upgini 1.2.38a3769.dev2__tar.gz → 1.2.38a3769.dev3__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Potentially problematic release.

This version of upgini might be problematic. Click here for more details.

Files changed (67) hide show
  1. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/PKG-INFO +1 -1
  2. upgini-1.2.38a3769.dev3/src/upgini/__about__.py +1 -0
  3. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/target_utils.py +3 -3
  4. upgini-1.2.38a3769.dev2/src/upgini/__about__.py +0 -1
  5. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/.gitignore +0 -0
  6. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/LICENSE +0 -0
  7. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/README.md +0 -0
  8. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/pyproject.toml +0 -0
  9. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/__init__.py +0 -0
  10. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/ads.py +0 -0
  11. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/ads_management/__init__.py +0 -0
  12. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/ads_management/ads_manager.py +0 -0
  13. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/autofe/__init__.py +0 -0
  14. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/autofe/all_operands.py +0 -0
  15. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/autofe/binary.py +0 -0
  16. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/autofe/date.py +0 -0
  17. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/autofe/feature.py +0 -0
  18. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/autofe/groupby.py +0 -0
  19. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/autofe/operand.py +0 -0
  20. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/autofe/unary.py +0 -0
  21. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/autofe/vector.py +0 -0
  22. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/data_source/__init__.py +0 -0
  23. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/data_source/data_source_publisher.py +0 -0
  24. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/dataset.py +0 -0
  25. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/errors.py +0 -0
  26. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/features_enricher.py +0 -0
  27. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/http.py +0 -0
  28. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/lazy_import.py +0 -0
  29. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/mdc/__init__.py +0 -0
  30. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/mdc/context.py +0 -0
  31. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/metadata.py +0 -0
  32. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/metrics.py +0 -0
  33. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/normalizer/__init__.py +0 -0
  34. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/normalizer/normalize_utils.py +0 -0
  35. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/resource_bundle/__init__.py +0 -0
  36. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/resource_bundle/exceptions.py +0 -0
  37. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/resource_bundle/strings.properties +0 -0
  38. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/resource_bundle/strings_widget.properties +0 -0
  39. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/sampler/__init__.py +0 -0
  40. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/sampler/base.py +0 -0
  41. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/sampler/random_under_sampler.py +0 -0
  42. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/sampler/utils.py +0 -0
  43. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/search_task.py +0 -0
  44. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/spinner.py +0 -0
  45. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/Roboto-Regular.ttf +0 -0
  46. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/__init__.py +0 -0
  47. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/base_search_key_detector.py +0 -0
  48. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/blocked_time_series.py +0 -0
  49. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/country_utils.py +0 -0
  50. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/custom_loss_utils.py +0 -0
  51. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/cv_utils.py +0 -0
  52. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/datetime_utils.py +0 -0
  53. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/deduplicate_utils.py +0 -0
  54. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/display_utils.py +0 -0
  55. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/email_utils.py +0 -0
  56. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/fallback_progress_bar.py +0 -0
  57. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/feature_info.py +0 -0
  58. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/features_validator.py +0 -0
  59. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/format.py +0 -0
  60. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/ip_utils.py +0 -0
  61. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/phone_utils.py +0 -0
  62. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/postal_code_utils.py +0 -0
  63. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/progress_bar.py +0 -0
  64. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/sklearn_ext.py +0 -0
  65. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/track_info.py +0 -0
  66. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/warning_counter.py +0 -0
  67. {upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/version_validator.py +0 -0
{upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/PKG-INFO RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.3
2
2
  Name: upgini
3
- Version: 1.2.38a3769.dev2
3
+ Version: 1.2.38a3769.dev3
4
4
  Summary: Intelligent data search & enrichment for Machine Learning
5
5
  Project-URL: Bug Reports, https://github.com/upgini/upgini/issues
6
6
  Project-URL: Homepage, https://upgini.com/
upgini-1.2.38a3769.dev3/src/upgini/__about__.py ADDED
@@ -0,0 +1 @@
1
+ __version__ = "1.2.38a3769.dev3"
{upgini-1.2.38a3769.dev2 → upgini-1.2.38a3769.dev3}/src/upgini/utils/target_utils.py RENAMED
@@ -306,7 +306,7 @@ def balance_undersample_time_series(
306
306
  id_counts.index = [ensure_tuple(i) for i in id_counts.index]
307
307
  id_counts = id_counts.sort_index(key=lambda x: [ids_sort[y] for y in x], ascending=False).cumsum()
308
308
  id_counts = id_counts[id_counts <= sample_size]
309
- min_different_ids = int(len(df[id_columns].drop_duplicates()) * min_different_ids_ratio)
309
+ min_different_ids = max(int(len(df[id_columns].drop_duplicates()) * min_different_ids_ratio), 1)
310
310
 
311
311
  def id_mask(sample_index: pd.Index) -> pd.Index:
312
312
  if isinstance(sample_index, pd.MultiIndex):
@@ -317,10 +317,10 @@ def balance_undersample_time_series(
317
317
  if len(id_counts) < min_different_ids:
318
318
  if logger is not None:
319
319
  logger.info(
320
- f"Different ids count {len(id_counts)} is less than min different ids {min_different_ids}, sampling time window"
320
+ f"Different ids count {len(id_counts)} for sample size {sample_size} is less than min different ids {min_different_ids}, sampling time window"
321
321
  )
322
322
  date_counts = df.groupby(id_columns)[date_column].nunique().sort_values(ascending=False)
323
- ids_to_sample = date_counts.index[:min_different_ids]
323
+ ids_to_sample = date_counts.index[:min_different_ids] if len(id_counts) > 0 else date_counts.index
324
324
  mask = id_mask(ids_to_sample)
325
325
  df = df[mask]
326
326
  sample_date_counts = df[date_column].value_counts().sort_index(ascending=False).cumsum()
upgini-1.2.38a3769.dev2/src/upgini/__about__.py DELETED
@@ -1 +0,0 @@
1
- __version__ = "1.2.38a3769.dev2"