medpipe 0.2.0.dev0__tar.gz → 1.0.0__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 medpipe might be problematic. Click here for more details.

Files changed (49) hide show
  1. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/PKG-INFO +1 -1
  2. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/pyproject.toml +1 -1
  3. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/__init__.py +1 -1
  4. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/data/sampler.py +85 -54
  5. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/metrics/core.py +4 -4
  6. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/pipeline/pipeline.py +16 -5
  7. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe.egg-info/PKG-INFO +1 -1
  8. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/test/test_data_sampler.py +132 -47
  9. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/LICENSE +0 -0
  10. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/README.md +0 -0
  11. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/requirements.txt +0 -0
  12. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/setup.cfg +0 -0
  13. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/_types.py +0 -0
  14. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/data/__init__.py +0 -0
  15. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/data/db.py +0 -0
  16. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/data/preprocessing.py +0 -0
  17. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/data/preprocessor.py +0 -0
  18. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/data/utils.py +0 -0
  19. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/data/weighting.py +0 -0
  20. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/metrics/__init__.py +0 -0
  21. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/metrics/plots.py +0 -0
  22. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/models/__init__.py +0 -0
  23. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/models/calibrators.py +0 -0
  24. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/models/core.py +0 -0
  25. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/models/predictors.py +0 -0
  26. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/pipeline/__init__.py +0 -0
  27. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/py.typed +0 -0
  28. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/utils/__init__.py +0 -0
  29. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/utils/config.py +0 -0
  30. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/utils/exceptions.py +0 -0
  31. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/utils/io.py +0 -0
  32. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe/utils/logger.py +0 -0
  33. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe.egg-info/SOURCES.txt +0 -0
  34. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe.egg-info/dependency_links.txt +0 -0
  35. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe.egg-info/requires.txt +0 -0
  36. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/src/medpipe.egg-info/top_level.txt +0 -0
  37. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/test/test_data_db.py +0 -0
  38. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/test/test_data_preprocessing.py +0 -0
  39. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/test/test_data_preprocessor.py +0 -0
  40. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/test/test_data_utils.py +0 -0
  41. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/test/test_data_weighting.py +0 -0
  42. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/test/test_models_calibrators.py +0 -0
  43. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/test/test_models_core.py +0 -0
  44. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/test/test_models_predictors.py +0 -0
  45. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/test/test_pipeline_pipeline.py +0 -0
  46. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/test/test_utils_config.py +0 -0
  47. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/test/test_utils_exceptions.py +0 -0
  48. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/test/test_utils_io.py +0 -0
  49. {medpipe-0.2.0.dev0 → medpipe-1.0.0}/test/test_utils_logger.py +0 -0
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: medpipe
3
- Version: 0.2.0.dev0
3
+ Version: 1.0.0
4
4
  Summary: A package for clinical AI pipeline development
5
5
  Author-email: Mathias Roesler <mathias.roesler@auckland.co.nz>
6
6
  License-Expression: Apache-2.0
@@ -4,7 +4,7 @@ requires = [ "setuptools>=80.9" ]
4
4
 
5
5
  [project]
6
6
  name = "medpipe"
7
- version = "0.2.0.dev0"
7
+ version = "1.0.0"
8
8
  description = "A package for clinical AI pipeline development"
9
9
  readme = "README.md"
10
10
  license = "Apache-2.0"
@@ -52,4 +52,4 @@ __all__ = [
52
52
  "extract_data_from_db",
53
53
  "extract_labels",
54
54
  ]
55
- __version__ = "v0.2.0.dev0"
55
+ __version__ = "v1.0.0"
@@ -44,7 +44,7 @@ if TYPE_CHECKING:
44
44
  def data_sampler(
45
45
  data: PredData,
46
46
  labels: Labels,
47
- target_ratio: float = 0.25,
47
+ reduction_factor: float = 0.25,
48
48
  sampler_fn: str = "random_undersampler",
49
49
  groups: npt.NDArray = np.array([]),
50
50
  **kwargs: Any,
@@ -54,9 +54,10 @@ def data_sampler(
54
54
 
55
55
  The majority class is assumed to have a False or 0 label.
56
56
  The new set will have an imbalance equal to:
57
- IR * target_ratio, where IR is the current imbalance ratio.
57
+ IR - IR * reduction_factor, where IR is the current imbalance ratio.
58
+ or a IR = 1, if reduction factor is 1.
58
59
 
59
- If the target ratio is too small, the algorithm defaults to
60
+ If the reduction factor is too small, the algorithm defaults to
60
61
  obtain a balanced dataset.
61
62
 
62
63
  Parameters
@@ -65,8 +66,8 @@ def data_sampler(
65
66
  Data to sample of shape (n_samples, n_features).
66
67
  labels : Labels
67
68
  Binary prediction labels of shape (n_samples, n_classes).
68
- target_ratio : float, default: 0.25
69
- Target ratio between the minority and majority classes.
69
+ reduction_factor : float, default: 0.25
70
+ Reduction factor to apply to the current imbalance ratio
70
71
  sampler_fn : str, default: "random_undersampler"
71
72
  Sampler function to use to sample the data.
72
73
  groups : npt.NDArray, default: np.array([])
@@ -88,25 +89,36 @@ def data_sampler(
88
89
  TypeError
89
90
  If labels is not npt.NDArray.
90
91
  ValueError
91
- If target_ratio is less than 0.0.
92
+ If reduction_factor is less than 0.0 or greater than 1.0
92
93
 
