Marcio Monteiro
commited on
Commit
•
9288bfa
1
Parent(s):
94a47f5
fix: removing train/test split (leaving this responsiblity to the data loader)
Browse files- README.md +0 -4
- convert.py +8 -27
- data/all/.~lock.train.csv# +0 -1
- data/all/test.csv +0 -3
- data/all/train.csv +2 -2
- data/{top4 → top4-balanced}/train.csv +2 -2
- data/top4/test.csv +0 -3
README.md
CHANGED
@@ -9,15 +9,11 @@ configs:
|
|
9 |
data_files:
|
10 |
- split: train
|
11 |
path: "data/all/train.csv"
|
12 |
-
- split: test
|
13 |
-
path: "data/all/test.csv"
|
14 |
default: true
|
15 |
- config_name: top4
|
16 |
data_files:
|
17 |
- split: train
|
18 |
path: "data/top4/train.csv"
|
19 |
-
- split: test
|
20 |
-
path: "data/top4/test.csv"
|
21 |
|
22 |
dataset_info:
|
23 |
features:
|
|
|
9 |
data_files:
|
10 |
- split: train
|
11 |
path: "data/all/train.csv"
|
|
|
|
|
12 |
default: true
|
13 |
- config_name: top4
|
14 |
data_files:
|
15 |
- split: train
|
16 |
path: "data/top4/train.csv"
|
|
|
|
|
17 |
|
18 |
dataset_info:
|
19 |
features:
|
convert.py
CHANGED
@@ -15,7 +15,6 @@ import pandas as pd
|
|
15 |
|
16 |
from bs4 import BeautifulSoup
|
17 |
|
18 |
-
from sklearn.model_selection import train_test_split
|
19 |
|
20 |
HEADER = [
|
21 |
"source",
|
@@ -28,10 +27,7 @@ HEADER = [
|
|
28 |
"pubdate",
|
29 |
]
|
30 |
|
31 |
-
|
32 |
-
TRAIN_OUTPUT_FILE = "train.csv"
|
33 |
-
TEST_OUTPUT_FILE = "test.csv"
|
34 |
-
|
35 |
|
36 |
def _clean_text(text):
|
37 |
text = text.replace("\\\n", "\n")
|
@@ -127,15 +123,15 @@ def run():
|
|
127 |
row["label"] = _categories.index(row["category"])
|
128 |
|
129 |
save_csv(rows)
|
130 |
-
|
131 |
-
split_csv_train_test(test_size=0.2, random_state=42)
|
132 |
-
filter_and_split_csv(["World", "Sports", "Business", "Sci/Tech"], "top4")
|
133 |
|
134 |
|
135 |
-
def save_csv(rows, fname=
|
136 |
"""
|
137 |
Save the processed data into a CSV file.
|
138 |
"""
|
|
|
|
|
139 |
with open(fname, "w", encoding="utf8") as f:
|
140 |
writer = csv.DictWriter(f, fieldnames=rows[0].keys())
|
141 |
writer.writeheader()
|
@@ -144,26 +140,13 @@ def save_csv(rows, fname=OUTPUT_FILE):
|
|
144 |
writer.writerow(row)
|
145 |
|
146 |
|
147 |
-
def
|
148 |
-
"""
|
149 |
-
Split the data into training and testing sets.
|
150 |
-
"""
|
151 |
-
df = pd.read_csv(OUTPUT_FILE)
|
152 |
-
train_df, test_df = train_test_split(df, **kwargs)
|
153 |
-
|
154 |
-
os.makedirs("data/all", exist_ok=True)
|
155 |
-
|
156 |
-
train_df.to_csv(os.path.join("data", "all", TRAIN_OUTPUT_FILE), index=False)
|
157 |
-
test_df.to_csv(os.path.join("data", "all", TEST_OUTPUT_FILE), index=False)
|
158 |
-
|
159 |
-
|
160 |
-
def filter_and_split_csv(categories, config_name, is_balanced=True, **kwargs):
|
161 |
"""
|
162 |
Filter the data by categories and split the data into training and testing
|
163 |
sets. If is_balanced is True, the data will be balanced to size of the
|
164 |
class with fewer examples.
