[ APSIS — SPEC D’INTÉGRATION RGPD ]
Taxonomy
34 labels, each mapped to a ner_type from the closed list accepted by arbitrate(), an RGPD regime, a token family, and the entity_type used when exporting.
PERORGLOCGPEFACPRODUCTMONEYLAWDATELANGUAGENORPQUANTITYMISC
[ IDENTITY ]
| apsis_label | ner_type | token_family | regime | entity_type | mask | p |
|---|
| PERSONNE | PER | PERSONNE | ART4 | PERSON | true | P0 |
| DATE_NAISSANCE | DATE | DATE_NAISSANCE | ART4 | DATE | true | P0 |
| MINEUR | PER | — | ART4 | PERSON | true | P0 |
- MINEUR — Masquer sur la personne rattachée.
[ CONTACT ]
| apsis_label | ner_type | token_family | regime | entity_type | mask | p |
|---|
| EMAIL | MISC | EMAIL | ART4 | EMAIL | true | P0 |
| TELEPHONE | MISC | TELEPHONE | ART4 | PHONE | true | P0 |
| ADRESSE | LOC | ADRESSE | ART4 | LOCATION | true | P0 |
[ IDENTIFIERS ]
| apsis_label | ner_type | token_family | regime | entity_type | mask | p |
|---|
| NIRART. 9 CRITICALITY | MISC | NIR | ART4 | ID_NUMBER | true | P0 |
| ID_OFFICIEL | MISC | ID | ART4 | ID_NUMBER | true | P0 |
| IBAN | MISC | IBAN | ART4 | OTHER | true | P0 |
| CARTE_BANCAIRE | MISC | CARTE | ART4 | CREDIT_CARD | true | P0 |
| SIREN | MISC | SIREN | ART4 | OTHER | true | P0 |
| SIRET | MISC | SIRET | ART4 | OTHER | false | P0 |
| ID_PRO_SANTE | MISC | ID_PRO | ART4 | ID_NUMBER | true | P0 |
| PLAQUE | MISC | PLAQUE | ART4 | OTHER | true | P1 |
| MATRICULE | MISC | MATRICULE | ART4 | ID_NUMBER | true | P1 |
- SIRET — true si personne physique (indépendant, micro-entrepreneur, nom + SIRET).
[ TECHNICAL ]
| apsis_label | ner_type | token_family | regime | entity_type | mask | p |
|---|
| IP | MISC | IP | ART4 | IP_ADDRESS | true | P0 |
| ID_TECH | MISC | ID_TECH | ART4 | OTHER | true | P1 |
| LOGIN | MISC | LOGIN | ART4 | OTHER | true | P1 |
| URL | MISC | URL | ART4 | URL | true | P1 |
- URL — Hors catalogue PDF — ajouté pour l’export Taks.md.
[ LOCATION ]
| apsis_label | ner_type | token_family | regime | entity_type | mask | p |
|---|
| LIEU | LOC / GPE / FAC | LIEU | ART4 | LOCATION | false | P1 |
| GEO | LOC | GEO | ART4 | LOCATION | true | P1 |
- LIEU — La whitelist geo_france décide (~1 098 communes).
[ PROFESSIONAL ]
| apsis_label | ner_type | token_family | regime | entity_type | mask | p |
|---|
| ORGANISATION | ORG | ENTREPRISE | ART4 | ORGANIZATION | false | P1 |
| MONTANT | MONEY | MONTANT | ART4 | OTHER | true | P1 |
| DATE | DATE | DATE | ART4 | DATE | true | P1 |
- ORGANISATION — Whitelist + mode décident.
- MONTANT — Masqué en strict, laissé à partir d’équilibré.
- DATE — Masqué en strict seulement.
[ ART. 9 — SPECIAL CATEGORY ]
| apsis_label | ner_type | token_family | regime | entity_type | mask | p |
|---|
| SANTE | MISC | SANTE | ART9 | OTHER | false | P0 |
| ORIGINE | NORP | ORIGINE | ART9 | OTHER | true | P0 |
| POLITIQUE | MISC | POLITIQUE | ART9 | OTHER | true | P0 |
| RELIGION | MISC | RELIGION | ART9 | OTHER | true | P0 |
| SYNDICAT | MISC | SYNDICAT | ART9 | OTHER | true | P0 |
| GENETIQUE | MISC | GENETIQUE | ART9 | OTHER | true | P0 |
| BIOMETRIE | MISC | BIOMETRIE | ART9 | OTHER | true | P0 |
| VIE_SEXUELLE | MISC | VIE_SEXUELLE | ART9 | OTHER | true | P0 |
| ORIENTATION | MISC | ORIENTATION | ART9 | OTHER | true | P0 |
- SANTE — PDF §4 : détecter, mais ne pas masquer le terme médical si aucune personne n’est attachée.
- BIOMETRIE — Uniquement la biométrie servant à identifier.
[ ART. 10 — CRIMINAL ]
| apsis_label | ner_type | token_family | regime | entity_type | mask | p |
|---|
| PENAL | MISC | PENAL | ART10 | OTHER | true | P0 |
- PENAL — Ne pas confondre avec un impayé commercial (art. 4 seulement).