93
94
  """
94
- if target_ratio < 0:
95
- raise ValueError(f"Target ratio should be positive, but got {target_ratio}")
95
+ if reduction_factor < 0:
96
+ raise ValueError(
97
+ f"Reduction factor should be positive, but got {reduction_factor}"
98
+ )
99
+ if reduction_factor > 1.0:
100
+ raise ValueError(
101
+ f"Reduction factor should be less than 1, but got {reduction_factor}"
102
+ )
103
+ array_check(labels)
96
104
 
97
- # Calculate new_ratio once
105
+ # Calculate target_ratio
98
106
  n_samples = len(labels)
99
- n_pos = np.sum(labels)
107
+ n_pos = np.sum(np.any(labels != 0, axis=1)) # Account for multilabel
100
108
 
101
- if target_ratio == 0 or n_pos == 0:
102
- new_ratio = 1.0
109
+ if reduction_factor == 0 or n_pos == 0:
110
+ return data, labels, groups # Exit early
111
+ elif reduction_factor == 1:
112
+ target_ratio = 1
103
113
  else:
104
114
  imbalance_ratio = (n_samples - n_pos) / n_pos
105
- new_ratio = max(1.0, 1 / (imbalance_ratio * target_ratio))
115
+ target_ratio = imbalance_ratio - imbalance_ratio * reduction_factor
116
+
117
+ target_ratio = 1 if target_ratio < 1 else target_ratio
106
118
 
107
119
  # Dispatch logic
108
120
  if sampler_fn == "smote":
109
- X_gen, y_gen = smote(data, labels, new_ratio, kwargs.get("k_neighbors", 5))
121
+ X_gen, y_gen = smote(data, labels, target_ratio, kwargs.get("k_neighbors", 5))
110
122
  if isinstance(data, DataFrame):
111
123
  return (
112
124
  concat((data, cast(DataFrame, X_gen))),
@@ -121,24 +133,24 @@ def data_sampler(
121
133
 
122
134
  if sampler_fn == "group_smote":
123
135
  return group_smote(
124
- data, labels, new_ratio, groups, kwargs.get("k_neighbors", 5)
136
+ data, labels, target_ratio, groups, kwargs.get("k_neighbors", 5)
125
137
  )
126
138
 
127
139
  # Index-based samplers
128
140
  sampler_map = {
129
- "random_undersampler": lambda: random_undersampler(labels, new_ratio),
141
+ "random_undersampler": lambda: random_undersampler(labels, target_ratio),
130
142
  "group_random_undersampler": lambda: group_random_undersampler(
131
- labels, new_ratio, groups
143
+ labels, target_ratio, groups
132
144
  ),
133
- "random_oversampler": lambda: random_oversampler(labels, new_ratio),
145
+ "random_oversampler": lambda: random_oversampler(labels, target_ratio),
134
146
  "group_random_oversampler": lambda: group_random_oversampler(
135
- labels, new_ratio, groups
147
+ labels, target_ratio, groups
136
148
  ),
137
149
  "mean_dist_sampler": lambda: mean_dist_sampler(
138
- data, labels, new_ratio, kwargs.get("hard_percent", 0.5)
150
+ data, labels, target_ratio, kwargs.get("hard_percent", 0.5)
139
151
  ),
140
152
  "group_mean_dist_sampler": lambda: group_mean_dist_sampler(
141
- data, labels, new_ratio, groups, kwargs.get("hard_percent", 0.5)
153
+ data, labels, target_ratio, groups, kwargs.get("hard_percent", 0.5)
142
154
  ),
143
155
  }
144
156
 
@@ -167,7 +179,7 @@ def random_undersampler(labels: Labels, target_ratio: float) -> npt.NDArray:
167
179
  labels : Labels
168
180
  Binary prediction labels of shape (n_samples, n_classes).
169
181
  target_ratio : float
170
- Ratio of minority over majority classes to achieve.
182
+ Ratio of majority over minority classes to achieve.
171
183
 
172
184
  Returns
173
185
  -------
@@ -179,13 +191,15 @@ def random_undersampler(labels: Labels, target_ratio: float) -> npt.NDArray:
179
191
  TypeError
180
192
  If labels is not npt.NDArray.
181
193
  ValueError
182
- If target_ratio is less than 0.0.
194
+ If target_ratio is less than 1.
183
195
 
184
196
  """
185
197
  array_check(labels)
186
198
 
187
- if target_ratio <= 0:
188
- raise ValueError(f"Target ratio should be positive, but got {target_ratio}")
199
+ if target_ratio < 1:
200
+ raise ValueError(
201
+ f"Target ratio should be greater than 1, but got {target_ratio}"
202
+ )
189
203
 
190
204
  # Any row with a 1 is minority; rows with all 0s are majority
191
205
  is_min = np.any(labels != 0, axis=1)
@@ -202,8 +216,7 @@ def random_undersampler(labels: Labels, target_ratio: float) -> npt.NDArray:
202
216
  return maj_potential_idx.astype(int)
203
217
 
204
218
  # Calculate target majority count
205
- # Ensure n_maj_target is an integer and doesn't exceed available samples
206
- n_maj_target = int(np.round(n_min_class / target_ratio))
219
+ n_maj_target = int(np.round(n_min_class * target_ratio))
207
220
  n_to_sample = min(n_maj_target, n_maj_available)
208
221
 
