docs: remove dangling L0/L1/L2 tier references - #191
Open
nwaughachukwuma wants to merge 3 commits into
Open
Conversation
The old l0/l1/l2 processing modes were replaced by --mode fast|accurate (with 'mm peek' covering the former metadata tier), but stale references survived the transition: - README, docs/user-guide, docs/pipelines, mm-skill: 'L2 cache key' -> 'extraction cache key' (the cache keyed on content hash + mode + merged model/extra_body) - AGENTS.md: cache.py is file-fingerprint memoization, not an 'L2 result cache' - benchmarks/bench_cli*.sh: 'L0' section labels -> 'metadata', '(L1 text)' -> '(fast text)' - benchmarks/bench_modes.py: 'via mm L1' docstrings -> 'via mm cat (fast mode)' - tests/test_cli.py: 'l2-*' fixture extraction IDs -> 'acc-*' The 'L2 (Euclidean)' mention in store/db.py is the distance norm, not the old tier, and is left untouched. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Cg3KtFq2ZWMck79bq6YvBV
|
Caution The consumer version of Gemini Code Assist on GitHub has been sunset. All code review activity has officially ceased. |
CI installs the dev extras with an unpinned 'ruff>=0.4', which now resolves to ruff 0.16.1 — its new/promoted rules flag 298 pre-existing errors and fail the test-python job. Pin to 0.15.4 (same fix as #181) so lint results are reproducible. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Cg3KtFq2ZWMck79bq6YvBV
The lock already resolved ruff to 0.15.4; only the recorded specifier needed updating to match pyproject. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Cg3KtFq2ZWMck79bq6YvBV
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
2 participants
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
The old
l0/l1/l2processing modes were replaced by--mode fast|accurate(withmm peekcovering the former metadata tier), but stale references survived the transition. This PR reconciles the last mile of that rename — no behavior changes, only docs, comments, labels, and test fixture IDs.Changes
README.md,docs/user-guide.md,docs/pipelines.md, and.claude/skills/mm-skill/SKILL.md(including the--no-cacheflag description). The cache is keyed on content hash + mode + mergedmodel/extra_bodyand serves both fast and accurate modes, so the mode-neutral name is used.AGENTS.md—cache.pywas annotated as "L2 result cache"; the module is actually file-fingerprint memoization (mtime/size-keyed LRU, in-memory + disk) for expensive helpers like Whisper and scene detection. The tree comment now says so. (CLAUDE.mdsymlinks toAGENTS.md.)bench_cli.sh,bench_cli_mini.sh,bench_cli_tiny.sh) — "L0:" section labels now read "metadata:" to match current tier naming; "(L1 text)" became "(fast text)".benchmarks/bench_modes.py— "via mm L1" docstrings now read "viamm cat(fast mode)".tests/python/test_cli.py—l2-*fixture extraction IDs renamed toacc-*(they all seedmode='accurate'rows).Deliberately kept
python/mm/store/db.pymentions "Distance metric is L2 (Euclidean)" — that's the mathematical L2 norm used by sqlite-vec, not the old processing tier, so it's untouched.Verification
l0/l1/l2token variants returns nothing except the Euclidean mention.bash -npasses on the three edited shell scripts.tests/python/test_cli.py: 80 passed, 4 skipped.🤖 Generated with Claude Code
https://claude.ai/code/session_01Cg3KtFq2ZWMck79bq6YvBV
Generated by Claude Code