dataflow-cv 2.0.0__tar.gz → 2.0.1__tar.gz

{dataflow_cv-2.0.0 → dataflow_cv-2.0.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dataflow-cv
-Version: 2.0.0
+Version: 2.0.1
 Summary: A computer vision dataset processing library — analyse, convert, visualize, and evaluate annotations across YOLO, LabelMe, and COCO formats
 Author: DataFlow-CV Team
 License: MIT
@@ -33,9 +33,7 @@ Requires-Dist: pycocotools>=2.0.0; extra == "coco"
 Provides-Extra: dev
 Requires-Dist: pytest>=7.0.0; extra == "dev"
 Requires-Dist: pytest-cov>=4.0.0; extra == "dev"
-Requires-Dist: black>=22.0.0; extra == "dev"
-Requires-Dist: isort>=5.12.0; extra == "dev"
-Requires-Dist: flake8>=6.0.0; extra == "dev"
+Requires-Dist: ruff>=0.16; extra == "dev"
 Requires-Dist: mypy>=1.0.0; extra == "dev"
 Dynamic: license-file
 
@@ -67,6 +65,7 @@ A computer vision dataset processing library — analyse, convert, visualize, an
 | 🎨 **Visualize** | OpenCV rendering with color-coded classes, display & save modes | `dataflow-cv visualize yolo ...` |
 | 📊 **Evaluate** | COCO mAP via pycocotools, single-threshold P/R/F1 per class | `dataflow-cv evaluate detection ...` |
 | 💻 **CLI + API** | Click-based CLI with rich `--help`; Python API for pipelines | `from dataflow.convert import ...` |
+| 🤖 **AI Skills** | Claude Code skill (`/dataflow:dataflow-cv`) for AI assistants — CLI/API reference & known gotchas | `claude plugin install dataflow@claude-skills` |
 
 ---
 

{dataflow_cv-2.0.0 → dataflow_cv-2.0.1}/README.md

@@ -26,6 +26,7 @@ A computer vision dataset processing library — analyse, convert, visualize, an
 | 🎨 **Visualize** | OpenCV rendering with color-coded classes, display & save modes | `dataflow-cv visualize yolo ...` |
 | 📊 **Evaluate** | COCO mAP via pycocotools, single-threshold P/R/F1 per class | `dataflow-cv evaluate detection ...` |
 | 💻 **CLI + API** | Click-based CLI with rich `--help`; Python API for pipelines | `from dataflow.convert import ...` |
+| 🤖 **AI Skills** | Claude Code skill (`/dataflow:dataflow-cv`) for AI assistants — CLI/API reference & known gotchas | `claude plugin install dataflow@claude-skills` |
 
 ---
 

{dataflow_cv-2.0.0 → dataflow_cv-2.0.1}/dataflow/__init__.py

@@ -1,4 +1,4 @@
 """DataFlow-CV: A computer vision dataset processing library."""
 
-__version__ = "2.0.0"
+__version__ = "2.0.1"
 __author__ = "DataFlow-CV Team"

{dataflow_cv-2.0.0 → dataflow_cv-2.0.1}/dataflow/analyse/filter.py

@@ -24,8 +24,7 @@ from .log_templates import (
     format_filter_result,
 )
 from .utils import create_handler, detect_format, load_class_names
-from dataflow.label.models import (DatasetAnnotations, ImageAnnotation,
-                                    ObjectAnnotation)
+from dataflow.label.models import DatasetAnnotations, ImageAnnotation, ObjectAnnotation
 
 
 class FilterAnalyser(BaseAnalyser):
@@ -99,17 +98,14 @@ class FilterAnalyser(BaseAnalyser):
         for new_id, name in new_classes.items():
             if name in name_to_old_id:
                 old_id = name_to_old_id[name]
-                mapping = CategoryMapping(
-                    new_id=new_id, old_id=old_id, name=name
-                )
+                mapping = CategoryMapping(new_id=new_id, old_id=old_id, name=name)
                 old_to_new[old_id] = mapping
                 kept.append(mapping)
             else:
                 missing.append(name)
                 if logger:
                     logger.warning(
-                        f'Category "{name}" in new class file not '
-                        f"found in source — skipping"
+                        f'Category "{name}" in new class file not found in source — skipping'
                     )
 
