refactor: adopt dead helpers identified by Titan grind

[carlos-alm]

Summary

Adopt dead helpers identified by the Titan GRIND phase — wiring previously-extracted-but-unused helpers into real call sites, and promoting/deduping a few forge-phase helpers that turned out to have further adoption opportunities:

• Adopt fget/iget from typed-array-helpers.ts in cpm.ts/modularity.ts (leiden).
• Wire db.busyTimeoutMs and community.capacityGrowthFactor config constants through their real call sites (db/connection, builder pipeline, watcher, communities, leiden).
• Adopt resolveMethodDefinitionName across 3 duplicate call sites in extractors/javascript.ts.
• Dedupe forge phase-11 call-resolution helpers (unwrapTypeEntry, resolveSameClassQualifiedMethod) across build-edges.ts/incremental.ts.
• Extract markExportedSymbols, dedupe a batch UPDATE in builder helpers.
• Promote timeMedian/median/round1 into a shared scripts/lib/bench-timing.ts, deduping the same logic across 3 benchmark scripts.

Titan Audit Context

• Phase: adoption
• Domain: cross-cutting (leiden, config/db/builder, extractors, resolver, scripts)
• Commits: 6
• Depends on: PR Bump tree-sitter-python from 0.23.6 to 0.25.0 #7 (warning improvements) — this PR's commits reference symbols extracted across the full forge stack (PRs Bump actions/setup-node from 4 to 6 #1-Bump tree-sitter-python from 0.23.6 to 0.25.0 #7), so it must merge after all of them

Changes

• src/graph/algorithms/leiden/{cpm,modularity}.ts
• src/db/connection.ts, src/domain/graph/builder/{pipeline,watcher}..., src/domain/graph/builder/stages/{native-db-lifecycle,native-orchestrator}.ts, src/features/communities.ts, src/graph/algorithms/leiden/{optimiser,partition}.ts, src/graph/algorithms/louvain.ts, src/infrastructure/config.ts, src/types.ts
• src/extractors/javascript.ts
• src/domain/graph/builder/{call-resolver,incremental}.ts, src/domain/graph/builder/stages/build-edges.ts, src/domain/graph/resolver/strategy.ts
• src/domain/graph/builder/{helpers,stages/insert-nodes,stages/native-orchestrator}.ts
• scripts/{benchmark,incremental-benchmark,query-benchmark,token-benchmark}.ts, new scripts/lib/bench-timing.ts

Metrics Impact

The Titan GRIND phase triaged 24 candidate dead-symbol targets: 6 were genuinely adoptable (this PR), 3 were promoted/deduped further (also this PR), and the remaining 15 were confirmed false positives caused by known codegraph role-classifier bugs (documented in generated/titan/ report and filed as GitHub issues).

Test plan

• CI passes (lint + build + tests)
• codegraph check --cycles --boundaries passes
• No new functions above complexity thresholds

[greptile-apps]

Greptile Summary

This PR adopts dead helpers identified during the Titan GRIND phase: wiring previously extracted-but-unused utilities into their real call sites and deduplicating recurring patterns across the codebase.

• Config constants wired end-to-end: busyTimeoutMs now flows from DEFAULTS through openDb/openReadonlyOrFail default parameters and is forwarded explicitly by every build/watch call site that already holds a resolved config; capacityGrowthFactor is threaded from communitiesData all the way to ensureCommCapacity in the Leiden partition.
• Structural deduplication: resolveSameClassQualifiedMethod, markExportedSymbols, unwrapTypeEntry, resolveMethodDefinitionName, fget/iget, and the bench-timing trio (median/round1/timeMedian) each had two-to-five byte-for-byte duplicates removed; the shared versions are functionally identical, chunk sizes and control-flow (return vs continue) preserved.
• Previous-review follow-up addressed: DEFAULT_CAPACITY_GROWTH_FACTOR is now exported from partition.ts and imported by optimiser.ts, closing the three-copy drift risk called out in the prior thread.

Confidence Score: 5/5

A clean adoption PR; every shared helper is functionally identical to the duplicates it replaces, and config values that were previously hardcoded are now forwarded from the already-resolved config objects that all call sites hold.

All deduplication is purely mechanical — chunk sizes, control flow, and default values are preserved exactly. The busyTimeoutMs wiring is consistent across every openDb call site that holds a resolved config, and the remaining ad-hoc callers fall back to the same 5000 ms default as before. No behavioral changes were introduced.

src/presentation/audit.ts has a double-cast that may warrant a follow-up type fix, but it does not affect runtime behavior.

