[resolve #8] /review finding by qa-specialist (report: .temp/output-review-from-inference-rle-mask-2026-04-14.md):
"No test covers mixed RLE + box-only predictions in the same batch. Produces misaligned masks (1 mask for 2 detections)..."

---
Co-authored-by: Claude Code <noreply@anthropic.com>

- Move supervision/keypoint/ → src/supervision/keypoint/ so shims are
  included in built wheels (project uses src layout)
- Replace raw DeprecationWarning with warn_deprecated (SupervisionWarnings)
  in all three shim files for consistent warning behaviour
- Fix wrong import path (supervision.utils.deprecation → .internal)
  and remove unsupported stacklevel arg from warn_deprecated call

[resolve #7] Review comment by @copilot (PR #2214):
"This warning uses the standard DeprecationWarning..."
[resolve #8] Review comment by @copilot (PR #2214):
"These deprecation shim modules are added under the top-level supervision/..."

---
Co-authored-by: Claude Code <noreply@anthropic.com>

- Move supervision/keypoint/ → src/supervision/keypoint/ so shims are
  included in built wheels (project uses src layout)
- Replace raw DeprecationWarning with warn_deprecated (SupervisionWarnings)
  in all three shim files for consistent warning behaviour
- Fix wrong import path (supervision.utils.deprecation → .internal)
  and remove unsupported stacklevel arg from warn_deprecated call

[resolve #7] Review comment by @copilot (PR #2214):
"This warning uses the standard DeprecationWarning..."
[resolve #8] Review comment by @copilot (PR #2214):
"These deprecation shim modules are added under the top-level supervision/..."

---
Co-authored-by: Claude Code <noreply@anthropic.com>

- Parametrized crash-regression test for 1/2/3/4 unique anchor
  positions (covers the splprep threshold boundary at len==4).
- Fallback-matches-raw assertion for <4 unique positions (verifies the
  new else branch uses unique_xy, same as smooth=False raw polyline).
- Single-frame smooth=True no-crash test (empty np.diff edge case).
- smooth=False stationary regression guard (ensures refactor left the
  non-smooth path untouched).

[resolve roboflow#5] /review finding by qa-specialist (report: boundary conditions untested)
[resolve roboflow#6] /review finding by qa-specialist (report: single-frame smooth=True untested)
[resolve roboflow#8] /review finding by qa-specialist (report: smooth=False regression guard)

---
Co-authored-by: Claude Code <noreply@anthropic.com>

Create a real 'subdir/' inside images/, use it as file_name, assert
ValueError matching 'resolves to directory' — exercises Branch 3 which
had zero coverage.

[resolve roboflow#8] /review finding by qa-specialist (report: .temp/output-review-fix-coco-loader-path-containment-2026-04-17.md):
"Branch 3 (is_dir guard) has zero test coverage"

---
Co-authored-by: Claude Code <noreply@anthropic.com>

- Add test_resize_dispatch_uses_cv2_for_dense: verifies _resize_crop takes
  the cv2 fallback for a checkerboard mask that exceeds _L3_DENSITY_THRESHOLD
- Add TestResizeParallelPath with two tests covering N>=8 masks (triggers
  ThreadPoolExecutor path): correctness and shape checks

[resolve #7] /review finding by qa-specialist (report: .temp/output-review-rle-resize-2026-04-22.md):
"No test for density-based dispatch threshold (_L3_DENSITY_THRESHOLD)"

[resolve #8] /review finding by qa-specialist (report: .temp/output-review-rle-resize-2026-04-22.md):
"No test for resize() with N >= 8 (thread pool code path never exercised)"

---
Co-authored-by: Claude Code <noreply@anthropic.com>

…alist (report: .reports/review/2026-05-19T10-05-17Z/review-report.md): use module-qualified patch targets

Change `patch("shutil.which", ...)` → `patch("supervision.utils.video.shutil.which", ...)`
and `patch("subprocess.run", ...)` → `patch("supervision.utils.video.subprocess.run", ...)`
to follow project convention and be robust to future `from shutil import which` refactors.

---
Co-authored-by: Claude Code <noreply@anthropic.com>

Public module-level function lacked any usage example, violating project
AGENTS.md convention. Added basic single-split export example.

[resolve roboflow#8] /review finding by foundry:doc-scribe (report: .reports/review/2026-05-22T18-44-11Z/review-report.md):

---
Co-authored-by: Claude Code <noreply@anthropic.com>

Add null-safety assertion (class_name is not None), full key-set + dtype parity check
across all data keys, and concatenation assertion covering the downstream mixed-batch
scenario. Replaces single dtype.kind check with structural invariant that guards future
key or dtype drift.

[resolve roboflow#4] [resolve roboflow#5] [resolve roboflow#8] /review findings by foundry:qa-specialist and foundry:challenger (report: .reports/review/2026-05-27T06-06-27Z/review-report.md)

---
Co-authored-by: Claude Code <noreply@anthropic.com>

- [resolve roboflow#5] oriented_box_non_max_suppression + oriented_box_non_max_merge:
  add Examples: doctest blocks (project convention for all public APIs).
  Challenge: evidence=VALID suggestion=VALID resolution=as-suggested
- [resolve roboflow#7] oriented_box_iou_batch: update summary/Returns from "Pairwise IoU"
  to overlap_metric-aware wording, matching box_iou_batch convention.
  Challenge: evidence=VALID suggestion=VALID resolution=as-suggested
- [resolve roboflow#8] oriented_box_iou_batch + public entry-points: add ndim/shape
  validation before reshape to catch flat (N,8) and wrong-ndim inputs early.
  Challenge: evidence=VALID suggestion=VALID resolution=as-suggested
- [resolve roboflow#9] oriented_box_non_max_merge: add assert 0 <= iou_threshold <= 1,
  matching the guard present on oriented_box_non_max_suppression and siblings.
  Challenge: evidence=VALID suggestion=VALID resolution=as-suggested

[resolve group] PR roboflow#2303 — items 5, 7, 8, 9

---
Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>
Co-authored-by: OpenAI Codex <codex@openai.com>