         # Build removed list
@@ -134,10 +130,7 @@ class FilterAnalyser(BaseAnalyser):
         total_after = 0
 
         for image_ann in dataset.images:
-            filtered = [
-                obj for obj in image_ann.objects
-                if obj.class_id in old_to_new
-            ]
+            filtered = [obj for obj in image_ann.objects if obj.class_id in old_to_new]
             for obj in filtered:
                 mapping = old_to_new[obj.class_id]
                 obj.class_id = mapping.new_id
@@ -202,9 +195,7 @@ class FilterAnalyser(BaseAnalyser):
             return result
 
         if not new_classes:
-            result.add_error(
-                f"No valid class names in new class file: {new_class_file}"
-            )
+            result.add_error(f"No valid class names in new class file: {new_class_file}")
             return result
 
         # ---- 3. Detect format + create handler ------------------------
@@ -259,9 +250,7 @@ class FilterAnalyser(BaseAnalyser):
         )
 
         if not old_to_new:
-            result.add_error(
-                "No matching categories between source and new class file"
-            )
+            result.add_error("No matching categories between source and new class file")
             return result
 
         # ---- 5. Ensure output directory --------------------------------
@@ -304,14 +293,16 @@ class FilterAnalyser(BaseAnalyser):
                             # rather than mutating the original (avoid
                             # aliasing — the original may be reused by the
                             # iterator or shared across images).
-                            filtered_objects.append(ObjectAnnotation(
-                                class_id=mapping.new_id,
-                                class_name=mapping.name,
-                                bbox=obj.bbox,
-                                segmentation=obj.segmentation,
-                                confidence=obj.confidence,
-                                is_crowd=obj.is_crowd,
-                            ))
+                            filtered_objects.append(
+                                ObjectAnnotation(
+                                    class_id=mapping.new_id,
+                                    class_name=mapping.name,
+                                    bbox=obj.bbox,
+                                    segmentation=obj.segmentation,
+                                    confidence=obj.confidence,
+                                    is_crowd=obj.is_crowd,
+                                )
+                            )
                     total_after += len(filtered_objects)
 
                     if filtered_objects:
@@ -328,9 +319,7 @@ class FilterAnalyser(BaseAnalyser):
                     wr = write_handler.write_one(filtered_img, output_dir)
                     if not wr.success:
                         for err in wr.errors:
-                            result.add_error(
-                                f"Write {image_ann.image_id}: {err}"
-                            )
+                            result.add_error(f"Write {image_ann.image_id}: {err}")
                         return result
             except Exception as e:
                 result.add_error(f"Failed during streaming filter: {e}")
@@ -342,23 +331,19 @@ class FilterAnalyser(BaseAnalyser):
             total_files = dataset.num_images
             total_before = dataset.num_objects
 
-            total_files_with_annotations, total_after = (
-                self._filter_dataset_images(dataset, old_to_new)
+            total_files_with_annotations, total_after = self._filter_dataset_images(
+                dataset, old_to_new
             )
 
             # Update categories
-            dataset.categories = {
-                km.new_id: km.name for km in kept_categories
-            }
+            dataset.categories = {km.new_id: km.name for km in kept_categories}
 
             try:
                 for image_ann in dataset.images:
                     wr = handler.write_one(image_ann, output_dir)
                     if not wr.success:
                         for err in wr.errors:
-                            result.add_error(
-                                f"Write {image_ann.image_id}: {err}"
-                            )
+                            result.add_error(f"Write {image_ann.image_id}: {err}")
                         return result
             except Exception as e:
                 result.add_error(f"Failed to write filtered output: {e}")
@@ -369,14 +354,12 @@ class FilterAnalyser(BaseAnalyser):
             total_files = dataset.num_images
             total_before = dataset.num_objects
 
-            total_files_with_annotations, total_after = (
-                self._filter_dataset_images(dataset, old_to_new)
+            total_files_with_annotations, total_after = self._filter_dataset_images(
+                dataset, old_to_new
             )
 
