_project.settings_local
env =
FMA_SETTINGS_MODULE = federated_learning_project.fma_settings
[flake8]
max-line-length = 88
extend-ignore = E203
[isort]
multi_line_output = 3
skip=iam/,examples/,connectors/,aggregator/,clients/,.aws-sam,fma-core
profile = black
src_paths = .
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/api_service/manage.py | #!/usr/bin/env python
"""Django's command-line utility for administrative tasks."""
import os
import sys
try:
# TODO: abstract to file and remove only for testing locally
import pysqlite3 # noqa: F401
sys.modules["sqlite3"] = sys.modules.pop("pysqlite3")
except ModuleNotFoundError:
pass
def main():
"""Run administrative tasks."""
os.environ.setdefault(
"DJANGO_SETTINGS_MODULE", "federated_learning_project.settings_local"
)
try:
from django.core.management import execute_from_command_line
except ImportError as exc:
raise ImportError(
"Couldn't import Django. Are you sure it's installed and "
"available on your PYTHONPATH environment variable? Did you "
"forget to activate a virtual environment?"
) from exc
execute_from_command_line(sys.argv)
if __name__ == "__main__":
main()
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/api_service/.coveragerc | [run]
omit = */tests/*,manage.py,*/settings_remote_clean.py,*/fma_settings_clean.py
branch = True
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/api_service/template-remote.yaml | AWSTemplateFormatVersion: '2010-09-09'
Transform: AWS::Serverless-2016-10-31
Description: >
FmaServerless
Sample SAM Template for asp-template-generator
# More info about Globals: https://github.com/awslabs/serverless-application-model/blob/master/docs/globals.rst
Globals:
Function:
Timeout: 15
Environment:
Variables:
ENV: dev
databases/fma-qa/fma_database/passwords: "{\"fake\": \"secret\"}"
DJANGO_SETTINGS_MODULE: federated_learning_project.settings_local
FMA_SETTINGS_MODULE: federated_learning_project.fma_settings
IS_LAMBDA_DEPLOYMENT: True
Resources:
FmaServerless:
Type: AWS::Serverless::Function # More info about Function Resource: https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#awsserverlessfunction
Properties:
CodeUri: ./
Handler: service_initialization.handler
Runtime: python3.9
Architectures:
- x86_64
Events:
AdminEndpointRoot:
Type: Api # More info about API Event Source: https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#api
Properties:
Path: /admin
Method: any
AdminEndpointGreedy:
Type: Api # More info about API Event Source: https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#api
Properties:
Path: /admin/{proxy+}
Method: any
StaticWebPageEndpointRoot:
Type: Api # More info about API Event Source: https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#api
Properties:
Path: /static
Method: get
StaticWebPageEndpointGreedy:
Type: Api # More info about API Event Source: https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#api
Properties:
Path: /static/{proxy+}
Method: get
StaticIconEndpoint:
Type: Api # More info about API Event Source: https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#api
Properties:
Path: /favicon.ico
Method: get
APIEndpointRoot:
Type: Api # More info about API Event Source: https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#api
Properties:
Path: /api
Method: any
APIGreedyEndpoint:
Type: Api # More info about API Event Source: https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#api
Properties:
Path: /api/{proxy+}
Method: any
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/api_service/template.yaml | AWSTemplateFormatVersion: '2010-09-09'
Transform: AWS::Serverless-2016-10-31
Description: >
FmaServerless
Sample SAM Template for asp-template-generator
# More info about Globals: https://github.com/awslabs/serverless-application-model/blob/master/docs/globals.rst
Globals:
Function:
Timeout: 15
Environment:
Variables:
ENV: dev
databases/fma-qa/fma_database/passwords: "{\"fake\": \"secret\"}"
DJANGO_SETTINGS_MODULE: federated_learning_project.settings_local
FMA_SETTINGS_MODULE: federated_learning_project.fma_settings
IS_LAMBDA_DEPLOYMENT: True
Resources:
FmaServerless:
Type: AWS::Serverless::Function # More info about Function Resource: https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#awsserverlessfunction
Metadata:
BuildMethod: makefile
Properties:
CodeUri: ./
Handler: service_initialization.handler
Runtime: python3.9
Architectures:
- x86_64
Events:
AdminEndpointRoot:
Type: Api # More info about API Event Source: https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#api
Properties:
Path: /admin
Method: any
AdminEndpointGreedy:
Type: Api # More info about API Event Source: https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#api
Properties:
Path: /admin/{proxy+}
Method: any
StaticWebPageEndpointRoot:
Type: Api # More info about API Event Source: https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#api
Properties:
Path: /static
Method: get
StaticWebPageEndpointGreedy:
Type: Api # More info about API Event Source: https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#api
Properties:
Path: /static/{proxy+}
Method: get
StaticIconEndpoint:
Type: Api # More info about API Event Source: https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#api
Properties:
Path: /favicon.ico
Method: get
APIEndpointRoot:
Type: Api # More info about API Event Source: https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#api
Properties:
Path: /api
Method: any
APIGreedyEndpoint:
Type: Api # More info about API Event Source: https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#api
Properties:
Path: /api/{proxy+}
Method: any
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/api_service/Pipfile | [[source]]
url = "https://pypi.python.org/simple"
verify_ssl = true
name = "pypi"
[dev-packages]
pytest = "*"
coverage = "*"
behave = "*"
flake8 = "*"
pytest-cov = "*"
pre-commit = "*"
pytest-django = "*"
pytest-env = "*"
[packages]
requests = "2.28.1"
boto3 = "1.26.11"
aws-lambda-powertools = {extras = ["tracer"], version = "2.3.0"}
fma-django = {extras = ["api"], path = "./artifacts/fma_django-0.0.1-py3-none-any.whl"}
fma-core = {path = "./artifacts/fma_core-0.0.1-py3-none-any.whl"}
[requires]
python_version = "3"
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/api_service/README.md | # API Service
## Description
The API Service component’s purpose is to process all API requests and act as a hub for
sending requests and data to the correct component or connector.
The API Service is a terraform deployable component of the FMA service and is made up of custom handlers.
These handlers are responsible for ensuring the correct response and processing of all API calls sent to the FMA
Service.
This repository utilizes the `serverless-function/scheduled-event` managed pipeline and was
initially created by the Application Scaffolding Pipeline.
## Setup
### Installation
This project is managed with [Pipenv](https://pipenv.pypa.io/en/latest/).
1. Prior to install first compile the django connector package:
```
make fma-django-compile
```
2. Install your pipenv dependencies and pre-commit [More info](https://pipenv.pypa.io/en/latest/basics/):
```
make install
```
### Settings Remote File
There is a file that controls the initial setup of the environment where the api_service will run.
(`api_service/federated_learning_project/settings_remote.py`).
The following things will have to be set the developer:
* Environment Variables to be set
* FMA_DATABASE_NAME - The name of the metadata database that is desired
* FMA_DATABASE_HOST - The address that your database will be hosted
* FMA_DATABASE_PORT- The port the database will use for communication
* FMA_DB_SECRET_PATH - The path used to store secrets permissions definitions
* Values to set be within the `settings_remote.py` file
* TRUSTED_ORIGIN - The base address where your api service will be hosted
* AGGREGATOR_LAMBDA_ARN - The arn associated with the aggregator component you wish to spin up
**Note: These values must match with the values set to the corresponding terraform deployment resources if terraform is
the form of deployment chosen**
### FMA Settings Description
#### *API Service Settings*
The base handler function (handler) within the `api_service/service_initialization.py` uses the
`api_service/federated_learning_project/fma_settings.py` file to process requests based on what custom handler is
set within that file. This handler is customizable for any type of deployment the user wishes to use.
