lovodkin93
commited on
Commit
•
d0542df
1
Parent(s):
c99e5ac
Controlled-Text-Reduction-dataset.py
CHANGED
@@ -392,31 +392,31 @@ class ControlledTectReduction(datasets.GeneratorBasedBuilder):
|
|
392 |
name=datasets.Split.TRAIN,
|
393 |
# These kwargs will be passed to _generate_examples
|
394 |
gen_kwargs={
|
395 |
-
"
|
396 |
},
|
397 |
),
|
398 |
datasets.SplitGenerator(
|
399 |
name=datasets.Split.VALIDATION,
|
400 |
# These kwargs will be passed to _generate_examples
|
401 |
gen_kwargs={
|
402 |
-
"
|
403 |
},
|
404 |
),
|
405 |
datasets.SplitGenerator(
|
406 |
name=datasets.Split.TEST,
|
407 |
# These kwargs will be passed to _generate_examples
|
408 |
gen_kwargs={
|
409 |
-
"
|
410 |
},
|
411 |
),
|
412 |
]
|
413 |
|
414 |
-
def _generate_examples(self,
|
415 |
|
416 |
""" Yields Controlled Text Reduction examples from a csv file. Each instance contains the document, the summary and the pre-selected spans."""
|
417 |
|
418 |
# merge annotations from sections
|
419 |
-
df = pd.
|
420 |
for counter, dic in enumerate(df.to_dict('records')):
|
421 |
columns_to_load_into_object = ["doc_text", "summary_text", "highlight_spans"]
|
422 |
# for key in columns_to_load_into_object:
|
|
|
392 |
name=datasets.Split.TRAIN,
|
393 |
# These kwargs will be passed to _generate_examples
|
394 |
gen_kwargs={
|
395 |
+
"filepath": corpora["train_DUC-2001-2002"],
|
396 |
},
|
397 |
),
|
398 |
datasets.SplitGenerator(
|
399 |
name=datasets.Split.VALIDATION,
|
400 |
# These kwargs will be passed to _generate_examples
|
401 |
gen_kwargs={
|
402 |
+
"filepath": corpora["dev_DUC-2001-2002"],
|
403 |
},
|
404 |
),
|
405 |
datasets.SplitGenerator(
|
406 |
name=datasets.Split.TEST,
|
407 |
# These kwargs will be passed to _generate_examples
|
408 |
gen_kwargs={
|
409 |
+
"filepath": corpora["test_DUC-2001-2002"],
|
410 |
},
|
411 |
),
|
412 |
]
|
413 |
|
414 |
+
def _generate_examples(self, filepath: List[str]):
|
415 |
|
416 |
""" Yields Controlled Text Reduction examples from a csv file. Each instance contains the document, the summary and the pre-selected spans."""
|
417 |
|
418 |
# merge annotations from sections
|
419 |
+
df = pd.read_csv(filepath)
|
420 |
for counter, dic in enumerate(df.to_dict('records')):
|
421 |
columns_to_load_into_object = ["doc_text", "summary_text", "highlight_spans"]
|
422 |
# for key in columns_to_load_into_object:
|