Commit graph

67 commits

Author SHA1 Message Date
5d772ef672 feat(metrics): recompute reads manual entries, methodology states them
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
2026-09-19 10:34:25 +02:00
47d19f9f38 fix(metrics): kpi caption honesty, manual-entry safety, extra coverage — Task 2 review findings
- Fix KPI row "Hours invested" caption to honestly show both session-derived and manually logged hours
- Add JSDoc warnings about manualMinutes being a discriminator (do not pass 0 for git commits)
- Extract isManual predicate to single source of truth, avoid triple derivation
- Add tests for mixed session+manual in same month and manual-only shape

Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
2026-09-19 10:32:40 +02:00
284385ca2b feat(metrics): fold manually logged minutes into project totals
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
2026-09-19 10:25:04 +02:00
4d1e6cf412 fix(commits): set source on ingest/backfill writes, tidy Task 1 review findings
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
2026-09-19 10:22:26 +02:00
65b09170af feat(commits): add source and manual_minutes fields
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
2026-09-19 10:14:51 +02:00
2cb745779a feat(nav): sidebar Kanban flyout, grouped by client 2026-09-17 10:42:50 +02:00
5dae0946eb feat(projects): add kind field (project | prototype) 2026-09-17 10:42:46 +02:00
6c41f4c662 fix(ai): raise max_tokens — inline reasoning was eating the whole budget
MiniMax M2.x reasons inside <think> tags in the normal content field, and
that reasoning is billed against max_tokens. At 4096 the heavier batches
spent the entire budget thinking, were truncated before writing a single
answer, and stripThinking correctly removed the unclosed block — leaving an
empty response.

The signature was distinctive: batches returned 25/25 or 0/25, never
anything partial. 193 of 779 commits went unsummarised with no error
anywhere. At 16384 the same 193 completed 8/8 batches with zero failures.

