fix: populate class_name in DetectionDataset annotations

[happyhj]

DetectionDataset.init now maps class_id to class names using CLASS_NAME_DATA_FIELD so that LabelAnnotator displays human-readable labels instead of raw integer IDs.

Closes #1772

Before submitting

• Self-reviewed the code
• Updated documentation, follow Google-style
• Added docs entry for autogeneration (if new functions/classes)
• Added/updated tests
• All tests pass locally

Description

DetectionDataset.__init__ now populates CLASS_NAME_DATA_FIELD in each annotation's data dict by mapping class_id to the dataset's classes list. This is done centrally in the constructor so all format loaders (from_yolo, from_coco, from_pascal_voc) benefit without individual changes.

Type of Change

• 🐛 Bug fix (non-breaking change which fixes an issue)

Motivation and Context

When loading a dataset via DetectionDataset.from_yolo(...) and annotating with LabelAnnotator, labels showed raw integer class IDs (e.g. 0, 1) instead of human-readable class names (e.g. dog, person). This is because the class_name data field was never populated on dataset-loaded Detections, unlike Detections.from_inference() which does set it.

Closes #1772

Changes Made

• Added CLASS_NAME_DATA_FIELD import to src/supervision/dataset/core.py
• Added 7 lines in DetectionDataset.__init__ to populate class_name from classes and class_id
• Added 5 test cases in tests/dataset/test_core.py covering init, empty detections, empty classes, merge, and YOLO integration

Testing

• I have tested this code locally
• I have added unit tests that prove my fix is effective or that my feature works
• All new and existing tests pass

Google Colab

Before fix: https://colab.research.google.com/drive/1ycOTk-zxUnjD_ZHcD8LBDlUVc9jij2Iu?usp=sharing
After fix: https://colab.research.google.com/drive/1kkz0YbwJv7X-9o9AVWu6v601sR4Bc5q9?usp=sharing

Screenshots/Videos

Additional Notes

No changes to individual format loaders or annotator logic were needed.

[CLAassistant]

All committers have signed the CLA.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 74%. Comparing base (d70bc4f) to head (e2b9f34).
⚠️ Report is 2 commits behind head on develop.

Additional details and impacted files

@@           Coverage Diff           @@
##           develop   #2156   +/-   ##
=======================================
  Coverage       74%     74%           
=======================================
  Files           62      62           
  Lines         7486    7492    +6     
=======================================
+ Hits          5547    5554    +7     
+ Misses        1939    1938    -1     

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Copilot]

Pull request overview

This PR fixes a bug where DetectionDataset displays raw class IDs (e.g., 0, 1) instead of human-readable class names (e.g., "dog", "cat") when using LabelAnnotator. The fix populates the class_name data field centrally in DetectionDataset.__init__ by mapping class_id values to class names from the dataset's classes list.

Changes:

• Added logic to populate CLASS_NAME_DATA_FIELD in DetectionDataset.__init__ for all annotations
• Added comprehensive test coverage including edge cases (empty annotations, merge operations, YOLO integration)

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
src/supervision/dataset/core.py	Added import of CLASS_NAME_DATA_FIELD and logic in __init__ to populate class_name in annotation data by indexing the classes list with class_id values
tests/dataset/test_core.py	Added TestClassNamePopulation test class with 5 test cases covering initialization, empty annotations, empty classes, merge operations, and YOLO format integration