Every commit scored for change-risk against this repo's own history, so 'elevated' means elevated here rather than on some global curve.
Needs review
98 commits sit in this repo's top risk tercile, which is 34% of the 288scored. The cut is drawn against this codebase's own history rather than a global curve, so a quiet repo still fills its top band, and here it starts at 8.3 out of 10. What pushes a commit up is size and spread together: a large change confined to one area scores below a smaller one scattered across a dozen files.
Commit categories over time, read off the subject line. Fixes carry the accent because that is the series this chart exists to show.
Began other-led, now leaning docs.
Ranked by change-risk, highest first. Priority is a tercile of this repo's own distribution, so a quiet repo still fills its top band.
| # | Commit | Author | When | Lines | Risk | Top driver |
|---|---|---|---|---|---|---|
| 1 | 94a11a26release: publish-readiness - LICENSE per crate and versioned internal deps (P50.1, P50.2) | mlengse | 2w ago | +21.7K -127 | 100%Elevated | more lines added than baseline |
| 2 | 79ef7899Rename kuzu-core to akar-core and update tools | mlengse | 4w ago | +105.2K -603 | 100%Elevated | more lines added than baseline |
| 3 | 120cb8cbinit | mlengse | 1mo ago | +9.7K -0 | 99%Elevated | more lines added than baseline |
| 4 | b93e9667fix: parser BOOLEAN type ordering + akar-python P53 translation layer (P53.1-P53.8) | mlengse | 1w ago | +1.7K -44 | 98%Elevated | more lines added than baseline |
| 5 | d79af77efeat(python): scaffold akar-python — PyO3 bindings (kuzu-compatible surface) | mlengse | 1w ago | +2.4K -0 | 98%Elevated | more lines added than baseline |
| 6 | fa344873feat(storage): implement data durability, file locking, and read-only mode (P45.4) | mlengse | 3w ago | +1.6K -37 | 98%Elevated | more lines added than baseline |
| 7 | 0d062928tools/rust_api Integration | mlengse | 1mo ago | +1.9K -4.6K | 98%Elevated | more lines added than baseline |
| 8 | 5547744bSprint 21: kairos-native migration (P55-P57) — 2 new crates, 27 tests | mlengse | 4d ago | +1.4K -4 | 96%Elevated | more lines added than baseline |
| 9 | f25c2a51fix: OPTIONAL MATCH edge probe executes add_bridge_batch UNWIND->CREATE chain (P53.25) | mlengse | 1w ago | +1.6K -741 | 96%Elevated | more lines added than baseline |
| 10 | 7c601a57fix: Sprint-19 batch E/F connectors + extension compile fixes (P52.29-P52.66) | mlengse | 1w ago | +1.5K -108 | 96%Elevated | more lines added than baseline |
| 11 | d4b30357migrate: post-Arrow API migration + aggregate col_indices guard (P50 prep) | mlengse | 2w ago | +1.6K -1.4K | 96%Elevated | more lines added than baseline |
| 12 | 8b1a19a2feat: implement WCOJ planner enumeration and embedded server mode (P46, P47) | mlengse | 2w ago | +2.4K -381 | 96%Elevated | more lines added than baseline |
| 13 | c60bf7cdtest(P41): implement crash recovery stress testing and MVCC enhancements | mlengse | 3w ago | +1.3K -161 | 96%Elevated | more lines added than baseline |
| 14 | a8ebb035Update akar-core implementation and add audit plan | mlengse | 4w ago | +2.0K -2.3K | 96%Elevated | more lines added than baseline |
| 15 | 61b5e0e6fix: Batch E storage/execution correctness + DML DELETE/SET row-id bug | mlengse | 1w ago | +1.1K -274 | 94%Elevated | more lines added than baseline |
