nacim/Presidio
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update default.yaml

Browse Source
This commit is contained in:
Nacim
2025-06-23 17:21:55 +02:00
committed by GitHub
parent a4b13fa396
commit ac3a7f03f0
1 changed files with 9 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
conf/default.yaml
View File

@@ -13,7 +13,6 @@ nlp_configuration:
model_name: en_core_web_lg model_name: en_core_web_lg
- lang_code: fr - lang_code: fr
model_name: fr_core_news_sm model_name: fr_core_news_sm
# Configuration fine du modèle NLP pour réduire les faux positifs
ner_model_configuration: ner_model_configuration:
labels_to_ignore: labels_to_ignore:
- MISC - MISC
@@ -36,15 +35,12 @@ nlp_configuration:
# 2. CONFIGURATION DU REGISTRE DES DÉTECTEURS # 2. CONFIGURATION DU REGISTRE DES DÉTECTEURS
# ===================================================================== # =====================================================================
recognizer_registry: recognizer_registry:
# A) CHARGER LES DÉTECTEURS PRÉDÉFINIS pour 'en' et 'fr'
load_predefined_recognizers: true load_predefined_recognizers: true
# B) AJOUTER VOS DÉTECTEURS PERSONNALISÉS qui s'ajoutent ou remplacent
# les versions par défaut pour la langue française
recognizers: recognizers:
- name: EmailRecognizer - name: EmailRecognizer
supported_language: fr supported_language: fr
entity_name: EMAIL_ADDRESS supported_entity: EMAIL_ADDRESS # CORRECTION: entity_name -> supported_entity
patterns: patterns:
- name: Email Pattern - name: Email Pattern
regex: "\\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\\.[A-Z|a-z]{2,}\\b" regex: "\\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\\.[A-Z|a-z]{2,}\\b"
@@ -53,7 +49,7 @@ recognizer_registry:
- name: PhoneRecognizer - name: PhoneRecognizer
supported_language: fr supported_language: fr
entity_name: PHONE_NUMBER supported_entity: PHONE_NUMBER # CORRECTION: entity_name -> supported_entity
patterns: patterns:
- name: Phone Pattern (FR/BE/LUX) - name: Phone Pattern (FR/BE/LUX)
regex: "\\b(?:(?:\\+|00)?(?:32|33|352)|0)\\s?[1-9](?:[\\s.-]?\\d{2}){3,4}\\b" regex: "\\b(?:(?:\\+|00)?(?:32|33|352)|0)\\s?[1-9](?:[\\s.-]?\\d{2}){3,4}\\b"
@@ -62,7 +58,7 @@ recognizer_registry:
- name: IbanRecognizer - name: IbanRecognizer
supported_language: fr supported_language: fr
entity_name: IBAN supported_entity: IBAN # CORRECTION: entity_name -> supported_entity
patterns: patterns:
- name: IBAN Pattern - name: IBAN Pattern
regex: "\\b[A-Z]{2}[0-9]{2}\\s?(?:[A-Z0-9]{4}\\s?){2,7}[A-Z0-9]{1,4}\\b" regex: "\\b[A-Z]{2}[0-9]{2}\\s?(?:[A-Z0-9]{4}\\s?){2,7}[A-Z0-9]{1,4}\\b"
@@ -71,7 +67,7 @@ recognizer_registry:
- name: CreditCardRecognizer - name: CreditCardRecognizer
supported_language: fr supported_language: fr
entity_name: CREDIT_CARD_NUMBER supported_entity: CREDIT_CARD_NUMBER # CORRECTION: entity_name -> supported_entity
patterns: patterns:
- name: Credit Card Pattern - name: Credit Card Pattern
regex: "\\b(?:4[0-9]{12}(?:[0-9]{3})?|5[1-5][0-9]{14}|6(?:011|5[0-9]{2})[0-9]{12}|3[47][0-9]{13}|3(?:0[0-5]|[68][0-9])[0-9]{11}|(?:2131|1800|35\\d{3})\\d{11})\\b" regex: "\\b(?:4[0-9]{12}(?:[0-9]{3})?|5[1-5][0-9]{14}|6(?:011|5[0-9]{2})[0-9]{12}|3[47][0-9]{13}|3(?:0[0-5]|[68][0-9])[0-9]{11}|(?:2131|1800|35\\d{3})\\d{11})\\b"
@@ -80,7 +76,7 @@ recognizer_registry:
- name: MoneyRecognizer - name: MoneyRecognizer
supported_language: fr supported_language: fr
entity_name: MONEY supported_entity: MONEY # CORRECTION: entity_name -> supported_entity
patterns: patterns:
- name: Money Pattern EUR - name: Money Pattern EUR
regex: "(?:EUR|€)\\s*\\d{1,3}(?:[.,\\s]\\d{3})*(?:[.,]\\d{2})?|\\d{1,3}(?:[.,\\s]\\d{3})*(?:[.,]\\d{2})?\\s*(?:EUR|€)" regex: "(?:EUR|€)\\s*\\d{1,3}(?:[.,\\s]\\d{3})*(?:[.,]\\d{2})?|\\d{1,3}(?:[.,\\s]\\d{3})*(?:[.,]\\d{2})?\\s*(?:EUR|€)"
@@ -88,7 +84,7 @@ recognizer_registry:
- name: BelgianNRNRecognizer - name: BelgianNRNRecognizer
supported_language: fr supported_language: fr
entity_name: BE_NATIONAL_REGISTER_NUMBER supported_entity: BE_NATIONAL_REGISTER_NUMBER # CORRECTION: entity_name -> supported_entity
patterns: patterns:
- name: NRN Pattern - name: NRN Pattern
regex: "\\b[0-9]{2}\\.[0-9]{2}\\.[0-9]{2}-[0-9]{3}\\.[0-9]{2}\\b" regex: "\\b[0-9]{2}\\.[0-9]{2}\\.[0-9]{2}-[0-9]{3}\\.[0-9]{2}\\b"
@@ -97,7 +93,7 @@ recognizer_registry:
- name: BelgianEnterpriseRecognizer - name: BelgianEnterpriseRecognizer
supported_language: fr supported_language: fr
entity_name: BE_ENTERPRISE_NUMBER supported_entity: BE_ENTERPRISE_NUMBER # CORRECTION: entity_name -> supported_entity
patterns: patterns:
- name: BE Enterprise Number - name: BE Enterprise Number
regex: "\\bBE\\s?0\\d{3}[\\.\\s]?\\d{3}[\\.\\s]?\\d{3}\\b" regex: "\\bBE\\s?0\\d{3}[\\.\\s]?\\d{3}[\\.\\s]?\\d{3}\\b"
@@ -106,7 +102,7 @@ recognizer_registry:
- name: FrenchINSEERecognizer - name: FrenchINSEERecognizer
supported_language: fr supported_language: fr
entity_name: FR_SOCIAL_SECURITY_NUMBER supported_entity: FR_SOCIAL_SECURITY_NUMBER # CORRECTION: entity_name -> supported_entity
patterns: patterns:
- name: INSEE Pattern - name: INSEE Pattern
regex: "\\b[12][\\s]?[0-9]{2}[\\s]?(?:0[1-9]|1[0-2])[\\s]?(?:2[ABab]|[0-9]{2})[\\s]?[0-9]{3}[\\s]?[0-9]{3}[\\s]?[0-9]{2}\\b" regex: "\\b[12][\\s]?[0-9]{2}[\\s]?(?:0[1-9]|1[0-2])[\\s]?(?:2[ABab]|[0-9]{2})[\\s]?[0-9]{3}[\\s]?[0-9]{3}[\\s]?[0-9]{2}\\b"
@@ -115,7 +111,7 @@ recognizer_registry:
- name: FrenchSIRENSIRETRecognizer - name: FrenchSIRENSIRETRecognizer
supported_language: fr supported_language: fr
entity_name: FR_SIREN_SIRET supported_entity: FR_SIREN_SIRET # CORRECTION: entity_name -> supported_entity
patterns: patterns:
- name: SIRET Pattern - name: SIRET Pattern
regex: "\\b[0-9]{3}[\\s]?[0-9]{3}[\\s]?[0-9]{3}[\\s]?[0-9]{5}\\b" regex: "\\b[0-9]{3}[\\s]?[0-9]{3}[\\s]?[0-9]{3}[\\s]?[0-9]{5}\\b"