209
222
  # Random selection
@@ -225,7 +238,7 @@ def group_random_undersampler(
225
238
  labels : Labels
226
239
  Binary prediction labels of shape (n_samples, n_classes).
227
240
  target_ratio : float
228
- Ratio of minority over majority classes to achieve.
241
+ Ratio of majority over minority classes to achieve.
229
242
  groups : npt.NDArray
230
243
  List of groups in which labels belong of shape (n_samples,).
231
244
 
@@ -240,14 +253,16 @@ def group_random_undersampler(
240
253
  If labels is not npt.NDArray.
241
254
  ValueError
242
255
  If labels and group do not have the same dimension.
243
- If target_ratio is less than 0.0.
256
+ If target_ratio is less than 1.
244
257
 
245
258
  """
246
259
  array_check(labels)
247
260
  array_dim_check(labels, groups, dim=0)
248
261
 
249
- if target_ratio <= 0:
250
- raise ValueError(f"Target ratio should be positive, but got {target_ratio}")
262
+ if target_ratio < 1:
263
+ raise ValueError(
264
+ f"Target ratio should be greater than 1, but got {target_ratio}"
265
+ )
251
266
 
252
267
  all_sampled_indices = [] # List for sampled indices
253
268
 
@@ -279,7 +294,7 @@ def random_oversampler(labels: Labels, target_ratio: float) -> npt.NDArray:
279
294
  labels : Labels
280
295
  Binary prediction labels of shape (n_samples, n_classes).
281
296
  target_ratio : float
282
- Ratio of minority over majority classes to achieve.
297
+ Ratio of majority over minority classes to achieve.
283
298
 
284
299
  Returns
285
300
  -------
@@ -291,14 +306,16 @@ def random_oversampler(labels: Labels, target_ratio: float) -> npt.NDArray:
291
306
  TypeError
292
307
  If labels is not npt.NDArray.
293
308
  ValueError
294
- If target_ratio is less than 0.0.
309
+ If target_ratio is less than 1.
295
310
  If there is no minority class.
296
311
 
297
312
  """
298
313
  array_check(labels)
299
314
 
300
- if target_ratio <= 0:
301
- raise ValueError(f"Target ratio should be positive, but got {target_ratio}")
315
+ if target_ratio < 1:
316
+ raise ValueError(
317
+ f"Target ratio should be greater than 1, but got {target_ratio}"
318
+ )
302
319
 
303
320
  # Generate class masks and indices once
304
321
  is_min = np.any(labels != 0, axis=1)
@@ -314,11 +331,9 @@ def random_oversampler(labels: Labels, target_ratio: float) -> npt.NDArray:
314
331
  raise ValueError("No minority examples found; cannot oversample.")
315
332
 
316
333
  # Calculate target size
317
- # We want (n_min_total / n_maj) = target_ratio
318
- target_n_min = int(np.round(n_maj_class * target_ratio))
334
+ target_n_min = int(np.round(n_maj_class / target_ratio))
319
335
 
320
336
  # Perform oversampling
321
- # This selects 'target_n_min' indices from the minority set with replacement.
322
337
  oversampled_min_idx = np.random.choice(min_indices, size=target_n_min, replace=True)
323
338
 
324
339
  # Concatenate and ensure integer type
@@ -337,7 +352,7 @@ def group_random_oversampler(
337
352
  labels : Labels
338
353
  Binary prediction labels of shape (n_samples, n_classes).
339
354
  target_ratio : float
340
- Ratio of minority over majority classes to achieve.
355
+ Ratio of majority over minority classes to achieve.
341
356
  groups : npt.NDArray
342
357
  List of groups in which labels belong of shape (n_samples,).
343
358
 
@@ -352,14 +367,16 @@ def group_random_oversampler(
352
367
  If labels is not npt.NDArray.
353
368
  ValueError
354
369
  If labels and group do not have the same dimension.
355
- If target_ratio is less than 0.0.
370
+ If target_ratio is less than 1.
356
371
 
357
372
  """
358
373
  array_check(labels)
359
374
  array_dim_check(labels, groups, dim=0)
360
375
 
361
- if target_ratio <= 0:
362
- raise ValueError(f"Target ratio should be positive, but got {target_ratio}")
376
+ if target_ratio < 1:
377
+ raise ValueError(
378
+ f"Target ratio should be greater than 1, but got {target_ratio}"
379
+ )
363
380
 
364
381
  all_sampled_indices = [] # List to contain sampled indices
365
382
  unique_groups = np.unique(groups)
@@ -405,7 +422,7 @@ def mean_dist_sampler(
405
422
  labels : Labels
406
423
  Binary prediction labels of shape (n_samples, n_classes).
407
424
  target_ratio : float
408
- Ratio of minority over majority classes to achieve.
425
+ Ratio of majority over minority classes to achieve.
409
426
  hard_percent : float, default: 0.5
410
427
  Percentage of examples that are considered hard, between 0 and 1.
411
428
  If hard_percent is 0.5, half of the examples are chosen from
@@ -422,15 +439,17 @@ def mean_dist_sampler(
422
439
  If labels is not npt.NDArray.
423
440
  ValueError
424
441
  If hard_percent is not between 0 and 1.
425
- If target_ratio is less than 0.0.
442
+ If target_ratio is less than 1.
426
443
 