The interview's question call had the same bug at 1024, where it would have
returned an empty question and silently ended the interview on the first
turn. Now 8192.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-17 09:19:34 +02:00
6e231374e2 feat(roadmap): loader scoped by the caller's own read rules 2026-09-17 07:17:18 +02:00
b164f3273c feat(roadmap): checklist-derived completion percentage 2026-09-17 07:16:57 +02:00
1c24687dae feat(roadmap): time axis, month ticks and greedy row packing 2026-09-17 07:16:47 +02:00
0454abb853 feat(roadmap): Now/Next/Later derived from board order and dates 2026-09-17 07:16:24 +02:00
05ce52a6bf feat(schema): card brief, transcript and size band 2026-09-16 18:04:22 +02:00
a74f3e79b5 feat(intake): stateless interview service 2026-09-16 17:55:36 +02:00
bac4694ab1 feat(intake): whitelist brief parsing that degrades to raw answers 2026-09-16 17:55:14 +02:00
1c6e739b16 feat(intake): question and brief prompts 2026-09-16 17:54:54 +02:00
04f38ee114 feat(intake): size band derived from decomposition and known areas 2026-09-16 17:54:39 +02:00
95c4c6c8fa feat(intake): question cap and transcript sanitisation 2026-09-16 17:54:27 +02:00
8711c36833 feat(card): progress narrative and humanised commit history 2026-09-16 16:25:19 +02:00
30446153c8 feat(ai): card narratives regenerated when the commit set changes 2026-09-16 16:24:38 +02:00
7ca2798561 feat(schema): card narrative fields 2026-09-16 16:24:18 +02:00
5cecc87010 feat(ai): narrative prompt and set-based staleness 2026-09-16 16:23:32 +02:00
6890a6d29e feat(ingest): summarise new commits after metrics 2026-09-16 16:16:27 +02:00
bf7d024cef feat(ai): commit summarisation service that never throws 2026-09-16 16:16:08 +02:00
8addf87eef feat(ai): index-matched batch prompting and parsing 2026-09-16 16:15:42 +02:00
bce71a666a feat(schema): commit summary fields 2026-09-16 16:15:18 +02:00
e837500c6f feat(ai): single point of provider selection 2026-09-16 15:53:59 +02:00
b233259f18 feat(ai): OpenAI-compatible adapter for MiniMax 2026-09-16 15:53:34 +02:00
b172079912 feat(ai): Anthropic adapter 2026-09-16 15:52:56 +02:00
f3e4b30d46 feat(ai): settings storage with a browser-safe projection 2026-09-16 15:52:12 +02:00
8f1ac43e98 feat(ai): strip MiniMax inline reasoning blocks 2026-09-16 15:51:37 +02:00
b5593457af feat(ai): non-editable preamble with an append-only owner prompt 2026-09-16 15:51:15 +02:00
cd0984595c test(metrics): cached figures agree with stored commits + vitest stubs
The integration project now aliases server-only to a tiny empty stub
so tests can import modules that carry 'import "server-only"' without
the throwing real implementation firing during prepare. Unit tests do
NOT get the alias — a pure module with a server import is exactly the
kind of leak the unit project's missing condition is meant to surface.
2026-09-16 15:19:32 +02:00
cc20cf4bfd feat(metrics): cached project metrics recomputed on ingest 2026-09-16 13:51:03 +02:00
dbf524b04c feat(metrics): activity calendar grid 2026-09-16 13:46:26 +02:00
b379669db3 feat(metrics): project metric aggregation from commits 2026-09-16 13:45:56 +02:00
b4d4a3bcb6 feat(metrics): session clustering with defined boundaries
The plan's summariseSessions test expected 2 sessions for four commits
spanning two 3-hour gaps, but with the default 120-minute gap three hours
is too far and produces 3 sessions. The implementation matches the boundary
rule defined in the plan (strictly less than the gap continues a session);
the test was off. Adjusted to a 90-minute follow-up which keeps the same
session and verifies total + longest correctly.
2026-09-16 13:45:01 +02:00
265d623220 feat(metrics): SA public holidays and workday counting 2026-09-16 13:44:01 +02:00
d7e0b97ff4 feat(linking): re-link on ingest and on card rule changes
- ingestProject now calls relinkProject after every store
- updateCard relinks when scopeScope rule is in the patch
- acceptCard, archiveCard, unlinkCommitArchiveCard relink after their change
- three owner Server Actions for the tray: relinkProjectAction, linkCommitAction, unlinkCommitAction
2026-09-16 13:24:17 +02:00
25b504c901 feat(linking): idempotent relink engine preserving manual links 2026-09-16 13:23:33 +02:00
9c27d34c3e feat(linking): commit-to-card resolution with date bounding 2026-09-16 13:21:26 +02:00
7970edf7c2 feat(linking): explicit [#42] card references 2026-09-16 13:20:56 +02:00
d06d1e99ec feat(cards): per-project card numbers for commit references
- number field on cards, unique per project
- backfilled existing cards in created order
- both create paths (owner action, client API) assign on creation
- seed script is idempotent across re-runs
2026-09-16 13:20:24 +02:00
198a7a8ceb feat(git): shared idempotent ingest core 2026-09-16 11:47:55 +02:00
809983ee5c feat(git): Forgejo commit API client
Verified against the live instance: status 200, total 787 commits, first
sha 1bc0da40... 'feat(customers): lift active/inactive toggle to top of
customer form'.
2026-09-16 11:47:22 +02:00
a05c1b8ae0 feat(schema): commits collection and project ingest fields 2026-09-16 11:46:57 +02:00
b8cd766c89 feat(git): timing-safe webhook signature verification 2026-09-16 11:45:32 +02:00
ff423fe3ba feat(git): strict conventional-commit parsing 2026-09-16 11:45:15 +02:00
38d3758d2d feat(brand): align default client brand colour with new teal palette 2026-09-16 11:34:51 +02:00
52e6bae17b fix(admin): extract default brand colour to .ts so the colour check stays clean 2026-09-16 11:32:50 +02:00