From f28e68441d287c5eb413f9299689b849ebe43521 Mon Sep 17 00:00:00 2001
From: Nacim <nacim@biqoz.com>
Date: Tue, 24 Jun 2025 15:22:40 +0200
Subject: [PATCH] Update default.yaml

---
 conf/default.yaml | 99 ++++++++++++++++++++++++++++++-----------------
 1 file changed, 64 insertions(+), 35 deletions(-)

diff --git a/conf/default.yaml b/conf/default.yaml
index 32afc9b..034aef8 100644
--- a/conf/default.yaml
+++ b/conf/default.yaml
@@ -1,5 +1,5 @@
 # =====================================================================
-# CONFIGURATION PRESIDIO - v31 (FINALE, ÉPURÉE ET FIABLE)
+# CONFIGURATION PRESIDIO - v24 (COMPLÈTE, AVEC VOTRE REGEX TÉLÉPHONE)
 # =====================================================================
 supported_languages: [en, fr]
 
@@ -36,7 +36,7 @@ recognizer_registry:
   load_predefined_recognizers: true
   
   recognizers:
-    # --- DÉTECTEURS FIABLES ET PRÉCIS UNIQUEMENT ---
+    # --- DÉTECTEURS FIABLES ET PRÉCIS ---
     - name: CustomDateRecognizer
       supported_language: fr
       supported_entity: CUSTOM_DATE
@@ -44,23 +44,59 @@ recognizer_registry:
         - name: Date JJ/MM/AAAA
           regex: "\\b(0[1-9]|[12][0-9]|3[01])[-/.](0[1-9]|1[012])[-/.](19|20)\\d{2}\\b"
           score: 1.0
-      context: ["date de naissance", "né le", "date"]
+      context: ["date de naissance", "né le"]
 
-    - name: BelgianPhoneRecognizer
+    - name: EmailRecognizer
       supported_language: fr
-      supported_entity: PHONE_NUMBER
+      supported_entity: EMAIL_ADDRESS
+      patterns: [{name: Email Pattern, regex: "\\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\\.[A-Z|a-z]{2,}\\b", score: 1.0}]
+      context: ["email", "courriel", "mail"]
+
+    # --- VOTRE REGEX TÉLÉPHONE, ADAPTÉE POUR PRESIDIO ---
+    - name: CustomBelgianPhoneRecognizer
+      supported_language: fr
+      supported_entity: BE_PHONE_NUMBER
       patterns:
-        - name: Belgian Mobile Pattern
-          regex: "\\b(?:\\+|00)32[\\s.-]?4[6-9][0-9](?:[\\s.-]?\\d{2}){3}\\b"
+        - name: Belgian Phone Pattern (votre regex adaptée)
+          # J'ai doublé les backslashes et remplacé ^/$ par \b
+          regex: "\\b(((\\+|00)32[ ]?(?:\\(0\\)[ ]?)?)|0)?4\\d{2}(\\s?\\d{2}[\\s\\.]?){3}\\b"
           score: 1.0
       context: ["téléphone", "tel", "gsm", "mobile"]
       
+    - name: CustomCreditCardRecognizer
+      supported_language: fr
+      supported_entity: CREDIT_CARD_NUMBER
+      patterns:
+        - name: Credit Card with spaces (Visa, Mastercard)
+          regex: "\\b(?:4[0-9]{3}(?:[ -]?[0-9]{4}){3}|5[1-5][0-9]{2}(?:[ -]?[0-9]{4}){3})\\b"
+          score: 0.95
+      context: ["carte", "visa", "mastercard"]
+      
     - name: IbanRecognizer
       supported_language: fr
       supported_entity: IBAN
       patterns: [{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", score: 0.95}]
       context: ["iban", "compte"]
+      
+    - name: SWIFTRecognizer
+      supported_language: fr
+      supported_entity: SWIFT_CODE
+      patterns:
+        - name: BIC/SWIFT Code
+          regex: "\\b[A-Z]{4}[A-Z]{2}[A-Z0-9]{2}(?:[A-Z0-9]{3})?\\b"
+          score: 0.5
+      context: ["bic", "swift"]
 
