sinapsis-data-visualization 0.1.12__tar.gz → 0.1.14__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.
Files changed (32) hide show
  1. {sinapsis_data_visualization-0.1.12/src/sinapsis_data_visualization.egg-info → sinapsis_data_visualization-0.1.14}/PKG-INFO +2 -1
  2. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/README.md +1 -0
  3. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/pyproject.toml +1 -1
  4. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/helpers/annotation_drawer_tools.py +14 -13
  5. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/helpers/plot_distributions.py +4 -3
  6. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/helpers/scikit_pca_analysis.py +1 -1
  7. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/helpers/tags.py +1 -1
  8. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/templates/__init__.py +1 -1
  9. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/templates/base_annotation_drawer.py +5 -2
  10. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/templates/base_visualization_template.py +11 -10
  11. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/templates/bbox_drawer.py +2 -1
  12. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/templates/data_distribution_visualization.py +17 -10
  13. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/templates/key_points_drawer.py +2 -1
  14. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/templates/label_drawer.py +5 -2
  15. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/templates/oriented_bbox_drawer.py +2 -1
  16. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/templates/segmentation_mask_drawer.py +4 -1
  17. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/templates/tabular_data_visualization.py +29 -25
  18. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14/src/sinapsis_data_visualization.egg-info}/PKG-INFO +2 -1
  19. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization.egg-info/SOURCES.txt +0 -2
  20. sinapsis_data_visualization-0.1.12/src/sinapsis_data_visualization/templates/data_distribution_visualization_template.py +0 -248
  21. sinapsis_data_visualization-0.1.12/src/sinapsis_data_visualization/templates/prediction_visualization.py +0 -213
  22. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/LICENSE +0 -0
  23. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/setup.cfg +0 -0
  24. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/__init__.py +0 -0
  25. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/__init__.py +0 -0
  26. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/helpers/__init__.py +0 -0
  27. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/helpers/annotation_drawer_types.py +0 -0
  28. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/helpers/color_utils.py +0 -0
  29. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization/helpers/detection_utils.py +0 -0
  30. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization.egg-info/dependency_links.txt +0 -0
  31. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization.egg-info/requires.txt +0 -0
  32. {sinapsis_data_visualization-0.1.12 → sinapsis_data_visualization-0.1.14}/src/sinapsis_data_visualization.egg-info/top_level.txt +0 -0
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: sinapsis-data-visualization
3
- Version: 0.1.12
3
+ Version: 0.1.14
4
4
  Summary: Templates to perform visualization of data distributions and image annotations.
5
5
  Author-email: SinapsisAI <dev@sinapsis.tech>
6
6
  Project-URL: Homepage, https://sinapsis.tech
@@ -19,6 +19,7 @@ Provides-Extra: all
19
19
  Requires-Dist: sinapsis-data-visualization[visualization-matplotlib]; extra == "all"
20
20
  Dynamic: license-file
21
21
 
22
+ [![sp](https://img.shields.io/badge/lang-sp-red.svg)](https://github.com/Sinapsis-AI/sinapsis-data-tools/blob/main/README.es.md)
22
23
  <h1 align="center">
23
24
  <br>
24
25
  <a href="https://sinapsis.tech/">
@@ -1,3 +1,4 @@
1
+ [![sp](https://img.shields.io/badge/lang-sp-red.svg)](https://github.com/Sinapsis-AI/sinapsis-data-tools/blob/main/README.es.md)
1
2
  <h1 align="center">
2
3
  <br>
3
4
  <a href="https://sinapsis.tech/">
@@ -1,6 +1,6 @@
1
1
  [project]
2
2
  name = "sinapsis-data-visualization"
3
- version = "0.1.12"
3
+ version = "0.1.14"
4
4
  description = "Templates to perform visualization of data distributions and image annotations."
5
5
  authors = [{ name = "SinapsisAI", email = "dev@sinapsis.tech" }]
6
6
 
@@ -236,17 +236,18 @@ def get_extra_ann_labels(
236
236
  """
237
237
  text_y_offsets = []
238
238
  extra_str_repr = []
239
- for label_k, label_v in annotation.extra_labels.items():
240
- extra_text: str = label_k
241
- if isinstance(label_v, (float, int)):
242
- extra_text = get_numeric_str(extra_text, label_v)
243
- else:
244
- extra_text = f"{label_k}: {label_v}"
245
- extra_text_w, extra_text_h = get_text_size(extra_text, text_style)
246
-
247
- extra_text_h += spacing
248
- text_h += extra_text_h
249
- text_w = max(extra_text_w, text_w)
250
- text_y_offsets.append(extra_text_h)
251
- extra_str_repr.append(extra_text)
239
+ if annotation.extra_labels is not None:
240
+ for label_k, label_v in annotation.extra_labels.items():
241
+ extra_text: str = label_k
242
+ if isinstance(label_v, (float, int)):
243
+ extra_text = get_numeric_str(extra_text, label_v)
244
+ else:
245
+ extra_text = f"{label_k}: {label_v}"
246
+ extra_text_w, extra_text_h = get_text_size(extra_text, text_style)
247
+
248
+ extra_text_h += spacing
249
+ text_h += extra_text_h
250
+ text_w = max(extra_text_w, text_w)
251
+ text_y_offsets.append(extra_text_h)
252
+ extra_str_repr.append(extra_text)
252
253
  return text_y_offsets, extra_str_repr, text_h, text_w
@@ -4,17 +4,18 @@ from typing import Sequence
4
4
  import matplotlib
5
5
  import matplotlib.pyplot as plt
6
6
  import numpy as np
7
+ from matplotlib.figure import Figure
7
8
  from sinapsis_core.utils.logging_utils import sinapsis_logger
8
9
  from sklearn.metrics import ConfusionMatrixDisplay
9
10
 
10
11
 
11
12
  def plot_distribution(
12
- figure: matplotlib.figure.Figure,
13
+ figure: Figure,
13
14
  labels: Sequence[str] | np.ndarray,
14
15
  counts: list[float] | np.ndarray,
15
16
  plot_type: str,
16
17
  kwargs: dict,
17
- ) -> matplotlib.figure.Figure:
18
+ ) -> Figure:
18
19
  """
19
20
  Base method to plot a distribution.
20
21
 
@@ -168,7 +169,7 @@ def plot_confusion_matrix(labels: Sequence[str], counts: np.ndarray, kwargs: dic
168
169
  kwargs (dict): Additional keyword arguments (cmap, etc.)
169
170
  """
170
171
 
