Skip to content

Comments

Explorer: performance overhaul, stream preview, and e2e testing#93

Merged
benbernard merged 2 commits intomasterfrom
feature/explorer-perf-and-polish
Feb 24, 2026
Merged

Explorer: performance overhaul, stream preview, and e2e testing#93
benbernard merged 2 commits intomasterfrom
feature/explorer-perf-and-polish

Conversation

@benbernard
Copy link
Owner

@benbernard benbernard commented Feb 23, 2026

Summary

  • Replace structuredClone with fast JSON clone and sample-based estimateSize (~4x pipeline speedup)
  • Stabilize wrappedDispatch with useRef and remove unused useState from useExecution to eliminate full-screen re-renders
  • Memoize 8 components with React.memo
  • Add stream preview to AddStageModal and EditStageModal (Tab to toggle panels, Enter to zoom into records)
  • Split useInput hooks to fix typing garble when TextInput is active
  • Fix fuzzy search ranking with name-first matching (+200 bonus) and minScore threshold
  • Fix fromps race condition, undo/redo stale cache, execution retry after error, disabled first stage, welcome screen
  • Rename tui → explorer throughout codebase
  • Add tmux-based e2e test harness with 10 smoke tests (RUN_E2E=1 bun test tests/explorer/e2e/)
  • Add 560+ unit tests: executor, reducer, fuzzy-match, unicode, cache-manager, sessions, etc.
  • Add executor benchmark script

Test plan

  • All 1383 tests pass (12 skipped: tmux e2e and slow fromps tests)
  • Lint and typecheck clean
  • Manual tmux smoke tests: npm run test:e2e

🤖 Generated with Claude Code

@benbernard @claude
feat: major explorer performance overhaul, stream preview, and e2e te…
7034fd8 
…sting

- Replace structuredClone with fast JSON clone (4x faster record cloning)
- Sample-based estimateSize instead of scanning all records (~600x faster)
- Stabilize wrappedDispatch with useRef to prevent cascading re-executions
- Remove unused useState from useExecution to eliminate full-screen re-renders
- Memoize 8 components with React.memo (TitleBar, ForkTabs, StageList, etc.)
- Add stream preview to AddStageModal and EditStageModal (Tab to toggle, Enter to zoom)
- Split useInput hooks to fix typing garble when TextInput is active
- Fix fuzzy search ranking: name-first matching with +200 bonus, minScore threshold
- Fix fromps race condition: skip UPDATE_STAGE_ARGS when args unchanged
- Fix undo/redo stale cache: clear cache and lastError on UNDO/REDO
- Fix execution not retrying after error + args edit (stageFingerprint dep)
- Fix disabled first stage showing 0 records
- Fix welcome screen 'n' key not working
- Rename tui → explorer throughout codebase
- Add tmux-based e2e test harness and 10 smoke tests (RUN_E2E=1 to enable)
- Add comprehensive unit tests: executor, reducer, fuzzy-match, unicode, etc.
- Add executor benchmark script

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
@github-actions
Copy link

github-actions bot commented Feb 23, 2026 •
edited
Loading

Performance Benchmark Results

⚠️ 10 regressions detected out of 103 benchmarks (threshold: 25%)

Benchmark Median Baseline Delta
Compiled KeySpec.setValue — nested (address/zip) 220.8µs 156.6µs +41.0% 🔴
grep — 10K records (string match) 828.1µs 443.6µs +86.7% 🔴
sort — 100 records (by score, numeric) 216.3µs 147.4µs +46.8% 🔴
chain — 5 ops (grep eval grep eval
new Record() — 10K objects 507.1µs 149.3µs +239.6% 🔴
Record.get — 10K records × 3 fields 63.1µs 47.7µs +32.2% 🔴
implicit — 3 ops (grep eval grep), 1K records 346.0µs
implicit — 5 ops (grep eval grep eval
TextDecoderStream — 10K lines 5.23ms 4.04ms +29.6% 🔴
bulk text + split — 100K lines 33.55ms 25.55ms +31.3% 🔴

103 benchmarks: 17 faster, 19 slower, 67 within noise (10%)

ℹ️ Note: Benchmarks are advisory-only. GitHub Actions shared runners have variable performance, so results may fluctuate ±25% between runs. For reliable benchmarking, run locally with bun run bench.