+    - name: SmartOrganizationRecognizer
+      supported_language: fr
+      supported_entity: ORGANIZATION
+      patterns:
+        - name: Company Name with Legal Form
+          regex: "\\b([A-Z][a-zà-ÿ]+(?:\\s[A-Z][a-zà-ÿ]+)*)\\s+(SPRL|SARL|SA|SCS|SNC)\\b"
+          score: 1.0
+      context: ["société", "entreprise"]
+
+    # --- NUMÉROS D'IDENTIFICATION ---
     - name: BelgianEnterpriseRecognizer
       supported_language: fr
       supported_entity: BE_ENTERPRISE_NUMBER
@@ -73,40 +109,29 @@ recognizer_registry:
       patterns: [{name: NRN Pattern, regex: "\\b[0-9]{2}\\.[0-9]{2}\\.[0-9]{2}-[0-9]{3}\\.[0-9]{2}\\b", score: 1.0}]
       context: ["registre national"]
       
+    - name: FrenchINSEERecognizer
+      supported_language: fr
+      supported_entity: FR_SOCIAL_SECURITY_NUMBER
+      patterns:
+        - name: INSEE Pattern with flexible spaces
+          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"
+          score: 0.95
+      context: ["sécurité sociale", "insee", "nir"]
+
 # 3. LISTE D'EXCLUSION (ALLOW LIST)
 # =====================================================================
 allow_list:
-  # Liste de sécurité pour les mots en majuscules des contrats
-  - CONTRAT
-  - PRÊT
-  - CONVERTIBLE
-  - PROJET
-  - CONSTITUE
-  - CONTIENT
-  - CONSEILS
-  - PARTIES
-  - UTILISANT
-  - DOCUMENT
-  - DEVRAIENT
-  - TOUJOURS
-  - ENVISAGER
-  - SOIGNEUSEMENT
-  - DEMANDER
-  - CONSEILLERS
-  - QUALIFIÉS
-  - ÉVALUER
-  - IMPLICATIONS
-  - UTILISATION
-  - ATTENDU
-  - QUE
-  # Liste standard
   - Adresse
   - ADRESSE
+  - Contrat
+  - Document
   - Société
   - Investisseur
   - Montant
+  - Prêt
   - Intérêt
   - Partie
+  - Parties
   - Annexe
   - Remboursement
   - Conversion
@@ -149,15 +174,17 @@ anonymizer_config:
     ORGANIZATION: replace
     DATE_TIME: replace 
     MONEY: replace
-    CREDIT_CARD_NUMBER: replace
-    EMAIL_ADDRESS: replace
 
     # Entités détectées par nos règles personnalisées
     CUSTOM_DATE: replace
-    PHONE_NUMBER: replace 
+    CREDIT_CARD_NUMBER: replace 
+    EMAIL_ADDRESS: replace
     IBAN: replace
     BE_ENTERPRISE_NUMBER: replace
     BE_NATIONAL_REGISTER_NUMBER: replace
+    FR_SOCIAL_SECURITY_NUMBER: replace
+    SWIFT_CODE: replace
+    BE_PHONE_NUMBER: replace
   
   replacements:
     PERSON: "<PERSONNE>"
@@ -168,7 +195,9 @@ anonymizer_config:
     CREDIT_CARD_NUMBER: "<NUMERO_CARTE_BANCAIRE>"
     MONEY: "<MONTANT>"
     EMAIL_ADDRESS: "<EMAIL>"
-    PHONE_NUMBER: "<TELEPHONE>"
     IBAN: "<IBAN>"
     BE_ENTERPRISE_NUMBER: "<NUM_ENTREPRISE_BE>"
     BE_NATIONAL_REGISTER_NUMBER: "<NRN_BELGE>"
+    FR_SOCIAL_SECURITY_NUMBER: "<NUM_SECU_FR>"
+    SWIFT_CODE: "<CODE_SWIFT>"
+    BE_PHONE_NUMBER: "<BE_PHONE_NUMBER>"