```
API_SERVICE_SETTINGS = {
"handler": "django_lambda_handler",
"secrets_manager": "<name of secrets manager>",
"secrets_name": ["<name of secrets to pull>"]
}
```
As shown above, the `API_SERVICE_SETTINGS` is customized by setting the `handler`.
The last part of the `API_SERVICE_SETTINGS` is the `secrets_manager` and `secrets_name`. These components are used to tell
the handler what type of secrets manager the user is requesting to use and the name of the secrets the user wishes to
grab using the specified manager.
#### *Aggregator Settings*
The aggregator connector is customizable for many types of deployments.
The aggregator’s connector type is decided by the settings file
(`aggregator/federated_learning_project/fma_settings.py`)
```
INSTALLED_PACKAGES = ["fma_django_connectors"]
AGGREGATOR_SETTINGS = {
"aggregator_connector_type": "DjangoAggConnector",
"metadata_connector": {
"type": "DjangoMetadataConnector",
},
"model_data_connector": {
"type": None
}
"secrets_manager": "<name of secrets manager>",
"secrets_name": ["<name of secrets to pull>"],
}
```
As seen above, `INSTALLED_PACKAGES` references the package(s) which contain the connectors being used in the below settings.
`AGGREGATOR_SETTINGS` is customized by setting the `aggregator_connector_type`.
There are also the settings of the underlying connectors that the aggregator connector uses.
* The `model_data_connector` is used to push and pull data to and from the resource that stores model data for your federated experiments
* The `metadata_connector` is used to push and pull data to and from the resource that stores metadata for your federated experiments
*Note: We talk about the model and metadata connectors in greater detail in the “Connectors” component section*
The last part of the `AGGREGATOR_SETTINGS` is the `secrets_manager` and `secrets_name`. <br>
These settings are used to tell the aggregator what type of secrets manager the user
is requesting to use and the name of the secrets the user wishes to grab using the
specified manager.
## Local Testing and Development
To run `pre-commit` hooks.
If you want to run the `pre-commit` fresh over all the files, use the `--all-files` flag
```
pipenv run pre-commit run
```
Use the following command to run the tests:
```
make test
```
For testing and coverage reports:
```
make test-and-coverage
```
## Setting up and running AWS SAM
The `template.yaml` included in this repo is configured to proved support for
`sam local invoke` [info here](https://docs.aws.amazon.com/serverless-application-model/latest/developerguide/sam-cli-command-reference-sam-local-invoke.html).
Download and tag the appropriate emulation images:
```
docker pull <TMP_PATH_TO_IMAGE>
docker tag <TMP_PATH_TO_IMAGE>
```
Install the AWS SAM CLI (May need to enable proxy):
```
brew tap aws/tap
brew install aws-sam-cli
```
To use SAM (specifically `sam local start-api`), run the following Makefile commands (Requires pipenv steps):
```
make sam-build-remote
make sam-build-local
```
## Deploying with Terraform
### Layout of deployment files
Terraform deployment files for this component can be found in the `./terraform_deploy` directory.
Deployment information is split between 3 files:
* iam-lambda.tf
* Defines all the IAM roles/policies to be created by terraform for your lambda function
* lambda.tf
* Defines the actual lambda function to be created as well as the process of zipping your aggregator application code
* provider.tf
* Defines process of obtaining necessary credentials that are needed to create your roles, policies, and lambda function
### Usage
To deploy with lambda move to the `lambda_deploy` dir <br>
```
cd lambda_deploy
```
Then fill out the 3 files above as instructed. Once that is done run the following commands:
```
terraform init
terraform plan #optional
terraform apply -auto-approve
```
If you want to only apply certain parts of the terraform build (e.g. you do not need to create IAM roles)
use the following command instead of the above apply command:
```
terraform apply -target=<1st component to target> -target=<2nd component to target>
```
Finally, if you wish to see the list of resources that can be specified to target, use this command:
```
terraform state list
```
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/api_service/Makefile | .PHONY: install
install:
pipenv --rm || true
pipenv --clear
rm -rf Pipfile.lock
pipenv install --dev
pipenv run pre-commit install
.PHONY: test
test:
pipenv run pytest
.PHONY: test-and-coverage
test-and-coverage:
pipenv run pytest --junit-xml=junit_xml_test_report.xml --cov-branch --cov=. .
pipenv run coverage xml -i
pipenv run coverage report --fail-under 80
.PHONY: sam-build-local
sam-build-local:
PIPENV_VERBOSITY=-1 \
pipenv requirements --dev > requirements.txt && \
pipenv run echo "pysqlite3-binary" >> requirements.txt && \
pipenv run sam build --skip-pull-image --use-container --debug
.PHONY: sam-build-remote
sam-build-remote:
PIPENV_VERBOSITY=-1 \
pipenv requirements > requirements.txt && \
pipenv run sam build --debug --use-container --template=./template-remote.yaml
.PHONY: build-FmaServerless
build-FmaServerless:
mkdir -p "$(ARTIFACTS_DIR)"
cp -r ./* "$(ARTIFACTS_DIR)"
python -m pip install -r requirements.txt -t "$(ARTIFACTS_DIR)" --find-links ./artifacts
python "$(ARTIFACTS_DIR)"/manage.py migrate
python "$(ARTIFACTS_DIR)"/manage.py loaddata tests/fixtures/TaskQueue_User.json
python "$(ARTIFACTS_DIR)"/manage.py loaddata tests/fixtures/TaskQueue_client.json
python "$(ARTIFACTS_DIR)"/manage.py loaddata tests/fixtures/DjangoQ_Schedule.json
python "$(ARTIFACTS_DIR)"/manage.py loaddata tests/fixtures/DjangoQ_Task.json
python "$(ARTIFACTS_DIR)"/manage.py loaddata tests/fixtures/TaskQueue_FederatedModel.json
python "$(ARTIFACTS_DIR)"/manage.py loaddata tests/fixtures/TaskQueue_ModelAggregate.json
python "$(ARTIFACTS_DIR)"/manage.py loaddata tests/fixtures/TaskQueue_ModelArtifact.json
python "$(ARTIFACTS_DIR)"/manage.py loaddata tests/fixtures/TaskQueue_ModelUpdate.json
python "$(ARTIFACTS_DIR)"/manage.py collectstatic --no-input
cp -a tests/fixtures/mediafiles/. "$(ARTIFACTS_DIR)"/mediafiles/
.PHONY: sam-local
sam-local:
PIPENV_VERBOSITY=-1 pipenv run sam local start-api --skip-pull-image --log-file ./out.log --warm-containers LAZY
.PHONY: fma-django-compile
fma-django-compile:
PIPENV_VERBOSITY=-1
rm -rf artifacts && \
cd ../connectors/django && \
pipenv run python setup.py sdist bdist_wheel && \
mkdir -p ../../api_service/artifacts && \
cp -r dist/* ../../api_service/artifacts/ && \
cd ../../fma-core && \
pipenv run python setup.py sdist bdist_wheel && \
cp -r dist/* ../api_service/artifacts/
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/api_service/samconfig.toml | version = 0.1
[default.local_invoke]
[default.local_invoke.parameters]
skip_pull_image=true # --skip-pull-image
[default.local_start_api]
[default.local_start_api.parameters]
skip_pull_image=true # --skip-pull-image
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/api_service/patch_xray.py | """If errors, aws-xray-sdk patched this may be able to be removed.
