🔬 Deep-Inspection Generator: Multi-Iteration Intelligence Report Architecture#1163
Merged
🔬 Deep-Inspection Generator: Multi-Iteration Intelligence Report Architecture#1163
Conversation
…ence reports - Add --depth CLI parameter (1-4) to config.ts, exported as analysisDepth - Add buildExecutiveSummary() for decision-maker briefs (depth ≥ 3), all 14 languages - Add buildPredictiveAssessment() with confidence %, coalition stability, risk scenarios (depth ≥ 2), all 14 languages - Add buildHistoricalContext() with precedent analysis and Nordic/EU benchmarking (depth ≥ 2), all 14 languages - Add buildMethodologySection() with data sources, iteration log, confidence score (depth ≥ 3), all 14 languages - Add deriveConfidence() heuristic (0-100) with named constants for weights - Extend buildStrategicImplications() from 5 languages to all 14 - Add new DEEP_SECTION_LABELS for executiveSummary, predictiveAssessment, historicalContext, methodology - Update generateDeepInspectionContent() to accept depth parameter with depth-gated sections - Create scripts/deep-inspection/index.ts with DeepInspectionPipeline class - Add 13 tests in tests/deep-inspection-pipeline.test.ts Co-authored-by: pethers <1726836+pethers@users.noreply.github.com>
Copilot
AI
changed the title
[WIP] Enhance deep-inspection generator for multi-iteration reports
🔬 Deep-Inspection Generator: Multi-Iteration Intelligence Report Architecture
Mar 13, 2026
github-actions
Bot
added
testing
Contributor
🔍 Lighthouse Performance Audit
📥 Download full Lighthouse report Budget Compliance: Performance budgets enforced via |
Contributor
🔍 Lighthouse Performance Audit
📥 Download full Lighthouse report Budget Compliance: Performance budgets enforced via |
Contributor
There was a problem hiding this comment.
Pull request overview
Refactors the deep-inspection generator into a depth-gated (1–4) multi-iteration intelligence report with new sections and expanded 14-language template coverage, and introduces a small DeepInspectionPipeline wrapper plus initial Vitest coverage.
Changes:
- Added
--depthCLI parsing (analysisDepth) and used it to gate new deep-inspection sections (executive summary, predictive assessment, historical context, methodology/confidence). - Expanded
buildStrategicImplications()templates from 5 to 14 languages. - Introduced
scripts/deep-inspection/index.tspipeline wrapper and added a new test suite.
Reviewed changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated 8 comments.
| File | Description |
|---|---|
scripts/generate-news-enhanced/config.ts |
Parses --depth into analysisDepth (1–4) for deep-inspection runs. |
scripts/generate-news-enhanced/generators.ts |
Adds depth-gated deep-inspection sections + 14-language strategic implications templates. |
scripts/deep-inspection/index.ts |
Adds DeepInspectionPipeline wrapper API around generateDeepInspection(). |
tests/deep-inspection-pipeline.test.ts |
Adds Vitest coverage for config depth, pipeline exports, and template compilation. |
| : Math.max(MIN_PASSAGE_PROBABILITY, BASE_PASSAGE_PROBABILITY - motCount * OPPOSITION_MOTION_PENALTY); | ||
| const blockPct = 100 - passagePct; | ||
|
|
||
| const topicStr = topic ? esc(topic) : (lang === 'sv' ? 'detta område' : 'this area'); |
Comment on lines
+1387
to
+1389
| da: `${topicStr ? `<strong>${topicStr}</strong> er del af` : 'Denne politik er del af'} en lang tradition for dansk parlamentarisk reform. ${domainList.length > 0 ? `I nordisk kontekst har ${domainList.join(', ')} historisk nydt gavn af tværpolitisk konsensus.` : ''} Trendanalyse viser, at politikken på dette område accelererer.`, | ||
| no: `${topicStr ? `<strong>${topicStr}</strong> er en del av` : 'Denne politikken er en del av'} en lang tradisjon for norsk parlamentarisk reform. ${domainList.length > 0 ? `I nordisk kontekst har ${domainList.join(', ')} historisk nytt godt av tverrpolitisk konsensus.` : ''} Trendanalyse indikerer at politikk på dette området akselererer.`, | ||
| fi: `${topicStr ? `<strong>${topicStr}</strong> on osa` : 'Tämä politiikka on osa'} pitkää suomalaista parlamentaarista uudistusperinnettä. ${domainList.length > 0 ? `Pohjoismaisessa kontekstissa ${domainList.join(', ')} aloilla on historiallisesti hyöty puolueiden välisestä yhteisymmärryksestä.` : ''} Trendanalyysi osoittaa, että tämän alan politiikka kiihtyy.`, |
Comment on lines
+1421
to
+1424
