mirror of
https://github.com/colbymchenry/codegraph.git
synced 2026-10-17 05:07:37 +08:00
main
12 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 Colby Mchenry
|
003305f1e7 |
fix(explore): claim complete source only for sections that are complete (#2077)
* fix(explore): claim complete source only for sections that are complete On the tiers with a completeness signal (>= 500 indexed files) every codegraph_explore response ended with "Complete source for N files is included above — do NOT re-read them", whatever the render had cut. A windowed spine method, a shrunk or dropped cluster and the per-symbol view all went out under that line; the oversize-spine window never even set the trim flag the small tiers key off, which is how a 62-line slice of vscode's 968-line rpcProtocol.ts was called complete. The line also said "Reserve Read for a single specific line range", which explore output must never say. Completeness is now measured at emission from the ranges actually sent (plus back-referenced spans) against the symbols each section set out to deliver. A trimmed response keeps the verbatim / already-Read guarantee, names the trimmed files and the most relevant elided symbols, and steers to another codegraph_explore. The note is offered from most to least specific and fitted to the leftover room (CG-26); a trimmed note yields the pointer list's first entry. The lost-pointer, cut and truncation notes drop "complete" the same way, at no greater length, so the epilogue floor is unchanged. The small tiers' trimmed note now also fires for the cuts the flag missed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(explore): offer an elided method as Owner.member in the completeness note The note offered `as_sql` for django's elided SQLCompiler.as_sql — one of 110 definitions — so agents appended a path and line to disambiguate it, and that query shape returns none of the body. `SQLCompiler.as_sql` alone returns most of it. Methods are now offered as `Owner.member` from the indexed qualified name; everything else keeps its bare name. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor(explore): fit the completeness note and pointer list in one tested function A trimmed note's optional detail (the files, the elided names) never costs a pointer entry the least specific note would have left. The fit moves out of handleExplore into fitExploreEpilogue so both rules are pinned by tests; the second-try branch it had could never fire and is gone. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test(explore): completeness note wording after #2068 flags the spine window Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(explore): record the completeness-note replay and agent A/B Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(explore): address review of the completeness note - Judge completeness AFTER the dedup fold: a remainder too small to fence is sent in neither call, so a section with one is trimmed, not complete. The whole-file render now declares its symbols too (clamped to the lines it prints; the file node is not one), so a fold there registers. - Label trimmed files unique among every path the response can name, the pointer list included. - "1 file", and no "0 files" when every section was held from an earlier call. - The lost-pointer, cut and truncation notes are one exported table, EXPLORE_FALLBACK_NOTES, with the length rule tested; the lost note picks its wording from trimmedShown instead of re-joining the response. - The design doc's follow-up list marks the overclaim fixed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
Colby Mchenry
|
724b5dee0c |
fix(explore): return named bodies from merged spine clusters; let named files use spare budget (#2068)
* fix(explore): return named bodies from merged spine clusters; let named files use spare budget Re-applied onto main after #2062 and #2063 landed in the same render path (originally e31cc9fc on claude/inspiring-grothendieck-8a7b3e): - the merged-spine hold-back and its top-up are dropped: #2062's hold-back (every cluster leaves room for the protected members ranked below it) subsumes them, and the past-the-pin-cap test passes on it; - #2063's "don't window a cluster holding an exact target" becomes the member rule's "don't window a spine method that is or holds one"; - the shrink orders exact targets ahead of spine members, and the exact branch prices member windows (sectionRangesOf) rather than whole bodies; - the final ceiling fit shortens the file header back to its estimate before it cuts source. With an exact target's overshoot (#2063) spending a named file to its funded line, a lower file's header running past its estimate cost it 109 chars of source (922 of 1,031; 995 on main). The valve test now asserts funding in full and delivery to the line, since #2062's exact rendering stops at whole lines. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(explore): record the agent A/B for the named-concentration fix Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(explore): record landing named concentration on main after #2062/#2063 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
