[data.tbx11k] Re-structure database to accomodate bbox annotation querying

.pre-commit-config.yaml

@@ -7,11 +7,11 @@
 # See https://pre-commit.com/hooks.html for more hooks
 repos:
   - repo: https://github.com/psf/black
-    rev: 23.1.0
+    rev: 23.7.0
     hooks:
       - id: black
   - repo: https://github.com/pycqa/docformatter
-    rev: v1.5.1
+    rev: v1.7.5
     hooks:
       - id: docformatter
   - repo: https://github.com/pycqa/isort
@@ -19,11 +19,11 @@ repos:
     hooks:
       - id: isort
   - repo: https://github.com/pycqa/flake8
-    rev: 6.0.0
+    rev: 6.1.0
     hooks:
       - id: flake8
   - repo: https://github.com/pre-commit/mirrors-mypy
-    rev: v1.0.0
+    rev: v1.4.1
     hooks:
     - id: mypy
       args: [
@@ -33,10 +33,15 @@ repos:
         --ignore-missing-imports,
       ]
   - repo: https://github.com/asottile/pyupgrade
-    rev: v3.3.1
+    rev: v3.10.1
     hooks:
     - id: pyupgrade
-      args: [--py38-plus]
+      args: [--py39-plus]
+  # - repo: https://github.com/pre-commit/mirrors-prettier
+  #   rev: v2.7.1
+  #   hooks:
+  #   - id: prettier
+  #     types_or: [json]
   - repo: https://github.com/pre-commit/pre-commit-hooks
     rev: v4.4.0
     hooks:
@@ -53,6 +58,7 @@ repos:
               ^doc/results/img/rad_sign_drop.png
               )
       - id: check-toml
+      - id: check-json
       - id: check-yaml
         exclude: |
           (?x)(
@@ -65,6 +71,6 @@ repos:
       - id: end-of-file-fixer
       - id: debug-statements
   - repo: https://github.com/fsfe/reuse-tool
-    rev: v1.1.2
+    rev: v2.1.0
     hooks:
       - id: reuse

src/ptbench/data/tbx11k/datamodule.py

@@ -4,6 +4,7 @@
 
 import importlib.resources
 import os
+import typing
 
 import PIL.Image
 
@@ -16,23 +17,32 @@ from ..typing import DatabaseSplit
 from ..typing import RawDataLoader as _BaseRawDataLoader
 from ..typing import Sample
 
-DatabaseSample = (
-    tuple[str, int] | tuple[str, int, list[tuple[int, int, int, int, int]]]
+BoundingBoxAnnotation: typing.TypeAlias = tuple[int, int, int, int, int]
+"""Location of TB radiological findings (latent or active)
+
+Objects of this type carry bounding-box information of radiological findings on
+the original 512x512 pixel images of TBX11k.  The radiological findings are
+defined as such:
+
+* 0/1: This sign is for latent TB (0), or active TB (1)
+* xmin: horizontal position of bounding box upper-left corner, in pixels
+* ymin: vertical position of bounding box upper-left corner, in pixels
+* width: width of the bounding box, in pixels
+* height: height of the bounding box, in pixels
+"""
+
+DatabaseSample: typing.TypeAlias = (
+    tuple[str, int] | tuple[str, int, list[BoundingBoxAnnotation]]
 )
 """Type of objects in our JSON representation for this database.
 
-For healthy/sick (no TB)/latent TB cases, each sample is represented by a
-filename, followed by the number 0 (negative class).
-
-For active TB cases, each sample is represented by a filename, followed by the
-number 1, and then 1 or more 5-tuples with radiological finding locations.  The
-radiological findings are defined as such:
+For healthy/sick (no TB)/latent TB cases, each sample is represented by
+a filename, relative to the root of the installed database, followed by
+the number 0 (negative class).
 
-* 0/1: Signs of latent TB (0), or active TB (1)
-* xmin
-* ymin
-* width
-* height
+For active TB cases, each sample is represented by a filename, followed
+by the number 1, and then by 1 or more 5-tuples with radiological
+finding locations, as described above.
 """
 
 
@@ -104,6 +114,29 @@ class RawDataLoader(_BaseRawDataLoader):
         """
         return sample[1]
 
+    def bbox_annotations(
+        self, sample: DatabaseSample
+    ) -> list[BoundingBoxAnnotation]:
+        """Loads a single image sample label from the disk.
+
+        Parameters
+        ----------
+
+        sample:
+            A tuple containing the path suffix, within the dataset root folder,
+            where to find the image to be loaded, an integer, representing the
+            sample label, and possible radiological findings represented by
+            bounding boxes.
+
+
+        Returns
+        -------
+
+        annotations
+            Bounding box annotations, if any available with the sample.
+        """
+        return sample[2] if len(sample) > 2 else []  # type: ignore
+
 
 def make_split(basename: str) -> DatabaseSplit:
     """Returns a database split for the Montgomery database."""

src/ptbench/data/tbx11k/make_splits_from_database.py

@@ -76,7 +76,7 @@ def reorder(data: dict) -> list:
     assert len(set(images.values())) == len(images), "Image ids are not unique"
 
     retval: dict[str, list[typing.Any]] = {
-        k["file_name"]: [-1] for k in data["images"]
+        k["file_name"]: [-1, []] for k in data["images"]
     }
 
     # we now "consume" all annotations and assign each to an image
@@ -85,7 +85,12 @@ def reorder(data: dict) -> list:
             categories[annotation["category_id"]],
             *[round(k) for k in annotation["bbox"]],
         ]
-        retval[images[annotation["image_id"]]].append(int_bbox)
+        retval[images[annotation["image_id"]]][1].append(int_bbox)
+
+    # remove empty bounding-box entries to save space on final JSON
+    for v in retval.values():
+        if not v[1]:
+            del v[1]
 
     return sorted([["imgs/" + k, *v] for k, v in retval.items()])
 
@@ -127,7 +132,7 @@ def normalize_labels(data: list) -> list:
                     f"but contains no tb bbox annotations"
                 )
                 return -1  # unknown diagnosis
-            bbx_labels: list[int] = [k[0] for k in s[2:]]
+            bbx_labels: list[int] = [k[0] for k in s[2]]
             tb_counts = collections.Counter(bbx_labels)
             assert 2 not in tb_counts, (
                 f"Label 2 (PulmonaryTuberculosis) was used in image {s[0]} "
@@ -149,7 +154,10 @@ def normalize_labels(data: list) -> list:
         else:
             raise RuntimeError("Cannot happen - please check")
 
-    return [[k[0], _set_label(k), *k[2:]] for k in data]
+    for k in data:
+        k[1] = _set_label(k)
+
+    return data
 
 
 def print_statistics(d: dict):