| da: ['Overfladeanalyse (hændelser og aktører identificeret)', 'Dybdeanalyse (motivationer og strategiske implikationer)', 'Prædiktiv analyse (prognoser og risikoscenarier)', 'Kvalitetsgennemgang (bias-tjek og fuldstændighedsverificering)'], | ||
| no: ['Overflateanalyse (hendelser og aktører identifisert)', 'Dybdeanalyse (motivasjoner og strategiske implikasjoner)', 'Prediktiv analyse (prognoser og risikoscenarier)', 'Kvalitetsgjennomgang (bias-sjekk og fullstendighetsverifisering)'], | ||
| fi: ['Pintaanalyysi (tapahtumat ja toimijat tunnistettu)', 'Syväanalyysi (motiivit ja strategiset vaikutukset)', 'Ennakoiva analyysi (ennusteet ja riskiskenaariot)', 'Laaduntarkistus (vinoutumien tarkistus ja kattavuuden varmennus)'], | ||
| de: ['Oberflächenanalyse (Ereignisse und Akteure identifiziert)', 'Tiefenanalyse (Motivationen und strategische Implikationen)', 'Prädiktive Analyse (Prognosen und Risikoszernarien)', 'Qualitätsprüfung (Bias-Prüfung und Vollständigkeitsverifikation)'], |
Comment on lines
+23
to
+27
| * languages: ['en', 'sv'], | ||
| * }); | ||
| * const result = await pipeline.run(); | ||
| * ``` | ||
| * |
Comment on lines
+80
to
+96
| // We test the confidence logic indirectly via buildMethodologySection exports. | ||
| // The function is module-private, so test observable behavior via content checks. | ||
| describe('confidence score derivation (via section HTML)', () => { | ||
| it('confidence is between 0 and 100', () => { | ||
| // Heuristic formula: min(100, Math.round(enrichmentRate * 70) + docBonus) | ||
| // Test boundary: 0 docs → 0 | ||
| const confOf = (enriched: number, total: number): number => { | ||
| if (total === 0) return 0; | ||
| const enrichmentRate = enriched / total; | ||
| const docBonus = Math.min(30, Math.round((total / 10) * 30)); | ||
| return Math.min(100, Math.round(enrichmentRate * 70) + docBonus); | ||
| }; | ||
| expect(confOf(0, 0)).toBe(0); | ||
| expect(confOf(5, 5)).toBe(85); // enrichment=100% (70pts) + docBonus=15 = 85 | ||
| expect(confOf(0, 5)).toBe(15); // enrichment=0% + docBonus=15 = 15 | ||
| expect(confOf(5, 10)).toBe(65); // enrichment=50% (35pts) + docBonus=30 = 65 | ||
| expect(confOf(10, 10)).toBe(100); // enrichment=100% (70pts) + docBonus=30 = 100 |
Comment on lines
+156
to
+161
| // It's async (returns Promise) | ||
| const result = pipeline.run(); | ||
| expect(result).toBeInstanceOf(Promise); | ||
| // We don't await it because it requires a live MCP connection; | ||
| // just verify it returns a Promise. | ||
| result.catch(() => {}); |
Comment on lines
+120
to
+133
| async run(): Promise<DeepInspectionResult> { | ||
| const depth = this.params.depth ?? 1; | ||
| const topic = this.params.focusTopic; | ||
|
|
||
| console.log(`🔬 DeepInspectionPipeline starting — depth ${depth}: ${this.phaseLabel(depth)}`); | ||
| if (topic) console.log(` Topic: ${topic}`); | ||
|
|
||
| const result = await generateDeepInspection(); | ||
|
|
||
| return { | ||
| ...result, | ||
| depth, | ||
| topic, | ||
| }; |
Comment on lines
+1274
to
+1277
| const headingPredictive = deepLabel('predictiveAssessment', lang); | ||
| const headingOutcome = deepLabel('likelyOutcome', lang); | ||
| const headingCoalition = deepLabel('coalitionStability', lang); | ||
| const headingRisk = deepLabel('riskScenarios', lang); |
Member
|
@copilot apply changes based on the comments in this thread and TypeScript & JavaScript Testing / Unit Tests (Vitest) (pull_request) |
- Add likelyOutcome/coalitionStability/riskScenarios labels to DEEP_SECTION_LABELS (all 14 langs) - Localize topicStr fallback per language in buildPredictiveAssessment (was English-only) - Fix da/no/fi historical context: use "svensk Riksdagsreform" / "Ruotsin valtiopäivien" (not Danish/Norwegian/Finnish) - Fix German typo: "Risikoszernarien" → "Risikoszenarien" - Fix pipeline example: remove non-existent languages property - Pipeline run() now uses config analysisDepth instead of constructor param - Fix pipeline test: check AsyncFunction without invoking run() - Fix CI: update agentic-workflow-mcp-queries.test.ts for merged workflow changes Co-authored-by: pethers <1726836+pethers@users.noreply.github.com>
Contributor
Author
All review comments addressed and CI failure fixed in fd4b967:
|
Contributor
🔍 Lighthouse Performance Audit
📥 Download full Lighthouse report Budget Compliance: Performance budgets enforced via |
Contributor
There was a problem hiding this comment.