Colby Mchenry
|
adad8407ff |
fix(explore): pay named members before incidental ones across clusters (#2062)
Clusters of equal max importance rank by density, and density counts every member, so a named function with unrelated helpers merged around it outranked a cluster holding a named function alone. Taken whole, the higher clusters spent the file's reservation on the helpers and the isolated named function rendered nothing (lib/response.ts repro: sendBody 0 of 36 lines). Reordering alone only moves the victim. Two rules, both scoped to incidental members (not named, not an entry point, not on the call path): - a cluster's incidental members may only use what is left once every lower-ranked cluster's protected members are paid; its own protected members are never held back, so named-vs-named stays rank's call; - in a cluster holding a protected member, incidental members are never kept past the render's ceiling in rendered size, so the source-order window back to the ceiling no longer cuts a named body's tail (gin's handleHTTPRequest) on selection or on any re-render. 47 deterministic queries over the 7 README repos plus express vs main: named bodies complete 198 -> 234 (17 responses up, 0 down), named lines +10%, responses over the 25K cap 5 -> 1. One new final-cut truncation (django-bag) comes from owedPayableBelow's MIN_CHARS threshold, recorded in the design doc. On main + the named-concentration branch: 235 -> 279. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
Colby McHenry
|
c65d56ceba |
docs: CG-22 — the epic's gate, re-run at CG-15's exact setup (#1500)
CG-21 fixed the unspent-reservation defect and re-ran the A/B itself. CG-22 is
the gate proper: CG-15's setup, unchanged, measured independently of the task
that wrote the fix. RUNS=3, both arms codegraph-on, sonnet/high,
CODEGRAPH_NO_PROMPT_HOOK=1 on both, baseline pinned to
|
||
|
Colby McHenry
|
abee46c5e4 |
docs: CG-21 A/B — the gate passes, all four bars (#1500)
Re-runs CG-15's agent A/B on the fixed build: same harness, same three prompts, same baseline ref, n=6 per arm on express and excalidraw. Read = 0 in all 15 new-arm runs. The express regression that routed the defect to CG-21 does not reproduce in 6 attempts, and the baseline now reads in 4 of 6 while the new arm reads in none (median 24.5s -> 21.5s), so the control beats the arm it previously lost to. client-go holds 92.7-96.2% answer share against a baseline run at 53.8%. Excalidraw's new arm is ~8s slower at the median and that is recorded as NOT attributable to the build rather than waved through: explore's own latency is 374ms vs 372ms on the same query and index, the deterministic responses differ by +2% with one byte-identical, and the unchanged main build's own median moved 34s -> 26.5s between the two sessions — the same magnitude as the gap. Bars were not re-baselined; they are CG-15's four, applied to a larger sample. The CG-15 section is kept intact and marked superseded, because its root-cause analysis is the record of why the fix looks like it does. |
||
|
Colby McHenry
|
51cd053d85 |
docs: CG-21 — spending the reservation (design record + CHANGELOG precision)
Records both levers, the funding-pool design (and the per-file version that dropped payslip_builder.go), the resolved memory-budget.ts exception, and the two hermetic fixtures with their mutation matrix. The CHANGELOG clause 'no longer trimmed while a smaller, weakly-related one is included whole' was imprecise after CG-21: the smaller file often IS still included whole now, when its share nearly covers it. Reworded to say what the fix actually guarantees. |
||
|
Colby McHenry
|
c7103c7f2f |
docs: CG-15 agent A/B of the #1500 allocation change — gate fails on the control