|
165 |
"""
|
166 |
-
df = pd.read_csv(
|
167 |
|
168 |
if is_balanced:
|
169 |
dfs = []
|
@@ -180,9 +163,7 @@ def filter_and_split_csv(categories, config_name, is_balanced=True, **kwargs):
|
|
180 |
|
181 |
os.makedirs(f"data/{config_name}", exist_ok=True)
|
182 |
|
183 |
-
|
184 |
-
train_df.to_csv(os.path.join("data", config_name, TRAIN_OUTPUT_FILE), index=False)
|
185 |
-
test_df.to_csv(os.path.join("data", config_name, TEST_OUTPUT_FILE), index=False)
|
186 |
|
187 |
|
188 |
def save_categories(categories, fname="categories.txt"):
|
|
|
15 |
|
16 |
from bs4 import BeautifulSoup
|
17 |
|
|
|
18 |
|
19 |
HEADER = [
|
20 |
"source",
|
|
|
27 |
"pubdate",
|
28 |
]
|
29 |
|
30 |
+
OUTPUT_FILE_PATH = os.path.join("data", "all", "train.csv")
|
|
|
|
|
|
|
31 |
|
32 |
def _clean_text(text):
|
33 |
text = text.replace("\\\n", "\n")
|
|
|
123 |
row["label"] = _categories.index(row["category"])
|
124 |
|
125 |
save_csv(rows)
|
126 |
+
save_csv_categories(["World", "Sports", "Business", "Sci/Tech"], "top4-balanced", is_balanced=True)
|
|
|
|
|
127 |
|
128 |
|
129 |
+
def save_csv(rows, fname=OUTPUT_FILE_PATH):
|
130 |
"""
|
131 |
Save the processed data into a CSV file.
|
132 |
"""
|
133 |
+
os.makedirs(os.path.join("data", "all"), exist_ok=True)
|
134 |
+
|
135 |
with open(fname, "w", encoding="utf8") as f:
|
136 |
writer = csv.DictWriter(f, fieldnames=rows[0].keys())
|
137 |
writer.writeheader()
|
|
|
140 |
writer.writerow(row)
|
141 |
|
142 |
|
143 |
+
def save_csv_categories(categories, config_name, is_balanced=True, **kwargs):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
144 |
"""
|
145 |
Filter the data by categories and split the data into training and testing
|
146 |
sets. If is_balanced is True, the data will be balanced to size of the
|
147 |
class with fewer examples.
|
148 |
"""
|
149 |
+
df = pd.read_csv(OUTPUT_FILE_PATH)
|
150 |
|
151 |
if is_balanced:
|
152 |
dfs = []
|
|
|
163 |
|
164 |
os.makedirs(f"data/{config_name}", exist_ok=True)
|
165 |
|
166 |
+
df.to_csv(os.path.join("data", config_name, "train.csv"), index=False)
|
|
|
|
|
167 |
|
168 |
|
169 |
def save_categories(categories, fname="categories.txt"):
|
data/all/.~lock.train.csv#
DELETED
@@ -1 +0,0 @@
|
|
1 |
-
,marcio,marcio-AX370-Gaming,10.04.2024 19:56,file:///home/marcio/.config/libreoffice/4;
|
|
|
|
data/all/test.csv
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:fb60f014952b4e1ae53fb3788e3e2b3a76e927ebf65931d87f2f8bcf5d22ad73
|
3 |
-
size 168573267
|
|
|
|
|
|
|
|
data/all/train.csv
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8e3bc4a1a85a56f46327a32c22f0661496b2db7cb18eefa689c9e6c06ea93518
|
3 |
+
size 843929665
|
data/{top4 → top4-balanced}/train.csv
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4a9a5712dd5bf47810bc6fc7df31b86a6eaba594cc34353efcf02fd7f6289771
|
3 |
+
size 344070837
|
data/top4/test.csv
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:8fb2c8e2218bd3e976e147bbefc3bfa6b5575bb29966910a7d835bd1e5cd5d6a
|
3 |
-
size 85898679
|
|
|
|
|
|
|
|