Pull request overview
Adds an analysis-depth concept to the deep-inspection generator to produce a more “intelligence report”-style output (executive summary, predictive assessment, historical context, methodology), and introduces a DeepInspectionPipeline wrapper module with accompanying test updates.
Changes:
- Added
--depthCLI parsing (analysisDepth1–4) and wired it into deep-inspection generation. - Expanded deep-inspection HTML generation with depth-gated intelligence sections + 14-language strategic-implications/templates and new DEEP_SECTION_LABELS.
- Added
scripts/deep-inspection/module entrypoint (DeepInspectionPipeline) and new/updated tests.
Reviewed changes
Copilot reviewed 5 out of 5 changed files in this pull request and generated 9 comments.
Show a summary per file
| File | Description |
|---|---|
| tests/deep-inspection-pipeline.test.ts | Adds coverage for analysisDepth parsing and new deep-inspection pipeline/module exports. |
| tests/agentic-workflow-mcp-queries.test.ts | Updates workflow pattern assertions to match the current workflow doc content. |
| scripts/generate-news-enhanced/generators.ts | Implements depth-gated deep-inspection sections, 14-language strategic implications, and logs depth. |
| scripts/generate-news-enhanced/config.ts | Adds --depth parsing and exports analysisDepth (1–4). |
| scripts/deep-inspection/index.ts | Introduces DeepInspectionPipeline wrapper class and types. |
Comment on lines
+1206
to
+1212
| const hasEnactedLaw = sfsDocs.length > 0; | ||
| const govLed = propCount > motCount; | ||
| const highScrutiny = betCount > 0; | ||
|
|
||
| const templates: Partial<Record<Language, string>> = { | ||
| en: `This deep-inspection intelligence report analyses ${docs.length} parliamentary document${docs.length !== 1 ? 's' : ''}${topic ? ` on <strong>${esc(topic)}</strong>` : ''}${domainPhrase ? `, spanning ${domainPhrase}` : ''}. Of these, ${enriched} ${enriched === 1 ? 'was' : 'were'} enriched with full text to enable substantive analysis. The legislative posture is ${govLed ? 'government-led' : 'opposition-driven'}, with ${propCount} proposition${propCount !== 1 ? 's' : ''} advancing the executive agenda${betCount > 0 ? `, ${betCount} committee report${betCount !== 1 ? 's' : ''} providing parliamentary scrutiny` : ''}${motCount > 0 ? `, and ${motCount} opposition motion${motCount !== 1 ? 's' : ''} challenging the direction` : ''}. ${hasEnactedLaw ? `${sfsDocs.length} statute${sfsDocs.length !== 1 ? 's' : ''} have already been enacted, establishing a legal baseline.` : highScrutiny ? 'Committee engagement indicates that the policy is under active parliamentary review, signalling that key decisions are imminent.' : 'The legislative pipeline remains at an early stage, requiring close monitoring for acceleration signals.'} ${domainPhrase ? `Policy domains engaged — ${domainPhrase} — reflect the cross-cutting nature of this initiative.` : 'The documents reflect focused policy engagement in this area.'} Decision-makers should prioritise tracking committee deliberations and chamber voting patterns as the most reliable forward indicators.`, | ||