427
444
  """
428
445
  array_check(labels)
429
446
 
430
447
  if not (0 <= hard_percent <= 1):
431
448
  raise ValueError(f"hard_percent must be between 0 and 1, got {hard_percent}")
432
- if target_ratio <= 0:
433
- raise ValueError(f"Target ratio should be positive, but got {target_ratio}")
449
+ if target_ratio < 1:
450
+ raise ValueError(
451
+ f"Target ratio should be greater than 1, but got {target_ratio}"
452
+ )
434
453
 
435
454
  # Identify minority and majority masks
436
455
  is_min = np.any(labels != 0, axis=1)
@@ -457,7 +476,7 @@ def mean_dist_sampler(
457
476
  sorted_rel_idx = np.argsort(dist)
458
477
 
459
478
  # Determine target counts
460
- n_maj_target = int(np.round(n_min / target_ratio))
479
+ n_maj_target = int(np.round(n_min * target_ratio))
461
480
  n_to_sample = min(n_maj_target, n_maj_available)
462
481
 
463
482
  n_hard = int(np.round(n_to_sample * hard_percent))
@@ -497,7 +516,7 @@ def group_mean_dist_sampler(
497
516
  labels : Labels
498
517
  Binary prediction labels of shape (n_samples, n_classes).
499
518
  target_ratio : float
500
- Ratio of minority over majority classes to achieve.
519
+ Ratio of majority over minority classes to achieve.
501
520
  groups : npt.NDArray
502
521
  List of groups in which labels belong of shape (n_samples,).
503
522
  hard_percent : float, default: 0.5
@@ -516,11 +535,16 @@ def group_mean_dist_sampler(
516
535
  If labels is not npt.NDArray.
517
536
  ValueError
518
537
  If labels and group do not have the same dimension.
538
+ If target ratio is less than 1.
519
539
 
520
540
  """
521
541
  array_check(labels)
522
542
  array_dim_check(labels, groups, dim=0)
523
543
 
544
+ if target_ratio < 1:
545
+ raise ValueError(
546
+ f"Target ratio should be greater than 1, but got {target_ratio}"
547
+ )
524
548
  all_sampled_indices = [] # List for sampled indices
525
549
  n_groups = np.unique(groups)
526
550
 
@@ -568,7 +592,7 @@ def smote(
568
592
  labels : Labels
569
593
  Binary prediction labels of shape (n_samples, n_classes).
570
594
  target_ratio : float
571
- Ratio of minority over majority classes to achieve.
595
+ Ratio of majority over minority classes to achieve.
572
596
  k_neighbors : int
573
597
  Number of neighbors to use for SMOTE knn.
574
598
 
@@ -584,13 +608,15 @@ def smote(
584
608
  TypeError
585
609
  If labels is not npt.NDArray.
586
610
  ValueError
587
- If target_ratio is less than 0.0.
611
+ It target_ratio is less than 1.
588
612
 
589
613
  """
590
614
  array_check(labels)
591
615
 
592
- if target_ratio <= 0:
593
- raise ValueError(f"Target ratio should be positive, but got {target_ratio}")
616
+ if target_ratio < 1:
617
+ raise ValueError(
618
+ f"Target ratio should be greater than 1, but got {target_ratio}"
619
+ )
594
620
 
595
621
  # Identify counts
596
622
  is_min = np.any(labels != 0, axis=1)
@@ -598,7 +624,7 @@ def smote(
598
624
  n_min_existing = np.sum(is_min)
599
625
 
600
626
  # Calculate how many NEW samples we need to generate
601
- n_target_total_min = int(np.round(n_maj * target_ratio))
627
+ n_target_total_min = int(np.round(n_maj / target_ratio))
602
628
  n_to_generate = n_target_total_min - n_min_existing
603
629
 
604
630
  if n_to_generate <= 0:
@@ -652,7 +678,7 @@ def group_smote(
652
678
  labels : Labels
653
679
  Binary prediction labels of shape (n_samples, n_classes).
654
680
  target_ratio : float
655
- Ratio of minority over majority classes to achieve.
681
+ Ratio of majority over minority classes to achieve.
656
682
  groups : npt.NDArray
657
683
  List of groups in which labels belong of shape (n_samples,).
658
684
  k_neighbors : int
@@ -673,11 +699,16 @@ def group_smote(
673
699
  If labels is not npt.NDArray
674
700
  ValueError
675
701
  If labels and group do not have the same dimension.
702
+ It target_ratio is less than 1.
676
703
 
677
704
  """
678
705
  array_check(labels)
679
706
  array_dim_check(labels, groups, dim=0)
680
707
 
708
+ if target_ratio < 1:
709
+ raise ValueError(
710
+ f"Target ratio should be greater than 1, but got {target_ratio}"
711
+ )
681
712
  unique_groups = np.unique(groups)
682
713
 
683
714
  # Start with original data to avoid repeated copying
@@ -129,8 +129,6 @@ def compute_all_CI(model_metrics: ModelMetrics, metric_list: list[str] = []) ->
129
129
  Model metrics for different folds.
130
130
  metric_list : list[str], default: []
131
131
  List of metrics to calculate confidence interval.
132
- **kwargs
133
- Extra arguments for the compute_CI function.
134
132
 