Three repos, both arms codegraph-on, sonnet/high, 3 runs per arm. PASS on the two medium repos: client-go (the reporter's Go shape, 2,001 of 2,454 files generated) and excalidraw hold Read 0 in every run of both arms, excalidraw goes 34s -> 24s at the median with one fewer explore call, and the generated clientsets/informers that took 10.5%% of a baseline envelope appear in no new run. FAIL on express, the small control, in 1 run of 3: 4 Reads and 52s against a baseline that read once. Not agent variance — replaying that run's query deterministically, lib/utils.js goes from 6,380 bytes whole to a 583-byte cluster stub and the envelope shrinks 13.8K -> 9.2K against an unchanged 13,000 budget. The diagnostic shows the allocator was right and the render loop was not: utils.js is the top-ranked file, was reserved 3,870 chars, and spent 583. The whole-file bound (allowance + grace = 4,450) lands just under the file's 5,293 bytes, so the whole-file render is declined and the unspent reservation is dropped rather than redistributed. Bar 1 is the hard gate, so per CG-15's acceptance rule the design goes back to CG-12 — the budget is not to be widened to compensate. Two candidate fixes are written up in the design doc. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> |
||
|
Colby McHenry
|
1d9206d2d0 |
test(explore): lock down proportional byte allocation (CG-14, #1500)
Coverage for the CG-12 allocator, built around "would this go red if the lever were removed" rather than line coverage — every way this regresses is silent, ending in an agent falling back to Read. Unit (`explore-proportional-allocation.test.ts`, 18 -> 38): calibration pins, envelope safety across every tier and 30 candidate shapes, the cliff boundary, spine weighting/trim survival, the diffuse control, and the degenerate inputs — identical scores, a lone file, a runaway top scorer, zero results, maxFiles 0, a non-finite score. End-to-end (`explore-allocation-e2e.test.ts`, new): CG-6's second regression fixture as a deterministic synthetic mirror — a large relevant file, a small helper that used to win by shipping whole, and an incidental `explore`/`BUDGET` collision — asserting per-file budget share, not file presence. Plus degenerate result sets and a survey-style diffuse control through the real render loop. The live self-query arm stays in probe-allocation.mjs, where drift is a number to re-baseline rather than a red suite. Reverting the render loop to the pre-CG-12 rules reproduces #1500 on the mirror exactly and takes 5 e2e + 2 payroll gates red: file score pre-CG-12 CG-12 src/mcp/allocator.ts 77.5 4,843 (39.7%) 9,335 (80.1%) src/util/budget-math.ts 36.0 6,079 (49.8%) 1,037 ( 8.9%) Two defects the invariants surfaced, both fixed in tools.ts: - rounded shares could sum past `pool`, so "reservations fit the envelope" was approximate rather than exact; both terms now floor - a non-finite score made every share Infinity/Infinity, handing the render loop a NaN allowance; `weightOf` now fails safe to 0 Also adds a hard-ceiling gate to the payroll fixture — at 19.3K against a 19.5K ceiling it is the only fixture that stresses the ~25K inline cap — and exports EXPLORE_ALLOCATION so invariant tests read the constants while one test pins the literals. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> |
||
|
Colby McHenry
|
5f7f5f59df |
feat(mcp): score-proportional byte allocation for explore, with a relative cliff (CG-12, #1500)
The explore envelope used to follow FILE SIZE, not relevance. Every admitted
file was capped at the same flat `maxCharsPerFile`, while the whole-file rule
handed anything under `maxCharsPerFile * 3` its entire contents — a 3x swing
decided by how big a file happened to be:
- self-query: `memory-budget.ts` (score 18) shipped whole and took 51.2% of
the response; `src/mcp/tools.ts` (score 41, 4x the graph mass, 3x the term
hits — it holds the allocator itself) was clipped at 3,800 and got 32.9%.
- #1500 Go fixture: two generated CRUD files shipped whole at ~4.5K each AND
consumed two of the tier's four file slots, so `BuildPayslip` — the
hand-written "calculate" half of the question — ranked #6 and never
rendered at all.
`allocateExploreBudget` now reserves each ranked file a share of the envelope
before anything renders, so the render loop spends a reservation instead of
racing for whatever the files above it left:
- weight = score x worth x (spine ? 2 : 1), where `worth` is `rankPenalty`
applied a SECOND time — ranking answers "is this file about the query",
allocation answers "will these bytes teach the agent anything", and
generated CRUD can legitimately rank while its bytes stay boilerplate;
- a relative cliff at 15% of the top weight (capped at SCORE_FLOOR_MAX, so a
god-file can't silence peers the score floor just admitted) gives a file
ZERO source — path, symbols and line numbers only — and crucially frees its
`maxFiles` slot for a file that earns its bytes;
- every admitted file gets MIN_CHARS, then the remainder splits by weight:
the floor keeps a diffuse survey question returning a spread, the remainder
concentrates a precise one;
- the flat per-file cap is retired as the primary guard, leaving a 70%-of-
envelope safety valve.