| sv: `Denna djupanalys granskar ${docs.length} riksdagsdokument${topic ? ` rörande <strong>${esc(topic)}</strong>` : ''}${domainPhrase ? ` inom ${domainPhrase}` : ''}. Av dessa berikades ${enriched} med fulltext. Det lagstiftande läget är ${govLed ? 'regeringsledet' : 'oppositionsdrivet'} med ${propCount} proposition${propCount !== 1 ? 'er' : ''}, ${betCount > 0 ? `${betCount} utskottsbetänkande${betCount !== 1 ? 'n' : ''} som ger parlamentarisk granskning, ` : ''}${motCount > 0 ? `och ${motCount} opposition${motCount !== 1 ? 'smotioner' : 'smotion'} som ifrågasätter inriktningen` : ''}. ${hasEnactedLaw ? `${sfsDocs.length} lag${sfsDocs.length !== 1 ? 'ar' : ''} har redan antagits och fastställt ett rättsligt ramverk.` : highScrutiny ? 'Utskottsengagemanget visar att policyn är under aktiv parlamentarisk granskning.' : 'Lagstiftningspipelinen befinner sig i ett tidigt skede.'} Beslutsfattare bör prioritera att följa utskottens arbete och omröstningar i kammaren.`, |
Comment on lines
+1280
to
+1285
| const sfsDocs = docs.filter(d => (d.doktyp || d.documentType) === 'sfs'); | ||
| const confidence = deriveConfidence(docs); | ||
|
|
||
| // Passage likelihood heuristic: if committee reports exceed motions → likely passage | ||
| const passageLikely = betCount > motCount || sfsDocs.length > 0; | ||
| const passagePct = passageLikely |
| const betCount = docs.filter(d => (d.doktyp || d.documentType) === 'bet').length; | ||
| const motCount = docs.filter(d => (d.doktyp || d.documentType) === 'mot').length; | ||
| const sfsDocs = docs.filter(d => (d.doktyp || d.documentType) === 'sfs'); | ||
| const enriched = docs.filter(d => d.contentFetched).length; |
| */ | ||
| function buildMethodologySection(docs: RawDocument[], topic: string | null, lang: Language, depth: number): string { | ||
| const esc = escapeHtml; | ||
| const enriched = docs.filter(d => d.contentFetched).length; |
Comment on lines
+74
to
+79
| it('pipeline run() is an async function without side effects on construction', async () => { | ||
| const { DeepInspectionPipeline } = await import('../scripts/deep-inspection/index.js'); | ||
| const pipeline = new DeepInspectionPipeline(); | ||
| // Verify run is an async function by checking its constructor name | ||
| expect(pipeline.run.constructor.name).toBe('AsyncFunction'); | ||
| }); |
Comment on lines
+1398
to
+1404
| const sfsDocs = docs.filter(d => (d.doktyp || d.documentType) === 'sfs'); | ||
| const propCount = docs.filter(d => (d.doktyp || d.documentType) === 'prop').length; | ||
| const allDomains = new Set<string>(); | ||
| docs.forEach(d => detectPolicyDomains(d, lang).forEach(dom => allDomains.add(dom))); | ||
| const domainList = [...allDomains].slice(0, 3).map(d => esc(d)); | ||
| const hasEnacted = sfsDocs.length > 0; | ||
| const topicStr = topic ? esc(topic) : null; |
Comment on lines
+1247
to
+1252
| function deriveConfidence(docs: RawDocument[]): number { | ||
| if (docs.length === 0) return 0; | ||
| const enriched = docs.filter(d => d.contentFetched).length; | ||
| const enrichmentRate = enriched / docs.length; // 0–1 | ||
| const docBonus = Math.min(MAX_DOCUMENT_BONUS, Math.round((docs.length / DOCUMENT_BONUS_DIVISOR) * MAX_DOCUMENT_BONUS)); | ||
| return Math.min(100, Math.round(enrichmentRate * ENRICHMENT_WEIGHT) + docBonus); |
Comment on lines
+53
to
+65
| /** | ||
| * Analysis depth (1–4): | ||
| * 1 — surface analysis (what happened) — default, fastest | ||
| * 2 — adds predictive assessment and historical context | ||
| * 3 — adds executive intelligence summary and methodology section | ||