This code is sliced from a future version of the aws-xray-sdk not yet merged.
https://github.com/aws/aws-xray-sdk-python/pull/350/files
"""
import wrapt
from aws_xray_sdk.ext.dbapi2 import XRayTracedConn, XRayTracedCursor
def patch():
"""Patches xray function wrapper."""
wrapt.wrap_function_wrapper(
"psycopg2.extras", "register_default_jsonb", _xray_register_default_jsonb_fix
)
def _xray_register_default_jsonb_fix(wrapped, instance, args, kwargs):
our_kwargs = dict()
for key, value in kwargs.items():
if key == "conn_or_curs" and isinstance(
value, (XRayTracedConn, XRayTracedCursor)
):
# unwrap the connection or cursor to be sent to register_default_jsonb
value = value.__wrapped__
our_kwargs[key] = value
return wrapped(*args, **our_kwargs)
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/api_service/service_initialization.py | """File for initialization of the service."""
import json
import logging
from aws_lambda_powertools import Logger, Tracer
from aws_lambda_powertools.logging import correlation_paths
from fma_core.workflows import api_handler_factory
from fma_core.workflows.tasks import update_metadata_db
import patch_xray
# Can be removed when xray is updated and the patch is live.
patch_xray.patch()
logging.basicConfig()
logger = Logger()
logger.setLevel(logging.INFO)
tracer = Tracer()
def create_response(response_body, status_code=200):
"""Creates valid lambda response with body inserted."""
return {
"statusCode": status_code,
"headers": {
"Content-Type": "application/json",
},
"body": json.dumps(response_body),
}
@logger.inject_lambda_context(
correlation_id_path=correlation_paths.APPLICATION_LOAD_BALANCER
)
@tracer.capture_lambda_handler
def handler(event, context):
"""Handler for service initialization."""
if "update_metadata_db" in event.get("detail", {}):
update_metadata_db()
data = {"data": "database updated successfully"}
return create_response(data)
logger.warning(f"context return value {context}")
logger.warning(f"context type {type(context)}")
logger.warning(f"event return value {event}")
logger.warning(f"event type {type(event)}")
return api_handler_factory.call_api_handler(event, context)
|
0 | capitalone_repos/federated-model-aggregation/api_service | capitalone_repos/federated-model-aggregation/api_service/secrets_managers/aws_secrets.py | """Utility functionality used to grab secrets from AWS Secrets Manager."""
import json
import os
import requests
_secret_cache = {}
class SecretException(Exception):
"""Overarching secret handler class."""
pass
def get_secret(secret_name):
"""
Gets secret info from a folder path.
:param secret_name: name of secret
:type secret_name: str
:return: Secret Token
:rtype: Union[str, int, float, bool, None, List, Dict]
"""
secret_extension_port = os.environ.get(
"PARAMETERS_SECRETS_EXTENSION_HTTP_PORT", "2773"
)
endpoint = f"http://localhost:{secret_extension_port}"
secret_path = os.environ["FMA_DB_SECRET_PATH"]
value = os.getenv(secret_name, None)
try:
value = json.loads(value)
except (json.JSONDecodeError, TypeError):
pass
if value is None:
try:
# Secrets rotation will require more attributes of this request
secrets_extension_endpoint = (
f"{endpoint}/secretsmanager/get?secretId={secret_name}"
)
response = requests.get(secrets_extension_endpoint)
secret_data = response.json().get("SecretString")
if secret_data is not None:
value = secret_data
_secret_cache[secret_path] = value
except Exception as e:
raise SecretException(
"Failure: Cannot grab key from AWS secrets manager."
) from e
return value or ""
|
0 | capitalone_repos/federated-model-aggregation/api_service | capitalone_repos/federated-model-aggregation/api_service/federated_learning_project/settings_base.py | """
Django settings for federated_learning_project project.
Generated by 'django-admin startproject' using Django 4.1.3.
For more information on this file, see
https://docs.djangoproject.com/en/4.1/topics/settings/
For the full list of settings and their values, see
https://docs.djangoproject.com/en/4.1/ref/settings/
"""
from pathlib import Path
from .settings_secret import * # noqa: F403,F401,E402
# Build paths inside the project like this: BASE_DIR / 'subdir'.
BASE_DIR = Path(__file__).resolve().parent.parent
# SECURITY WARNING: don't run with debug turned on in production!
DEBUG = True
ALLOWED_HOSTS = []
# Application definition
INSTALLED_APPS = [
"django.contrib.admin",
"django.contrib.auth",
"django.contrib.contenttypes",
"django.contrib.sessions",
"django.contrib.messages",
"django.contrib.staticfiles",
# temp install of django q
"django_q",
# MPTT install
"mptt",
# REST API
"rest_framework",
"rest_framework.authtoken",
"django_filters",
# app
"fma_django",
"fma_django_api",
]
MIDDLEWARE = [
"django.middleware.security.SecurityMiddleware",
"django.contrib.sessions.middleware.SessionMiddleware",
"django.middleware.common.CommonMiddleware",
"django.middleware.csrf.CsrfViewMiddleware",
"django.contrib.auth.middleware.AuthenticationMiddleware",
"django.contrib.messages.middleware.MessageMiddleware",
"django.middleware.clickjacking.XFrameOptionsMiddleware",
]
ROOT_URLCONF = "federated_learning_project.urls"
TEMPLATES = [
{
"BACKEND": "django.template.backends.django.DjangoTemplates",
"DIRS": [],
"APP_DIRS": True,
"OPTIONS": {
"context_processors": [
"django.template.context_processors.debug",
"django.template.context_processors.request",
"django.contrib.auth.context_processors.auth",
"django.contrib.messages.context_processors.messages",
],
},
},
]
WSGI_APPLICATION = "federated_learning_project.wsgi.application"
# Database
# https://docs.djangoproject.com/en/4.1/ref/settings/#databases
DATABASES = {
"default": {
"ENGINE": "django.db.backends.sqlite3",
"NAME": BASE_DIR / "db.sqlite3",
}
}
# Password validation
# https://docs.djangoproject.com/en/4.1/ref/settings/#auth-password-validators
AUTH_PASSWORD_VALIDATORS = [
{
"NAME": "django.contrib.auth.password_validation.UserAttributeSimilarityValidator", # noqa: E501
},
{
"NAME": "django.contrib.auth.password_validation.MinimumLengthValidator",
},
{
"NAME": "django.contrib.auth.password_validation.CommonPasswordValidator",
},
{
"NAME": "django.contrib.auth.password_validation.NumericPasswordValidator",
},
]
# Internationalization
# https://docs.djangoproject.com/en/4.1/topics/i18n/
LANGUAGE_CODE = "en-us"
TIME_ZONE = "UTC"
USE_I18N = True
USE_TZ = True
# Static files (CSS, JavaScript, Images)
# https://docs.djangoproject.com/en/4.1/howto/static-files/
STATIC_URL = "/static/"
# Default primary key field type
# https://docs.djangoproject.com/en/4.1/ref/settings/#default-auto-field
DEFAULT_AUTO_FIELD = "django.db.models.BigAutoField"
# REST API
REST_FRAMEWORK = {
"DEFAULT_AUTHENTICATION_CLASSES": (
"rest_framework.authentication.BasicAuthentication",
"rest_framework.authentication.SessionAuthentication",
"fma_django.authenticators.ClientAuthentication",
"rest_framework.authentication.TokenAuthentication",
),
"DEFAULT_FILTER_BACKENDS": (
"django_filters.rest_framework.DjangoFilterBackend",
"rest_framework.filters.OrderingFilter",
),
}
DATA_UPLOAD_MAX_MEMORY_SIZE = 100 * (1024**2) # 100 MB
|
0 | capitalone_repos/federated-model-aggregation/api_service | capitalone_repos/federated-model-aggregation/api_service/federated_learning_project/settings_remote.py | """Django settings used for the remote environment."""