171
- cmap = kwargs.pop("cmap", plt.cm.Blues)
172
+ cmap = kwargs.pop("cmap", matplotlib.colormaps["Blues"])
172
173
  disp = ConfusionMatrixDisplay(confusion_matrix=counts, display_labels=labels)
173
174
  disp.plot(cmap=cmap, **kwargs)
174
175
  plt.tight_layout()
@@ -29,7 +29,7 @@ def perform_k_means_analysis(feature_vec: np.ndarray) -> tuple[np.ndarray, np.nd
29
29
  return labels, reduced_vectors
30
30
 
31
31
 
32
- def pre_process_images(images: list[ImagePacket], size: int = 28) -> np.ndarray | None:
32
+ def pre_process_images(images: list[ImagePacket], size: int = 28) -> np.ndarray:
33
33
  """Resizes and flatten images to homogenize them
34
34
 
35
35
  Args:
@@ -2,7 +2,7 @@
2
2
  from enum import Enum
3
3
 
4
4
 
5
- class Tags(Enum):
5
+ class Tags(str, Enum):
6
6
  ANNOTATIONS = "annotations"
7
7
  BBOX = "bbox"
8
8
  CHARTS = "charts"
@@ -15,7 +15,7 @@ _template_lookup = {
15
15
  }
16
16
 
17
17
 
18
- def __getattr__(name: str) -> Callable: # type:ignore
18
+ def __getattr__(name: str) -> Callable:
19
19
  if name in _template_lookup:
20
20
  module = importlib.import_module(_template_lookup[name])
21
21
  return getattr(module, name)
@@ -53,6 +53,8 @@ class BaseAnnotationDrawer(Template, abc.ABC):
53
53
  overwrite: bool = False
54
54
  randomized_color: bool = True
55
55
 
56
+ attributes: AttributesBaseModel
57
+
56
58
  def __init__(self, attributes: TemplateAttributeType) -> None:
57
59
  """Initializes the Image Annotation Drawer with the given attributes."""
58
60
  super().__init__(attributes)
@@ -88,7 +90,7 @@ class BaseAnnotationDrawer(Template, abc.ABC):
88
90
  """
89
91
  return get_color_rgb_tuple(
90
92
  color_map=self.COLOR_MAP,
91
- class_id=ann.label,
93
+ class_id=int(ann.label) if ann.label is not None else None,
92
94
  randomized=self.attributes.randomized_color,
93
95
  )
94
96
 
@@ -154,7 +156,7 @@ class BaseAnnotationDrawer(Template, abc.ABC):
154
156
  ImagePacket(
155
157
  content=annotated_image,
156
158
  color_space=image_packet.color_space,
157
- source=f"a_{Path(image_packet.source).name}",
159
+ source=f"a_{Path(getattr(image_packet, 'source')).name}",
158
160
  )
159
161
  )
160
162
  container.images.extend(annotated_images)
@@ -169,6 +171,7 @@ class BaseAnnotationDrawer(Template, abc.ABC):
169
171
  Returns:
170
172
  DataContainer: The updated data container with images annotated.
171
173
  """
174
+
172
175
  self.process_images(container)
173
176
 
174
177
  return container
@@ -5,10 +5,10 @@ from abc import ABC, abstractmethod
5
5
  from dataclasses import dataclass
6
6
  from typing import Any
7
7
 
8
- import matplotlib
9
8
  import matplotlib.pyplot as plt
10
9
  import numpy as np
11
10
  import PIL.Image
11
+ from matplotlib.figure import Figure
12
12
  from pydantic import Field
13
13
  from sinapsis_core.data_containers.data_packet import DataContainer, ImagePacket
14
14
  from sinapsis_core.template_base import Template
@@ -90,6 +90,7 @@ class BaseVisualizationTemplate(Template, ABC):
90
90
  tags=[Tags.IMAGE, Tags.MATPLOTLIB, Tags.VISUALIZATION],
91
91
  )
92
92
  AttributesBaseModel = BasePlotAttributes
93
+ attributes: BasePlotAttributes
93
94
 
94
95
  def __init__(self, attributes: TemplateAttributeType) -> None:
95
96
  super().__init__(attributes)
@@ -97,7 +98,7 @@ class BaseVisualizationTemplate(Template, ABC):
97
98
  if self.attributes.save_image_dir and not os.path.exists(self.attributes.save_image_dir):
98
99
  os.makedirs(self.attributes.save_image_dir)
99
100
 
100
- def create_figure(self, text: str | None = None, rotate_x_ticks: bool = False) -> matplotlib.figure.Figure:
101
+ def create_figure(self, text: str | None = None, rotate_x_ticks: bool = False) -> Figure:
101
102
  """
102
103
  Creates a matplotlib figure with basic configuration.
103
104
 
@@ -106,7 +107,7 @@ class BaseVisualizationTemplate(Template, ABC):
106
107
  rotate_x_ticks (bool): Whether to rotate x-axis tick labels.
107
108
 
108
109
  Returns:
109
- matplotlib.figure.Figure: The configured figure.
110
+ Figure: The configured figure.
110
111
  """
111
112
  figure = plt.figure(figsize=(self.attributes.fig_width, self.attributes.fig_height))
112
113
  plt.xlabel(self.attributes.x_label or "X-Axis")
@@ -123,7 +124,7 @@ class BaseVisualizationTemplate(Template, ABC):
123
124
 
124
125
  def save_figure(
125
126
  self,
126
- figure: matplotlib.figure.Figure,
127
+ figure: Figure,
127
128
  name: str,
128
129
  container_id: str | None = None,
129
130
  ) -> str:
@@ -131,7 +132,7 @@ class BaseVisualizationTemplate(Template, ABC):
131
132
  Saves the figure to disk and records the path.
132
133
 
133
134
  Args:
134
- figure (matplotlib.figure.Figure): The figure to save.
135
+ figure (Figure): The figure to save.
135
136
  name (str): Base name for the figure file.
136
137
  container_id (str | None): Optional container ID to append to filename.
137
138
 
@@ -156,13 +157,13 @@ class BaseVisualizationTemplate(Template, ABC):
156
157
 
157
158
  return fig_path
158
159
 
159
- def add_figure_to_container(self, container: DataContainer, figure: matplotlib.figure.Figure, name: str) -> None:
160
+ def add_figure_to_container(self, container: DataContainer, figure: Figure, name: str) -> None:
160
161
  """
161
162
  Adds the figure to the container as an image packet if requested.
162
163
 
