feat(metrics): add near-duplicate block detection - #18

Closed

aspala wants to merge 76 commits into

mainfrom

feat/near-duplicate-blocks

Commits on Mar 19, 2026

chore: add tooling config, ci workflow, and dependency updates

aspala
and
claude
committed
861c365
refactor(engine): extract engine layer and reorganize metrics namespaces

aspala
and
claude
committed
dad7ca1
feat(ast): add AST parsing system with lexing, signals, and node classification

aspala
and
claude
committed
653e830
feat(languages): add multi-language definitions and test fixtures

aspala
and
claude
committed
62ed12d
feat(analysis): add OTP-based analysis run context and servers

aspala
and
claude
committed
fb4135e
feat(combined-metrics): add combined metrics scoring framework

aspala
and
claude
committed
110ffd2
feat(near-duplicate-blocks): add near-duplicate block detection and impact analysis

aspala
and
claude
committed
5d67df7
feat(diagnostics): add config and diagnostics modules

aspala
and
claude
committed
ef91d6b
feat(health-report): update health report with new categories and grading

aspala
and
claude
committed
3a465c9
feat(cli): add diagnose command and update existing commands

aspala
and
claude
committed
13f8b3c
feat(scalar-tuner): add scalar tuner UI tool

aspala
and
claude
committed
397fd8f
docs: update README and action configuration

aspala
and
claude
committed
5250d22
test(analysis): add tests for OTP analysis servers

aspala
and
claude
committed
d49f463
test(ast): add tests for AST parsing, signals, nodes, and enrichment

aspala
and
claude
committed
6678ca7
chore(gitignore): ignore scripts/*.exs and docs plan directories

aspala
and
claude
committed
45ae2b2

Commits on Mar 20, 2026

feat(combined-metrics): detect and persist language coverage per behavior

aspala
and
claude
committed
b58c8bf
feat(combined-metrics): filter diagnose_aggregate by language

aspala
and
claude
committed
a8267d3
fix(combined-metrics): improve behavior_language_applies? guards and docs

aspala
and
claude
committed
4e12620
feat(combined-metrics): filter score_aggregate by project languages

aspala
and
claude
committed
bca01a2
feat(file-scorer): detect file language and filter behaviors accordingly

aspala
and
claude
committed
a9f017b
feat: thread project languages through all codebase-level scoring call sites

aspala
and
claude
committed
5556a30
fix(block-impact-analyzer): avoid redundant language detection, clarify project_languages param

aspala
and
claude
committed
d8d448a
feat(cli): add --apply-languages flag to sample_report mix task

aspala
and
claude
committed
1e57187
chore(combined-metrics): populate _languages coverage from samples
aspala
committed
76fc86c
feat(ci): add workflow to auto-sync behavior language coverage and scalars
aspala
committed
db764fc
chore(combined-metrics): sync language coverage and scalar vectors [skip ci]
github-actions[bot]
committed
1d10f53

Commits on Mar 21, 2026

Commits on Mar 22, 2026

fix(near-duplicate-blocks): return zeroed map when blocks is nil

aspala
and
claude
committed
0787134
chore(combined-metrics): annotate language coverage for all behaviors

aspala
and
claude
committed
20d1773
perf(block-impact): precompute log metrics and cache behaviors for cosine scoring

aspala
and
claude
committed
e68cbd3
refactor(health-report): pre-compute cosines outside Grader

aspala
and
claude
committed
70302cf
refactor(engine): integrate block impact computation into the analyzer pipeline

aspala
and
claude
committed
71c8f0a
test: update tests for refactored grader and near-duplicate-blocks interfaces

aspala
and
claude
committed
3abbc0a
refactor: fix dialyzer, credo nesting, and compiler warnings

aspala
and
claude
committed
3b67dbb
fix(test): pass path to build_file_context in NearDuplicateBlocksFileTest

aspala
and
claude
committed
03cce19
chore(combined-metrics): sync language coverage and scalar vectors [skip ci]
github-actions[bot]
committed
ddff5c1

Commits on Mar 26, 2026

docs: add multi-part PR comment design spec

aspala
and
claude
committed
4e2eeb0
docs: address spec review issues in multi-part PR comments design

aspala
and
claude
committed
9a220dd
docs: fix sentinel and part-count clarifications in multi-part spec

aspala
and
claude
committed
d2438f5
docs: final spec clarifications — part-count loop and placeholder ownership

aspala
and
claude
committed
757114e
feat(multi-part-comments): split health report into multiple PR comments
aspala
committed
ffa7400
fix(action): pass GITHUB_TOKEN to run.sh for PR comment posting
aspala
committed
d198001
chore: remove docs/ and ignore test/ in codeqa analysis
aspala
committed
570812a
chore: ignore README.md and action.yml in codeqa analysis
aspala
committed
4b43621
feat(blocks): show top 10 blocks by impact with source code
aspala
committed
fe643c3
chore: trigger CI
aspala
committed
228aef2
feat(blocks): filter blocks by line count (3-20 lines default)
aspala
committed
9b2d184
feat(blocks): filter blocks by diff line overlap when --base-ref provided
aspala
committed
bfa9e3e
feat(report): show worst offending block per cosine category
aspala
committed
c9704ed
chore(combined-metrics): sync language coverage and scalar vectors [skip ci]
github-actions[bot]
committed
09c0901
fix(report): display blocks from 1-15 lines instead of 4-10
aspala
committed
8925e76

Commits on Apr 14, 2026