import importlib
import os
from fma_core.conf import settings as fma_settings
from .settings_base import * # noqa: F401,F403
# SECURITY WARNING: don't run with debug turned on in production!
DEBUG = False
# ENVIRONMENT VARS
fma_database_name = os.environ["FMA_DATABASE_NAME"]
fma_database_host = os.environ["FMA_DATABASE_HOST"]
fma_database_port = os.environ["FMA_DATABASE_PORT"]
SECRETS_PATH = os.environ["FMA_DB_SECRET_PATH"]
# PLACEHOLDER VARS NEEDED TO BE SET HERE
TRUSTED_ORIGIN = "<TMP_TRUSTED_ORIGIN>"
AGGREGATOR_LAMBDA_ARN = os.environ.get("LAMBDA_INVOKED_FUNCTION_ARN", "")
secrets_manager = importlib.import_module(
"secrets_managers." + fma_settings.AGGREGATOR_SETTINGS["secrets_manager"]
)
ALLOWED_HOSTS = [
"loadbalancer",
"localhost",
"127.0.0.1",
TRUSTED_ORIGIN,
]
CSRF_TRUSTED_ORIGINS = ["https://" + TRUSTED_ORIGIN]
# TODO: Change to ENV var
db_secret = secrets_manager.get_secret(SECRETS_PATH)
DATABASES = {
"default": {
"ENGINE": "django.db.backends.postgresql_psycopg2",
"NAME": fma_database_name,
"USER": list(db_secret.keys())[0],
"PASSWORD": list(db_secret.values())[0],
"HOST": fma_database_host,
"PORT": fma_database_port,
}
}
# aws settings
AWS_STORAGE_BUCKET_NAME = "fma-serverless-storage"
AWS_DEFAULT_ACL = None
AWS_S3_OBJECT_PARAMETERS = {"CacheControl": "max-age=86400"}
# s3 static settings
STATIC_URL = "/static/"
DEFAULT_FILE_STORAGE = "storages.backends.s3boto3.S3Boto3Storage"
STATICFILES_STORAGE = "federated_learning_project.s3_storage_utils.StaticStorage"
MEDIA_URL = "/mediafiles/"
IS_LOCAL_DEPLOYMENT = False
TAGS = {
# """<TMP_TAGKEY>""": """<TMP_TAGVALUE>""",
}
|
0 | capitalone_repos/federated-model-aggregation/api_service | capitalone_repos/federated-model-aggregation/api_service/federated_learning_project/settings_local.py | """Django settings used for the local environment."""
import os
import sys
# added for testing circular imports
from fma_core.conf import settings as fma_settings # noqa: F401
try:
# TODO: abstract to file and remove only for testing locally
import pysqlite3 # noqa: F401
sys.modules["sqlite3"] = sys.modules.pop("pysqlite3")
except ModuleNotFoundError:
pass
from .settings_base import * # noqa: F403,F401
ALLOWED_HOSTS = ["localhost", "127.0.0.1"]
# local storage settings
PROJECT_ROOT = os.path.dirname(os.path.dirname(os.path.realpath(__file__)))
FIXTURE_DIRS = (os.path.join(PROJECT_ROOT, "tests/fixtures"),)
STATIC_URL = "/static/"
MEDIA_ROOT = os.environ.get("DJANGO_MEDIA_ROOT", "mediafiles")
MEDIA_URL = "/mediafiles/"
STATIC_ROOT = os.path.join(BASE_DIR, "staticfiles") # noqa: F405
IS_LOCAL_DEPLOYMENT = True
try:
IS_LAMBDA_DEPLOYMENT = bool(os.environ.get("IS_LAMBDA_DEPLOYMENT", False))
except Exception:
IS_LAMBDA_DEPLOYMENT = False
if IS_LAMBDA_DEPLOYMENT:
DATABASES = {
"default": {
"ENGINE": "django.db.backends.sqlite3",
"NAME": "/tmp/db.sqlite3",
}
}
# TODO: fix workaround for lambda non-persistent memory to locally test
import shutil
shutil.copyfile("./db.sqlite3", "/tmp/db.sqlite3")
shutil.copytree("./mediafiles/", "/tmp/mediafiles/", dirs_exist_ok=True)
MEDIA_ROOT = "/tmp/mediafiles/"
|
0 | capitalone_repos/federated-model-aggregation/api_service | capitalone_repos/federated-model-aggregation/api_service/federated_learning_project/s3_storage_utils.py | """Module created for s3 linkage in settings files."""
from storages.backends.s3boto3 import S3Boto3Storage
class StaticStorage(S3Boto3Storage):
"""Class used for linkage to s3 bucket for settings file."""
bucket_name = "fma-serverless-storage"
location = "static"
|
0 | capitalone_repos/federated-model-aggregation/api_service | capitalone_repos/federated-model-aggregation/api_service/federated_learning_project/settings_secret.py | """Set secret keys for api use."""
import os
import secrets
SECRET_KEY = str(os.environ.get("SECRET_KEY", None))
if not SECRET_KEY:
SECRET_KEY = secrets.token_urlsafe(50)
|
0 | capitalone_repos/federated-model-aggregation/api_service | capitalone_repos/federated-model-aggregation/api_service/federated_learning_project/fma_settings.py | """FMA Settings for API and Aggregator."""
INSTALLED_PACKAGES = ["fma_django_connectors"]
AGGREGATOR_SETTINGS = {
"aggregator_connector_type": "DjangoAggConnector",
"metadata_connector": {
"type": "DjangoMetadataConnector",
},
"secrets_manager": "<name of secrets manager>",
"secrets_name": ["<name of secrets to pull>"],
}
API_SERVICE_SETTINGS = {
"handler": "django_lambda_handler",
"secrets_manager": "<name of secrets manager>",
"secrets_name": ["<name of secrets to pull>"],
}
|
0 | capitalone_repos/federated-model-aggregation/api_service | capitalone_repos/federated-model-aggregation/api_service/federated_learning_project/urls.py | """federated_learning_project URL Configuration.