163
164
  Args:
164
165
  container (DataContainer): The data container to update.
165
- figure (matplotlib.figure.Figure): The figure to add.
166
+ figure (Figure): The figure to add.
166
167
  name (str): Name key for the figure.
167
168
  """
168
169
  if self.attributes.add_to_image_packets:
@@ -185,7 +186,7 @@ class BaseVisualizationTemplate(Template, ABC):
185
186
  counts: list[float] | np.ndarray,
186
187
  text: str | None = None,
187
188
  rotate_x_ticks: bool = False,
188
- ) -> matplotlib.figure.Figure:
189
+ ) -> Figure:
189
190
  """
190
191
  Creates a figure, plots data, saves it, and adds it to the container.
191
192
 
@@ -198,7 +199,7 @@ class BaseVisualizationTemplate(Template, ABC):
198
199
  rotate_x_ticks (bool): Whether to rotate x-axis tick labels.
199
200
 
200
201
  Returns:
201
- matplotlib.figure.Figure: The created figure.
202
+ Figure: The created figure.
202
203
  """
203
204
  figure = self.create_figure(text=text, rotate_x_ticks=rotate_x_ticks)
204
205
 
@@ -211,7 +212,7 @@ class BaseVisualizationTemplate(Template, ABC):
211
212
  )
212
213
 
213
214
  plot_name = f"{plot_type}" if not text else f"{plot_type}_{text}"
214
- self.save_figure(figure, plot_name, container.container_id)
215
+ self.save_figure(figure, plot_name, str(container.container_id))
215
216
  self.add_figure_to_container(container, figure, plot_name)
216
217
  return figure
217
218
 
@@ -8,7 +8,8 @@ from sinapsis_data_visualization.helpers.tags import Tags
8
8
  from sinapsis_data_visualization.templates.label_drawer import LabelDrawer
9
9
 
10
10
  BBoxDrawerUIProperties = LabelDrawer.UIProperties
11
- BBoxDrawerUIProperties.tags.extend([Tags.BBOX])
11
+ if BBoxDrawerUIProperties.tags is not None:
12
+ BBoxDrawerUIProperties.tags.extend([Tags.BBOX])
12
13
 
13
14
 
14
15
  class BBoxDrawer(LabelDrawer):
@@ -31,7 +31,8 @@ class DataDistributionAttributes(BasePlotAttributes):
31
31
 
32
32
 
33
33
  DataDistributionUIProperties = BaseVisualizationTemplate.UIProperties
34
- DataDistributionUIProperties.tags.extend([Tags.CHARTS, Tags.CLUSTERING, Tags.DISTRIBUTION, Tags.PLOTS])
34
+ if DataDistributionUIProperties.tags is not None:
35
+ DataDistributionUIProperties.tags.extend([Tags.CHARTS, Tags.CLUSTERING, Tags.DISTRIBUTION, Tags.PLOTS])
35
36
 
36
37
 
37
38
  class DataDistributionVisualization(BaseVisualizationTemplate):
@@ -46,8 +47,9 @@ class DataDistributionVisualization(BaseVisualizationTemplate):
46
47
 
47
48
  AttributesBaseModel = DataDistributionAttributes
48
49
  UIProperties = DataDistributionUIProperties
50
+ attributes: DataDistributionAttributes
49
51
 
50
- def retrieve_labels_from_images(self, container: DataContainer) -> tuple[list[str | int], list[int]]:
52
+ def retrieve_labels_from_images(self, container: DataContainer) -> tuple[list[str], list[int]]:
51
53
  """
52
54
  Iterates through the ImagePackets, extracting each of the
53
55
  annotations for all of them.
@@ -55,18 +57,23 @@ class DataDistributionVisualization(BaseVisualizationTemplate):
55
57
  with that label.
56
58
 
57
59
  Returns:
58
- tuple[list[str | int], list[int]]: list of labels and number of coincidences for each label
60
+ tuple[list[str], list[int]]: list of labels and number of coincidences for each label
59
61
  """
60
- labels: list[str | int] = [ann.label_str for image in container.images for ann in image.annotations]
61
62
 
62
- label_count: dict[str | int, int] = {}
63
+ labels: list[str] = []
64
+ for image in container.images:
65
+ for ann in image.annotations:
66
+ if ann.label_str is not None:
67
+ labels.extend([ann.label_str])
68
+
69
+ label_count: dict[str, int] = {}
63
70
  for label in labels:
64
71
  label_count[label] = label_count.get(label, 0) + 1
65
72
 
66
73
  labels, counts = list(label_count.keys()), list(label_count.values())
67
74
  return labels, counts
68
75
 
69
- def process_cluster(self, container: DataContainer) -> tuple[list[int | str], np.ndarray]:
76
+ def process_cluster(self, container: DataContainer) -> tuple[list[str], np.ndarray]:
70
77
  """
71
78
  For the k-means clustering, the images need to be flattened first
72
79
  and then returned as a reduced vector. This method calls the pre_process_images
@@ -79,15 +86,15 @@ class DataDistributionVisualization(BaseVisualizationTemplate):
79
86
  container (DataContainer): Container to extract the images from
80
87
 
81
88
  Returns:
82
- list[str | int]: list of labels
89
+ list[str]: list of labels
83
90
  np.ndarray : transformed values from the pca analysis
84
91
  """
85
92
  images = container.images
86
93
  feature_arr = pre_process_images(images)
87
94
  label, counts = perform_k_means_analysis(feature_arr)
88
- return label, counts
95
+ return label.tolist(), counts.tolist()
89
96
 
90
- def get_data_for_visualization(self, container: DataContainer) -> tuple[list[str | int], list[int]]:
97
+ def get_data_for_visualization(self, container: DataContainer) -> tuple[list[str], list[int]]:
91
98
  """
92
99
  Gets labels and counts from images in the container.
93
100
 
@@ -99,7 +106,7 @@ class DataDistributionVisualization(BaseVisualizationTemplate):
99
106
  """
100
107
  return self.retrieve_labels_from_images(container)
101
108
 
102
- def generate_visualizations(self, container: DataContainer, data: tuple[list[str | int], list[int]]) -> None:
109
+ def generate_visualizations(self, container: DataContainer, data: tuple[list[str], list[int | float]]) -> None:
103
110
  """
104
111
  Generates all requested visualizations.
105
112
 
@@ -11,7 +11,8 @@ from sinapsis_data_visualization.helpers.tags import Tags
11
11
  from sinapsis_data_visualization.templates.bbox_drawer import BBoxDrawer
12
12
 
13
13
  KeyPointsDrawerUIProperties = BBoxDrawer.UIProperties
14
- KeyPointsDrawerUIProperties.tags.extend([Tags.KEYPOINTS])
14
+ if KeyPointsDrawerUIProperties.tags is not None:
15
+ KeyPointsDrawerUIProperties.tags.extend([Tags.KEYPOINTS])
15
16
 