             # Update categories to match new class file
-            dataset.categories = {
-                km.new_id: km.name for km in kept_categories
-            }
+            dataset.categories = {km.new_id: km.name for km in kept_categories}
 
             output_path = output_dir / label_path.name
             try:
@@ -413,29 +396,20 @@ class FilterAnalyser(BaseAnalyser):
         if missing_categories:
             s = "y" if len(missing_categories) == 1 else "ies"
             result.add_warning(
-                f"{len(missing_categories)} categor{s} in new class "
-                f"file not found in source"
+                f"{len(missing_categories)} categor{s} in new class file not found in source"
             )
 
         if total_after == 0 and total_before > 0:
-            result.add_warning(
-                "All annotations were filtered out — output files are empty"
-            )
+            result.add_warning("All annotations were filtered out — output files are empty")
 
         # ---- 9. Log output ---------------------------------------------
         self._log_info(
-            format_analyse_header(
-                "Category Filter", label_path, f"{fmt} (auto-detected)"
-            )
+            format_analyse_header("Category Filter", label_path, f"{fmt} (auto-detected)")
         )
         self._log_info(
-            f"  Original class: {original_class_file.name} "
-            f"({len(original_classes)} categories)"
-        )
-        self._log_info(
-            f"  New class:      {new_class_file.name} "
-            f"({len(new_classes)} categories)\n"
+            f"  Original class: {original_class_file.name} ({len(original_classes)} categories)"
         )
+        self._log_info(f"  New class:      {new_class_file.name} ({len(new_classes)} categories)\n")
         self._log_info(
             format_filter_result(
                 total_files=total_files,
@@ -449,8 +423,6 @@ class FilterAnalyser(BaseAnalyser):
             )
         )
         if result.log_path:
-            self._log_info(
-                format_analyse_result("✓ Success", result.log_path)
-            )
+            self._log_info(format_analyse_result("✓ Success", result.log_path))
 
         return result

{dataflow_cv-2.0.0 → dataflow_cv-2.0.1}/dataflow/analyse/log_templates.py

@@ -55,8 +55,9 @@ def format_analyse_header(
         lines.append(format_kv("Source", label))
     else:
         lines.append(
-            format_kv("Sources", f"{', '.join(str(p) for p in path_list)}"
-                      f"    ({len(path_list)} paths)")
+            format_kv(
+                "Sources", f"{', '.join(str(p) for p in path_list)}    ({len(path_list)} paths)"
+            )
         )
 
     if class_file is not None:
@@ -81,11 +82,7 @@ def format_stats_path_breakdown(path_stats) -> str:
         label = f"{ps['path']}"
         if ps.get("recursive"):
             label += " (recursive)"
-        lines.append(
-            f"  {label:<40} "
-            f"{ps['files']:>5} files, "
-            f"{ps['annotations']:>5} annotations"
-        )
+        lines.append(f"  {label:<40} {ps['files']:>5} files, {ps['annotations']:>5} annotations")
     lines.append("")
     return "\n".join(lines)
 
@@ -131,19 +128,23 @@ def format_stats_result(
                 for name, count in per_class.items()
             ]
             rows.append(["─" * 15, "─" * 4, "─" * 7])
-            rows.append([
-                f"Total ({len(per_class)})",
-                "",
-                str(sum(per_class.values())),
-            ])
+            rows.append(
+                [
+                    f"Total ({len(per_class)})",
+                    "",
+                    str(sum(per_class.values())),
+                ]
+            )
         else:
             headers = ["Class", "Count"]
             rows = [[name, str(count)] for name, count in per_class.items()]
             rows.append(["─" * 15, "─" * 7])
-            rows.append([
-                f"Total ({len(per_class)})",
-                str(sum(per_class.values())),
-            ])
+            rows.append(
+                [
+                    f"Total ({len(per_class)})",
+                    str(sum(per_class.values())),
+                ]
+            )
         lines.append(format_section("Per-Class"))
         lines.append(format_table(headers, rows))
     else:
@@ -253,7 +254,7 @@ def format_filter_result(
         for name in missing_categories:
             lines.append(f'    "{name}"')
     else:
-        lines.append(f"  Not found in source: 0 categories")
+        lines.append("  Not found in source: 0 categories")
     lines.append("")
 
     # ── Filter Summary ──
@@ -262,10 +263,12 @@ def format_filter_result(
     lines.append(format_kv("Files with annotations", str(total_files_with_annotations)))
     lines.append(format_kv("Annotations before", str(annotations_before)))
     lines.append(format_kv("Annotations after", str(annotations_after)))
-    lines.append(format_kv(
-        "Output",
-        f"{total_files_with_annotations} files → {output_dir}",
-    ))
+    lines.append(
+        format_kv(
+            "Output",
+            f"{total_files_with_annotations} files → {output_dir}",
+        )
+    )
 
     return "\n".join(lines)
 