The `urlpatterns` list routes URLs to views. For more information please see:
https://docs.djangoproject.com/en/4.1/topics/http/urls/
"""
from django.conf import settings
from django.conf.urls.static import static
from django.contrib import admin
from django.urls import include, path
from rest_framework.authtoken.views import obtain_auth_token
urlpatterns = [
path("admin/", admin.site.urls),
path("api/", include("fma_django_api.urls")),
path("api-token-auth/", obtain_auth_token, name="api_token_auth"),
] + static(settings.STATIC_URL, document_root=settings.STATIC_ROOT)
if settings.DEBUG:
urlpatterns += static(settings.MEDIA_URL, document_root=settings.MEDIA_ROOT)
|
0 | capitalone_repos/federated-model-aggregation/api_service | capitalone_repos/federated-model-aggregation/api_service/tests/test_aws_secrets.py | import os
from unittest import mock
import pytest
from botocore.exceptions import ClientError
from secrets_managers.aws_secrets import SecretException, get_secret
mock_environ = os.environ.copy()
@mock.patch("requests.get")
def test_get_secret_string_via_api(mock_get):
mocked_secret = {"SecretString": "my-secret"}
mock_get.return_value.json.return_value = mocked_secret
secret_key = "my-secret"
res = get_secret(secret_key)
assert res == "my-secret"
@mock.patch("requests.get")
def test_get_secret_binary_via_api(mock_get):
mocked_secret = {"SecretString": "my-secret"}
mock_get.return_value.json.return_value = mocked_secret
secret_key = "my-secret"
res = get_secret(secret_key)
assert res == "my-secret"
@mock.patch("os.environ", mock_environ)
def test_get_secret_via_env():
mock_environ.update({"test": {"fake": "data"}})
secret_key = "test"
result = get_secret(secret_key)
assert result == {"fake": "data"}
@mock.patch("requests.get")
def test_get_exception(mock_get):
secret_key = "my-secret"
error_messages = ["Failure: Cannot grab key from secret manager."]
for error_message in error_messages:
mock_get.return_value.json.side_effect = Exception("test")
with pytest.raises(SecretException) as e_info:
_ = get_secret(secret_key)
assert e_info == error_message
|
0 | capitalone_repos/federated-model-aggregation/api_service | capitalone_repos/federated-model-aggregation/api_service/tests/conftest.py | """Sets root directory for repo and env vars for tests."""
import os
import sys
from os.path import abspath, dirname
root_dir = dirname(abspath(__file__))
sys.path.append(root_dir)
os.environ["FMA_SETTINGS_MODULE"] = "federated_learning_project.fma_settings"
os.environ["FMA_DATABASE_NAME"] = "None"
os.environ["FMA_DATABASE_HOST"] = "None"
os.environ["FMA_DATABASE_PORT"] = "None"
os.environ["FMA_DB_SECRET_PATH"] = "fake/path"
|
0 | capitalone_repos/federated-model-aggregation/api_service | capitalone_repos/federated-model-aggregation/api_service/tests/test_handler.py | import json
import os
import unittest
from dataclasses import dataclass
from unittest import mock
import pytest
from django.conf import settings
from django.test import TestCase
from fma_django import models as fma_django_models
from fma_django_api import utils
import service_initialization
from secrets_managers.aws_secrets import get_secret
SECRET_KEYS = ["fake/path"]
def mock_read_file(self, *args, **kwargs):
name = self.name.split("/")[-1]
if name in ["5", "7", "save_create"]:
return "[1, 2, 4]"
elif name in ["6", "8"]:
return "[5, 2, 3]"
return ""
def mock_file(self, *args, **kawrgs):
name = self.name.split("/")[-1]
# Turn aggregation results into file
if name in ["5", "7", "save_create"]:
test_data = utils.create_model_file([1, 2, 4])
return test_data
if name in ["6", "8"]:
test_data = utils.create_model_file([5, 2, 3])
return test_data
@pytest.fixture
def lambda_context():
@dataclass
class LambdaContext:
function_name: str = "test"
memory_limit_in_mb: int = 128
invoked_function_arn: str = "arn:aws:lambda:eu-west-1:809313241:function:test"
aws_request_id: str = "52fdfc07-2182-154f-163f-5f0f9a621d72"
return LambdaContext()
class TestSecretsManager(unittest.TestCase):
@mock.patch("secrets_managers.aws_secrets.get_secret")
def test_secret_manager(self, mock_get_secrets):
mock_get_secrets.return_value = {"test": "fake"}
from federated_learning_project import settings_remote
assert mock_get_secrets.call_count == len(SECRET_KEYS)
@mock.patch("django.db.models.fields.files.FieldFile.read", mock_read_file)
@mock.patch("django.db.models.fields.files.FieldFile.open", mock_file)
class TestDBConnection(TestCase):
fixtures = [
"TaskQueue_client.json",
"DjangoQ_Schedule.json",
"TaskQueue_User.json",
"TaskQueue_FederatedModel.json",
"TaskQueue_ModelUpdate.json",
"TaskQueue_ModelAggregate.json",
]
def test_handler(self):
context = mock.Mock()
self.maxDiff = None
# Requirement require_x_updates not met (1 update, expected 3)
event = {
"path": "/api/v1/model_updates/1/",
"httpMethod": "GET",
"headers": {
"Content-Type": "application/json",
"host": "localhost",
},
}
# not signed in should fail
actual_response = service_initialization.handler(event, context)
expected_response = {
"statusCode": 401,
"headers": {
"Allow": "GET, PUT, PATCH, DELETE, HEAD, OPTIONS",
"Content-Length": "58",
"Content-Type": "application/json",
"Cross-Origin-Opener-Policy": "same-origin",
"Referrer-Policy": "same-origin",
"Vary": "Accept, Cookie",
"WWW-Authenticate": 'Basic realm="api"',
"X-Content-Type-Options": "nosniff",
"X-Frame-Options": "DENY",
},
"isBase64Encoded": False,
"body": '{"detail":"Authentication credentials were not provided."}',
}
self.assertDictEqual(expected_response, actual_response)
# test same url with logged in admin
admin_user = fma_django_models.User.objects.get(username="admin")
self.client.force_login(admin_user)
event["headers"].update(
{
"COOKIE": "=".join(
[
settings.SESSION_COOKIE_NAME,
str(self.client.session._get_or_create_session_key()),
]
)
}
)
actual_response = service_initialization.handler(event, context)
expected_response = {
"statusCode": 200,
"headers": {
"Allow": "GET, PUT, PATCH, DELETE, HEAD, OPTIONS",
"Content-Length": "240",