Full benchmark results

JSON Parsing

Benchmark Median Baseline Delta Throughput
Record.fromJSON — 100 lines 159.4µs 153.2µs +4.1% 627.40K rec/s
Record.fromJSON — 10K lines 14.52ms 13.09ms +10.9% 🔴 688.54K rec/s, 203.9 MB/s
InputStream.fromString — 100 records 220.2µs 214.6µs +2.6% 454.13K rec/s
InputStream.fromString — 10K records 19.26ms 18.15ms +6.2% 519.15K rec/s, 153.7 MB/s
JSON.parse baseline — 10K lines (no Record) 13.22ms 12.79ms +3.4% 756.26K rec/s, 224.0 MB/s
JSON.parse single array — 10K records 12.98ms 12.50ms +3.8% 770.48K rec/s, 228.2 MB/s

JSON Serialization

Benchmark Median Baseline Delta Throughput
Record.toString — 100 records 82.1µs 86.6µs -5.2% 1.22M rec/s
Record.toString — 10K records 8.56ms 8.10ms +5.7% 1.17M rec/s, 346.1 MB/s
Record.toJSON — 10K records 293.4µs 279.0µs +5.2% 34.08M rec/s
JSON.stringify baseline — 10K objects (no Record) 8.05ms 7.92ms +1.6% 1.24M rec/s, 368.0 MB/s
Batch join — 10K records (map+join) 8.82ms 8.27ms +6.6% 1.13M rec/s, 335.8 MB/s

KeySpec Access

Benchmark Median Baseline Delta Throughput
KeySpec — simple key (name) 218.2µs 214.4µs +1.8% 45.82M rec/s
KeySpec — nested key (address/zip) 542.9µs 530.2µs +2.4% 18.42M rec/s
KeySpec — deep nested (address/coords/lat) 571.6µs 551.8µs +3.6% 17.49M rec/s
KeySpec — array index (tags/#0) 545.9µs 528.8µs +3.2% 18.32M rec/s
Direct property access baseline (rec['name']) 44.1µs 52.4µs -15.8% 🟢 226.59M rec/s
Direct nested access baseline (rec.address.coords.lat) 82.9µs 100.9µs -17.8% 🟢 120.63M rec/s
KeySpec construction — cached (same spec 10K times) 281.9µs 295.2µs -4.5% 35.48M rec/s
KeySpec construction — unique specs (10K different) 2.75ms 3.38ms -18.6% 🟢 3.63M rec/s
Compiled KeySpec.resolveValue — nested (address/zip) 165.3µs 312.4µs -47.1% 🟢 60.50M rec/s
Compiled KeySpec.resolveValue — deep (address/coords/lat) 123.9µs 176.9µs -30.0% 🟢 80.72M rec/s
Compiled KeySpec.resolveValue — array (tags/#0) 154.1µs 267.2µs -42.3% 🟢 64.88M rec/s
Compiled KeySpec.setValue — nested (address/zip) 220.8µs 156.6µs +41.0% 🔴 45.28M rec/s

Core Operations

Benchmark Median Baseline Delta Throughput
grep — 10K records (r.age > 50) 449.4µs 447.3µs +0.5% 22.25M rec/s
grep — 10K records (string match) 828.1µs 443.6µs +86.7% 🔴 12.08M rec/s
eval — 10K records (add computed field) 2.51ms 2.41ms +4.3% 3.99M rec/s
xform — 10K records (push each record) 2.45ms 2.81ms -12.7% 🟢 4.08M rec/s
sort — 100 records (by score, numeric) 216.3µs 147.4µs +46.8% 🔴 462.22K rec/s
sort — 10K records (by score, numeric) 19.29ms 18.48ms +4.4% 518.38K rec/s
sort — 10K records (by name, lexical) 12.71ms 12.27ms +3.6% 786.87K rec/s
collate — 100 records (count by city) 391.9µs 362.5µs +8.1% 255.20K rec/s
collate — 10K records (count by city) 13.81ms 11.77ms +17.3% 🔴 724.20K rec/s
fromcsv — 10K rows (parse CSV to records) 14.50ms 15.04ms -3.6% 689.86K rec/s, 45.3 MB/s