| * 4 — full multi-iteration intelligence report with all sections | ||
| * | ||
| * **Note**: This must match the `--depth` CLI arg passed at process start, | ||
| * since config.ts reads it at module load time. Setting this here only | ||
| * affects log labels, not the actual generation depth. | ||
| */ | ||
| depth?: 1 | 2 | 3 | 4; | ||
| } |
Comment on lines
+119
to
+132
| async run(): Promise<DeepInspectionResult> { | ||
| const depth = analysisDepth; | ||
| const topic = this.params.focusTopic; | ||
|
|
||
| console.log(`🔬 DeepInspectionPipeline starting — depth ${depth}: ${this.phaseLabel(depth)}`); | ||
| if (topic) console.log(` Topic: ${topic}`); | ||
|
|
||
| const result = await generateDeepInspection(); | ||
|
|
||
| return { | ||
| ...result, | ||
| depth, | ||
| topic, | ||
| }; |
Member
|
@copilot apply changes based on the comments in this thread |
…eline API, test mock
- Fix SFS detection in buildExecutiveSummary, buildPredictiveAssessment, buildHistoricalContext,
buildStrategicImplications to use dokumentnamn.startsWith('SFS') fallback (aligns with
buildDeepInspectionSections pattern)
- Fix enrichment counting: use `fullText || fullContent` instead of `contentFetched` in
deriveConfidence, buildExecutiveSummary, buildStrategicImplications, buildMethodologySection
(contentFetched can be true without full text)
- Remove unused `depth`/`focusTopic` from DeepInspectionPipelineParams — these were
misleading since the effective values come from CLI config
- Pipeline run() now sources topic from config.focusTopic instead of constructor params
- Replace fragile AsyncFunction constructor name test with mock-based approach: vi.mock
generators at top level, assert pipeline.run() returns Promise and result has valid depth
Co-authored-by: pethers <1726836+pethers@users.noreply.github.com>
Member
|
@copilot apply changes based on the comments in this thread |
Contributor
🔍 Lighthouse Performance Audit
📥 Download full Lighthouse report Budget Compliance: Performance budgets enforced via |
Co-authored-by: pethers <1726836+pethers@users.noreply.github.com>
Contributor
Author
Implemented all requested updates in 02056de.
Validation run:
No UI changes were made in this update, so there is no UI screenshot. |
Contributor
🔍 Lighthouse Performance Audit
📥 Download full Lighthouse report Budget Compliance: Performance budgets enforced via |
Contributor
There was a problem hiding this comment.
Pull request overview
This PR extends the deep-inspection generator to support configurable multi-iteration “intelligence report” output (depth 1–4), adds new depth-gated report sections (executive summary, historical context, predictive assessment, methodology/confidence), and introduces a thin programmatic DeepInspectionPipeline wrapper with updated tests to validate the new behavior.
Changes:
- Add
--depth=1..4CLI parsing (analysisDepth) and pass it into deep-inspection rendering to gate advanced sections. - Expand deep-inspection content with new localized section labels and new section builders (executive summary, historical context, predictive assessment, methodology/confidence).
- Add/adjust Vitest coverage to assert depth-gating behavior and MCP workflow documentation requirements; introduce
scripts/deep-inspection/index.tswrapper.
Reviewed changes
Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.