@@ -295,14 +298,12 @@ def format_partition_result(
     Returns:
         Formatted partition summary.
     """
-    mode_label = {"images": "Images only", "labels": "Labels only",
-                  "both": "Labels + Images"}[mode]
+    mode_label = {"images": "Images only", "labels": "Labels only", "both": "Labels + Images"}[mode]
 
     lines = [
         format_kv("Mode", mode_label),
         format_kv("Partitions", str(num_partitions)),
-        format_kv("Shuffle", f"{'Yes' if shuffle else 'No'}"
-                  f"{f' (seed={seed})' if shuffle else ''}"),
+        format_kv("Shuffle", f"{'Yes' if shuffle else 'No'}{f' (seed={seed})' if shuffle else ''}"),
         format_kv("Move", "Yes" if move else "No"),
         format_kv("Total files", str(total_files)),
         "",
@@ -310,10 +311,7 @@ def format_partition_result(
     ]
 
     for i in range(num_partitions):
-        lines.append(
-            f"  Part {i + 1}:  {partition_sizes[i]:>6} files → "
-            f"{partition_dirs[i]}"
-        )
+        lines.append(f"  Part {i + 1}:  {partition_sizes[i]:>6} files → {partition_dirs[i]}")
 
     return "\n".join(lines)
 

{dataflow_cv-2.0.0 → dataflow_cv-2.0.1}/dataflow/analyse/partition.py

@@ -26,7 +26,6 @@ from .utils import (
     _IMAGE_EXTENSIONS,
     create_handler,
     detect_format,
-    load_class_names,
 )
 
 
@@ -93,15 +92,11 @@ class PartitionAnalyser(BaseAnalyser):
         # 1. Validate inputs
         # ------------------------------------------------------------------
         if num < 2:
-            result.add_error(
-                f"Number of partitions must be at least 2, got: {num}"
-            )
+            result.add_error(f"Number of partitions must be at least 2, got: {num}")
             return result
 
         if label_dir is None and image_dir is None:
-            result.add_error(
-                "At least one of label_dir or image_dir must be provided"
-            )
+            result.add_error("At least one of label_dir or image_dir must be provided")
             return result
 
         # ------------------------------------------------------------------
@@ -246,15 +241,9 @@ class PartitionAnalyser(BaseAnalyser):
                         wr = handler.write_one(img_ann, part_dir)
                         if not wr.success:
                             for err in wr.errors:
-                                result.add_error(
-                                    f"Write {part_dir.name}/"
-                                    f"{img_ann.image_id}: {err}"
-                                )
+                                result.add_error(f"Write {part_dir.name}/{img_ann.image_id}: {err}")
                     except Exception as e:
-                        result.add_error(
-                            f"Write {part_dir.name}/"
-                            f"{img_ann.image_id}: {e}"
-                        )
+                        result.add_error(f"Write {part_dir.name}/{img_ann.image_id}: {e}")
 
                 # Move label files if in move mode
                 if move:
@@ -265,9 +254,7 @@ class PartitionAnalyser(BaseAnalyser):
                         else:  # labelme
                             src_label = label_dir / f"{img_ann.image_id}.json"
                         if src_label.exists():
-                            _copy_or_move_file(
-                                src_label, part_dir, move=True, logger=self.logger
-                            )
+                            _copy_or_move_file(src_label, part_dir, move=True, logger=self.logger)
 
             elif mode == "both":
                 labels_subdir = part_dir / "labels"
@@ -282,25 +269,18 @@ class PartitionAnalyser(BaseAnalyser):
                         if not wr.success:
                             for err in wr.errors:
                                 result.add_error(
-                                    f"Write {part_dir.name}/labels/"
-                                    f"{img_ann.image_id}: {err}"
+                                    f"Write {part_dir.name}/labels/{img_ann.image_id}: {err}"
                                 )
                     except Exception as e:
-                        result.add_error(
-                            f"Write {part_dir.name}/labels/"
-                            f"{img_ann.image_id}: {e}"
-                        )
+                        result.add_error(f"Write {part_dir.name}/labels/{img_ann.image_id}: {e}")
 
