Skip to content

fix(corpus): replace hardcoded .corpus/ paths with VMAF_CORPUS_DIR env var - #212

Merged
lusoris merged 2 commits into
masterfrom
fix/corpus-path-hardcoding-clean
May 30, 2026
Merged

lusoris merged 2 commits into
masterfrom
fix/corpus-path-hardcoding-clean

Conversation

@lusoris

@lusoris lusoris commented May 29, 2026 •

Copy link
Copy Markdown
Contributor

Summary

  • ai/scripts/calibrate_nr_threshold.py: _DEFAULT_CORPUS now reads VMAF_CORPUS_DIR env var (fallback .corpus/netflix).
  • tools/vmaf-tune/tests/test_bbb_e2e_v5_bug_cluster.py: BBB corpus reads VMAF_CORPUS_DIR (fallback /workspace/.corpus/bbb_e2e).
  • .gitignore: .corpus/ entry added (was only in .git/info/exclude).

Supersedes #122 which was contaminated.

no docs needed: internal scripts + test fixtures + .gitignore, no user-discoverable surface.

Test plan

  • VMAF_CORPUS_DIR=... resolves correctly + falls back to default
  • git check-ignore resolves .corpus/ from .gitignore
  • CI green

Deep-dive deliverables (ADR-0108)

  • Research digest — no digest needed: trivial path-source audit
  • Decision matrix — no alternatives: env var + fallback is the standard pattern (already used by validate_ensemble_seeds.py)
  • AGENTS.md invariant note — no rebase-sensitive invariants
  • Reproducer — VMAF_CORPUS_DIR=/my/corpus python ai/scripts/calibrate_nr_threshold.py --dry-run
  • Changelog fragment — no fragment needed: dev-tooling tweak
  • Rebase-notes entry — no rebase impact

Lint clean (CLAUDE §12 r12)

  • 4 files touched (1 script + 1 test + 2 config); all paths corrected

State drift (CLAUDE §12 r13)

  • No bug-tracking change — pure path-config hygiene

FFmpeg-patch sync (CLAUDE §12 r14)

  • No public-API change — internal scripts only

🤖 Generated with Claude Code

…v var

- ai/scripts/calibrate_nr_threshold.py: _DEFAULT_CORPUS now reads
  VMAF_CORPUS_DIR env var (falling back to .corpus/netflix), matching
  the pattern already used by validate_ensemble_seeds.py; adds missing
  import os.
- tools/vmaf-tune/tests/test_bbb_e2e_v5_bug_cluster.py: BBB corpus path
  reads VMAF_CORPUS_DIR (falling back to /workspace/.corpus/bbb_e2e)
  instead of a hardcoded /workspace prefix.
- .gitignore: add .corpus/ so the exclusion is visible to all contributors
  (was only in .git/info/exclude).

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
@lusoris
lusoris marked this pull request as ready for review May 30, 2026 02:52
@lusoris
lusoris enabled auto-merge (squash) May 30, 2026 02:52
@lusoris
lusoris merged commit b200a21 into master May 30, 2026
92 of 102 checks passed
@lusoris
lusoris deleted the fix/corpus-path-hardcoding-clean branch May 30, 2026 03:07
lusoris added a commit that referenced this pull request May 30, 2026
…DR-0839) (#258)

Mechanical cleanup in 3 post-C++23-wave cpp files:
- fex_ctx_vector.cpp: rename local 'capacity' → 'new_capacity' (shadow fix)
- feature_collector.cpp: every (T*) C-style cast → static_cast<T*>
- sycl/common.cpp: same C-style → static_cast<> pattern

Recreates closed PR #196's 6-file targeted change from current master
(the original had cascading rebase conflicts from #197+#212+#256 etc).
docs/adr/README.md index row already present on master; skipping that file.

Co-authored-by: lusoris <lusoris@pm.me>
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
lusoris added a commit that referenced this pull request May 30, 2026
…ADR-0840) (#259)

Recreates closed PR #199's 6-file targeted change from current master
(the original had cascading rebase conflicts from #197+#212+#256+#258).

- gpu_dispatch_env.c: TOCTOU window between getenv() and snapshot
- vmaf.c: cu_state leak on import failure path

Co-authored-by: lusoris <lusoris@pm.me>
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
@lusoris lusoris added this to the 1.0.0 — First release milestone Sep 4, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant