index

Start at Dashboard

Catalog of every note in this wiki. One line per note. Add new entries under the matching category. Categories are fixed: see Wiki Conventions and Vault Style Guide.

Maps of Content

• Build Inventory MOC — search before building anything. 37 indexed code units (14 backend files, 14 frontend pages, 4 query modules, 5 MSW handlers + the registry index) with status, gap_ref, mock_ref. Schema locked by repo-side ADR-0010. Walker script + CI hook scheduled next.
• Frontend MOC — kundkort frontend (17 notes; legacy — see banner there)
• KB MOC — KB legal subproject (13 notes)
• Tests MOC — test suite + coverage gaps (10 notes)
• History MOC — git eras and notable commits (12 notes)
• Compliance MOC — GDPR, opt-out, RoPA, Article 14 (8 notes)

Architecture

• System Overview — what DBPOC is, who it serves, the four pipeline stages at a glance.
• Stack — Bun, TypeScript, PostgreSQL, Redis, BullMQ, Keycloak — versions and roles.
• Repository Layout — top-level folders in ./ and what lives where.
• Repository Layout Complete — 89 TypeScript files mapped with exports and purposes.
• Pipeline — Scrape → Enrich → Update queues, worker boundaries, retry policy.
• Decision Register — 12 ADRs with context + consequences (yesterday’s vault format; cross-references the canonical docs/adr/ in repo).
• Symbol Map — code file → wiki note navigation for 40+ source files.

Enrichment

• Crawlee Scraper — direct site crawler, rate limits, contact extraction quirks.
• Domain Discovery — how a company gets matched to a website (Serper + heuristics).
• Google Places — Maps fallback for phone, address, and existence checks.
• Firecrawl — third-party scraper used as a fallback path.
• Name Validation — NER and blocklist filtering for contact names.
• Lead Scoring — scoring rubric and the four validation layers.
• EnrichV7 — top-level orchestrator function called by Enrich_Job.

Data

• Bolagsverket Import — bulk ingest of ~1.8M Swedish companies.
• SCB Import — bulk ingest of ~1.6M foundations from SCB TSV.
• Database Schema — tables in enrichnodedb, ownership and relationships.
• Database Schema Complete — full catalog (30+ tables, 50+ indexes, full migration history).
• Schema Migrations — migrations/ folder, how they are run, current state.

Compliance

• GDPR Legitimate Interest — Article 6(1)(f) basis and the balancing test.
• GDPR Audit Findings — full compliance audit; 7 critical gaps, 8 high-priority gaps.
• Bisnode case — IMY 2019 decision; the lesson behind Article 14 timing.
• Reklamspärr — Swedish marketing opt-out registry handling.
• Article 14 — notification obligations when data is collected from third parties.
• RoPA Log — record of processing activities, where it lives in the schema.
• Opt-Out Hashes — hashed opt-out blocklist and matching procedure.
• Blocklists — domain blocklist + opt-out blocklist composition.
• Domain Blocklist — domain-level rejection rules.

Operations

• Known Issues — open bugs, broken assumptions, things that bite.
• Autoresearch Loop — how autoresearch/ runs experiments and persists results, plus the three runner variants.
• Experiment Results — index into autoresearch/results/ with one-line summaries.
• Autoresearch Result Types — the three structurally distinct artefact families under autoresearch/results/.
• Environment Variables — full inventory of env vars (58 docs).
• Scripts Reference — every script under scripts/ with purpose + invocation.
• Lint Report 2026-04-27 — last lint pass output.
• Local Development — Docker Compose, env vars, common workflows.

Scripts

• Import Scripts — Bolagsverket / SCB / merge npm aliases, PRV trademark importer, IIS .se zone loader.
• Archive Scripts — archive-inactive, archive-non-ab, verify-archive, restore-from-archive runbook.
• Backup Scripts — pg_dump wrapper, when to run, restore path.
• Debug Scripts — every test-* / check-* / clear-* and the manual enrichment probes.
• One-off Scripts — generate-reports, the standalone SQL files, frontend build helper.

Frontend

• Frontend Overview — what frontend/kundkort/ is, stack, top-level entry, data flow.
• Kundkort Overview — yesterday’s framing of the React SPA; tech stack, auth flow, data flow.
• Components Reference — full catalog of 25 components (pages, sections, UI, charts).
• Hooks Reference — useAuth, useKundkort, useEcoApi, useSearch.
• Services Reference — ECOAPI client (insights, gaps, enrichment).
• Frontend Build — how the SPA is bundled and served by src/api/index.ts.
• Auth Flow — useAuth, LoginModal, the DEV_MODE auto-login flag.
• Kundkort API Client — every backend endpoint the SPA calls.
• EcoAPI Integration — secondary insights backend at port 3100, hard-coded auth.
• Search Page — type-ahead search, recent searches, advanced filters, CSV export.
• Kundkort Page — detail view shell: tabs, sticky bar, Berika action.
• Identity Card — Översikt tab, basic facts row.
• Contacts Section — ContactInfoCard and per-role ContactsSection with ContactCard.
• Summary Section — sammanfattning text plus source pill.
• Financials Section — three-row table (anställda, omsättning, resultat) per year.
• Koncern Section — parent and subsidiaries from data.koncern.
• Varumarken Section — trademarks from data.agda_varumarken.
• Tremor Charts — FinancialChart, ContactDistributionChart, DataCompletenessChart.
• Gaps Panel — client-derived data-gap analysis with “Fyll gap” button.
• Insights Panel — metrics tiles, growth and risk signals.
• UI Primitives — SectionHeading, KvRow, GapBadge, ContactCard, SkeletonLoader, LoadingCard, ErrorState, ErrorPanel, Footer, CompanyHeader, formatOrgNr.