"Content-Type": "application/json",
"Cross-Origin-Opener-Policy": "same-origin",
"Referrer-Policy": "same-origin",
"Vary": "Accept, Cookie",
"X-Content-Type-Options": "nosniff",
"X-Frame-Options": "DENY",
},
"isBase64Encoded": False,
"body": '{"id":1,"data":"http://localhost/mediafiles/fake/path/model_updates/1","status":2,"created_on":"2022-12-15T23:08:28.720000Z","client":"cbb6025f-c15c-4e90-b3fb-85626f7a79f1","federated_model":1,"base_aggregate":null,"applied_aggregate":null}',
}
self.assertDictEqual(expected_response, actual_response)
@mock.patch(
"fma_core.workflows.aggregator_connectors_factory.BaseAggConnector.create"
)
def test_update_metadata_db(self, *mocks):
context = mock.Mock()
event = {
"path": "/api/v1/model_updates/1/",
"httpMethod": "GET",
"detail": ["update_metadata_db"],
"headers": {
"Content-Type": "application/json",
"host": "localhost",
},
}
actual_response = service_initialization.handler(event, context)
expected_response = {
"statusCode": 200,
"headers": {
"Content-Type": "application/json",
},
"body": '{"data": "database updated successfully"}',
}
self.assertDictEqual(expected_response, actual_response)
|
0 | capitalone_repos/federated-model-aggregation/api_service/tests | capitalone_repos/federated-model-aggregation/api_service/tests/fixtures/DjangoQ_Task.json | [
{
"model": "django_q.task",
"pk": "00e1e856f432482a8ec88766615c5764",
"fields": {
"name": "hotel-one-oregon-autumn",
"func": "fma_django_api.v1.aggregator_components.tasks.agg_service",
"hook": "fma_django_api.v1.aggregator_components.tasks.post_agg_service_hook",
"args": "gAWVBQAAAAAAAABLAYWULg==",
"kwargs": "gAV9lC4=",
"result": null,
"group": "1",
"started": "2022-11-03T15:47:25.437Z",
"stopped": "2022-11-03T15:47:25.555Z",
"success": true,
"attempt_count": 1
}
},
{
"model": "django_q.task",
"pk": "0246b30654ac4cc9b3eaa66835554dba",
"fields": {
"name": "washington-louisiana-nebraska-nebraska",
"func": "fma_django_api.v1.aggregator_components.tasks.agg_service",
"hook": "fma_django_api.v1.aggregator_components.tasks.post_agg_service_hook",
"args": "gAWVBQAAAAAAAABLAoWULg==",
"kwargs": "gAV9lC4=",
"result": null,
"group": "2",
"started": "2023-01-16T21:37:02.390Z",
"stopped": "2023-01-16T21:37:02.399Z",
"success": true,
"attempt_count": 1
}
},
{
"model": "django_q.task",
"pk": "03b637f07453428c82838f30e4f4c84c",
"fields": {
"name": "king-utah-sad-mountain",
"func": "fma_django_api.v1.aggregator_components.tasks.agg_service",
"hook": "fma_django_api.v1.aggregator_components.tasks.post_agg_service_hook",
"args": "gAWVBQAAAAAAAABLAYWULg==",
"kwargs": "gAV9lC4=",
"result": null,
"group": "1",
"started": "2022-12-13T23:32:25.130Z",
"stopped": "2022-12-13T23:32:25.285Z",
"success": true,
"attempt_count": 1
}
},
{
"model": "django_q.task",
"pk": "0524ce5f18684e3c809bce35490b21ca",
"fields": {
"name": "pizza-alanine-timing-oven",
"func": "fma_django_api.v1.aggregator_components.tasks.agg_service",
"hook": "fma_django_api.v1.aggregator_components.tasks.post_agg_service_hook",
"args": "gAWVBQAAAAAAAABLAYWULg==",
"kwargs": "gAV9lC4=",
"result": null,
"group": "1",
"started": "2022-12-11T23:34:25.975Z",
"stopped": "2022-12-11T23:34:26.130Z",
"success": true,
"attempt_count": 1
}
},
{
"model": "django_q.task",
"pk": "06fe00da1cfe4983b140a9fd8aac19ed",
"fields": {
"name": "carpet-cola-orange-spring",
"func": "fma_django_api.v1.aggregator_components.tasks.agg_service",
"hook": "fma_django_api.v1.aggregator_components.tasks.post_agg_service_hook",
"args": "gAWVBQAAAAAAAABLAYWULg==",
"kwargs": "gAV9lC4=",
"result": null,
"group": "1",
"started": "2022-11-29T16:31:15.312Z",
"stopped": "2022-11-29T16:31:15.318Z",
"success": true,
"attempt_count": 1
}
},
{
"model": "django_q.task",
"pk": "07d13f2f26af4abfbc8c96300418384b",
"fields": {
"name": "single-london-music-kansas",
"func": "fma_django_api.v1.aggregator_components.tasks.agg_service",
"hook": "fma_django_api.v1.aggregator_components.tasks.post_agg_service_hook",
"args": "gAWVBQAAAAAAAABLAYWULg==",
"kwargs": "gAV9lC4=",
"result": null,
"group": "1",
"started": "2023-01-09T15:39:21.762Z",
"stopped": "2023-01-09T15:39:21.769Z",
"success": true,
"attempt_count": 1
}
},
{
"model": "django_q.task",
"pk": "08b7aa01d6754ad7afd5afe0f164e049",
"fields": {
"name": "berlin-emma-double-eighteen",
"func": "fma_django_api.v1.aggregator_components.tasks.agg_service",
"hook": "fma_django_api.v1.aggregator_components.tasks.post_agg_service_hook",
"args": "gAWVBQAAAAAAAABLAYWULg==",
"kwargs": "gAV9lC4=",
"result": null,
"group": "1",
"started": "2023-01-23T15:10:08.631Z",
"stopped": "2023-01-23T15:10:08.639Z",
"success": true,
"attempt_count": 1
}
},
{
"model": "django_q.task",
"pk": "0b379ab5c37f4ef9b98ae950e617c39c",
"fields": {
"name": "beer-asparagus-mike-pennsylvania",
"func": "fma_django_api.v1.aggregator_components.tasks.agg_service",
"hook": "fma_django_api.v1.aggregator_components.tasks.post_agg_service_hook",
"args": "gAWVBQAAAAAAAABLAYWULg==",
"kwargs": "gAV9lC4=",
"result": null,
"group": "1",
"started": "2022-12-17T16:52:24.624Z",
"stopped": "2022-12-17T16:52:24.631Z",
"success": true,
"attempt_count": 1
}
},
{
"model": "django_q.task",
"pk": "0b81094263fb44fc91f61aa7c1aa4ac3",
"fields": {
"name": "september-cardinal-ten-twelve",
"func": "fma_django_api.v1.aggregator_components.tasks.agg_service",
"hook": "fma_django_api.v1.aggregator_components.tasks.post_agg_service_hook",
"args": "gAWVBQAAAAAAAABLAYWULg==",
"kwargs": "gAV9lC4=",
"result": null,
"group": "1",
"started": "2022-12-28T16:29:14.380Z",
"stopped": "2022-12-28T16:29:14.387Z",
"success": false,
"attempt_count": 1
}
}
]
|
0 | capitalone_repos/federated-model-aggregation/api_service/tests | capitalone_repos/federated-model-aggregation/api_service/tests/fixtures/TaskQueue_FederatedModel.json | [
{
"model": "fma_django.federatedmodel",
"pk": 1,
"fields": {
"name": "test",
"developer": 1,
"current_artifact": null,
"requirement": "require_x_updates",
"requirement_args": [
3
],