Show a summary per file
| File | Description |
|---|---|
scripts/generate-news-enhanced/config.ts |
Adds --depth parsing and exports analysisDepth (1–4, default 1). |
scripts/generate-news-enhanced/generators.ts |
Implements depth-gated deep-inspection sections + 14-language strategic implications; exports test hooks. |
scripts/deep-inspection/index.ts |
Adds DeepInspectionPipeline wrapper delegating to generateDeepInspection(). |
tests/deep-inspection-pipeline.test.ts |
Adds behavioral tests for depth gating, labels, and pipeline exports. |
tests/agentic-workflow-mcp-queries.test.ts |
Tightens/extends workflow-doc assertions for post-query filtering and cross-referencing. |
Comment on lines
+176
to
+179
| // Should have explicit >= fromDate filtering guidance | ||
| expect(content).toMatch(/>=\s*fromDate/); | ||
| // Should reference Filter results instruction | ||
| expect(content).toMatch(/Filter results/i); |
Comment on lines
+171
to
+173
| it('depth 4 includes the quality-review methodology iteration', async () => { | ||
| const html = await render(4, 'en'); | ||
| expect(html).toContain('Quality review (bias check and completeness verification)'); |
Comment on lines
+166
to
+168
| expect(html).toContain('Likely Outcome'); | ||
| expect(html).toContain('Coalition Stability Forecast'); | ||
| expect(html).toContain('Risk Scenarios'); |
Member
|
@copilot apply changes based on the comments in this thread |
Co-authored-by: pethers <1726836+pethers@users.noreply.github.com>
Contributor
Author
Addressed in commit I made the requested assertions less brittle by switching to semantic/structural checks:
Validation run:
No UI files were changed in this update, so no screenshot applies. |
Contributor
🔍 Lighthouse Performance Audit
📥 Download full Lighthouse report Budget Compliance: Performance budgets enforced via |
Signed-off-by: James Pether Sörling <pethers@users.noreply.github.com>
Contributor
🔍 Lighthouse Performance Audit
📥 Download full Lighthouse report Budget Compliance: Performance budgets enforced via |
![github-code-quality[bot]](https://avatars.githubusercontent.com/u/9919?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Pull request overview
Enhances the deep-inspection generator toward a multi-iteration “intelligence report” model by introducing depth-gated narrative sections, expanding section labels/templates to all 14 languages, and adding a thin DeepInspectionPipeline programmatic entrypoint plus new behavioral tests.
Changes:
- Add
--depth=1..4parsing and depth-gated deep-inspection narrative sections (executive summary, historical context, predictive assessment, methodology/confidence). - Expand
DEEP_SECTION_LABELSand strategic-implications templates to cover all 14 supported languages. - Add a
scripts/deep-inspection/entrypoint wrapper and new/updated Vitest tests for depth gating and workflow documentation patterns.
Reviewed changes
Copilot reviewed 5 out of 5 changed files in this pull request and generated 1 comment.
Show a summary per file
| File | Description |
|---|---|
| tests/deep-inspection-pipeline.test.ts | Adds behavioral tests for depth parsing, pipeline exports, and depth-gated section rendering. |
| tests/agentic-workflow-mcp-queries.test.ts | Strengthens assertions around post-query date filtering and cross-referencing guidance in workflow docs. |
| scripts/generate-news-enhanced/generators.ts | Introduces new deep-inspection section builders, depth-gated rendering, and 14-language labels/templates (but currently contains duplicate imports/vars). |
| scripts/generate-news-enhanced/config.ts | Adds --depth parsing (but currently collides with existing analysisDepth export). |
| scripts/deep-inspection/index.ts | Adds a thin DeepInspectionPipeline wrapper around generateDeepInspection(). |
Comments suppressed due to low confidence (2)
scripts/generate-news-enhanced/generators.ts:2213
contentis declared twice in the same scope (first calling generateDeepInspectionContent(..., analysisDepth), then again calling generateDeepInspectionContent(...)). This is a compile-time error and also discards the new depth-aware rendering. Remove the old duplicate declaration and keep the depth-aware call.
// Topic-focused deep-inspection content (NOT generic content)
const content: string = generateDeepInspectionContent(enrichedDocs, sanitizedTopic, lang, analysisDepth);
// Topic-focused deep-inspection content (template-driven body text).
// At this stage the AI pipeline supplies structured sections (SWOT,
// dashboard, mindmap, Sankey, watch points) while the main narrative
// body is still produced by generateDeepInspectionContent(). Wiring
// AI-derived narrative into the article body is the next phase.
const content: string = generateDeepInspectionContent(enrichedDocs, sanitizedTopic, lang);
scripts/generate-news-enhanced/config.ts:116
config.tsnow exportsanalysisDepthtwice with different types/meanings: a numeric deep-inspection depth (1–4) and the existing AI iteration depth (AnalysisDepth= quick/standard/deep). This name collision will fail compilation and makes imports ambiguous. Rename one of them (e.g.deepInspectionDepthfor 1–4, keepanalysisDepthfor AI iterations) and update all imports/usages accordingly.
/**
* Analysis depth for deep-inspection (1–4).
