holylovenia
commited on
Commit
•
993a88b
1
Parent(s):
10e6e92
Upload idk_mrc.py with huggingface_hub
Browse files- idk_mrc.py +17 -16
idk_mrc.py
CHANGED
@@ -21,9 +21,9 @@ from typing import Dict, List, Tuple
|
|
21 |
|
22 |
import datasets
|
23 |
|
24 |
-
from
|
25 |
-
from
|
26 |
-
from
|
27 |
|
28 |
_CITATION = """\
|
29 |
@misc{putri2022idk,
|
@@ -78,23 +78,23 @@ _SUPPORTED_TASKS = [Tasks.QUESTION_ANSWERING]
|
|
78 |
|
79 |
_SOURCE_VERSION = "1.0.0"
|
80 |
|
81 |
-
|
82 |
|
83 |
|
84 |
-
def
|
85 |
"""
|
86 |
-
Construct
|
87 |
idk_mrc_baseline_{name}_{schema} format for the baseline datasets.
|
88 |
Suported dataset names: see _ALL_DATASETS
|
89 |
"""
|
90 |
-
if schema != "source" and schema != "
|
91 |
raise ValueError(f"Invalid schema: {schema}")
|
92 |
|
93 |
if name not in _ALL_DATASETS:
|
94 |
raise ValueError(f"Invalid dataset name: {name}")
|
95 |
|
96 |
if name == "idk_mrc":
|
97 |
-
return
|
98 |
name="idk_mrc_{schema}".format(schema=schema),
|
99 |
version=datasets.Version(version),
|
100 |
description="IDK-MRC with {schema} schema".format(schema=schema),
|
@@ -102,7 +102,7 @@ def nusantara_config_constructor(name, schema, version):
|
|
102 |
subset_id="idk_mrc",
|
103 |
)
|
104 |
else:
|
105 |
-
return
|
106 |
name="idk_mrc_baseline_{name}_{schema}".format(name=name, schema=schema),
|
107 |
version=datasets.Version(version),
|
108 |
description="IDK-MRC baseline ({name}) with {schema} schema".format(name=name, schema=schema),
|
@@ -115,11 +115,11 @@ class IdkMrc(datasets.GeneratorBasedBuilder):
|
|
115 |
"""IDK-MRC is an Indonesian MRC dataset that covers answerable and unanswerable questions"""
|
116 |
|
117 |
SOURCE_VERSION = datasets.Version(_SOURCE_VERSION)
|
118 |
-
|
119 |
|
120 |
BUILDER_CONFIGS = [
|
121 |
-
|
122 |
-
for name in _ALL_DATASETS for schema, version in zip(["source", "
|
123 |
]
|
124 |
|
125 |
DEFAULT_CONFIG_NAME = "idk_mrc_source"
|
@@ -145,7 +145,7 @@ class IdkMrc(datasets.GeneratorBasedBuilder):
|
|
145 |
}
|
146 |
)
|
147 |
|
148 |
-
elif self.config.schema == "
|
149 |
features = schemas.qa_features
|
150 |
|
151 |
return datasets.DatasetInfo(
|
@@ -158,14 +158,14 @@ class IdkMrc(datasets.GeneratorBasedBuilder):
|
|
158 |
|
159 |
def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]:
|
160 |
"""Returns SplitGenerators."""
|
161 |
-
if self.config.name == "idk_mrc_source" or self.config.name == "
|
162 |
data_name = "idk_mrc"
|
163 |
train_data_path = dl_manager.download_and_extract(_URLS[_DATASETNAME]["train"])
|
164 |
validation_data_path = dl_manager.download_and_extract(_URLS[_DATASETNAME]["validation"])
|
165 |
test_data_path = dl_manager.download_and_extract(_URLS[_DATASETNAME]["test"])
|
166 |
else:
|
167 |
try:
|
168 |
-
data_name = re.search("baseline_(.+?)_(source|
|
169 |
except AttributeError:
|
170 |
raise ValueError(f"Invalid config name: {self.config.name}")
|
171 |
|
@@ -215,7 +215,7 @@ class IdkMrc(datasets.GeneratorBasedBuilder):
|
|
215 |
for key, example in enumerate(examples):
|
216 |
yield key, example
|
217 |
|
218 |
-
elif self.config.schema == "
|
219 |
for key, example in enumerate(examples):
|
220 |
for qa in example["qas"]:
|
221 |
# Use question ID as key
|
@@ -228,4 +228,5 @@ class IdkMrc(datasets.GeneratorBasedBuilder):
|
|
228 |
"choices": [],
|
229 |
"context": example["context"],
|
230 |
"answer": [ans["text"] for ans in qa["answers"]],
|
|
|
231 |
}
|
|
|
21 |
|
22 |
import datasets
|
23 |
|
24 |
+
from seacrowd.utils import schemas
|
25 |
+
from seacrowd.utils.configs import SEACrowdConfig
|
26 |
+
from seacrowd.utils.constants import Tasks
|
27 |
|
28 |
_CITATION = """\
|
29 |
@misc{putri2022idk,
|
|
|
78 |
|
79 |
_SOURCE_VERSION = "1.0.0"
|
80 |
|
81 |
+
_SEACROWD_VERSION = "2024.06.20"
|
82 |
|
83 |
|
84 |
+
def seacrowd_config_constructor(name, schema, version):
|
85 |
"""
|
86 |
+
Construct SEACrowdConfig with idk_mrc_{schema} format for the main dataset &
|
87 |
idk_mrc_baseline_{name}_{schema} format for the baseline datasets.