                     # Match and copy/move image
                     stem = img_ann.image_id
                     if stem in image_stems:
-                        _copy_or_move_file(
-                            image_stems[stem], images_subdir, move, self.logger
-                        )
+                        _copy_or_move_file(image_stems[stem], images_subdir, move, self.logger)
                     else:
                         self._log_warning(
-                            f"No matching image found for label "
-                            f"'{img_ann.image_id}' in {image_dir}"
+                            f"No matching image found for label '{img_ann.image_id}' in {image_dir}"
                         )
 
                 # Move label files if in move mode
@@ -312,21 +292,16 @@ class PartitionAnalyser(BaseAnalyser):
                             src_label = label_dir / f"{img_ann.image_id}.json"
                         if src_label.exists():
                             _copy_or_move_file(
-                                src_label, labels_subdir,
-                                move=True, logger=self.logger
+                                src_label, labels_subdir, move=True, logger=self.logger
                             )
 
                 # Report unmatched images (in image_dir but not in labels)
                 if not move:  # Only warn for copy mode; move mode self-resolves
-                    label_stems = {
-                        img_ann.image_id
-                        for img_ann in items[start:end]
-                    }
+                    label_stems = {img_ann.image_id for img_ann in items[start:end]}
                     for stem, img_path in image_stems.items():
                         if stem not in label_stems:
                             self._log_warning(
-                                f"Image '{img_path.name}' has no matching "
-                                f"label — skipped"
+                                f"Image '{img_path.name}' has no matching label — skipped"
                             )
 
             # Copy class_file to partition directory
@@ -336,10 +311,7 @@ class PartitionAnalyser(BaseAnalyser):
                     if not target_cf.exists():
                         shutil.copy2(str(class_file), str(target_cf))
                 except OSError as e:
-                    result.add_warning(
-                        f"Could not copy class file to "
-                        f"{part_dir.name}: {e}"
-                    )
+                    result.add_warning(f"Could not copy class file to {part_dir.name}: {e}")
 
         # ------------------------------------------------------------------
         # 6. Build result
@@ -361,8 +333,9 @@ class PartitionAnalyser(BaseAnalyser):
         # ------------------------------------------------------------------
         # 7. Log output
         # ------------------------------------------------------------------
-        mode_label = {"images": "Images Only", "labels": "Labels Only",
-                      "both": "Labels + Images"}[mode]
+        mode_label = {"images": "Images Only", "labels": "Labels Only", "both": "Labels + Images"}[
+            mode
+        ]
         label_paths = label_dir if label_dir else image_dir
         self._log_info(
             format_analyse_header(
@@ -385,8 +358,6 @@ class PartitionAnalyser(BaseAnalyser):
             )
         )
         if result.log_path:
-            self._log_info(
-                format_analyse_result("✓ Success", result.log_path)
-            )
+            self._log_info(format_analyse_result("✓ Success", result.log_path))
 
         return result

{dataflow_cv-2.0.0 → dataflow_cv-2.0.1}/dataflow/analyse/sample.py

@@ -103,15 +103,11 @@ class SampleAnalyser(BaseAnalyser):
         # 1. Validate inputs
         # ------------------------------------------------------------------
         if label_dir is None and image_dir is None:
-            result.add_error(
-                "At least one of label_dir or image_dir must be provided"
-            )
+            result.add_error("At least one of label_dir or image_dir must be provided")
             return result
 
         if count < 1:
-            result.add_error(
-                f"Count must be at least 1, got: {count}"
-            )
+            result.add_error(f"Count must be at least 1, got: {count}")
             return result
 
         # ------------------------------------------------------------------
@@ -190,8 +186,7 @@ class SampleAnalyser(BaseAnalyser):
         actual_count = min(count, total)
         if count > total:
             result.add_warning(
-                f"Requested {count} files but only {total} available — "
-                f"collecting all"
+                f"Requested {count} files but only {total} available — collecting all"
             )
 