Pipeline Overhead

Benchmark Median Baseline Delta Throughput
chain — single op (grep), 10K records 7.49ms 7.04ms +6.3% 1.34M rec/s
chain — 3 ops (grep eval grep), 10K records 8.17ms 7.27ms
chain — 5 ops (grep eval grep eval grep), 10K records
passthrough baseline — 10K records (direct collector) 6.31ms 5.97ms +5.6% 1.59M rec/s

Record Creation & Serialization

Benchmark Median Baseline Delta Throughput
new Record() — 10K objects 507.1µs 149.3µs +239.6% 🔴 19.72M rec/s
new Record() empty — 10K 151.4µs 135.5µs +11.7% 🔴 66.07M rec/s
Record.get — 10K records × 3 fields 63.1µs 47.7µs +32.2% 🔴 475.79M rec/s
Record.set — 10K records × 1 field 67.8µs 60.8µs +11.6% 🔴 147.45M rec/s
Record.toJSON — 10K records 285.1µs 279.0µs +2.2% 35.08M rec/s
Record.toString — 10K records 8.76ms 8.10ms +8.2% 1.14M rec/s
Record.clone — 10K records 6.37ms 56.23ms -88.7% 🟢 1.57M rec/s
Record.fromJSON — 10K lines 13.90ms 13.09ms +6.2% 719.48K rec/s, 213.1 MB/s
Record.dataRef — 10K records (zero-copy) 40.1µs 38.2µs +4.9% 249.47M rec/s
Record.sort — 10K records (numeric field) 11.83ms 11.49ms +3.0% 845.29K rec/s
Record.sort — 10K records (lexical field) 6.19ms 6.09ms +1.7% 1.62M rec/s
Record.cmp — 1M comparisons (single field) 104.83ms 112.05ms -6.4% 9.54M rec/s
Record.sort — 10K records (nested field numeric) 16.06ms 15.62ms +2.9% 622.48K rec/s
Record.cmp — 1M comparisons (multi-field cached) 91.41ms 87.21ms +4.8% 10.94M rec/s
Record.sort — 10K records (cached comparator reuse) 11.66ms 11.74ms -0.7% 857.76K rec/s

Chain vs Pipe

Benchmark Median Baseline Delta Throughput
chain — 2 ops (grep eval), 100 records 132.1µs 153.0µs -13.7% 🟢
pipe — 2 ops (grep eval), 100 records 450.51ms 419.85ms +7.3%
implicit — 2 ops (grep eval), 100 records 115.6µs 144.1µs -19.8% 🟢
chain — 2 ops (grep eval), 1K records 330.0µs 301.8µs +9.3%
pipe — 2 ops (grep eval), 1K records 393.17ms 384.03ms +2.4%
implicit — 2 ops (grep eval), 1K records 216.0µs 209.2µs +3.3%
chain — 2 ops (grep eval), 10K records 1.11ms 1.07ms +3.4%
pipe — 2 ops (grep eval), 10K records 376.96ms 396.94ms -5.0%
implicit — 2 ops (grep eval), 10K records 1.14ms 1.22ms -6.4%
chain — 3 ops (grep eval grep), 100 records 149.3µs 159.5µs
pipe — 3 ops (grep eval grep), 100 records 566.74ms 546.67ms
implicit — 3 ops (grep eval grep), 100 records 134.3µs 162.2µs
chain — 3 ops (grep eval grep), 1K records 220.2µs 203.0µs
pipe — 3 ops (grep eval grep), 1K records 563.20ms 578.88ms
implicit — 3 ops (grep eval grep), 1K records 346.0µs 218.9µs
chain — 3 ops (grep eval grep), 10K records 1.13ms 1.06ms
pipe — 3 ops (grep eval grep), 10K records 570.96ms 564.71ms
implicit — 3 ops (grep eval grep), 10K records 1.17ms 1.40ms
chain — 5 ops (grep eval grep eval grep), 100 records
pipe — 5 ops (grep eval grep eval grep), 100 records
implicit — 5 ops (grep eval grep eval grep), 100 records
chain — 5 ops (grep eval grep eval grep), 1K records
pipe — 5 ops (grep eval grep eval grep), 1K records
implicit — 5 ops (grep eval grep eval grep), 1K records
chain — 5 ops (grep eval grep eval grep), 10K records
pipe — 5 ops (grep eval grep eval grep), 10K records
implicit — 5 ops (grep eval grep eval grep), 10K records