135
133
  Returns
136
134
  -------
@@ -282,10 +280,12 @@ def compute_pred_metrics(
282
280
  # Returns array of scores for each label
283
281
  scores = func(y_true, y_pred, average=None, zero_division=0.0)
284
282
  # Append the weighted average as the last element
285
- weighted = func(y_true, y_pred, average="weighted")
283
+ weighted = func(y_true, y_pred, average="weighted", zero_division=0.0)
286
284
  metric_dict[metric] = np.append(scores, weighted)
287
285
  else:
288
- metric_dict[metric] = [func(y_true, y_pred, average="binary")]
286
+ metric_dict[metric] = [
287
+ func(y_true, y_pred, average="binary", zero_division=0.0)
288
+ ]
289
289
 
290
290
  return metric_dict
291
291
 
@@ -157,7 +157,7 @@ class Pipeline:
157
157
  self.n_labels = len(self.label_list)
158
158
 
159
159
  print_message(
160
- f"Setting up Pipeline v{self.version} ({self.n_labels} labels)",
160
+ f"Setting up Pipeline {self.version} ({self.n_labels} labels)",
161
161
  self.logger,
162
162
  SCRIPT_NAME,
163
163
  )
@@ -480,7 +480,9 @@ class Pipeline:
480
480
  weights = self._weight_data(y_tr_label)
481
481
 
482
482
  print_message(
483
- f"Label: {label} | Fold {i+1}/{n_folds}", self.logger, SCRIPT_NAME
483
+ f"Label: {label} | {fold_key} | Fold {i+1}/{n_folds}",
484
+ self.logger,
485
+ SCRIPT_NAME,
484
486
  )
485
487
 
486
488
  # Execute training and evaluation
@@ -511,6 +513,12 @@ class Pipeline:
511
513
  X_final, y_work[:, k : k + 1], cv_groups
512
514
  )
513
515
  weights = self._weight_data(y_tr_final)
516
+ print_message(
517
+ f"Train set size: {len(X_tr_final):,}", self.logger, SCRIPT_NAME
518
+ )
519
+ print_message(
520
+ f"Calibration set size: {len(X_cal_set):,}", self.logger, SCRIPT_NAME
521
+ )
514
522
  self.fit_model(X_tr_final, y_tr_final, "predictor", label, weights)
515
523
 
516
524
  if self.calibrator_type:
@@ -550,14 +558,17 @@ class Pipeline:
550
558
  Label associated with the model to train.
551
559
  fold_key : Any
552
560
  Key representing the current fold.
553
- X_cal : PosProba, default: np.np.array([])
561
+ X_cal : PosProba, default: np.array([])
554
562
  Calibration data of shape (n_samples,) for the calibrator.
555
- y_cal : Labels, default: np.np.array([])
563
+ y_cal : Labels, default: np.array([])
556
564
  Calibration labels of shape (n_samples,) for the calibrator.
557
- weights : npt.NDArray, default: np.np.array([])
565
+ weights : npt.NDArray, default: np.array([])
558
566
  Weights to address class imbalance.
559
567
 
560
568
  """
569
+ print_message(f"Train set size: {len(X_train):,}", self.logger, SCRIPT_NAME)
570
+ print_message(f"Calibration set size: {len(X_cal):,}", self.logger, SCRIPT_NAME)
571
+ print_message(f"Test set size: {len(X_test):,}", self.logger, SCRIPT_NAME)
561
572
  # Fit predictor on train set
562
573
  self.fit_model(X_train, y_train, "predictor", label, weights)
563
574
 
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: medpipe
3
- Version: 0.2.0.dev0
3
+ Version: 1.0.0
4
4
  Summary: A package for clinical AI pipeline development
5
5
  Author-email: Mathias Roesler <mathias.roesler@auckland.co.nz>
6
6
  License-Expression: Apache-2.0
@@ -12,6 +12,7 @@ import pandas as pd
12
12
  import pytest
13
13
 
14
14
  from medpipe.data.sampler import (
15
+ data_sampler,
15
16
  group_mean_dist_sampler,
16
17
  group_random_oversampler,
17
18
  group_random_undersampler,
@@ -111,13 +112,101 @@ labels_groups_no_maj = np.array(
111
112
  )
112
113
 