         sampled = items[:actual_count]
@@ -221,13 +216,10 @@ class SampleAnalyser(BaseAnalyser):
                 _copy_or_move_file(src_path, label_subdir, move, self.logger)
                 # Match and copy image
                 if stem in image_stems:
-                    _copy_or_move_file(
-                        image_stems[stem], image_subdir, move, self.logger
-                    )
+                    _copy_or_move_file(image_stems[stem], image_subdir, move, self.logger)
                 else:
                     self._log_warning(
-                        f"No matching image found for label '{stem}' "
-                        f"in image directory"
+                        f"No matching image found for label '{stem}' in image directory"
                     )
                     unmatched_image_warnings += 1
             else:
@@ -238,10 +230,7 @@ class SampleAnalyser(BaseAnalyser):
             sampled_stems = {stem for _, stem in sampled}
             for stem, img_path in image_stems.items():
                 if stem not in sampled_stems:
-                    self._log_warning(
-                        f"Image '{img_path.name}' has no matching "
-                        f"label — skipped"
-                    )
+                    self._log_warning(f"Image '{img_path.name}' has no matching label — skipped")
 
         # ------------------------------------------------------------------
         # 7. Copy class_file to output_dir if provided
@@ -255,9 +244,7 @@ class SampleAnalyser(BaseAnalyser):
                     else:
                         shutil.copy2(str(class_file), str(target_cf))
             except OSError as e:
-                result.add_warning(
-                    f"Could not copy class file: {e}"
-                )
+                result.add_warning(f"Could not copy class file: {e}")
 
         # ------------------------------------------------------------------
         # 8. Build result
@@ -306,8 +293,6 @@ class SampleAnalyser(BaseAnalyser):
             )
         )
         if result.log_path:
-            self._log_info(
-                format_analyse_result("✓ Success", result.log_path)
-            )
+            self._log_info(format_analyse_result("✓ Success", result.log_path))
 
         return result

{dataflow_cv-2.0.0 → dataflow_cv-2.0.1}/dataflow/analyse/split.py

@@ -101,15 +101,11 @@ class SplitAnalyser(BaseAnalyser):
         # 1. Validate inputs
         # ------------------------------------------------------------------
         if label_dir is None and image_dir is None:
-            result.add_error(
-                "At least one of label_dir or image_dir must be provided"
-            )
+            result.add_error("At least one of label_dir or image_dir must be provided")
             return result
 
         if not 0.0 < ratio < 1.0:
-            result.add_error(
-                f"Ratio must be between 0 and 1 (exclusive), got: {ratio}"
-            )
+            result.add_error(f"Ratio must be between 0 and 1 (exclusive), got: {ratio}")
             return result
 
         # ------------------------------------------------------------------
@@ -238,10 +234,7 @@ class SplitAnalyser(BaseAnalyser):
             label_stems = {stem for _, stem in train_items + val_items}
             for stem, img_path in image_stems.items():
                 if stem not in label_stems:
-                    self._log_warning(
-                        f"Image '{img_path.name}' has no matching "
-                        f"label — skipped"
-                    )
+                    self._log_warning(f"Image '{img_path.name}' has no matching label — skipped")
 
         # ------------------------------------------------------------------
         # 7. Copy class_file to both output dirs if provided
@@ -302,9 +295,7 @@ class SplitAnalyser(BaseAnalyser):
             )
         )
         if result.log_path:
-            self._log_info(
-                format_analyse_result("✓ Success", result.log_path)
-            )
+            self._log_info(format_analyse_result("✓ Success", result.log_path))
 
         return result
 
@@ -343,26 +334,16 @@ def _split_files(
         for label_path, stem in train_items:
             _copy_or_move_file(label_path, train_label_dir, move, logger)
             if stem in image_stems:
-                _copy_or_move_file(
-                    image_stems[stem], train_image_dir, move, logger
-                )
+                _copy_or_move_file(image_stems[stem], train_image_dir, move, logger)
             else:
-                logger.warning(
-                    f"No matching image found for label '{stem}' in "
-                    f"image directory"
-                )
+                logger.warning(f"No matching image found for label '{stem}' in image directory")
 
         for label_path, stem in val_items:
             _copy_or_move_file(label_path, val_label_dir, move, logger)
             if stem in image_stems:
-                _copy_or_move_file(
-                    image_stems[stem], val_image_dir, move, logger
-                )
+                _copy_or_move_file(image_stems[stem], val_image_dir, move, logger)
             else:
-                logger.warning(
-                    f"No matching image found for label '{stem}' in "
-                    f"image directory"
-                )
+                logger.warning(f"No matching image found for label '{stem}' in image directory")
     else:
         train_dir = output_dir / "train"
         val_dir = output_dir / "val"