* 1 — surface analysis (what happened) — default, fastest
* 2 — adds predictive assessment and historical context
* 3 — adds executive intelligence summary and methodology (3 iterations)
* 4 — full report: adds quality-review iteration in methodology (4 iterations)
*/
const rawDepth: string = parseArgValue(depthArg);
const depthArgProvided: boolean = !!depthArg;
const parsedDepthNum: number = rawDepth ? Number(rawDepth) : NaN;
const depthIsValid: boolean = Number.isInteger(parsedDepthNum) && parsedDepthNum >= 1 && parsedDepthNum <= 4;
if (depthArgProvided && !depthIsValid) {
console.warn(`Invalid --depth value "${rawDepth}" (expected integer 1–4), falling back to default 1.`);
}
const safeDepth: number = depthIsValid ? parsedDepthNum : 1;
export const analysisDepth: 1 | 2 | 3 | 4 =
safeDepth === 4 ? 4 :
safeDepth === 3 ? 3 :
safeDepth === 2 ? 2 :
1;
// ---------------------------------------------------------------------------
// Analysis depth (controls number of AI analysis iterations)
// ---------------------------------------------------------------------------
/**
* --analysis-depth=<quick|standard|deep>
*
* Controls how many AI analysis iterations run for deep-inspection articles:
* - `quick` — 1 pass (initial analysis only; fast)
* - `standard` — 2 passes (initial + SWOT refinement; default)
* - `deep` — 3 passes (initial + refinement + stakeholder validation)
*/
const analysisDepthArg: string | undefined = args.find(arg => arg.startsWith('--analysis-depth='));
const rawAnalysisDepth: string = parseArgValue(analysisDepthArg ?? '').toLowerCase();
const VALID_ANALYSIS_DEPTHS: readonly AnalysisDepth[] = ['quick', 'standard', 'deep'];
function parseAnalysisDepth(raw: string): AnalysisDepth {
if ((VALID_ANALYSIS_DEPTHS as readonly string[]).includes(raw)) {
return raw as AnalysisDepth;
}
if (raw && raw !== '') {
console.warn(`⚠️ Unknown --analysis-depth value "${raw}", falling back to "standard". Valid values: ${VALID_ANALYSIS_DEPTHS.join(', ')}`);
}
return 'standard';
}
export const analysisDepth: AnalysisDepth = parseAnalysisDepth(rawAnalysisDepth);
| import type { Language } from '../types/language.js'; | ||
| import type { GenerationResult, DateRange, ArticleCategory, TemplateSection, SwotEntry } from '../types/article.js'; | ||
| import type { TitleSet } from './types.js'; | ||
| import { languages, stats, getSharedClient, requireMcp, toISODate, documentIds, documentUrls, focusTopic, analysisDepth } from './config.js'; |
Signed-off-by: James Pether Sörling <pethers@users.noreply.github.com>
Member
|
@copilot apply changes based on the comments in this thread |
Contributor
🔍 Lighthouse Performance Audit
📥 Download full Lighthouse report Budget Compliance: Performance budgets enforced via |
…ressions Co-authored-by: pethers <1726836+pethers@users.noreply.github.com>
Contributor
Author
Implemented in
Validation:
No UI files changed in this update. |
1 similar comment
Contributor
Author
Implemented in
Validation:
No UI files changed in this update. |
Contributor
🔍 Lighthouse Performance Audit
📥 Download full Lighthouse report Budget Compliance: Performance budgets enforced via |
Signed-off-by: James Pether Sörling <pethers@users.noreply.github.com>
Contributor
🔍 Lighthouse Performance Audit
📥 Download full Lighthouse report Budget Compliance: Performance budgets enforced via |
Contributor
There was a problem hiding this comment.
Pull request overview
This PR extends the deep-inspection generator into a depth-gated “intelligence report” flow by adding new report sections, introducing a --depth=1..4 CLI parameter, and adding a thin DeepInspectionPipeline programmatic entrypoint, with tests covering the new behavior.
Changes:
- Add numeric deep-inspection
analysisDepthparsing (--depth=1..4) and rename the existing quick/standard/deep setting toanalysisMode. - Expand deep-inspection content generation with new depth-gated sections (executive summary, historical context, predictive assessment, methodology/confidence) and 14-language labels/templates.
- Add
scripts/deep-inspection/index.tswrapper and new Vitest coverage for depth gating + exports.
Reviewed changes
Copilot reviewed 5 out of 5 changed files in this pull request and generated 5 comments.