16
17
 
17
18
  class KeyPointsDrawer(BBoxDrawer):
@@ -28,7 +28,8 @@ from sinapsis_data_visualization.templates.base_annotation_drawer import (
28
28
  )
29
29
 
30
30
  LabelDrawerUIProperties = BaseAnnotationDrawer.UIProperties
31
- LabelDrawerUIProperties.tags.extend([Tags.LABELS, Tags.CLASSIFICATION])
31
+ if LabelDrawerUIProperties.tags is not None:
32
+ LabelDrawerUIProperties.tags.extend([Tags.LABELS, Tags.CLASSIFICATION])
32
33
 
33
34
 
34
35
  class LabelDrawer(BaseAnnotationDrawer):
@@ -97,6 +98,8 @@ class LabelDrawer(BaseAnnotationDrawer):
97
98
  if isinstance(self.text_style, dict):
98
99
  self.text_style = TextStyle(**self.text_style)
99
100
 
101
+ attributes: AttributesBaseModel
102
+
100
103
  def set_drawing_strategy(self) -> None:
101
104
  """Appends the annotation to be drawn (i.e., label)
102
105
  to the list of tasks in the execute method"""
@@ -270,7 +273,7 @@ class LabelDrawer(BaseAnnotationDrawer):
270
273
  image = image_info[0]
271
274
  font_scale, font_thickness = get_dynamic_text_properties(image)
272
275
  self.attributes.text_style.font_scale = font_scale
273
- self.attributes.text_style.thickness = font_thickness
276
+ self.attributes.text_style.thickness = int(font_thickness)
274
277
 
275
278
  image = self.add_label(image_info=image_info, class_label=class_label, spacing=int(5 * font_scale))
276
279
 
@@ -12,7 +12,8 @@ from sinapsis_data_visualization.helpers.tags import Tags
12
12
  from sinapsis_data_visualization.templates.bbox_drawer import BBoxDrawer
13
13
 
14
14
  OrientedBBoxDrawerUIProperties = BBoxDrawer.UIProperties
15
- OrientedBBoxDrawerUIProperties.tags.extend([Tags.ORIENTED_BBOX])
15
+ if OrientedBBoxDrawerUIProperties.tags is not None:
16
+ OrientedBBoxDrawerUIProperties.tags.extend([Tags.ORIENTED_BBOX])
16
17
 
17
18
 
18
19
  class OrientedBBoxDrawer(BBoxDrawer):
@@ -7,7 +7,8 @@ from sinapsis_data_visualization.helpers.tags import Tags
7
7
  from sinapsis_data_visualization.templates.bbox_drawer import BBoxDrawer
8
8
 
9
9
  SegmentationMaskDrawerUIProperties = BBoxDrawer.UIProperties
10
- SegmentationMaskDrawerUIProperties.tags.extend([Tags.SEGMENTATION, Tags.MASKS])
10
+ if SegmentationMaskDrawerUIProperties.tags is not None:
11
+ SegmentationMaskDrawerUIProperties.tags.extend([Tags.SEGMENTATION, Tags.MASKS])
11
12
 
12
13
 
13
14
  class SegmentationMaskDrawer(BBoxDrawer):
@@ -59,6 +60,8 @@ class SegmentationMaskDrawer(BBoxDrawer):
59
60
 
60
61
  alpha: float = 0.5
61
62
 
63
+ attributes: AttributesBaseModel
64
+
62
65
  def set_drawing_strategy(self) -> None:
63
66
  """
64
67
  Appends the method to draw the segmentation masks on the image, to
@@ -1,10 +1,10 @@
1
1
  # -*- coding: utf-8 -*-
2
2
  from typing import Sequence
3
3
 
4
- import matplotlib
5
4
  import matplotlib.pyplot as plt
6
5
  import numpy as np
7
6
  import pandas as pd
7
+ from matplotlib.figure import Figure
8
8
  from pydantic import Field
9
9
  from sinapsis_core.data_containers.data_packet import DataContainer
10
10
  from sinapsis_data_visualization.helpers.tags import Tags
@@ -40,7 +40,8 @@ class TabularDataVisualizationAttributes(BasePlotAttributes):
40
40
 
41
41
 
42
42
  TabularDataVisualizationUIProperties = BaseVisualizationTemplate.UIProperties
43
- TabularDataVisualizationUIProperties.tags.extend([Tags.CHARTS, Tags.PLOTS, Tags.CORRELATION, Tags.PANDAS])
43
+ if TabularDataVisualizationUIProperties.tags is not None:
44
+ TabularDataVisualizationUIProperties.tags.extend([Tags.CHARTS, Tags.PLOTS, Tags.CORRELATION, Tags.PANDAS])
44
45
 
45
46
 
46
47
  class TabularDataVisualization(BaseVisualizationTemplate):
@@ -53,6 +54,8 @@ class TabularDataVisualization(BaseVisualizationTemplate):
53
54
 
54
55
  AttributesBaseModel = TabularDataVisualizationAttributes
55
56
 
57
+ attributes: TabularDataVisualizationAttributes
58
+
56
59
  def get_data_for_visualization(self, container: DataContainer) -> pd.DataFrame | None:
57
60
  """
58
61
  Extracts and converts tabular data from the container.
@@ -67,20 +70,21 @@ class TabularDataVisualization(BaseVisualizationTemplate):
67
70
  self.logger.warning("No generic_field_key specified")
68
71
  return None
69
72
 
70
- data_dict = self._get_generic_data(container, self.attributes.generic_field_key)
73
+ dataframes = getattr(container, "data_frames", None)
71
74
 
72
- if data_dict is None:
75
+ if dataframes is None:
73
76
  self.logger.warning(f"No data found with key '{self.attributes.generic_field_key}'")
74
77
  return None
78
+ dataframe = dataframes[0].content
75
79
 
76
- if isinstance(data_dict, pd.DataFrame):
77
- dataset = data_dict
80
+ if isinstance(dataframe, pd.DataFrame):
81
+ data = dataframe
78
82
  else:
79
- self.logger.warning(f"Data format not recognized: {type(data_dict)}")
83
+ self.logger.warning(f"Data format not recognized: {type(dataframe)}")
80
84
  return None
81
85
 
82
- self.logger.info(f"Available columns: {dataset.columns}")
83
- return dataset
86
+ self.logger.info(f"Available columns: {data.columns}")
87
+ return data
84
88
 
85
89
  def identify_feature_types(self, df: pd.DataFrame) -> tuple[list[str], list[str]]:
86
90
  """
@@ -114,7 +118,7 @@ class TabularDataVisualization(BaseVisualizationTemplate):
114
118
 