|
88 |
Suported dataset names: see _ALL_DATASETS
|
89 |
"""
|
90 |
+
if schema != "source" and schema != "seacrowd_qa":
|
91 |
raise ValueError(f"Invalid schema: {schema}")
|
92 |
|
93 |
if name not in _ALL_DATASETS:
|
94 |
raise ValueError(f"Invalid dataset name: {name}")
|
95 |
|
96 |
if name == "idk_mrc":
|
97 |
+
return SEACrowdConfig(
|
98 |
name="idk_mrc_{schema}".format(schema=schema),
|
99 |
version=datasets.Version(version),
|
100 |
description="IDK-MRC with {schema} schema".format(schema=schema),
|
|
|
102 |
subset_id="idk_mrc",
|
103 |
)
|
104 |
else:
|
105 |
+
return SEACrowdConfig(
|
106 |
name="idk_mrc_baseline_{name}_{schema}".format(name=name, schema=schema),
|
107 |
version=datasets.Version(version),
|
108 |
description="IDK-MRC baseline ({name}) with {schema} schema".format(name=name, schema=schema),
|
|
|
115 |
"""IDK-MRC is an Indonesian MRC dataset that covers answerable and unanswerable questions"""
|
116 |
|
117 |
SOURCE_VERSION = datasets.Version(_SOURCE_VERSION)
|
118 |
+
SEACROWD_VERSION = datasets.Version(_SEACROWD_VERSION)
|
119 |
|
120 |
BUILDER_CONFIGS = [
|
121 |
+
seacrowd_config_constructor(name, schema, version)
|
122 |
+
for name in _ALL_DATASETS for schema, version in zip(["source", "seacrowd_qa"], [_SOURCE_VERSION, _SEACROWD_VERSION])
|
123 |
]
|
124 |
|
125 |
DEFAULT_CONFIG_NAME = "idk_mrc_source"
|
|
|
145 |
}
|
146 |
)
|
147 |
|
148 |
+
elif self.config.schema == "seacrowd_qa":
|
149 |
features = schemas.qa_features
|
150 |
|
151 |
return datasets.DatasetInfo(
|
|
|
158 |
|
159 |
def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]:
|
160 |
"""Returns SplitGenerators."""
|
161 |
+
if self.config.name == "idk_mrc_source" or self.config.name == "idk_mrc_seacrowd_qa":
|
162 |
data_name = "idk_mrc"
|
163 |
train_data_path = dl_manager.download_and_extract(_URLS[_DATASETNAME]["train"])
|
164 |
validation_data_path = dl_manager.download_and_extract(_URLS[_DATASETNAME]["validation"])
|
165 |
test_data_path = dl_manager.download_and_extract(_URLS[_DATASETNAME]["test"])
|
166 |
else:
|
167 |
try:
|
168 |
+
data_name = re.search("baseline_(.+?)_(source|seacrowd_qa)", self.config.name).group(1)
|
169 |
except AttributeError:
|
170 |
raise ValueError(f"Invalid config name: {self.config.name}")
|
171 |
|
|
|
215 |
for key, example in enumerate(examples):
|
216 |
yield key, example
|
217 |
|
218 |
+
elif self.config.schema == "seacrowd_qa":
|
219 |
for key, example in enumerate(examples):
|
220 |
for qa in example["qas"]:
|
221 |
# Use question ID as key
|
|
|
228 |
"choices": [],
|
229 |
"context": example["context"],
|
230 |
"answer": [ans["text"] for ans in qa["answers"]],
|
231 |
+
"meta": {}
|
232 |
}
|