Two changes were needed to make the reservation bite: an oversize cluster now
shrinks by whole MEMBER symbol ranges (a single-cluster god-file previously
took ~40% more than allotted, and the file below it was dropped for lack of
room), and the arrival-order budget stops are gone — they cut files by the
order they were reached rather than by merit.
Measured: payroll-go answer group 25.6% -> 78.7%, generated 57.4% -> 0%, and
`func (s *Service) BuildPayslip` now delivered; self-query `tools.ts` 18.5% ->
60.6%, past the epic's >50% bar. Controls hold: cobra/gin diffuse survey
queries keep their file spread (3->3, 3->4), express's middleware query is
byte-identical, and gin's flow query moves its top file from the thin `ginS`
singleton wrapper to `routergroup.go`.
One documented exception to "no previously-unclipped file becomes clipped":
`memory-budget.ts` was unclipped-whole at 5,672 and now clusters within its
3.1K reservation. That is the epic's own diagnosis of the bug — it scored 18
against 58 and was taking the larger slice purely for being small.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Colby McHenry
|
a3898cdc70 |
feat(mcp): relevance scoring overhaul for explore — kill incidental name-collision matches (CG-10, #1500)
Explore's per-file relevance awarded +50/+10/+3/+1 by match class and admitted
anything scoring >= 3. Neither half held up: the tier said HOW a symbol reached
us, never whether the match was evidence, and an absolute floor admits noise on
any repo where the top file scores 50+. Three scripts/agent-eval/*.mjs harnesses
took 63% of this repo's own "how does explore allocate its output budget" answer
on nothing but an unused `const explore` and a `const BUDGET`.
Four levers:
- KIND WEIGHT (RELEVANCE_KIND_WEIGHT): callables and types 1.0, members ~0.5,
variable/constant/parameter 0.15-0.35. A weak-kind symbol with no usage edge
anywhere in the graph (`contains` excluded — nesting is not usage) drops to
0.08. Only weak kinds in the top two tiers pay for the DB probe; the subgraph's
own edges answer most cases free. No measurable latency change (210 vs 211
ms/call, n=12 interleaved).
- PERIPHERAL CAP: nodes >=2 hops from any match accumulate into a bucket capped
at 5. Uncapped they added a flat +1 each, so a file grew more relevant by being
bigger — parse-session.mjs reached 22 off one constant plus twelve unrelated
symbols.
- RANK PENALTY: generated files x0.3, low-value x0.5, applied to the score AND
the graph mass. Score alone would not have fixed #1500 — the generated CRUD
carries MORE graph mass than the hand-written use-case, and graph mass outranks
score in the comparator. Self-normalizing, never a hard exclusion.
- RELATIVE FLOOR: clamp(topScore * 0.2, 1, 10). Capped at one full-strength
direct match so concentration elsewhere can never exclude one (without it a
named-seed-heavy file pushed the floor to 21 and dropped a file the agent had
named by class name). Backfills to 3 candidates when it would leave fewer, and
drops the evidence requirement rather than return nothing at all.
excludeLowValueFiles was dead config — declared per tier, read nowhere; the
test/spec exclusion has been unconditional for a while. Removed. The real gap was
the detector: `isLowValue` anchored on a leading `/`, so a repo-ROOT `test/` dir
(express, cobra, most of npm and Go) never matched — express's routing question
spent 59% of its envelope on three test files. Anchored at `^` too, and the
filter now runs before the floor and judges "are there other candidates?" on the
whole gather.