115
119
  return numeric_cols, categorical_cols
116
120
 
117
- def prepare_histogram_data(self, df: pd.DataFrame, feature: str) -> tuple[list[float], list[float]]:
121
+ def prepare_histogram_data(self, df: pd.DataFrame, feature: str) -> tuple[list[str], list[float]]:
118
122
  """
119
123
  Prepares data for a histogram plot.
120
124
 
@@ -131,9 +135,9 @@ class TabularDataVisualization(BaseVisualizationTemplate):
131
135
  bins = np.linspace(min_val, max_val, 20)
132
136
  counts, edges = np.histogram(df[feature].to_numpy(), bins=bins)
133
137
 
134
- labels = [(edges[i] + edges[i + 1]) / 2 for i in range(len(edges) - 1)]
138
+ labels = [str((edges[i] + edges[i + 1]) / 2) for i in range(len(edges) - 1)]
135
139
 
136
- return labels, counts
140
+ return labels, counts.tolist()
137
141
 
138
142
  def prepare_boxplot_data(self, df: pd.DataFrame, feature: str) -> tuple[list[str], list[np.ndarray]]:
139
143
  """
@@ -187,7 +191,7 @@ class TabularDataVisualization(BaseVisualizationTemplate):
187
191
 
188
192
  return features_to_correlate, corr_matrix
189
193
 
190
- def prepare_categorical_data(self, df: pd.DataFrame, feature: str) -> tuple[list[str], list[int]]:
194
+ def prepare_categorical_data(self, df: pd.DataFrame, feature: str) -> tuple[list[str], list[int | float]]:
191
195
  """
192
196
  Prepares data for a categorical distribution plot.
193
197
 
@@ -203,7 +207,7 @@ class TabularDataVisualization(BaseVisualizationTemplate):
203
207
  counts = value_counts.values.tolist()
204
208
  return labels, counts
205
209
 
206
- def prepare_target_pie_data(self, df: pd.DataFrame) -> tuple[list[str], list[int]]:
210
+ def prepare_target_pie_data(self, df: pd.DataFrame) -> tuple[list[str], list[int | float]]:
207
211
  """
208
212
  Prepares data for a target distribution pie chart.
209
213
 
@@ -224,7 +228,7 @@ class TabularDataVisualization(BaseVisualizationTemplate):
224
228
 
225
229
  def create_boxplot(
226
230
  self, labels: Sequence[str], data_points: Sequence[np.ndarray], title: str | None = None
227
- ) -> matplotlib.figure.Figure:
231
+ ) -> Figure:
228
232
  """
229
233
  Creates a boxplot figure.
230
234
 
@@ -237,21 +241,21 @@ class TabularDataVisualization(BaseVisualizationTemplate):
237
241
  matplotlib.figure.Figure: The created boxplot.
238
242
  """
239
243
  fig = self.create_figure(text=title)
240
- plt.boxplot(data_points, labels=labels)
244
+ plt.boxplot(data_points, labels=labels) # ty: ignore[unknown-argument]
241
245
  plt.ylabel("Value")
242
246
  plt.xticks(rotation=45, ha="right")
243
247
  plt.tight_layout()
244
248
  return fig
245
249
 
246
- def generate_visualizations(self, container: DataContainer, dataset: pd.DataFrame) -> None:
250
+ def generate_visualizations(self, container: DataContainer, data: pd.DataFrame) -> None:
247
251
  """
248
252
  Generates all requested visualizations for the tabular data.
249
253
 
250
254
  Args:
251
255
  container (DataContainer): Data container.
252
- dataset (pd.DataFrame): The tabular data.
256
+ data (pd.DataFrame): The tabular data.
253
257
  """
254
- numeric_cols, categorical_cols = self.identify_feature_types(dataset)
258
+ numeric_cols, categorical_cols = self.identify_feature_types(data)
255
259
  self.logger.info(f"Found {len(numeric_cols)} numeric features and {len(categorical_cols)} categorical features")
256
260
 
257
261
  if not numeric_cols and not categorical_cols:
@@ -260,7 +264,7 @@ class TabularDataVisualization(BaseVisualizationTemplate):
260
264
 
261
265
  if self.attributes.histogram:
262
266
  for feature in numeric_cols:
263
- labels, counts = self.prepare_histogram_data(dataset, feature)
267
+ labels, counts = self.prepare_histogram_data(data, feature)
264
268
  if len(labels) > 0 and len(counts) > 0:
265
269
  self.plot_and_save(
266
270
  container=container,
@@ -272,14 +276,14 @@ class TabularDataVisualization(BaseVisualizationTemplate):
272
276
 
273
277
  if self.attributes.box_plot:
274
278
  for feature in numeric_cols:
275
- box_labels, data_points = self.prepare_boxplot_data(dataset, feature)
279
+ box_labels, data_points = self.prepare_boxplot_data(data, feature)
276
280
  if len(box_labels) > 0 and len(data_points) > 0:
277
281
  fig = self.create_boxplot(labels=box_labels, data_points=data_points, title=f"{feature}")
278
- self.save_figure(fig, f"boxplot_{feature}", container.container_id)
282
+ self.save_figure(fig, f"boxplot_{feature}", str(container.container_id))
279
283
  self.add_figure_to_container(container, fig, f"boxplot_{feature}")
280
284
 
281
285
  if self.attributes.correlation and len(numeric_cols) > 1:
282
- corr_labels, corr_matrix = self.prepare_correlation_data(dataset, numeric_cols)
286
+ corr_labels, corr_matrix = self.prepare_correlation_data(data, numeric_cols)
283
287
  if len(corr_labels) > 0 and corr_matrix.size > 0:
284
288
  self.plot_and_save(
285
289
  container=container,
@@ -291,7 +295,7 @@ class TabularDataVisualization(BaseVisualizationTemplate):
291
295
 
292
296
  if self.attributes.categorical_dist and categorical_cols:
293
297
  for feature in categorical_cols:
294
- cat_labels, cat_counts = self.prepare_categorical_data(dataset, feature)
298
+ cat_labels, cat_counts = self.prepare_categorical_data(data, feature)
295
299
  if len(cat_labels) > 0 and len(cat_counts) > 0:
296
300
  self.plot_and_save(
297
301
  container=container,
@@ -303,7 +307,7 @@ class TabularDataVisualization(BaseVisualizationTemplate):
303
307
  )
304
308
 