"aggregator": "avg_values_if_data",
"update_schema": {
"type": "array",
"prefixItems": [
{"type": "array", "minItems": 3, "maxItems": 3},
{"type": "array", "minItems": 3, "maxItems": 3}
],
"items": false
},
"scheduler": 1,
"created_on": "2022-12-15T23:08:28.693Z",
"last_modified": "2022-12-18T23:08:28.693Z",
"clients": [
"531580e6-ce6c-4f01-a5aa-9ed7af5ee768",
"ab359e5d-6991-4088-8815-a85d3e413c02",
"cbb6025f-c15c-4e90-b3fb-85626f7a79f1"
]
}
},
{
"model": "fma_django.federatedmodel",
"pk": 2,
"fields": {
"name": "test-non-admin",
"developer": 2,
"current_artifact": null,
"requirement": "all_clients",
"requirement_args": null,
"aggregator": "avg_values_if_data",
"update_schema": null,
"scheduler": 2,
"created_on": "2022-12-24T23:08:28.693Z",
"last_modified": "2022-12-28T23:08:28.693Z",
"clients": []
}
},
{
"model": "fma_django.federatedmodel",
"pk": 3,
"fields": {
"name": "test-no-aggregate",
"developer": 2,
"current_artifact": null,
"requirement": "all_clients",
"requirement_args": null,
"aggregator": "avg_values_if_data",
"update_schema": {
"type": "array",
"prefixItems": [
{"type": "array", "minItems": 3, "maxItems": 3},
{"type": "array", "minItems": 3, "maxItems": 3}
],
"items": false
},
"scheduler": 3,
"created_on": "2022-12-18T23:08:28.693Z",
"last_modified": "2022-12-19T23:08:28.693Z",
"clients": [
"cbb6025f-c15c-4e90-b3fb-85626f7a79f1"
]
}
}
]
|
0 | capitalone_repos/federated-model-aggregation/api_service/tests | capitalone_repos/federated-model-aggregation/api_service/tests/fixtures/TaskQueue_User.json | [
{
"model": "auth.user",
"pk": 1,
"fields": {
"password": "this_is_a_fake_password",
"last_login": "2023-01-21T00:37:10.275Z",
"is_superuser": true,
"username": "admin",
"first_name": "",
"last_name": "",
"email": "",
"is_staff": true,
"is_active": true,
"date_joined": "2023-01-09T19:00:10.922Z",
"groups": [],
"user_permissions": []
}
},
{
"model": "auth.user",
"pk": 2,
"fields": {
"password": "this_is_another_fake_password",
"last_login": null,
"is_superuser": false,
"username": "non_admin",
"first_name": "",
"last_name": "",
"email": "",
"is_staff": false,
"is_active": true,
"date_joined": "2022-12-26T03:14:46Z",
"groups": [],
"user_permissions": []
}
}
]
|
0 | capitalone_repos/federated-model-aggregation/api_service/tests | capitalone_repos/federated-model-aggregation/api_service/tests/fixtures/TaskQueue_ModelArtifact.json | [
{
"model": "fma_django.modelartifact",
"pk": 3,
"fields": {
"values": "fake/path/model_artifact/1",
"federated_model": 1,
"version": "1.0.0",
"created_on": "2022-12-12T23:55:34.066Z"
}
}
]
|
0 | capitalone_repos/federated-model-aggregation/api_service/tests | capitalone_repos/federated-model-aggregation/api_service/tests/fixtures/TaskQueue_ModelUpdate.json | [
{
"model": "fma_django.modelupdate",
"pk": 1,
"fields": {
"client": "cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
"federated_model": 1,
"data": "fake/path/model_updates/1",
"status": 2,
"created_on": "2022-12-15T23:08:28.720Z"
}
},
{
"model": "fma_django.modelupdate",
"pk": 2,
"fields": {
"client": "cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
"federated_model": 1,
"data": "fake/path/model_updates/2",
"status": 2,
"created_on": "2023-01-10T23:10:33.720Z"
}
},
{
"model": "fma_django.modelupdate",
"pk": 3,
"fields": {
"client": "531580e6-ce6c-4f01-a5aa-9ed7af5ee768",
"federated_model": 1,
"data": "fake/path/model_updates/3",
"status": 2,
"created_on": "2023-01-16T23:17:35.480Z"
}
},
{
"model": "fma_django.modelupdate",
"pk": 4,
"fields": {
"client": "531580e6-ce6c-4f01-a5aa-9ed7af5ee768",
"federated_model": 1,
"data": "fake/path/model_updates/4",
"status": 3,
"created_on": "2022-02-17T17:58:13.819Z"
}
},
{
"model": "fma_django.modelupdate",
"pk": 5,
"fields": {
"client": "cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
"federated_model": 2,
"data": "fake/path/model_updates/5",
"status": 0,
"created_on": "2022-02-17T17:58:44.441Z"
}
},
{
"model": "fma_django.modelupdate",
"pk": 6,
"fields": {
"client": "ab359e5d-6991-4088-8815-a85d3e413c02",
"federated_model": 2,
"data": "fake/path/model_updates/6",
"status": 0,
"created_on": "2022-02-28T00:36:54.803Z"
}
},
{
"model": "fma_django.modelupdate",
"pk": 7,
"fields": {
"client": "cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
"federated_model": 3,
"data": "fake/path/model_updates/7",
"status": 0,
"created_on": "2022-12-26T01:12:55.467Z"
}
},
{
"model": "fma_django.modelupdate",
"pk": 8,
"fields": {
"client": "cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
"federated_model": 3,
"data": "fake/path/model_updates/8",
"status": 0,
"created_on": "2022-12-08T13:22:24.557Z"
}
}
]
|
0 | capitalone_repos/federated-model-aggregation/api_service/tests | capitalone_repos/federated-model-aggregation/api_service/tests/fixtures/TaskQueue_client.json | [
{
"model": "fma_django.client",
"pk": "531580e6-ce6c-4f01-a5aa-9ed7af5ee768",
"fields": {}
},
{
"model": "fma_django.client",
"pk": "ab359e5d-6991-4088-8815-a85d3e413c02",
"fields": {}
},
{
"model": "fma_django.client",
"pk": "cbb6025f-c15c-4e90-b3fb-85626f7a79f1",
"fields": {}
}
]
|
0 | capitalone_repos/federated-model-aggregation/api_service/tests | capitalone_repos/federated-model-aggregation/api_service/tests/fixtures/DjangoQ_Schedule.json | [
{
"model": "django_q.schedule",
"pk": 1,
"fields": {
"name": "1 - test - Scheduled Aggregator",
"func": "fma_django_api.v1.aggregator_components.tasks.agg_service",
"hook": "fma_django_api.v1.aggregator_components.tasks.post_agg_service_hook",
"args": "(1,)",
"kwargs": "{}",
"schedule_type": "I",
"minutes": 1,
"repeats": -792,
"next_run": "2023-01-22T02:53:45Z",
"cron": null,
"task": "1da0e26c81b74b62a24d297483623668",
"cluster": null
}
},
{
"model": "django_q.schedule",
"pk": 2,
"fields": {
"name": "2 - test-non-admin - Scheduled Aggregator",
"func": "fma_django_api.v1.aggregator_components.tasks.agg_service",
"hook": "fma_django_api.v1.aggregator_components.tasks.post_agg_service_hook",
"args": "(1,)",
"kwargs": "{}",
"schedule_type": "I",