113
114
 
115
+ @pytest.mark.parametrize(
116
+ "labels, data_len",
117
+ [
118
+ (single_labels, 12),
119
+ (bi_labels, 13),
120
+ (tri_labels, 13),
121
+ ],
122
+ )
123
+ def test_data_sampler_success(labels, data_len):
124
+ data = pd.DataFrame(np.random.rand(20, 2))
125
+ X, y, grps = data_sampler(data, labels, 0.5)
126
+
127
+ assert len(X) == data_len
128
+ assert len(y) == data_len
129
+ assert len(grps) == 0
130
+
131
+
132
+ @pytest.mark.parametrize(
133
+ "labels, reduction_factor, data_len",
134
+ [
135
+ (single_labels, 1.0, 8),
136
+ (single_labels, 0.25, 16),
137
+ (bi_labels, 1.0, 12),
138
+ (bi_labels, 0.25, 16),
139
+ (tri_labels, 1.0, 10),
140
+ (tri_labels, 0.25, 16),
141
+ ],
142
+ )
143
+ def test_data_sampler_reduction_factor(labels, reduction_factor, data_len):
144
+ data = pd.DataFrame(np.random.rand(20, 2))
145
+ X, y, grps = data_sampler(data, labels, reduction_factor)
146
+
147
+ assert len(X) == data_len
148
+ assert len(y) == data_len
149
+ assert len(grps) == 0
150
+
151
+
152
+ @pytest.mark.parametrize(
153
+ "sampler_fn, data_len",
154
+ [
155
+ ("random_undersampler", [8, 12, 10]),
156
+ ("random_oversampler", [32, 28, 30]),
157
+ ("mean_dist_sampler", [8, 12, 10]),
158
+ ("smote", [32, 28, 20]),
159
+ ],
160
+ )
161
+ def test_data_sampler_sampler_fn(sampler_fn, data_len):
162
+ data = pd.DataFrame(np.random.rand(20, 2))
163
+ labels = (single_labels, bi_labels, tri_labels)
164
+ for i in range(3):
165
+ if sampler_fn == "smote":
166
+ X, y, grps = data_sampler(
167
+ data,
168
+ labels[i],
169
+ reduction_factor=1.0,
170
+ sampler_fn=sampler_fn,
171
+ k_neighbors=1,
172
+ )
173
+ else:
174
+ X, y, grps = data_sampler(
175
+ data, labels[i], reduction_factor=1.0, sampler_fn=sampler_fn
176
+ )
177
+
178
+ assert len(X) == data_len[i]
179
+ assert len(y) == data_len[i]
180
+ assert len(grps) == 0
181
+
182
+
183
+ @pytest.mark.parametrize(
184
+ "reduction_factor",
185
+ [1.5, -1.0],
186
+ )
187
+ def test_data_sampler_value_error(reduction_factor):
188
+ data = pd.DataFrame(np.random.rand(20, 2))
189
+ with pytest.raises(ValueError):
190
+ data_sampler(data, single_labels, reduction_factor)
191
+
192
+
193
+ @pytest.mark.parametrize(
194
+ "labels",
195
+ ["str", (1, 2, 3), {"a": 1}],
196
+ )
197
+ def test_data_sampler_type_error(labels):
198
+ data = pd.DataFrame(np.random.rand(20, 2))
199
+ with pytest.raises(TypeError):
200
+ data_sampler(data, labels)
201
+
202
+
114
203
  # Basic functionality tests
115
204
  @pytest.mark.parametrize(
116
205
  "target_ratio, labels, n_min_class",
117
206
  [
118
- (0.5, single_labels, 4),
119
- (0.5, bi_labels, 6),
120
- (0.5, tri_labels, 5),
207
+ (2, single_labels, 4),
208
+ (1, bi_labels, 6),
209
+ (3, tri_labels, 5),
121
210
  ],
122
211
  )
123
212
  def test_random_undersampler_success(target_ratio, labels, n_min_class):
@@ -131,22 +220,22 @@ def test_random_undersampler_success(target_ratio, labels, n_min_class):
131
220
 
132
221
  assert minority_count == n_min_class
133
222
  assert (
134
- len(sampled_labels) == n_min_class + (1 / target_ratio) * n_min_class
223
+ len(sampled_labels) == n_min_class + target_ratio * n_min_class
135
224
  ) # Total samples length based on target_ratio and class imbalance
136
225
 
137
226
 
138
227
  @pytest.mark.parametrize(
139
228
  "target_ratio, labels, n_maj_class",
140
229
  [
141
- (0.5, single_labels, 8),
230
+ (2.0, single_labels, 8),
142
231
  (1.0, single_labels, 4),
143
- (2.0, single_labels, 2),
144
- (0.5, bi_labels, 12),
232
+ (4.0, single_labels, 16),
233
+ (2.0, bi_labels, 12),
145
234
  (1.0, bi_labels, 6),
146
- (3.0, bi_labels, 2),
147
- (0.5, tri_labels, 10),
235
+ (1.5, bi_labels, 9),
236
+ (2.0, tri_labels, 10),
148
237
  (1.0, tri_labels, 5),
149
- (5.0, tri_labels, 1),
238
+ (3.0, tri_labels, 15),
150
239
  ],
151
240
  )
152
241
  def test_random_undersampler_target_ratio(target_ratio, labels, n_maj_class):
@@ -163,9 +252,7 @@ def test_random_undersampler_target_ratio(target_ratio, labels, n_maj_class):
163
252
  # Check if the ratio is close to target ratio
164
253
  if minority_count > 0:
165
254
  ratio = majority_count / minority_count
166
- assert np.isclose(
167
- ratio, 1 / target_ratio, atol=0.1
168
- ) # Allow for slight imprecision
255
+ assert np.isclose(ratio, target_ratio, atol=0.1) # Allow for slight imprecision
169
256
 
170
257
  # Assert that the target ratio is close to what we selected
171
258
  assert majority_count == n_maj_class # Should match expected number of majority
@@ -173,7 +260,7 @@ def test_random_undersampler_target_ratio(target_ratio, labels, n_maj_class):
173
260
 
174
261
  # Edge case: No minority class samples
175
262
  def test_random_undersampler_no_minority():
176
- sample_idx = random_undersampler(labels_no_minority, target_ratio=0.5)
263
+ sample_idx = random_undersampler(labels_no_minority, target_ratio=1)
177
264
  sampled_labels = labels_no_minority[sample_idx]
178
265
 
179
266
  # If no minority class, only majority class samples should be selected
@@ -186,7 +273,7 @@ def test_random_undersampler_no_minority():
186
273
  [
187
274
  (0.5, labels_empty),
188
275
  (0.0, single_labels),
189
- (0.0, bi_labels),
276
+ (0.9, bi_labels),
190
277
  (-0.1, tri_labels),
191
278
  ],
192
279
  )
@@ -198,11 +285,11 @@ def test_random_undersampler_value_error(target_ratio, labels):
198
285
  # Raises a TypeError
199
286
  @pytest.mark.parametrize(
200
287
  "labels",
201
- [0.5, "string"],
288
+ [1.0, "string"],
202
289
  )
203
290
  def test_random_undersampler_type_error(labels):
204
291
  with pytest.raises(TypeError):
205
- random_undersampler(labels, target_ratio=0.5)
292
+ random_undersampler(labels, target_ratio=1)
206
293
 