305
309
  if self.attributes.pie_chart:
306
- pie_labels, pie_counts = self.prepare_target_pie_data(dataset)
310
+ pie_labels, pie_counts = self.prepare_target_pie_data(data)
307
311
  if len(pie_labels) > 0 and len(pie_counts) > 0:
308
312
  self.plot_and_save(
309
313
  container=container,
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: sinapsis-data-visualization
3
- Version: 0.1.12
3
+ Version: 0.1.14
4
4
  Summary: Templates to perform visualization of data distributions and image annotations.
5
5
  Author-email: SinapsisAI <dev@sinapsis.tech>
6
6
  Project-URL: Homepage, https://sinapsis.tech
@@ -19,6 +19,7 @@ Provides-Extra: all
19
19
  Requires-Dist: sinapsis-data-visualization[visualization-matplotlib]; extra == "all"
20
20
  Dynamic: license-file
21
21
 
22
+ [![sp](https://img.shields.io/badge/lang-sp-red.svg)](https://github.com/Sinapsis-AI/sinapsis-data-tools/blob/main/README.es.md)
22
23
  <h1 align="center">
23
24
  <br>
24
25
  <a href="https://sinapsis.tech/">
@@ -21,10 +21,8 @@ src/sinapsis_data_visualization/templates/base_annotation_drawer.py
21
21
  src/sinapsis_data_visualization/templates/base_visualization_template.py
22
22
  src/sinapsis_data_visualization/templates/bbox_drawer.py
23
23
  src/sinapsis_data_visualization/templates/data_distribution_visualization.py
24
- src/sinapsis_data_visualization/templates/data_distribution_visualization_template.py
25
24
  src/sinapsis_data_visualization/templates/key_points_drawer.py
26
25
  src/sinapsis_data_visualization/templates/label_drawer.py
27
26
  src/sinapsis_data_visualization/templates/oriented_bbox_drawer.py
28
- src/sinapsis_data_visualization/templates/prediction_visualization.py
29
27
  src/sinapsis_data_visualization/templates/segmentation_mask_drawer.py
30
28
  src/sinapsis_data_visualization/templates/tabular_data_visualization.py
@@ -1,248 +0,0 @@
1
- # -*- coding: utf-8 -*-
2
- """Data distribution visualization template"""
3
-
4
- import os
5
- from typing import Any
6
-
7
- import matplotlib
8
- import matplotlib.pyplot as plt
9
- import numpy as np
10
- from pydantic import Field
11
- from pydantic.dataclasses import dataclass
12
- from sinapsis_core.data_containers.data_packet import DataContainer
13
- from sinapsis_core.template_base import Template
14
- from sinapsis_core.template_base.base_models import (
15
- OutputTypes,
16
- TemplateAttributes,
17
- TemplateAttributeType,
18
- UIPropertiesMetadata,
19
- )
20
- from sinapsis_core.utils.env_var_keys import SINAPSIS_CACHE_DIR
21
- from sinapsis_data_visualization.helpers.plot_distributions import (
22
- plot_distribution,
23
- plot_map,
24
- )
25
- from sinapsis_data_visualization.helpers.scikit_pca_analysis import (
26
- perform_k_means_analysis,
27
- pre_process_images,
28
- )
29
-
30
-
31
- class PlotAttributes(TemplateAttributes):
32
- """Base attributes for all plot types.
33
-
34
- Attributes:
35
- fig_title (str|None): Title for the figure.
36
- y_label (str|None): Label of the y-axis.
37
- x_label (str|None): Label of the x-axis.
38
- fig_width (int): Width of the figure.
39
- fig_height (int): Height of the figure.
40
- x_position (float): X-position for the legend and/or texts.
41
- y_position (float): Y-position for the legend and/or texts.
42
- save_image_dir (str): Directory to save the image.
43
- fig_name (str): Name of the figure.
44
- """
45
-
46
- fig_title: str | None = None
47
- y_label: str | None = None
48
- x_label: str | None = None
49
- fig_width: int = 10
50
- fig_height: int = 10
51
- x_position: float = 0.5
52
- y_position: float = 0.5
53
- text: str | None = None
54
- save_image_dir: str | None = SINAPSIS_CACHE_DIR
55
- fig_name: str
56
- kwargs: dict[str, Any] = Field(default_factory=dict)
57
-
58
-
59
- @dataclass
60
- class PlotTypes:
61
- """Dataclass with different types of plots.
62
- HISTOGRAM: Key for the histogram plot
63
- BOX_PLOT: Key for the box plot
64
- PIE_CHART: Key for the pie chart plot
65
- CLUSTERING: Key for the k-means custering plot
66
- """
67
-
68
- HISTOGRAM: str = "histogram"
69
- BOX_PLOT: str = "box_plot"
70
- PIE_CHART: str = "pie_chart"
71
- CLUSTERING: str = "k_means_clustering"
72
-
73
-
74
- class DataDistributionVisualization(Template):
75
- """
76
- This template plots the distribution for image data, and can be
77
- extended to plot the distribution of different Packets, provided
78
- they have labels.
79
- The template allows for the drawing of histograms, box plots, pie charts
80
- and PCA from k-means cluster
81
- The template returns a figure in the container as an ImagePacket.
82
-
83
- Usage example:
84
-
85
- agent:
86
- name: my_test_agent
87
- templates:
88
- - template_name: InputTemplate
89
- class_name: InputTemplate
90
- attributes: {}
91
- - template_name: DataDistributionVisualization
92
- class_name: DataDistributionVisualization
93
- template_input: InputTemplate
94
- attributes:
95
- fig_title: null
96
- y_label: null
97
- x_label: null
98
- fig_width: 10
99
- fig_height: 10
100
- x_position: 0.5
101
- y_position: 0.5
102
- text: null
103
- save_image_dir: $SINAPSIS_CACHE_DIR
104
- fig_name: 'desired_name_for_the_figure'
105
- kwargs: '{}'
106
- histogram: false
107
- box_plot: false
108
- pie_chart: false
109
- k_means_clustering: false
110
-
111
-
112
- """
113
-
114
- UIProperties = UIPropertiesMetadata(output_type=OutputTypes.IMAGE)
115
-
116
- class AttributesBaseModel(PlotAttributes):
117
- """
118
- Attributes for data analysis using matplotlib.
119
-
120
- Attributes:
121
- histogram (bool): Whether to plot a histogram. Defaults to False.
122
- box_plot (bool): Whether to plot a box plot. Defaults to False.
123
- pie_chart (bool): Whether to plot a pie chart. Defaults to False.
124
- k_means_clustering (bool): Whether to plot k-means clustering. Defaults to False.
125
- """
126
-
127
- histogram: bool = False
128
- box_plot: bool = False
129
- pie_chart: bool = False
130
- k_means_clustering: bool = False
131
-
132
- def __init__(self, attributes: TemplateAttributeType) -> None:
133
- """Initializes the DataDistributionVisualization
