docs: update build performance benchmarks (3.3.1)

[github-actions]

Automated build benchmark update for 3.3.1 from workflow run #397.

[greptile-apps]

Greptile Summary

Automated benchmark update adding 3.3.1 measurements across both changed files. All data is internally consistent — per-file averages, 50K-file extrapolations, percentage deltas, and the raw JSON block all cross-validate correctly against the build numbers in the JSON payload.

Notable observations from the data:

• Significant overall regression vs 3.1.4: Native build speed increased from 3.5 → 6.1 ms/file (+74%), query time from 3ms → 7ms (+122%), and the 50K-file build estimate nearly doubled (175s → 305s).
• Complexity analysis phase: ~10× increase. The phase breakdown shows the native Complexity phase jumped from 15.6ms to 170.6ms (WASM: 76.5ms → 215.5ms). This single phase now accounts for a disproportionate share of total build time and is the largest driver of the overall regression.
• File and graph growth: The analysed codebase grew from 398 → 442 files (+11%), but nodes/file increased 31% (13.4 → 17.6) and edges/file 27% (28.8 → 36.5), suggesting complexity analysis now covers substantially more graph surface area — consistent with the Complexity phase spike.
• Version gap in benchmarks: No 3.2.x or 3.3.0 entries exist in the history; 3.3.1 records follow directly after 3.1.4. This is worth verifying that intermediate releases either did not trigger the benchmark workflow or the data was intentionally omitted.
• The README summary table correctly rounds the 7.1ms native query time to 7ms, consistent with the rounding convention used in previous entries.

Confidence Score: 5/5

• Safe to merge — purely automated documentation update with no code changes and internally consistent data.
• All benchmark values cross-validate (per-file averages, extrapolations, percentage deltas, and JSON block are consistent). No logic, no runtime code, and no risk of breakage. The notable performance regressions are already shipped in 3.3.1; this PR only records them.
• No files require special attention. The Complexity phase regression noted in the summary is pre-existing in 3.3.1 and outside the scope of this documentation PR.

Important Files Changed

Filename	Overview
README.md	Updates 8 performance metrics in the README benchmark table to reflect 3.3.1 measurements; all values are consistent with the detailed benchmarks file.
generated/benchmarks/BUILD-BENCHMARKS.md	Adds 3.3.1 rows to all benchmark tables (build, incremental, query), updates raw totals and phase breakdowns, adds the full structured JSON data block; all values cross-validate correctly.

Sequence Diagram

sequenceDiagram
    participant WF as CI Workflow #397
    participant CG as Codegraph CLI
    participant FS as Filesystem
    participant PR as This PR (Bot)

    WF->>CG: codegraph build (native + WASM)
    CG-->>WF: buildTimeMs, nodes, edges, dbSize, phases
    WF->>CG: run query benchmarks (fn-deps, path, roles, …)
    CG-->>WF: queryTimeMs per query type
    WF->>CG: run incremental rebuilds (no-op + 1-file)
    CG-->>WF: noopRebuildMs, oneFileRebuildMs
    WF->>FS: write generated/benchmarks/BUILD-BENCHMARKS.md
    WF->>FS: update README.md summary table
    FS-->>PR: automated commit with 3.3.1 benchmark data

Loading

_{Last reviewed commit: "docs: update build p..."}