| 16 | e36d46b9fix: P48 correctness fixes - multi-hop join, rel COPY, pushdown, migration idempotency, date/timestamp scan (P48.1-48.3, 48.5, 48.7) | mlengse | 2w ago | +820 -70 | 94%Elevated | more lines added than baseline |
| 17 | 656f8714feat: implement Sprint 13 bug fixes and performance optimizations, including radixsort, row-level OCC, plan caching, and vectorized Arrow operations. | mlengse | 3w ago | +938 -139 | 94%Elevated | more lines added than baseline |
| 18 | b148bdcatest: resolve audit issues 6.3 and 6.4 | mlengse | 3w ago | +914 -187 | 94%Elevated | more lines added than baseline |
| 19 | df8561e4fase 20 | mlengse | 1mo ago | +1.1K -250 | 94%Elevated | more lines added than baseline |
| 20 | f9014250fix: bug batch P51.28-P51.36 (drop-table sequences, two-phase commit, prepared params, CLI/wasm) | mlengse | 1w ago | +657 -88 | 93%Elevated | more lines added than baseline |
| 21 | 2498794cfix: storage correctness batch - PK index sync on soft delete, vector index save/load, ART string escape, MVCC versioned updates, zone map typing (P52.16-P52.22) | mlengse | 1w ago | +521 -186 | 93%Elevated | more lines added than baseline |
| 22 | d62c55b0fix: audit batch 4 (P52.1, P52.3, P52.8, P52.9, P52.10, P52.11) - OPTIONAL MATCH NULL padding, filter 0-size chunk, scalar agg on empty, sort elision, merge sort ties | mlengse | 1w ago | +638 -68 | 93%Elevated | more lines added than baseline |
| 23 | c0b5b4ecdocs: update SPEC, STATUS, and implementation plan with actual audit numbers | mlengse | 2w ago | +782 -59 | 93%Elevated | more lines added than baseline |
| 24 | d1f799c8Column data structure | mlengse | 1mo ago | +519 -0 | 93%Elevated | more lines added than baseline |
| 25 | fa6ac0adfeat: P54.5 — PyO3 expose kNN/LSTM/spread/louvain submodules | mlengse | 4d ago | +664 -6 | 90%Elevated | more lines added than baseline |
| 26 | b26ed5eefeat: P54.5 — PyO3 expose kNN/LSTM/spread/louvain submodules | mlengse | 4d ago | +664 -6 | 90%Elevated | more lines added than baseline |
| 27 | 9a477cdafeat: P54.1-P54.3 — graph algorithms upgrade (Sprint 20) | mlengse | 5d ago | +736 -52 | 90%Elevated | more lines added than baseline |
| 28 | 300a5cd5feat: parquet export improvements + pyo3 0.29.2 upgrade | mlengse | 5d ago | +464 -68 | 90%Elevated | more lines added than baseline |
| 29 | 0e0cf991fix: UNWIND binding resolves in MATCH/CREATE/SET; map columns survive joins (P53.26) | mlengse | 6d ago | +595 -120 | 90%Elevated | more lines added than baseline |
| 30 | 6c59b3a3fix: P53.14 G3 chain clauses — parse/bind/plan arbitrary Kairos chains | mlengse | 1w ago | +520 -186 | 90%Elevated | more lines added than baseline |
| 31 | e3c7b5d0fix: connection txn safety + export/import round-trip + remote drain (P52.12-P52.28) | mlengse | 1w ago | +617 -192 | 90%Elevated | more lines added than baseline |
| 32 | e4952863fix: MERGE edge with label-less nodes resolves MATCH-bound vars; OPTIONAL MATCH allows existing vars (P53.20/P53.21) | mlengse | 1w ago | +569 -56 | 87%Elevated | more lines added than baseline |
| 33 | 7c34381dP53.17 (G6) SET arithmetic reads pre-update values; multi-item SET is atomic | mlengse | 1w ago | +514 -94 | 87%Elevated | more lines added than baseline |
| 34 | ef4f1430fix: Sprint-19 batch G P52 audit reconciliation (P52.51-P52.61) | mlengse | 1w ago | +568 -292 | 87%Elevated | more lines added than baseline |