134
- template with given attributes.
135
-
136
- Args:
137
- attributes (TemplateAttributeType): The attributes for
138
- configuring the data analysis.
139
- """
140
- super().__init__(attributes)
141
-
142
- if not os.path.exists(self.attributes.save_image_dir):
143
- os.makedirs(self.attributes.save_image_dir)
144
-
145
- def create_figure(self, text: str | None = None, rotate_x_ticks: bool = False) -> matplotlib.figure.Figure:
146
- """
147
- Creates the matplotlib.figure.Figure object and adds relevant
148
- information, such as the x-label, y-label and title.
149
- It also adds available text to the desired x and y positions and rotated
150
- the x-ticks is flag is True
151
-
152
- Args:
153
- text (str | None): Optional text to add to the figure. Defaults to None.
154
- rotate_x_ticks (bool): Flag to rotate x-ticks 90 degrees. Defaults to False
155
- Returns:
156
- The initialized figure with corresponding parameters
157
-
158
- """
159
- figure = plt.figure(figsize=(self.attributes.fig_width, self.attributes.fig_height))
160
- plt.xlabel(self.attributes.x_label or "X-Axis")
161
- plt.ylabel(self.attributes.y_label or "Y-Axis")
162
- plt.title(self.attributes.fig_title or "Plot Title")
163
-
164
- # Prepare the text for annotation
165
- if text:
166
- plt.text(self.attributes.x_position, self.attributes.y_position, text)
167
-
168
- if rotate_x_ticks:
169
- plt.xticks(rotation=90)
170
- return figure
171
-
172
- @staticmethod
173
- def retrieve_labels_from_images(
174
- container: DataContainer,
175
- ) -> tuple[list[str | int], list[int]]:
176
- """
177
- Iterates through the ImagePackets, extracting each of the
178
- annotations for all of them.
179
- Generates a dictionary with the label and number of points associated
180
- with that label.
181
- Returns:
182
- tuple[list[str | int], list[int]]: list of labels and number of coincidences for each label
183
- """
184
- labels: list[str | int] = [ann.label_str for image in container.images for ann in image.annotations]
185
-
186
- label_count: dict[str | int, int] = {}
187
- for label in labels:
188
- label_count[label] = label_count.get(label, 0) + 1
189
- labels, counts = list(label_count.keys()), list(label_count.values())
190
- return labels, counts
191
-
192
- @staticmethod
193
- def process_cluster(container: DataContainer) -> tuple[list[int | str], np.ndarray]:
194
- """
195
- For the k-means clustering, the images need to be flattened first
196
- and then returned as a reduced vector. This method calls the pre_process_images
197
- method from scikit-pca analysis to perform such preprocessing.
198
- Then, perform_k_means_analysis instantiates a KMeans class from scikit-learn
199
- and perform Principal Component Analysis (PCA) on the reduced vector.
200
-
201
-
202
- Args:
203
- container (DataContainer): Container to extract the images from
204
-
205
- Returns:
206
- list[str | int]: list of labels
207
- np.ndarray : transformed values from the pca analysis
208
- """
209
- images = container.images
210
- feature_arr = pre_process_images(images)
211
- label, counts = perform_k_means_analysis(feature_arr)
212
- return label, counts
213
-
214
- def execute(self, container: DataContainer) -> DataContainer:
215
- figure = self.create_figure()
216
- labels_to_plot, counts_to_plot = self.retrieve_labels_from_images(container)
217
- fig_name = self.attributes.fig_name
218
-
219
- if self.attributes.k_means_clustering:
220
- plot_type = PlotTypes.CLUSTERING
221
- fig_name = f"{PlotTypes.CLUSTERING}_{self.attributes.fig_name}"
222
- labels, feature_vec = self.process_cluster(container)
223
- figure = plot_distribution(
224
- figure=figure,
225
- labels=labels,
226
- counts=feature_vec,
227
- plot_type=plot_type,
228
- kwargs=self.attributes.kwargs,
229
- )
230
- else:
231
- for plot_type in plot_map:
232
- if not getattr(self.attributes, plot_type):
233
- continue
234
- figure = plot_distribution(
235
- figure=figure,
236
- labels=labels_to_plot,
237
- counts=counts_to_plot,
238
- plot_type=plot_type,
239
- kwargs=self.attributes.kwargs,
240
- )
241
-
242
- fig_name = f"{plot_type}_{self.attributes.fig_name}"
243
- fig_path = os.path.join(self.attributes.save_image_dir, fig_name)
244
- if figure:
245
- plt.savefig(f"{fig_path}_{container.container_id}")
246
- plt.close()
247
- container.generic_data[plot_type] = figure
248
- return container
@@ -1,213 +0,0 @@
1
- # -*- coding: utf-8 -*-
2
- import numpy as np
3
- from sinapsis_core.data_containers.data_packet import DataContainer
4
- from sinapsis_data_visualization.helpers.tags import Tags
5
- from sinapsis_data_visualization.templates.base_visualization_template import (
6
- BasePlotAttributes,
7
- BaseVisualizationTemplate,
8
- PlotTypes,
9
- )
10
- from sklearn.metrics import auc, confusion_matrix, precision_recall_curve, roc_curve
11
-
12
-
13
- class PredictionPlotTypes(PlotTypes):
14
- """Extends plot types with model prediction visualization types"""
15
-
16
- CONFUSION_MATRIX = "confusion_matrix"
17
- ROC_CURVE = "roc_curve"
18
- PRECISION_RECALL_CURVE = "precision_recall_curve"
19
-
20
-
21
- class PredictionVisualizationAttributes(BasePlotAttributes):
22
- """
23
- Attributes for prediction visualization.
24
-
25
- Attributes:
26
- generic_field_key (str): Key to access prediction results in generic_data (inherited from BasePlotAttributes).
27
- generic_training_key (str): Key to access training data in generic_data.
28
- target_column (str): Name of the target column if needed.
29
- confusion_matrix (bool): Whether to create a confusion matrix.
30
- roc_curve (bool): Whether to create a ROC curve.
31
- precision_recall_curve (bool): Whether to create a precision-recall curve.
32
- """
33
-
34
- generic_training_key: str = "fetch_openmlWrapper_viz"
35
- target_column: str = "target"
36
- confusion_matrix: bool = True
37
- roc_curve: bool = True
38
- precision_recall_curve: bool = True
39
-
40
-
41
- PredictionVisualizationUIProperties = BaseVisualizationTemplate.UIProperties
42
- PredictionVisualizationUIProperties.tags.extend([Tags.CONFUSION, Tags.CURVES, Tags.METRICS])
