Datasets:

GiliGold commited on
Commit
f5d9f50
·
verified ·
1 Parent(s): 0ad2c45

Upload knessetCorpus.py

Browse files
Files changed (1) hide show
  1. knessetCorpus.py +37 -12
knessetCorpus.py CHANGED
@@ -336,11 +336,36 @@ class KnessetCorpus(datasets.GeneratorBasedBuilder):
336
  urls = self.config.data_urls
337
  downloaded_files = dl_manager.download_and_extract(urls)
338
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
339
  return [
340
  datasets.SplitGenerator(
341
  name=datasets.Split.TRAIN,
342
  gen_kwargs={
343
- "data_files": downloaded_files
344
  },
345
  ),
346
  ]
@@ -373,17 +398,17 @@ class KnessetCorpus(datasets.GeneratorBasedBuilder):
373
  sample[feature] = row[feature]
374
 
375
  # Handle nested structures (e.g., `protocol_sentences`)
376
- if "protocol_sentences" in row:
377
- protocol_sentences = row.get("protocol_sentences", [])
378
- for sentence in protocol_sentences:
379
- sentence_id = sentence.get("sentence_id", None)
380
- if sentence_id:
381
- yield sentence_id, sentence
382
- else:
383
- try:
384
- yield id_, sample
385
- except Exception as e:
386
- print(f'couldnt yield sample. error: {e}. sample is: {sample}.', flush=True)
387
  except Exception as e:
388
  print(f"Error opening file '{filepath}': {e}. Skipping.")
389
 
 
336
  urls = self.config.data_urls
337
  downloaded_files = dl_manager.download_and_extract(urls)
338
 
339
+ # Handle single file case
340
+ if not isinstance(downloaded_files, list):
341
+ downloaded_files = [downloaded_files]
342
+
343
+ # Only process if we detect zip files
344
+ if any('zip://' in str(file) for file in downloaded_files):
345
+ extracted_files = []
346
+ for file_path in downloaded_files:
347
+ # Remove the zip:// prefix and download the actual zip content
348
+ actual_url = str(file_path).replace('zip://::', '')
349
+ extracted = dl_manager.download_and_extract(actual_url)
350
+
351
+ # If it's a directory, get all files inside
352
+ if os.path.isdir(extracted):
353
+ for root, _, files in os.walk(extracted):
354
+ for filename in files:
355
+ full_path = os.path.join(root, filename)
356
+ extracted_files.append(full_path)
357
+ else:
358
+ extracted_files.append(extracted)
359
+
360
+ data_files = extracted_files
361
+ else:
362
+ data_files = downloaded_files
363
+
364
  return [
365
  datasets.SplitGenerator(
366
  name=datasets.Split.TRAIN,
367
  gen_kwargs={
368
+ "data_files": data_files,
369
  },
370
  ),
371
  ]
 
398
  sample[feature] = row[feature]
399
 
400
  # Handle nested structures (e.g., `protocol_sentences`)
401
+ # if "protocol_sentences" in row:
402
+ # protocol_sentences = row.get("protocol_sentences", [])
403
+ # for sentence in protocol_sentences:
404
+ # sentence_id = sentence.get("sentence_id", None)
405
+ # if sentence_id:
406
+ # yield sentence_id, sentence
407
+ # else:
408
+ try:
409
+ yield id_, sample
410
+ except Exception as e:
411
+ print(f'couldnt yield sample. error: {e}. sample is: {sample}.', flush=True)
412
  except Exception as e:
413
  print(f"Error opening file '{filepath}': {e}. Skipping.")
414