Testing

• Test Strategy — what runs under bun test, env vars, what’s mocked, what hits the DB or network.
• API Tests — tests/api/*.test.ts: REST integration tests for auth, CRUD, search, kundkort.
• Enrichment Tests — tests/enrichment/* plus src/enrichmentEngine.v7.test.ts: extractors, processors, board-member integration.
• Fetcher Tests — tests/fetchers/* plus the colocated mapper test: SCB PxWebApi, Bolagsverket bulk download, iXBRL parser.
• Domain Discovery Tests — tests/domainDiscovery.test.ts: hosting and municipality rejection.
• Speed Tests — tests/speed.test.ts: 50 ms registry-lookup budget.
• Compliance Tests — src/compliance.test.ts, src/integration.test.ts, Article 14 section: hashing, RoPA, opt-out, queues.
• Validation Tests — src/lib/validation.test.ts and src/validationEngine.test.ts: input validators and four-layer engine.
• Regression Guard Tests — autoresearch/regression.test.ts: must-pass guard for the autoresearch loop.
• Test Coverage Gaps — source modules that have no automated test.

Process

• Memory Rules — what belongs in ../Memory/ versus the wiki.
• Wiki Conventions — filename, heading, frontmatter, callout rules.
• Vault Style Guide — the controlled tag vocabulary, callout palette, MOC pattern.
• Lint Checklist — mechanical checks every edit must pass.
• LLM Wiki Best Practices — 15 patterns for LLM-optimized knowledge bases.
• QA Report 2026-04-28 — first QA pass of vault claims.
• QA Report Final 2026-04-28 — final QA validation snapshot.
• QA Gate Full Report 2026-04-28 — the audit that drove today’s work.
• Test Coverage Report — vault-level test coverage snapshot.

History

• History Overview — top-level timeline of the 50 commits in master, grouped into eras.
• Notable Commits — top 20 commits with hash, date, and why they matter.
• Git History — yesterday’s vault take on the 8-week timeline (parallel framing to History Overview).
• Experiment History — all 29 quality-loop rounds of the Crawlee experiment.
• History Foundation Era — security hardening and pre-rewrite restore point (2026-03-11 → 2026-03-24).
• History Phase Refactor Era — Phase 1 / 2 / 3 of the Enterprise Enrichment Plan (2026-03-24 → 2026-03-25).
• History Domain Discovery Era — single-day binge that produced the current discovery pipeline (2026-03-25).
• History Firecrawl Era — feature-flagged LLM extractor experiment (2026-03-26 → 2026-03-27).
• History Kundkort Era — React frontend, BV VärdefullaDatamängder API, role normalisation (2026-03-31 → 2026-04-01).
• History Crawlee Era — Crawlee scraper and the 29-round autoresearch quality loop (2026-04-01 → 2026-04-02).
• History Frontend Era — Tremor dashboard and ECOAPI (2026-04-06).
• History Archive Era — backup script, archive tables, archive runs (2026-04-06).
• History Docs Era — SYSTEM_OVERVIEW, root-files archive, name-validation regression (2026-04-22 → 2026-04-27).
• History Migrations Era — schema_migrations tracking added late (2026-04-27).

KB Subproject

• Knowledge Base Overview — yesterday’s framing of the KB subproject (architecture, tech stack, build process).
• Article Index — yesterday’s complete catalog of all GDPR/compliance articles.
• KB Overview — what KB/ is, why it exists, dev workflow, port 3001.
• KB Architecture — server → frontend → kb modules → Anthropic SDK call graph.
• KB Content Index — the 10 live articles, plus the unwired drafts and the 10-vs-29 mismatch.
• KB GDPR Articles — Art. 6 / 14 / 17 live + Art. 30 / overview drafts.
• KB Swedish Law — Dataskyddslagen, IMY, Bolagsverket modules.
• KB B2B Enrichment — LIA, data sources, the five enrichment tiers, Art. 14 ops, retention.
• KB IMY Decisions — IMY structure, 2019–25 timeline, 2025 written-LIA requirement.
• KB Chat Flow — useChat + Anthropic agentic loop with search_web tool.
• KB Search — Fuse.js client-side fuzzy search (no body indexing).
• KB Credibility Scoring — scoreUrl() tiers and the EU-DPA-centric domain lists.
• KB UI Components — Sidebar, ArticleViewer, ChatPanel, TOC, SourceBadge, etc.
• KB Settings — useSettings, modal, model options, theme handling.
• KB Legal Disclaimer — text, persistence, why it exists.

Lessons

• Failed Approaches — 11 lessons from experiments that didn’t work (latest: Path 3B PDF+LLM, 2026-05-05).
• Technical Debt — P0/P1/P2 register with 12 items and remediation plans.