Important Files Changed

Filename	Overview
scripts/lib/bench-timing.ts	New shared module extracting median/round1/timeMedian from duplicate copies across four benchmark scripts; well-typed, empty-array guard added to median, runs parameter is now mandatory.
src/db/connection.ts	Threads busyTimeoutMs through openDb/openReadonlyOrFail defaults and resolveDbSettings; loadConfig is now always invoked inside resolveDbSettings regardless of whether engineOpt is provided, which is intentional to also fetch busyTimeoutMs.
src/domain/graph/builder/helpers.ts	Extracts markExportedSymbols using a module-level WeakMap cache (same BATCH_CHUNK=500 as the old per-call EXPORT_CHUNK=500 constants), correctly deduping the chunked UPDATE loop.
src/domain/graph/builder/call-resolver.ts	Extracts resolveSameClassQualifiedMethod as an exported function; logic is identical to the private duplicates removed from build-edges.ts and incremental.ts.
src/graph/algorithms/leiden/partition.ts	Exports DEFAULT_CAPACITY_GROWTH_FACTOR (resolving the previous-review's three-copy concern), adds capacityGrowthFactor to PartitionState and MakePartitionOptions, wires through ensureCommCapacity.
src/graph/algorithms/leiden/optimiser.ts	Imports DEFAULT_CAPACITY_GROWTH_FACTOR from partition.ts (previous-review addressed), adds capacityGrowthFactor to LeidenOptions/NormalizedOptions, passes it through to makePartition at both call sites.
src/extractors/javascript.ts	Replaces three identical computed_property_name inline blocks with calls to the file-local resolveMethodDefinitionName; control flow (return vs continue) is preserved correctly at each call site.
src/presentation/audit.ts	Adds an as-unknown-as cast to satisfy outputResult's parameter type; the double escape hints at a structural type mismatch between AuditResult and Record<string, unknown> that may deserve a proper fix.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A["communitiesData()\nsrc/features/communities.ts"] -->|capacityGrowthFactor| B["louvainCommunities()\nsrc/graph/algorithms/louvain.ts"]
    B -->|JS path| C["louvainJS() → detectClusters()"]
    C --> D["runLouvainUndirectedModularity()\noptimiser.ts"]
    D -->|capacityGrowthFactor| E["makePartition()\npartition.ts"]
    E --> F["ensureCommCapacity()\nuses s.capacityGrowthFactor"]
    B -->|native path| G["native Rust louvain\n⚠ ignores capacityGrowthFactor"]

    H["DEFAULTS.db.busyTimeoutMs\nconfig.ts"] -->|default param| I["openDb() / openReadonlyOrFail()\nconnection.ts"]
    J["resolveDbSettings()\nconnection.ts"] -->|busyTimeoutMs| I
    I --> K["pipeline.ts / watcher.ts\nnative-orchestrator.ts\nnative-db-lifecycle.ts"]

    P["resolveSameClassQualifiedMethod()\ncall-resolver.ts (exported)"] --> Q["build-edges.ts\nresolveFallbackTargets"]
    P --> R["incremental.ts\napplyThisReceiverFallbacks"]

    S["markExportedSymbols()\nhelpers.ts (exported)"] --> T["insert-nodes.ts\ninsertDefinitionsAndExports"]
    S --> U["native-orchestrator.ts\ninsertBackfilledNodes"]

Loading

%%{init: {'theme': 'base', 'themeVariables': {"darkMode": true, "background": "#0d1117", "primaryColor": "#21262d", "primaryTextColor": "#e6edf3", "primaryBorderColor": "#8b949e", "lineColor": "#8b949e", "textColor": "#e6edf3", "edgeLabelBackground": "#161b22", "actorBkg": "#21262d", "actorBorder": "#8b949e", "actorTextColor": "#e6edf3", "actorLineColor": "#8b949e", "signalColor": "#8b949e", "signalTextColor": "#e6edf3", "noteBkgColor": "#373320", "noteBorderColor": "#d4a72c", "noteTextColor": "#f0e6c0", "labelBoxBkgColor": "#21262d", "labelBoxBorderColor": "#8b949e", "labelTextColor": "#e6edf3", "loopTextColor": "#e6edf3", "activationBkgColor": "#30363d", "activationBorderColor": "#8b949e"}}}%%
flowchart TD
    A["communitiesData()\nsrc/features/communities.ts"] -->|capacityGrowthFactor| B["louvainCommunities()\nsrc/graph/algorithms/louvain.ts"]
    B -->|JS path| C["louvainJS() → detectClusters()"]
    C --> D["runLouvainUndirectedModularity()\noptimiser.ts"]
    D -->|capacityGrowthFactor| E["makePartition()\npartition.ts"]
    E --> F["ensureCommCapacity()\nuses s.capacityGrowthFactor"]
    B -->|native path| G["native Rust louvain\n⚠ ignores capacityGrowthFactor"]