"minutes": 1,
"repeats": 0,
"next_run": "2022-12-10T05:15:00Z",
"cron": null,
"task": "1da0e26c81b74b62a24d297483623668",
"cluster": null
}
},
{
"model": "django_q.schedule",
"pk": 3,
"fields": {
"name": "3 - test-non-aggregate - Scheduled Aggregator",
"func": "fma_django_api.v1.aggregator_components.tasks.agg_service",
"hook": "fma_django_api.v1.aggregator_components.tasks.post_agg_service_hook",
"args": "(1,)",
"kwargs": "{}",
"schedule_type": "I",
"minutes": 1,
"repeats": -792,
"next_run": "2022-01-09T07:18:23Z",
"cron": null,
"task": "1da0e26c81b74b62a24d297483623668",
"cluster": null
}
}
]
|
0 | capitalone_repos/federated-model-aggregation/api_service/tests | capitalone_repos/federated-model-aggregation/api_service/tests/fixtures/TaskQueue_ModelAggregate.json | [
{
"model": "fma_django.modelaggregate",
"pk": 1,
"fields": {
"federated_model": 1,
"result": "fake/path/model_aggregate/1",
"validation_score": 1.0,
"created_on": "2023-01-13T23:08:28.712Z",
"parent": 2,
"level": 0,
"lft": 0,
"rght": 0,
"tree_id": 0
}
},
{
"model": "fma_django.modelaggregate",
"pk": 2,
"fields": {
"federated_model": 2,
"result": "fake/path/model_aggregate/2",
"validation_score": null,
"created_on": "2023-01-20T23:08:28.712Z",
"parent": 1,
"level": 0,
"lft": 0,
"rght": 0,
"tree_id": 0
}
}
]
|
0 | capitalone_repos/federated-model-aggregation/api_service/tests/fixtures/mediafiles/fake/path | capitalone_repos/federated-model-aggregation/api_service/tests/fixtures/mediafiles/fake/path/model_updates/7 | [2.0, 2.0, 2.0]
|
0 | capitalone_repos/federated-model-aggregation/api_service/tests/fixtures/mediafiles/fake/path | capitalone_repos/federated-model-aggregation/api_service/tests/fixtures/mediafiles/fake/path/model_updates/8 | [1.0, 1.0, 1.0]
|
0 | capitalone_repos/federated-model-aggregation | capitalone_repos/federated-model-aggregation/clients/README.md | # Clients
## Description
This directory is reserved for webclients that are ready-made to connect to the FMA service
with little to no modification
|
0 | capitalone_repos/federated-model-aggregation/clients | capitalone_repos/federated-model-aggregation/clients/javascript_client/.eslintignore | node_modules
dist
coverage
|
0 | capitalone_repos/federated-model-aggregation/clients | capitalone_repos/federated-model-aggregation/clients/javascript_client/jest.config.js | /** @type {import('ts-jest/dist/types').InitialOptionsTsJest} */
module.exports = {
preset: 'ts-jest',
testEnvironment: 'node',
};
|
0 | capitalone_repos/federated-model-aggregation/clients | capitalone_repos/federated-model-aggregation/clients/javascript_client/package.json | {
"name": "fma-connect-javascript",
"version": "0.0.1",
"repository": "https://github.com/capitalone/federated-model-aggregation/tree/main/clients/javascript_client",
"devDependencies": {
"@types/jest": "^27.5.1",
"@types/node": "^17.0.33",
"@typescript-eslint/eslint-plugin": "^5.62.0",
"@typescript-eslint/parser": "^5.62.0",
"eslint": "^8.46.0",
"eslint-config-prettier": "^8.9.0",
"eslint-config-standard-with-typescript": "^37.0.0",
"eslint-formatter-table": "^7.32.1",
"eslint-plugin-import": "^2.28.0",
"eslint-plugin-n": "^16.0.1",
"eslint-plugin-prettier": "^5.0.0",
"eslint-plugin-promise": "^6.1.1",
"husky": "^8.0.3",
"jest": "^28.1.3",
"lint-staged": "^13.2.3",
"prettier": "^3.0.0",
"ts-jest": "^28.0.3",
"typescript": "^5.1.6"
},
"dependencies": {
"@tensorflow/tfjs": "^4.10.0",
"@tensorflow/tfjs-core": "^4.10.0",
"argparse": "^2.0.1",
"fetch-mock": "^9.11.0",
"jest-fetch-mock": "^3.0.3",
"node-fetch": "^2.6.1",
"ts-node": "^10.8.1"
},
"scripts": {
"compile": "rm -rfv dist/lib && tsc && tsc --build tsconfig.es5.json",
"dev": "tsc --watch",
"fix": "prettier --ignore-path .gitignore --write \"**/*.+(js|ts|json)\"",
"format": "prettier --ignore-path .gitignore --check .",
"lint": "eslint fma_connect --ignore-path .eslintignore --ext .ts --format table",
"prepare": "cd ../../ && husky install clients/javascript_client/.husky",
"test": "jest --coverage --passWithNoTests"
}
}
|
0 | capitalone_repos/federated-model-aggregation/clients | capitalone_repos/federated-model-aggregation/clients/javascript_client/.prettierignore | node_modules
dist
coverage
# Ignore all JS files, this is only for TS
**/*.js
|
0 | capitalone_repos/federated-model-aggregation/clients | capitalone_repos/federated-model-aggregation/clients/javascript_client/README.md | # Javascript Client for FMA service
An api client wrapper for connecting to the federated-learning-service.
## Setup
Run `npm setup` in `clients/javascript_client` directory
## Testing
To run the tests of API interface functions run `npm test` in `clients/javascript_client` directory
|
0 | capitalone_repos/federated-model-aggregation/clients | capitalone_repos/federated-model-aggregation/clients/javascript_client/tsconfig.json | {
"compilerOptions": {
"outDir": "./dist/lib/es6",
"target": "ES2020",
"module": "CommonJS",
// Recommended: Compiler complains about expressions implicitly typed as 'any'
"noImplicitAny": true
},
"include": ["fma_connect/**/*"],
"exclude": ["node_modules", "dist", "coverage"]
}
|
0 | capitalone_repos/federated-model-aggregation/clients | capitalone_repos/federated-model-aggregation/clients/javascript_client/.prettierrc | {
"semi": true,
"trailingComma": "all",
"printWidth": 120,
"singleQuote": true,
"arrowParens": "always",
"proseWrap": "preserve",
"quoteProps": "consistent"
}
|
0 | capitalone_repos/federated-model-aggregation/clients | capitalone_repos/federated-model-aggregation/clients/javascript_client/.eslintrc.json | {
"parser": "@typescript-eslint/parser",
"parserOptions": {
"ecmaVersion": 12,
"sourceType": "module"
},
"plugins": ["@typescript-eslint"],
"extends": ["eslint:recommended", "plugin:@typescript-eslint/recommended", "prettier"],
"rules": {
// values can be off, warn, error
"@typescript-eslint/no-unused-vars": "error",
"@typescript-eslint/consistent-type-definitions": ["error", "type"]
},
"env": {
"browser": true,
"es2021": true
}
}
|