fix: sort YOLO class names by numeric keys#2296
Merged
Borda merged 9 commits intoJun 7, 2026
Merged
Conversation
|
|
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@
## develop #2296 +/- ##
=======================================
Coverage 79% 79%
=======================================
Files 66 66
Lines 8622 8640 +18
=======================================
+ Hits 6846 6863 +17
- Misses 1776 1777 +1 🚀 New features to boost your workflow:
|
Contributor
There was a problem hiding this comment.
Pull request overview
This PR fixes YOLO data.yaml class-name extraction to sort names mapping keys numerically when they represent class indices, preventing lexicographic mis-ordering (e.g., "10" before "2"), and adds a regression test to lock the behavior.
Changes:
- Update
_extract_class_namesto sort dict keys using numeric ordering when possible. - Add a regression test covering quoted numeric YAML keys (
'10'vs'2').
Assessment (n/5)
- Code quality: 3/5
- Testing: 4/5
- Documentation: 5/5
Reviewed changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.
| File | Description |
|---|---|
src/supervision/dataset/formats/yolo.py |
Adjusts sorting logic for names dict keys to avoid lexicographic ordering for numeric-like keys. |
tests/dataset/formats/test_yolo.py |
Adds a regression test to ensure quoted numeric class keys are ordered by numeric index. |
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
`stripped.lstrip("-").isdigit()` stripped ALL leading hyphens, so "--1"
passed _is_int_like() but int("--1") raised ValueError at sort time.
Changed to `stripped.isdigit()` which is consistent with int() and also
rejects negative class indices (non-negative by YOLO spec).
---
Co-authored-by: Claude Code <noreply@anthropic.com>
When a data.yaml names dict has both int-like keys (e.g. "0", "2") and non-int-like keys (e.g. "foo"), the previous code silently fell back to lexicographic sort, producing wrong class-index order for numeric keys. Now raises ValueError with a clear message listing example keys from each category, consistent with the function's existing validation style. --- Co-authored-by: Claude Code <noreply@anthropic.com>
Convert the one-off numeric-string test to a parametrized suite covering: - Quoted string numeric keys (original bug scenario, '10' after '2') - Native int keys (most common real YOLO format from Ultralytics/Roboflow) - Non-numeric string keys (lexicographic fallback path) - Empty names dict (vacuously correct edge case) - Mixed numeric/non-numeric keys (ValueError from new mixed-key guard) --- Co-authored-by: Claude Code <noreply@anthropic.com>
…able sorted(..., key=lambda key: int(key)) — the lambda parameter shadows the outer `keys` list variable name. Using `k` makes the scope boundary clear. --- Co-authored-by: Claude Code <noreply@anthropic.com>
Without this guard, isinstance(True, int) == True causes YAML `true`/`false` keys to be treated as class indices 1 and 0. --- Co-authored-by: Claude Code <noreply@anthropic.com>
Documents the two-branch sort logic (numeric vs lexicographic), the bool exclusion invariant, and the new mixed-key ValueError — non-obvious after the recent changes. --- Co-authored-by: Claude Code <noreply@anthropic.com>
Borda
previously approved these changes
Jun 7, 2026
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
data.yamlmapping keys numerically when they are numeric-like strings."10", which previously sorted before"2".Root cause
_extract_class_namessorted mapping keys directly, so quoted numeric YAML keys used lexicographic order instead of class index order.Tests
uv run pytest tests/dataset/formats/test_yolo.py::test_extract_class_names_sorts_numeric_string_keys -quv run pytest tests/dataset/formats -quv run pytest -quv run pre-commit run --files src/supervision/dataset/formats/yolo.py tests/dataset/formats/test_yolo.pygit diff --check