new full
This commit is contained in:
@@ -3,7 +3,7 @@ recognizer_registry:
|
||||
recognizers:
|
||||
- name: BelgianAddressRecognizer
|
||||
supported_language: fr
|
||||
supported_entity: BE_ADDRESS
|
||||
supported_entity: ADRESSE_BELGE
|
||||
patterns:
|
||||
# Pattern principal : numéro + rue + code postal + ville (SANS contexte)
|
||||
- name: Adresse complète avec numéro devant
|
||||
|
||||
@@ -3,7 +3,7 @@ recognizer_registry:
|
||||
recognizers:
|
||||
- name: BelgianIDCardRecognizer
|
||||
supported_language: fr
|
||||
supported_entity: BE_ID_CARD
|
||||
supported_entity: CARTE_IDENTITE_BELGE
|
||||
patterns:
|
||||
- name: Carte d'identité belge
|
||||
regex: "\\b[0-9]{3}\\-[0-9]{7}\\-[0-9]{2}\\b"
|
||||
@@ -15,7 +15,7 @@ recognizer_registry:
|
||||
|
||||
- name: BelgianPassportRecognizer
|
||||
supported_language: fr
|
||||
supported_entity: BE_PASSPORT
|
||||
supported_entity: PASSEPORT_BELGE
|
||||
patterns:
|
||||
- name: Passeport belge
|
||||
regex: "\\b[A-Z]{2}[0-9]{6}\\b"
|
||||
|
||||
@@ -3,7 +3,7 @@ recognizer_registry:
|
||||
recognizers:
|
||||
- name: BelgianNRNRecognizer
|
||||
supported_language: fr
|
||||
supported_entity: BE_NATIONAL_REGISTER_NUMBER
|
||||
supported_entity: REGISTRE_NATIONAL_BELGE
|
||||
patterns:
|
||||
- name: NRN avec points et tiret
|
||||
regex: "\\b[0-9]{2}\\.[0-9]{2}\\.[0-9]{2}-[0-9]{3}\\.[0-9]{2}\\b"
|
||||
|
||||
@@ -3,7 +3,7 @@ recognizer_registry:
|
||||
recognizers:
|
||||
- name: BelgianPhoneRecognizer
|
||||
supported_language: fr
|
||||
supported_entity: BE_PHONE_NUMBER
|
||||
supported_entity: TELEPHONE_BELGE
|
||||
patterns:
|
||||
# Patterns avec contexte Tel: et Tél:
|
||||
- name: Téléphone fixe avec contexte Tel
|
||||
@@ -20,7 +20,7 @@ recognizer_registry:
|
||||
score: 0.99
|
||||
# Patterns généraux (sans contexte spécifique)
|
||||
- name: Téléphone fixe belge
|
||||
regex: '(?<!BCE\s*:?\s*)\b0[1-9](?:[./\s]?\d{2,3}){3}(?=\b|\)|$|[.,;])(?!.*BCE)'
|
||||
regex: '(?<!BCE\s*:?\s*)(?<!FR[0-9]{2}\s[0-9]{4}\s[0-9]{4}\s)\b0[1-9](?:[./\s]?\d{2,3}){3}(?=\b|\)|$|[.,;])(?!.*BCE)(?!\s[0-9A-Z]{4}\s[0-9A-Z]{3})'
|
||||
score: 0.95
|
||||
- name: Mobile belge
|
||||
regex: '\b04[0-9]{2}[./\s]?[0-9]{2}[./\s]?[0-9]{2}[./\s]?[0-9]{2}(?=\b|\)|$|[.,;])'
|
||||
|
||||
20
conf/recognizers/PII/belgian/vat_numbers.yaml
Normal file
20
conf/recognizers/PII/belgian/vat_numbers.yaml
Normal file
@@ -0,0 +1,20 @@
|
||||
# Recognizer pour numéros de TVA belges
|
||||
recognizer_registry:
|
||||
recognizers:
|
||||
- name: BelgianVATRecognizer
|
||||
supported_language: fr
|
||||
supported_entity: TVA_BELGE
|
||||
patterns:
|
||||
- name: TVA belge format standard
|
||||
regex: "\\bBE[0-9]{4}\\.[0-9]{3}\\.[0-9]{3}\\b"
|
||||
score: 1.0
|
||||
- name: TVA belge compact
|
||||
regex: "\\bBE[0-9]{10}\\b"
|
||||
score: 0.95
|
||||
- name: TVA avec contexte
|
||||
regex: "\\b(?:TVA|tva)\\s*:?\\s*BE[0-9]{4}\\.[0-9]{3}\\.[0-9]{3}\\b"
|
||||
score: 1.0
|
||||
- name: Numéro d'entreprise BCE
|
||||
regex: "\\b(?:BCE|bce)\\s*:?\\s*BE[0-9]{4}\\.[0-9]{3}\\.[0-9]{3}\\b"
|
||||
score: 1.0
|
||||
context: ["TVA", "BCE", "numéro d'entreprise", "identification", "intracommunautaire"]
|
||||
Reference in New Issue
Block a user