207
294
 
208
295
  # Basic functionality test
@@ -244,9 +331,9 @@ def test_group_random_undersampler_success(
244
331
  (1.0, single_labels, single_groups, [1, 3]),
245
332
  (1.0, bi_labels, bi_groups, [1, 3, 2]),
246
333
  (1.0, tri_labels, tri_groups, [3, 2]),
247
- (0.5, single_labels, single_groups, [2, 6]),
248
- (0.5, bi_labels, bi_groups, [2, 6, 4]),
249
- (0.5, tri_labels, tri_groups, [6, 4]),
334
+ (2, single_labels, single_groups, [2, 6]),
335
+ (2, bi_labels, bi_groups, [2, 6, 4]),
336
+ (2, tri_labels, tri_groups, [6, 4]),
250
337
  ],
251
338
  )
252
339
  def test_group_random_undersampler_target_ratio(
@@ -270,7 +357,7 @@ def test_group_random_undersampler_target_ratio(
270
357
  if minority_count > 0:
271
358
  ratio = majority_count / minority_count
272
359
  assert np.isclose(
273
- ratio, 1 / target_ratio, atol=0.1
360
+ ratio, target_ratio, atol=0.1
274
361
  ) # Allow for slight imprecision
275
362
 
276
363
  assert majority_count == n_maj_class # Should match expected number of minority
@@ -283,20 +370,20 @@ def test_group_random_undersampler_target_ratio(
283
370
  )
284
371
  def test_group_random_undersampler_type_error(labels):
285
372
  with pytest.raises(TypeError):
286
- group_random_undersampler(labels, target_ratio=0.5, groups=single_groups)
373
+ group_random_undersampler(labels, target_ratio=1.0, groups=single_groups)
287
374
 
288
375
 
289
376
  # Labels and groups should have the same number of samples
290
377
  def test_group_random_undersampler_mismatched_dimensions():
291
378
  with pytest.raises(ValueError):
292
379
  groups = np.array([0, 1])
293
- group_random_undersampler(single_labels, target_ratio=0.5, groups=groups)
380
+ group_random_undersampler(single_labels, target_ratio=1.0, groups=groups)
294
381
 
295
382
 
296
383
  # Edge case: No minority class samples in any group
297
384
  def test_group_random_undersampler_no_minority():
298
385
  sample_idx = group_random_undersampler(
299
- labels_no_minority, target_ratio=0.5, groups=labels_groups_no_maj
386
+ labels_no_minority, target_ratio=1.0, groups=labels_groups_no_maj
300
387
  )
301
388
  sampled_labels = labels_no_minority[sample_idx]
302
389
 
@@ -307,9 +394,9 @@ def test_group_random_undersampler_no_minority():
307
394
  @pytest.mark.parametrize(
308
395
  "target_ratio, labels, n_maj_class",
309
396
  [
310
- (0.5, single_labels, 16),
311
- (0.5, bi_labels, 14),
312
- (0.5, tri_labels, 15),
397
+ (1.0, single_labels, 16),
398
+ (1.0, bi_labels, 14),
399
+ (1.0, tri_labels, 15),
313
400
  ],
314
401
  )
315
402
  def test_random_oversampler_success(target_ratio, labels, n_maj_class):
@@ -320,7 +407,6 @@ def test_random_oversampler_success(target_ratio, labels, n_maj_class):
320
407
  is_maj = np.all(sampled_labels == 0, axis=1)
321
408
  majority_count = np.sum(is_maj)
322
409
 
323
- # In the optimized version, we keep all original majority samples
324
410
  assert majority_count == n_maj_class
325
411
  # Total length should be n_maj + (n_maj * target_ratio)
326
412
  expected_len = n_maj_class + int(np.round(n_maj_class * target_ratio))
@@ -330,8 +416,8 @@ def test_random_oversampler_success(target_ratio, labels, n_maj_class):
330
416
  @pytest.mark.parametrize(
331
417
  "target_ratio, labels, expected_min_count",
332
418
  [
333
- (0.5, tri_labels, 8), # 15 maj * 0.5 = 7.5 -> 8
334
- (1.0, tri_labels, 15), # 15 maj * 1.0 = 15
419
+ (2.0, tri_labels, 8), # 15 maj / 2.0 = 7.5 -> 8
420
+ (1.0, tri_labels, 15), # 15 maj / 1.0 = 15
335
421
  ],
336
422
  )
337
423
  def test_random_oversampler_target_ratio(target_ratio, labels, expected_min_count):
@@ -344,9 +430,9 @@ def test_random_oversampler_target_ratio(target_ratio, labels, expected_min_coun
344
430
  minority_count = np.sum(is_min)
345
431
  majority_count = np.sum(is_maj)
346
432
 
347
- # Check ratio: min / maj should be target_ratio
348
- calculated_ratio = minority_count / majority_count
349
- assert np.isclose(calculated_ratio, target_ratio, atol=0.05)
433
+ # Check ratio: maj / min should be target_ratio
434
+ calculated_ratio = majority_count / minority_count
435
+ assert np.isclose(calculated_ratio, target_ratio, atol=0.4)
350
436
  assert minority_count == expected_min_count
351
437
 