    H["DEFAULTS.db.busyTimeoutMs\nconfig.ts"] -->|default param| I["openDb() / openReadonlyOrFail()\nconnection.ts"]
    J["resolveDbSettings()\nconnection.ts"] -->|busyTimeoutMs| I
    I --> K["pipeline.ts / watcher.ts\nnative-orchestrator.ts\nnative-db-lifecycle.ts"]

    P["resolveSameClassQualifiedMethod()\ncall-resolver.ts (exported)"] --> Q["build-edges.ts\nresolveFallbackTargets"]
    P --> R["incremental.ts\napplyThisReceiverFallbacks"]

    S["markExportedSymbols()\nhelpers.ts (exported)"] --> T["insert-nodes.ts\ninsertDefinitionsAndExports"]
    S --> U["native-orchestrator.ts\ninsertBackfilledNodes"]

Loading

_{Reviews (5): Last reviewed commit: "fix: import performance explicitly in be..." | Re-trigger Greptile}

[greptile-apps]

Three copies of the same 1.5 default now exist independently: DEFAULT_CAPACITY_GROWTH_FACTOR here, DEFAULT_CAPACITY_GROWTH_FACTOR in partition.ts, and DEFAULTS.community.capacityGrowthFactor in config.ts. The value in optimiser.ts normalises the option before it reaches makePartition, but partition.ts also keeps its own copy as a last-resort fallback for callers that skip the optimiser layer. If the tuning value ever changes, all three sites must be updated atomically — any miss silently produces mismatched growth behaviour depending on which code path ran. Importing the constant from one canonical location would close that gap.

Suggested change

	const DEFAULT_CAPACITY_GROWTH_FACTOR: number = 1.5;
	import { DEFAULT_CAPACITY_GROWTH_FACTOR } from './partition.js';

Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!

[carlos-alm]

Fixed — exported DEFAULT_CAPACITY_GROWTH_FACTOR from partition.ts and imported it in optimiser.ts instead of keeping an independent literal. The third copy (DEFAULTS.community.capacityGrowthFactor in config.ts) is intentionally kept separate — it follows the same established pattern as DEFAULT_MAX_LEVELS/DEFAULT_MAX_LOCAL_PASSES in this file (module-level fallback mirrored from the user-facing config default, documented via comment), so no further action needed there.

[github-actions]

Codegraph Impact Analysis

48 functions changed → 231 callers affected across 114 files

• median in scripts/lib/bench-timing.ts:25 (13 transitive callers)
• round1 in scripts/lib/bench-timing.ts:35 (10 transitive callers)
• timeMedian in scripts/lib/bench-timing.ts:45 (5 transitive callers)
• runPerfBenchmarks in scripts/token-benchmark.ts:262 (2 transitive callers)
• openDb in src/db/connection.ts:161 (21 transitive callers)
• openReadonlyOrFail in src/db/connection.ts:333 (126 transitive callers)
• ResolvedDbSettings.engine in src/db/connection.ts:360 (0 transitive callers)
• ResolvedDbSettings.busyTimeoutMs in src/db/connection.ts:361 (0 transitive callers)
• resolveDbSettings in src/db/connection.ts:374 (24 transitive callers)
• openRepo in src/db/connection.ts:427 (28 transitive callers)
• openReadonlyWithNative in src/db/connection.ts:482 (8 transitive callers)
• resolveSameClassQualifiedMethod in src/domain/graph/builder/call-resolver.ts:48 (8 transitive callers)
• getExportStmt in src/domain/graph/builder/helpers.ts:453 (15 transitive callers)
• markExportedSymbols in src/domain/graph/builder/helpers.ts:470 (7 transitive callers)
• applyThisReceiverFallbacks in src/domain/graph/builder/incremental.ts:600 (5 transitive callers)
• setupPipeline in src/domain/graph/builder/pipeline.ts:173 (4 transitive callers)
• runPipelineStages in src/domain/graph/builder/pipeline.ts:277 (4 transitive callers)
• resolveReflectionKeyExprFallback in src/domain/graph/builder/stages/build-edges.ts:1137 (3 transitive callers)
• resolveDefinePropertyAccessorFallback in src/domain/graph/builder/stages/build-edges.ts:1182 (3 transitive callers)
• insertDefinitionsAndExports in src/domain/graph/builder/stages/insert-nodes.ts:258 (3 transitive callers)

[carlos-alm]

Addressed Greptile's review feedback: exported DEFAULT_CAPACITY_GROWTH_FACTOR from partition.ts and imported it in optimiser.ts (commit da0c267) instead of keeping an independently-drifting literal. See inline reply for details on why the third copy in config.ts is intentionally kept separate.

[carlos-alm]

@greptileai

[carlos-alm]

Addressed the remaining item from Greptile's review summary: added the missing import { performance } from 'node:perf_hooks'; in scripts/lib/bench-timing.ts — every other benchmark script imports it explicitly rather than relying on the global.

[carlos-alm]

@greptileai