Factory blueprint
North Star en huidige verantwoordelijkheid
Every vacancy follows one evidence-first pipeline: source -> extraction -> normalization -> taxonomy -> requirement records -> profile evidence matching -> scoring -> review -> human decision.
Nu vastleggen
Foundation
Lock the technically/procedurally complete 89-vacancy extraction scope, run final content-acceptance sampling, and prepare discovery/import plus normalization/taxonomy without mixing pipeline layers.
Factory
71 modules
Uit de Miro/PDF-modulekaarten als volledige backlog.
Volgende bouwlus
Post-extraction acceptance and intake expansion
The 89-vacancy extraction scope is content-acceptance ready, new real vacancies can enter a compliance-safe source staging lane, and the next normalization/taxonomy contracts are explicit without activating matching or scoring.
Fasepoorten
Van brondata naar menselijke beslissing
| Fase | Status | Poort |
|---|---|---|
| Foundation | complete | Control Room shows data, catalog, owner rules and no critical context risks. |
| Vacancy intake | current | Canonical vacancy source stores raw text and source identity only. |
| Extraction | technically-complete-content-pending | Every extracted claim has source quote/provenance and validation state. |
| Normalization | next | Accepted candidates become comparable values without adding new facts. |
| Taxonomy | next | Concepts, aliases, requirement dimensions and OR groups are mapped with confidence. |
| Matching | future-gated | No positive match without profile evidence. |
| Scoring | future-gated | Hard disqualifiers cannot be overridden by a score. |
| Review/output | future | Human decision, notes and learning feedback are explicit and traceable. |
Roadmap
Poortjes tot engine-code
| Stap | Status | Wat moet zichtbaar zijn |
|---|---|---|
| 1. Bronmapping | klaar | 10 primaire tabs, 89 vacatures, XLSX-naamverschillen gecontroleerd. |
| 2. Bronrijen staging | klaar | 2999 / 2999 ruwe sheetrijen in sheet_import_rows. |
| 3. Eerste profielreview | klaar | Preferences, experiences, skills en capability bundles zichtbaar en bewerkbaar in staging. |
| 4. Vacaturereview | klaar | 89 / 89 vacatures zichtbaar met titel, bedrijf, URL, reviewstatus en details. |
| 5. Architectuurcatalogus | klaar | 83 modules vastgelegd met eigenaar, niet-eigenaar, afhankelijkheden en code-owner. |
| 6. Vacature-extractie foundation | volgende | Raw text naar secties, requirement candidates, source quotes en requirement records. Nog geen matching of scoring. |
| 7. Edit workflows | later | Toevoegen, corrigeren, archiveren en bronbehoud voor profiel en vacatures. |
Eerste bouwlus
Sheet-import, profielbeeld en vacaturebeheer
Deze lus maakt de bestaande sheet-data zichtbaar en reviewbaar. Daarna pas gaan extractie, normalisatie, matching en scoring erop aansluiten.
Nog niet canonical
Staged profile rows are not yet copied into final profile tables | Staged vacancies are not yet copied into canonical vacancies | Legacy vacancy analysis fields remain reference material, not final truth | Stefan content approval for all 89 extraction outputs is still pending | Discovery/import automation beyond manual/email/export/API-approved capture is not active | Normalization, taxonomy, matching, scoring and AI verdicts remain gated
- 03-import/google-sheet-source Google Sheet Source / planned Use the audited Stefan_Jobsearch_System_v1.1 mapping as the import contract.
- 04-profile/profile-import Profile Import / planned Build copy/reference/ignore import preview for primary profile tabs.
- 04-profile/profile-editor Profile Editor / planned Design read-only-first profile views, then add guarded edit flows.
- 05-vacancy/vacancy-import Vacancy Import / planned Stage 89 sheet vacancy records and show source/status columns first.
- 05-vacancy/vacancy-editor Vacancy Editor / planned Design manual add/edit/delete workflow with source preservation.
Guardrails
Bestandsgrootte en context-risico
Bestanden432
Grootste bestand759 regels
Hoogste schatting9158 tokens
| Bestand | Status | Regels | Tokens |
|---|---|---|---|
_private/bootstrap.php |
risk | 759 | 6620 |
public-site/assets/design/account.css |
risk | 741 | 4313 |
public-site/modules/accounts.php |
risk | 712 | 9158 |
public-site/modules/vacancy_repository.php |
risk | 664 | 7236 |
src/modules/06-extraction/requirement-extraction/RequirementExtraction.php |
risk | 520 | 5762 |
public-site/index.php |
risk | 513 | 7115 |
_private/pages/discovery_pipeline_audit_section.php |
risk | 354 | 6119 |
src/modules/06-extraction/role-signal-extraction/RoleSignalTaskLikeRules.php |
risk | 319 | 6634 |
_private/vacancy_discovery_full_text_search.php |
ok | 499 | 5345 |
_private/vacancy_discovery_pipeline_audit_review.php |
ok | 491 | 5158 |