352
438
 
@@ -357,7 +443,7 @@ def test_random_oversampler_target_ratio(target_ratio, labels, expected_min_coun
357
443
  )
358
444
  def test_random_oversampler_type_error(labels):
359
445
  with pytest.raises(TypeError):
360
- random_oversampler(labels, target_ratio=0.5)
446
+ random_oversampler(labels, target_ratio=1.0)
361
447
 
362
448
 
363
449
  @pytest.mark.parametrize(
@@ -391,7 +477,7 @@ def test_group_random_oversampler_success(
391
477
 
392
478
  def test_group_random_oversampler_no_minority():
393
479
  # if no minority, just return majority.
394
- sample_idx = group_random_oversampler(labels_no_minority, 0.5, labels_groups_no_maj)
480
+ sample_idx = group_random_oversampler(labels_no_minority, 1.5, labels_groups_no_maj)
395
481
  assert len(sample_idx) == len(labels_no_minority)
396
482
 
397
483
 
@@ -413,14 +499,14 @@ def test_random_oversampler_value_error(target_ratio, labels):
413
499
  )
414
500
  def test_group_random_oversampler_type_error(labels):
415
501
  with pytest.raises(TypeError):
416
- group_random_oversampler(labels, target_ratio=0.5, groups=single_groups)
502
+ group_random_oversampler(labels, target_ratio=1.0, groups=single_groups)
417
503
 
418
504
 
419
505
  # Labels and groups should have the same number of samples
420
506
  def test_group_random_oversampler_mismatched_dimensions():
421
507
  with pytest.raises(ValueError):
422
508
  groups = np.array([0, 1])
423
- group_random_oversampler(single_labels, target_ratio=0.5, groups=groups)
509
+ group_random_oversampler(single_labels, target_ratio=1.0, groups=groups)
424
510
 
425
511
 
426
512
  def test_mean_dist_sampler_basic():
@@ -430,10 +516,10 @@ def test_mean_dist_sampler_basic():
430
516
  labels = np.zeros((10, 2))
431
517
  labels[8:] = 1
432
518
 
433
- # If target_ratio is 0.5, and we have 2 minority samples,
434
- # we need 2/0.5 = 4 majority samples.
435
- res_idx = mean_dist_sampler(data, labels, target_ratio=0.5, hard_percent=0.5)
436
- assert len(res_idx) == 6 # 2 minority + 4 majority
519
+ # If target_ratio is 3, and we have 2 minority samples,
520
+ # we need 2*3 = 6 majority samples.
521
+ res_idx = mean_dist_sampler(data, labels, target_ratio=3.0, hard_percent=0.5)
522
+ assert len(res_idx) == 8 # 2 minority + 4 majority
437
523
  assert isinstance(res_idx, np.ndarray)
438
524
 
439
525
 
@@ -442,9 +528,9 @@ def test_mean_dist_sampler_exceptions():
442
528
  labels = np.zeros((5, 1))
443
529
 
444
530
  with pytest.raises(ValueError):
445
- mean_dist_sampler(data, labels, target_ratio=0.5, hard_percent=1.5)
531
+ mean_dist_sampler(data, labels, target_ratio=1.0, hard_percent=1.5)
446
532
 
447
- with pytest.raises(ValueError, match="Target ratio should be positive"):
533
+ with pytest.raises(ValueError):
448
534
  mean_dist_sampler(data, labels, target_ratio=-0.1)
449
535
 
450
536
 
@@ -494,9 +580,8 @@ def test_smote_invalid_ratio():
494
580
  with pytest.raises(ValueError):
495
581
  smote(data, labels, target_ratio=0, k_neighbors=1)
496
582
 
497
- # --- Mock Data Setup ---
498
-
499
583
 
584
+ # --- Mock Data Setup ---
500
585
  def create_smote_data():
501
586
  # Group 0: 20 majority, 5 minority (Needs oversampling)
502
587
  # Group 1: 20 majority, 20 minority (Already balanced)
@@ -523,7 +608,7 @@ def test_group_smote_logic():
523
608
  target_ratio = 1.0 # Want 1:1 ratio
524
609
  k_neighbors = 2 # Small k to ensure it works on smaller groups
525
610
 
526
- X_res, y_res, g_res = group_smote(data, labels, target_ratio, groups, k_neighbors)
611
+ _, y_res, g_res = group_smote(data, labels, target_ratio, groups, k_neighbors)
527
612
 
528
613
  # Check Group 0 (Was 20 maj, 5 min -> Should now be 20 maj, 20 min)
529
614
  g0_mask = g_res == 0
@@ -545,7 +630,7 @@ def test_group_smote_pandas_compatibility():
545
630
  data, labels, groups = create_smote_data()
546
631
  df_data = pd.DataFrame(data, columns=[f"feat_{i}" for i in range(10)])
547
632
 
548
- X_res, y_res, g_res = group_smote(df_data, labels, 0.5, groups, k_neighbors=3)
633
+ X_res, y_res, _ = group_smote(df_data, labels, 1.0, groups, k_neighbors=3)
549
634
 
550
635
  # Verify X_res is still a DataFrame
551
636
  assert isinstance(X_res, pd.DataFrame)
File without changes
File without changes
File without changes
File without changes