43
-
44
-
45
- class PredictionVisualization(BaseVisualizationTemplate):
46
- """
47
- Template for visualizing model predictions with evaluation metrics.
48
-
49
- This template supports confusion matrices, ROC curves, and precision-recall curves
50
- for classification models.
51
- """
52
-
53
- AttributesBaseModel = PredictionVisualizationAttributes
54
- UIProperties = PredictionVisualizationUIProperties
55
-
56
- def get_data_for_visualization(self, container: DataContainer) -> tuple[object, object]:
57
- """
58
- Extracts prediction and training data from the container.
59
-
60
- Args:
61
- container (DataContainer): Data container.
62
-
63
- Returns:
64
- tuple[object, object]: Tuple containing prediction results and training data.
65
- """
66
- predictions_data = self._get_generic_data(container, self.attributes.generic_field_key)
67
-
68
- if not predictions_data:
69
- self.logger.warning(f"No prediction data found with key '{self.attributes.generic_field_key}'")
70
- return (None, None)
71
-
72
- training_data = self._get_generic_data(container, self.attributes.generic_training_key)
73
-
74
- return predictions_data, training_data
75
-
76
- def _extract_true_and_predicted_values(self, data: tuple[object, object]) -> tuple[np.ndarray, np.ndarray]:
77
- """
78
- Extracts true and predicted values from data in a flexible way.
79
-
80
- Args:
81
- data (tuple[object, object]): Tuple containing prediction results and training data.
82
-
83
- Returns:
84
- tuple[np.ndarray, np.ndarray]: True values and predicted values.
85
- """
86
- predictions, training_data = data
87
- y_true = None
88
- y_pred = None
89
-
90
- if hasattr(predictions, "predictions"):
91
- y_pred = predictions.predictions
92
-
93
- if training_data is not None and hasattr(training_data, "y_test"):
94
- y_test = training_data.y_test
95
- y_true = y_test.values.flatten() if hasattr(y_test, "values") else np.array(y_test)
96
-
97
- return y_true, y_pred
98
-
99
- def prepare_confusion_matrix_data(self, data: tuple[object, object]) -> tuple[list[str], np.ndarray]:
100
- """
101
- Prepares data for a confusion matrix plot.
102
-
103
- Args:
104
- data (tuple[object, object]): Tuple containing prediction results and training data.
105
-
106
- Returns:
107
- tuple[list[str], np.ndarray]: Class labels and confusion matrix array.
108
- """
109
- y_true, y_pred = self._extract_true_and_predicted_values(data)
110
-
111
- if y_true is None or y_pred is None:
112
- self.logger.warning("Could not extract true and predicted values for confusion matrix")
113
- return [], np.array([])
114
-
115
- class_labels = sorted(set(np.concatenate([np.unique(y_true), np.unique(y_pred)])))
116
- class_labels = [str(label) for label in class_labels]
117
-
118
- cm = confusion_matrix(y_true, y_pred)
119
-
120
- return class_labels, cm
121
-
122
- def prepare_roc_curve_data(self, data: tuple[object, object]) -> tuple[list[str], np.ndarray]:
123
- """
124
- Prepares data for a ROC curve plot.
125
-
126
- Args:
127
- data (tuple[object, object]): Tuple containing prediction results and training data.
128
-
129
- Returns:
130
- tuple[list[str], np.ndarray]: Curve label and FPR/TPR values.
131
- """
132
- y_true, y_pred = self._extract_true_and_predicted_values(data)
133
-
134
- if y_true is None or y_pred is None:
135
- self.logger.warning("Could not extract true and predicted values for ROC curve")
136
- return [], np.array([])
137
-
138
- fpr, tpr, _ = roc_curve(y_true, y_pred)
139
- roc_auc = auc(fpr, tpr)
140
-
141
- curve_label = [f"ROC Curve (AUC = {roc_auc:.3f})"]
142
-
143
- curve_data = np.column_stack((fpr, tpr))
144
-
145
- return curve_label, curve_data
146
-
147
- def prepare_precision_recall_data(self, data: tuple[object, object]) -> tuple[list[str], np.ndarray]:
148
- """
149
- Prepares data f or a precision-recall curve plot.
150
-
151
- Args:
152
- data (tuple[object, object]): Tuple containing prediction results and training data.
153
-
154
- Returns:
155
- tuple[list[str], np.ndarray]: Curve label and precision/recall values.
156
- """
157
- y_true, y_pred = self._extract_true_and_predicted_values(data)
158
-
159
- if y_true is None or y_pred is None:
160
- self.logger.warning("Could not extract true and predicted values for precision-recall curve")
161
- return [], np.array([])
162
-
163
- precision, recall, _ = precision_recall_curve(y_true, y_pred)
164
-
165
- ap = np.mean(precision)
166
-
167
- curve_label = [f"Precision-Recall (AP = {ap:.3f})"]
168
-
169
- curve_data = np.column_stack((precision, recall))
170
-
171
- return curve_label, curve_data
172
-
173
- def generate_visualizations(self, container: DataContainer, data: tuple[object, object]) -> None:
174
- """
175
- Generates all requested visualizations for model predictions.
176
-
177
- Args:
178
- container (DataContainer): Data container.
179
- data (tuple[object, object]): Tuple containing prediction results and training data.
180
- """
181
- if not data or len(data) != 2 or not hasattr(data[0], "predictions"):
182
- self.logger.warning("No prediction data available for visualization")
183
- return
184
-
185
- if self.attributes.confusion_matrix:
186
- labels, cm_data = self.prepare_confusion_matrix_data(data)
187
- if len(labels) > 0 and cm_data.size > 0:
188
- self.plot_and_save(
189
- container=container,
190
- plot_type=PredictionPlotTypes.CONFUSION_MATRIX,
191
- labels=labels,
192
- counts=cm_data,
193
- )
194
-
195
- if self.attributes.roc_curve:
196
- labels, roc_data = self.prepare_roc_curve_data(data)
197
- if len(labels) > 0 and roc_data.size > 0:
198
- self.plot_and_save(
199
- container=container,
200
- plot_type=PredictionPlotTypes.ROC_CURVE,
201
- labels=labels,
202
- counts=roc_data,
203
- )
204
-
205
- if self.attributes.precision_recall_curve:
206
- labels, pr_data = self.prepare_precision_recall_data(data)
207
- if len(labels) > 0 and pr_data.size > 0:
208
- self.plot_and_save(
209
- container=container,
210
- plot_type=PredictionPlotTypes.PRECISION_RECALL_CURVE,
211
- labels=labels,
212
- counts=pr_data,
213
- )