Fix Essay from grader's data (#3)
Browse files- update essay origin (3a1043ed0f2d921f2dd5de07994d0c76f1d73ca4)
- update dataset (1c97a2b405be5dc5c2edc4f2b57ff0b95a5f971f)
- aes_enem_dataset.py +12 -12
- propor2024.tar.gz +0 -0
aes_enem_dataset.py
CHANGED
|
@@ -340,14 +340,14 @@ class AesEnemDataset(datasets.GeneratorBasedBuilder):
|
|
| 340 |
if self.config.name == "sourceAWithGraders":
|
| 341 |
grader_a, grader_b = self._parse_graders_data(dirname)
|
| 342 |
grader_a_data = pd.merge(
|
| 343 |
-
train_df[["id", "id_prompt"]],
|
| 344 |
-
grader_a,
|
| 345 |
on=["id", "id_prompt"],
|
| 346 |
how="inner",
|
| 347 |
)
|
| 348 |
grader_b_data = pd.merge(
|
| 349 |
-
train_df[["id", "id_prompt"]],
|
| 350 |
-
grader_b,
|
| 351 |
on=["id", "id_prompt"],
|
| 352 |
how="inner",
|
| 353 |
)
|
|
@@ -355,14 +355,14 @@ class AesEnemDataset(datasets.GeneratorBasedBuilder):
|
|
| 355 |
train_df = pd.concat([train_df, grader_b_data])
|
| 356 |
|
| 357 |
grader_a_data = pd.merge(
|
| 358 |
-
val_df[["id", "id_prompt"]],
|
| 359 |
-
grader_a,
|
| 360 |
on=["id", "id_prompt"],
|
| 361 |
how="inner",
|
| 362 |
)
|
| 363 |
grader_b_data = pd.merge(
|
| 364 |
-
val_df[["id", "id_prompt"]],
|
| 365 |
-
grader_b,
|
| 366 |
on=["id", "id_prompt"],
|
| 367 |
how="inner",
|
| 368 |
)
|
|
@@ -370,14 +370,14 @@ class AesEnemDataset(datasets.GeneratorBasedBuilder):
|
|
| 370 |
val_df = pd.concat([val_df, grader_b_data])
|
| 371 |
|
| 372 |
grader_a_data = pd.merge(
|
| 373 |
-
test_df[["id", "id_prompt"]],
|
| 374 |
-
grader_a,
|
| 375 |
on=["id", "id_prompt"],
|
| 376 |
how="inner",
|
| 377 |
)
|
| 378 |
grader_b_data = pd.merge(
|
| 379 |
-
test_df[["id", "id_prompt"]],
|
| 380 |
-
grader_b,
|
| 381 |
on=["id", "id_prompt"],
|
| 382 |
how="inner",
|
| 383 |
)
|
|
|
|
| 340 |
if self.config.name == "sourceAWithGraders":
|
| 341 |
grader_a, grader_b = self._parse_graders_data(dirname)
|
| 342 |
grader_a_data = pd.merge(
|
| 343 |
+
train_df[["id", "id_prompt","essay"]],
|
| 344 |
+
grader_a.drop(columns=['essay']),
|
| 345 |
on=["id", "id_prompt"],
|
| 346 |
how="inner",
|
| 347 |
)
|
| 348 |
grader_b_data = pd.merge(
|
| 349 |
+
train_df[["id", "id_prompt","essay"]],
|
| 350 |
+
grader_b.drop(columns=['essay']),
|
| 351 |
on=["id", "id_prompt"],
|
| 352 |
how="inner",
|
| 353 |
)
|
|
|
|
| 355 |
train_df = pd.concat([train_df, grader_b_data])
|
| 356 |
|
| 357 |
grader_a_data = pd.merge(
|
| 358 |
+
val_df[["id", "id_prompt","essay"]],
|
| 359 |
+
grader_a.drop(columns=['essay']),
|
| 360 |
on=["id", "id_prompt"],
|
| 361 |
how="inner",
|
| 362 |
)
|
| 363 |
grader_b_data = pd.merge(
|
| 364 |
+
val_df[["id", "id_prompt","essay"]],
|
| 365 |
+
grader_b.drop(columns=['essay']),
|
| 366 |
on=["id", "id_prompt"],
|
| 367 |
how="inner",
|
| 368 |
)
|
|
|
|
| 370 |
val_df = pd.concat([val_df, grader_b_data])
|
| 371 |
|
| 372 |
grader_a_data = pd.merge(
|
| 373 |
+
test_df[["id", "id_prompt","essay"]],
|
| 374 |
+
grader_a.drop(columns=['essay']),
|
| 375 |
on=["id", "id_prompt"],
|
| 376 |
how="inner",
|
| 377 |
)
|
| 378 |
grader_b_data = pd.merge(
|
| 379 |
+
test_df[["id", "id_prompt","essay"]],
|
| 380 |
+
grader_b.drop(columns=['essay']),
|
| 381 |
on=["id", "id_prompt"],
|
| 382 |
how="inner",
|
| 383 |
)
|
propor2024.tar.gz
CHANGED
|
Binary files a/propor2024.tar.gz and b/propor2024.tar.gz differ
|
|
|