Show a summary per file
| File | Description |
|---|---|
scripts/generate-news-enhanced/generators.ts |
Adds depth-gated intelligence sections, labels, and wires depth into deep-inspection generation + metadata writing. |
scripts/generate-news-enhanced/config.ts |
Introduces --depth parsing (analysisDepth 1–4) and renames the existing --analysis-depth output to analysisMode. |
scripts/deep-inspection/index.ts |
Adds a thin DeepInspectionPipeline class delegating to generateDeepInspection(). |
tests/deep-inspection-pipeline.test.ts |
Adds behavioral tests for depth parsing, exported hooks, section gating, and pipeline exports. |
tests/agentic-workflow-mcp-queries.test.ts |
Tightens workflow-doc assertions around post-query date filtering and cross-referencing guidance. |
Comment on lines
40
to
43
| import { languages, stats, getSharedClient, requireMcp, toISODate, documentIds, documentUrls, focusTopic, analysisDepth, METADATA_DIR } from './config.js'; | ||
| import { runAnalysisPipeline } from '../ai-analysis/pipeline.js'; | ||
| import type { AnalysisResult, AnalysisIterationMetadata, AnalysisDepth } from '../ai-analysis/types.js'; | ||
| import { languages, stats, getSharedClient, requireMcp, toISODate, documentIds, documentUrls, focusTopic, analysisIterations } from './config.js'; |
| focusTopic: sanitizedTopic, | ||
| }); | ||
| const pipelineDuration = iterationDurationsMs.reduce((a, b) => a + b, 0); | ||
| console.log(` 🌐 Generating ${lang.toUpperCase()} version...`); |
Comment on lines
+2146
to
+2149
| focusTopic: sanitizedTopic, | ||
| completedAt: analysis.completedAt, | ||
| }; | ||
| writeAnalysisMetadata(slug, iterationMetadata); |
Comment on lines
2151
to
2155
| // Topic-focused deep-inspection content (NOT generic content) | ||
| const content: string = generateDeepInspectionContent(enrichedDocs, sanitizedTopic, lang, analysisDepth); | ||
| // Topic-focused deep-inspection content (uses AI strategic implications & takeaways) | ||
| const content: string = generateDeepInspectionContent(enrichedDocs, sanitizedTopic, lang, aiResult); | ||
|
|
Comment on lines
2116
to
2133
| @@ -1564,6 +2132,24 @@ export async function generateDeepInspection(): Promise<GenerationResult> { | |||
| console.log(` 🤖 AI analysis: ${aiResult.iterations} iteration(s), analysis score ${aiResult.analysisScore}`); | |||
| } | |||
Member
|
@copilot apply changes based on the comments in this thread |
Co-authored-by: pethers <1726836+pethers@users.noreply.github.com>
Contributor
🔍 Lighthouse Performance Audit
📥 Download full Lighthouse report Budget Compliance: Performance budgets enforced via |
Contributor
Author
Applied in I addressed the review thread items by:
Validation on this commit:
No UI files were changed in this update. |
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.
Original prompt
This section details on the original issue you should resolve
<issue_title>🔬 Deep-Inspection Generator Enhancement — Multi-Iteration AI Intelligence Reports</issue_title>
<issue_description>## 📋 Issue Type
Feature / Intelligence Enhancement
🎯 Objective
Enhance the deep-inspection generator (
generators.ts:642-1690) to produce comprehensive AI-driven political intelligence reports with multi-iteration analysis, document cross-referencing, historical context, and predictive insights. Transform deep-inspection from a document summary tool into a full political intelligence assessment platform.📊 Current State
generateDeepInspection()performs single-pass document analysisgenerateDeepInspectionContent()creates 5 sections via template logic:generateDeepPolicyAnalysis()generateDeepAnalysisSection()cross-document synthesisbuildDeepInspectionSections()creates 4 template sections:🚀 Desired State
Deep Intelligence Report Architecture
Phase 1: Document Intelligence Collection
Phase 2: Multi-Iteration AI Analysis
Phase 3: Multi-Perspective Intelligence Sections
Architecture Changes
generateDeepInspection()into modular pipeline:generators.tsto dedicated module:scripts/deep-inspection/pipeline.ts,analysis.ts,sections.ts,types.ts--depthCLI parameter:quick(1 iteration),standard(3),deep(4+)🔧 Implementation Approach
scripts/deep-inspection/module directoryDeepInspectionPipelineclass with configurable depth📍 Connect Copilot coding agent with Jira, Azure Boards or Linear to delegate work to Copilot in one click without leaving your project management tool.