| 35 | a2210e28feat: MERGE/CREATE DML multi-pattern + relationship support (P51.7, P51.8, P51.37) | mlengse | 1w ago | +517 -846 | 87%Elevated | more lines added than baseline |
| 36 | e4457108fix: re-open WCOJ benchmark with correct edge setup; fix equality-filter drop and CREATE node-predicate (P48.4, BUG-A/BUG-B) | mlengse | 2w ago | +443 -33 | 87%Elevated | more lines added than baseline |
| 37 | 451b2a63perf: add comprehensive benchmark suite for scalability testing and recovery time analysis | mlengse | 3w ago | +478 -59 | 87%Elevated | more lines added than baseline |
| 38 | 18e5b59ffeat: implement centralized AkarError and apply it across crates | mlengse | 4w ago | +532 -229 | 87%Elevated | more lines added than baseline |
| 39 | 3ecba585ART Index Implementation for Kuzu Rust: Persistence via BufferManager, Catalog Integration, Storage Integration | mlengse | 1mo ago | +382 -0 | 87%Elevated | more lines added than baseline |
| 40 | d3b20a07Parser — CREATE VECTOR INDEX DDL | mlengse | 1mo ago | +380 -0 | 87%Elevated | more lines added than baseline |
| 41 | bd2addffOn-Disk HashIndex | mlengse | 1mo ago | +363 -0 | 87%Elevated | more lines added than baseline |
| 42 | 70950977Connect PhysicalScan to real table data | mlengse | 1mo ago | +384 -0 | 87%Elevated | more lines added than baseline |
| 43 | 34a29b63Gap Fase 1 | mlengse | 1mo ago | +406 -0 | 87%Elevated | more lines added than baseline |
| 44 | 3bc228e9feat: P54.4 — LSTM 1-layer local ML crate (akar-ml) | mlengse | 5d ago | +643 -8 | 83%Elevated | more lines added than baseline |
| 45 | 0b648d5cfeat: P54.4 — LSTM 1-layer local ML crate (akar-ml) | mlengse | 5d ago | +643 -7 | 83%Elevated | more lines added than baseline |
| 46 | 70a7eb40fix: UNWIND source survives join reorder; Extend keeps map columns (P53.37a) | mlengse | 5d ago | +372 -26 | 83%Elevated | more lines added than baseline |
| 47 | fd8f88ccfix: SET/MERGE/DELETE drop-in write correctness; rustfmt normalize (P53.29-P53.32) | mlengse | 6d ago | +536 -154 | 83%Elevated | more lines added than baseline |
| 48 | 5651ac8ffix(P53.15, G4): nested aggregate COALESCE(MAX/SUM(x),0) binds and executes | mlengse | 1w ago | +373 -33 | 83%Elevated | more lines added than baseline |
| 49 | 46de39d0fix: complex-type query output (List/Struct/Map) round-trip (P53.12) | mlengse | 1w ago | +550 -37 | 83%Elevated | more lines added than baseline |
| 50 | 8564faa6fix: optimizer audit - neutralize unsafe flat passes, fix ART range scan | mlengse | 1w ago | +546 -346 | 83%Elevated | more lines added than baseline |
Two views of the same model: where the cuts fall, and what commit shape lands you above them.
Every scored commit, binned on the raw 0 to 10 score rather than the percentile. Percentile ranks are uniform by construction, so that axis has no shape to draw. The dashed lines are the tercile cuts behind each row's priority pill.
The 200 most recent commits, on their own recency sample rather than the feed above: that defaults to risk-sorted, so reusing it would plot only the top tercile and call it the spread. Big and scattered is what the model penalises. Click a dot to open it.
mlengse/akar has 310 commits in its history from 2 contributors, the first of them Jun 28, 2026. In the last 90 days 355 files were touched, 1,186 times in total, most often query.rs. Every commit is scored for change risk from its size, spread and the history of the files it touches.
