Repository navigation
feat(slice-67g): correct and explainable dashboard numbers — tips, glossary, planned totals - #273
Merged
Merged
Conversation
Allows rebuilding a single app container without restarting the DB or the other service — useful when only one layer changed (e.g. frontend-only fix). ./start-services.sh rebuild server ./start-services.sh rebuild frontend ./start-services.sh rebuild all Uses `docker compose build <svc>` + `up -d --no-deps <svc>` so the DB container and sibling app containers are left untouched.
…ords have one definition
Add named pure-function explain sentences for every Detail and Stats inventory row, wire them into StatTile/StatCard via the existing explain prop, and render one tip per tile carrying both explain and hint text. Changes: - detailExplain.ts (new): totalExplain, successfulExplain, failedExplain, interruptedExplain, inProgressExplain, discardedExplain, notStartedExplain, durationExplain, bayesianTrialsExplain, exploreHint — fixed templates only - detailExplain.test.ts (new): 14 GWT tests with full coverage - estimateLabels.ts: add experimentsExplain, queryResultsExplain, dbUsedExplain, chunksEstimateExplain — each names scope (whole store) - estimateLabels.test.ts: new describe for the four explain helpers - StatTile.tsx: TileTips renders one InfoTip when explain or hint is set - StatTile.test.tsx: one-tip-per-tile test - VectorDbStatsPanel.tsx: explain on Experiments, Query Results, DB Used, Chunks (est. MB) tiles - VectorDbStatsPanel.test.tsx: tooltip text verified per tile - ExperimentDetailScreen.tsx: explain on all StatCards; Bayesian trials tip; Explore InfoTip replaces native title=; native title= removed where explanatory (truncation-aid uses kept) - ExperimentDetailScreen.test.tsx: 2 new tests (partial Bayesian, running grid) - deterministicNumbers.guard.test.ts: detailExplain.ts and StatTile.tsx added to PURE_MODULES guard
… retire observedTotal - explorerPlan.ts: pure fn mapping sweep_plan → plannedTotal/isBayesian/formula/note; strategy from sweep_plan.search_strategy first, config.execution second; Bayesian M = planned_runs ?? bayesian_summary.planned_trials ?? run_count; grid M = full_grid_count - explorerLabels.ts: pure fixed-template sentences — plannedResultsLine, tiedConfigCount (server tied_count preferred), tiedExplain, maxScoreExplain (normalisation note), queryAvgExplain, chunkAvgExplain, matchesExplain, resultsColumnExplain - ExplorePanels: HyperparametersTab accepts plan/configsWithResults/topKFinal; header prints "(N of M planned runs have results)"; Cartesian line renders FormulaLine from plan.formula; native title= replaced by InfoTip on tied badge, score columns and config card; "combination" retired in favour of "configuration" - SearchExplorerScreen: fetches getExperiment on mount; passes plan, configsWithResults and topKFinal to HyperparametersTab; N MATCHES span gets matchesExplain InfoTip - plannedTotalAgreement.test.tsx: 5-fixture agreement test (grid/failed/partial/ Bayesian/empty-padding) — Total tile equals Explorer M, no Detail/Explorer disagreement - deterministicNumbers.guard.test.ts: explorerLabels.ts, explorerPlan.ts added to PURE_MODULES; SearchExplorerScreen.tsx added to SCREENS; 667 tests pass
…creens Replace ambiguous native title= attributes with InfoTip components in ExperimentsScreen, CompareScreen, CrossStoreLeaderboard and ImprovementPanel so every count, rank and score value has an accessible explanation. Key fixes: - plannedRunCount() returns bayesian_summary.planned_trials for Bayesian sweeps (not run_count), eliminating planned-vs-executed ambiguity - bayesianNotStartedCount avoids double subtraction; bayesianTotalsLine reused for the Bayesian sentence in the list row - duplicates_collapsed wired from CompareResponse through dedupNote() so a silent dedup produces a visible sentence (0, 1, N, absent field handled) - rankScope prop on CrossStoreLeaderboard distinguishes group rank (Compare) from export rank (Walkthru) - storyTileExplain / storyRunsExplain explain percentage baselines and primary-store-only scope - Pure modules explainLabels.ts and scoreLabels additions added to deterministicNumbers.guard.test.ts All 689 frontend tests pass; coverage 97.4/93.2/98.5/98.6 (floors 95/90/95/95). Backend 1335 passed. ruff and mypy clean. ESLint 0 warnings.
…etion - CHANGELOG: 67G entry (Glossary, tips on every screen, Explorer planned total, List/Compare/Walkthru tips, determinism guard) - PROGRESS: 67D/67E/67F/67G rows → ✅ COMPLETE with PR refs - PROGRESS Decision Log: DECISIONS #315 (bundled PR strategy) and #316 (one-tip-per-tile, "combination" → "configuration") - GROWTH: 67F and 67G Abstraction Views deltas
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
GlossaryButtonmounted on every screen viaAppPageChrome; 10 fixed terms (Run, Planned runs, Configuration, Trial, Query result, Golden query/record, Similarity score, Lone run = 100, Observed vs planned, Est.) live in pureglossary.ts; Escape closes, focus returnsdetailExplain.ts(10 pure explain functions) +estimateLabels.tsadditions;StatTilerenders oneInfoTipper tile;VectorDbStatsPanelandExperimentDetailScreencarry scope-annotated explanationsexplorerPlan.tsmapssweep_plan→{plannedTotal, isBayesian, formula};ExplorePanelsheader prints(N of M planned runs have results);FormulaLinereplacesobservedTotal; Explorer and Detail totals verified against 5 fixturesexplainLabels.tsfor the Experiments list (count, pagination, Bayesian trials, delete-when-running);CompareScreenwiresduplicates_collapsed+rankScope="group";CrossStoreLeaderboardandImprovementPanelcarry score, rank, overlap and baseline tipsdeterministicNumbers.guard.test.tscovers all new pure modules and screensCloses
Closes #315
Test plan
Sub-slice commits
b49c8c5cf73fd17cee8a5dd4a78d🤖 Generated with Claude Code
Complexity
Automatically refreshed by CI for product-code changes.
Functions above the JavaScript threshold
ExperimentDetailScreen.tsx:67— CC 119SearchExplorerScreen.tsx:39— CC 39ExplorePanels.tsx:454— CC 29LoadingFeedbackPanel.tsx:41— CC 28labels.ts:43— CC 25ExperimentsScreen.tsx:74— CC 22VectorDbStatsPanel.tsx:91— CC 19ExperimentControlButtons.tsx:54— CC 14ExperimentsScreen.tsx:480— CC 14loadWalkthruResults.ts:37— CC 14Functions above the cognitive-complexity threshold
ExperimentDetailScreen.tsx:67— cognitive 59LoadingFeedbackPanel.tsx:41— cognitive 27labels.ts:43— cognitive 25ExplorePanels.tsx:454— cognitive 23markdown.ts:59— cognitive 20SearchExplorerScreen.tsx:39— cognitive 18steps.ts:63— cognitive 16