Measured before/after on the same indexes (baseline
|
||
|
Colby McHenry
|
bd86ad2061 |
test(explore): #1500 regression fixtures for budget allocation (CG-6)
Two permanent fixtures pinning the failure mode from issue #1500 — explore spending its byte envelope on files that merely name-collide with the query. BOTH FAIL TODAY, by design: they document the bug and become the pass gate for CG-10 (scoring) + CG-12 (proportional allocation). __tests__/fixtures/payroll-go/ — a synthetic Go service mirroring the reporter's shape: generated FKIT CRUD beside a hand-written payroll use-case, entered from an HTTP route. Half the generated tree carries ORDINARY names detectable only by their `// Code generated ... DO NOT EDIT.` header (the #1500 case, and end-to-end cover for CG-5); `payrollpb/*.pb.go` covers the path-detectable channel. BuildPayslip, Upsert and Store each exist twice, generated and hand-written. cycle.go sits above the whole-file window so it clips; the generated files sit below it so they ship whole. Asking "how does payroll cycle create and calculate payslips?" — naming none of the answering symbols — the generated CRUD delivers 57.4% of the envelope against the hand-written layer's 25.6%, all of the latter domain types. cycle.go is allocated the single largest slice (30.6%) and delivers ZERO: the hard ceiling drops its whole section. runPayrollCycleAll, the hand-written BuildPayslip and the real Upsert never reach the agent. The second fixture is this repo, "how does explore allocate its output budget across files", where scripts/agent-eval/*.mjs take 71.8% against tools.ts's 18.5% despite scoring 4.6x lower. It reads the live index, so its assertions are relative rather than fixed percentages. - scripts/agent-eval/probe-allocation.mjs — per-file budget-share probe, driving the CG-4 diagnostic through a JSONL sidecar so it measures the shipping allocator. Fixture entries are hermetic (copy + re-index per run, verified byte-identical across runs); exits 1 while any assertion fails. - scripts/agent-eval/allocation-fixtures.json — both fixtures declared, with the 2026-08-03 baselines. - __tests__/explore-allocation-1500.test.ts — fixture-shape assertions green today; the allocation assertions held as `it.fails` so the suite stays green while the bug is open and goes RED the moment it is fixed. Also documented and deliberately left unfixed: runPayrollCycleAll's `s.store.Upsert` edge resolves to the GENERATED Store.Upsert, not the hand-written one — same-name method resolution across two packages picks the wrong receiver. It is upstream of the allocation bug, so it belongs with CG-10's scoring work. Refs #1500 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> |
||
|
Colby McHenry
|
b37f191f5a |
feat(mcp): per-file allocation diagnostic for explore (CG-4)
How codegraph_explore divides its byte envelope among files was unobservable — you could read a response and guess, but not say "this file took 16% and that one took 20%." Nothing else in the budget- allocation epic is measurable without that. CODEGRAPH_EXPLORE_DEBUG now emits one report per explore call (stderr table, stderr JSON, or a JSONL sidecar path). Per file: relevance score, graph mass, term hits, ranking flags, render mode, bytes allocated vs delivered, both shares, and whether it was clipped — plus why a ranked candidate never rendered. Totals cover envelope vs maxOutputChars vs the hard ceiling, the source/meta split, the selection funnel, and the score floor and relevance-gate thresholds applied. Allocated and delivered are reported separately on purpose: they diverge exactly when the 25K ceiling truncates, and conflating them is how a dropped trailing file goes unnoticed. Off by default and byte-identical when off — it ships in the product binary, and a diagnostic that perturbs the response by one byte would invalidate every A/B taken with it on. ExploreDiagnostics.start() returns null unless the env var is set, so every call site is a `diag?.` no-op. Baseline recorded in docs/design/explore-budget-allocation.md: on this repo, src/mcp/tools.ts gets 15.8% of the envelope while three weakly- relevant agent-eval scripts take 61% between them — despite tools.ts carrying 5.4x the score and 2.6x the graph mass of any of them. Small files ship whole; the large answer file is clipped at maxCharsPerFile. Rank ordering is correct and buys nothing. The loop also allocated 23,193 chars against an 18,000 budget. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> |