Line Reading

Benchmark Median Baseline Delta Throughput
InputStream.fromFile — 100 lines 518.5µs 513.2µs +1.0% 192.85K rec/s, 57.0 MB/s
InputStream.fromString — 100 lines 195.6µs 192.9µs +1.4% 511.32K rec/s, 151.0 MB/s
manual buffer (isolated) — 100 lines 264.5µs 266.4µs -0.7% 378.02K rec/s, 111.7 MB/s
bulk text + split — 100 lines 96.6µs 78.0µs +23.8% 🔴 1.04M rec/s, 305.9 MB/s
node readline — 100 lines 479.9µs 513.3µs -6.5% 208.37K rec/s, 61.5 MB/s
TextDecoderStream — 100 lines 279.6µs 277.7µs +0.7% 357.68K rec/s, 105.6 MB/s
binary newline scan — 100 lines 287.4µs 368.1µs -21.9% 🟢 347.99K rec/s, 102.8 MB/s
bun native stdin — 100 lines 26.81ms 24.93ms +7.5% 3.73K rec/s, 1.1 MB/s
InputStream.fromFile — 10K lines 24.94ms 23.85ms +4.6% 400.92K rec/s, 118.7 MB/s
InputStream.fromString — 10K lines 18.99ms 17.71ms +7.2% 526.70K rec/s, 156.0 MB/s
manual buffer (isolated) — 10K lines 6.58ms 6.27ms +5.0% 1.52M rec/s, 450.0 MB/s
bulk text + split — 10K lines 2.48ms 2.37ms +4.8% 4.03M rec/s, 1194.1 MB/s
node readline — 10K lines 9.94ms 9.33ms +6.6% 1.01M rec/s, 297.8 MB/s
TextDecoderStream — 10K lines 5.23ms 4.04ms +29.6% 🔴 1.91M rec/s, 565.9 MB/s
binary newline scan — 10K lines 7.92ms 9.85ms -19.6% 🟢 1.26M rec/s, 374.0 MB/s
bun native stdin — 10K lines 44.48ms 41.98ms +5.9% 224.84K rec/s, 66.6 MB/s
InputStream.fromFile — 100K lines 283.89ms 258.30ms +9.9% 352.25K rec/s, 104.7 MB/s
InputStream.fromString — 100K lines 229.04ms 220.03ms +4.1% 436.60K rec/s, 129.7 MB/s
manual buffer (isolated) — 100K lines 31.13ms 36.57ms -14.9% 🟢 3.21M rec/s, 954.7 MB/s
bulk text + split — 100K lines 33.55ms 25.55ms +31.3% 🔴 2.98M rec/s, 885.7 MB/s
node readline — 100K lines 90.18ms 85.49ms +5.5% 1.11M rec/s, 329.5 MB/s
TextDecoderStream — 100K lines 37.48ms 36.52ms +2.6% 2.67M rec/s, 792.9 MB/s
binary newline scan — 100K lines 73.16ms 77.17ms -5.2% 1.37M rec/s, 406.2 MB/s
bun native stdin — 100K lines 131.44ms 113.90ms +15.4% 🔴 760.83K rec/s, 226.1 MB/s

@benbernard @claude
feat: colorful Catppuccin UI overhaul for explorer
76b1dd7 
Apply the full Catppuccin Mocha palette across all explorer components.
Previously almost everything rendered as gray text; now each UI element
uses semantic accent colors for a bright, friendly appearance.

- TitleBar: mauve title, peach session, blue input, green fork, teal counts
- PipelineBar: lavender cursor, mauve pipes, teal badges
- StatusBar: lavender keybinds, teal undo/redo, green status messages
- InspectorHeader: blue label, peach operation name, teal counts
- RecordTable: lavender bold headers, colored column action hints
- SchemaView: per-type coloring (green/teal/yellow/peach/mauve), pct gradient
- HelpPanel: structured sections with colored headers and lavender keys
- WelcomeScreen: mauve title, colored sessions and key hints
- Modals: colored titles, green $ prompts, category-colored headers
- ConfirmDialog: yellow border, green confirm, red cancel
- Replace StageList/StageRow with compact PipelineBar, add VimTextInput

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
@benbernard benbernard merged commit ab82d9c into master Feb 24, 2026
4 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@benbernard