diff --git a/conf/default.yaml b/conf/default.yaml index 3489444..a3c1966 100644 --- a/conf/default.yaml +++ b/conf/default.yaml @@ -1,5 +1,5 @@ # ===================================================================== -# CONFIGURATION PRESIDIO - v4 (PROPRE ET AFFINÉE) +# CONFIGURATION PRESIDIO - v6 (PUISSANCE MAXIMALE) # ===================================================================== supported_languages: [en, fr] @@ -11,7 +11,8 @@ nlp_configuration: - lang_code: en model_name: en_core_web_lg - lang_code: fr - model_name: fr_core_news_lg + # MODIFIÉ : Passage au modèle le plus large et le plus précis + model_name: fr_core_news_lg ner_model_configuration: labels_to_ignore: - MISC @@ -24,12 +25,11 @@ nlp_configuration: - PRODUCT - QUANTITY - WORK_OF_ART - # Abaisser le seuil pour les entités qui n'étaient pas détectées confidence_thresholds: DEFAULT_CONFIDENCE: 0.65 PERSON: 0.85 - LOCATION: 0.4 # MODIFIÉ : Accepte les adresses même avec une faible confiance - DATE_TIME: 0.4 # MODIFIÉ : Accepte les dates même avec une faible confiance + LOCATION: 0.4 + DATE_TIME: 0.4 ORGANIZATION: 0.7 # 2. CONFIGURATION DU REGISTRE DES DÉTECTEURS @@ -38,47 +38,25 @@ recognizer_registry: load_predefined_recognizers: true recognizers: - # --- DÉTECTEURS EXISTANTS --- + # --- DÉTECTEURS FONDAMENTAUX --- + - name: CustomDateRecognizer + supported_language: fr + supported_entity: CUSTOM_DATE + patterns: + - name: Date DD/MM/YYYY + 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"] - name: EmailRecognizer supported_language: fr 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"] - - name: PhoneRecognizer supported_language: fr supported_entity: PHONE_NUMBER patterns: [{name: Phone Pattern (FR/BE/LUX), regex: "\\b(?:(?:\\+|00)?(?:32|33|352)|0)\\s?[1-9](?:[\\s.-]?\\d{2}){3,4}\\b", score: 0.8}] context: ["téléphone", "tel", "mobile", "gsm", "contact"] - - - 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", "bancaire"] - - - name: BelgianNRNRecognizer - supported_language: fr - supported_entity: BE_NATIONAL_REGISTER_NUMBER - 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", "nrn", "niss"] - - - name: BelgianEnterpriseRecognizer - supported_language: fr - supported_entity: BE_ENTERPRISE_NUMBER - patterns: [{name: BE Enterprise Number, regex: "\\bBE\\s?0\\d{3}[\\.\\s]?\\d{3}[\\.\\s]?\\d{3}\\b", score: 0.95}] - context: ["numéro d'entreprise", "btw", "tva", "BCE", "KBO"] - - # --- CORRECTION MAJEURE : Règle BIC/SWIFT stricte --- - - name: BICRecognizer - supported_language: fr - supported_entity: BE_BIC - patterns: - - name: BIC/SWIFT Code - regex: "\\b([A-Z]{4}[A-Z]{2}[A-Z0-9]{2}([A-Z0-9]{3})?)\\b" - score: 1.0 - context: ["bic", "swift"] - - name: JobTitleRecognizer supported_language: fr supported_entity: JOB_TITLE @@ -88,14 +66,87 @@ recognizer_registry: score: 0.8 context: ["représentée par", "fonction", "poste", "attention de"] - - name: CompanyFormRecognizer + # --- DÉTECTEURS FINANCIERS --- + - name: CreditCardRecognizer supported_language: fr - supported_entity: LEGAL_FORM + supported_entity: CREDIT_CARD_NUMBER patterns: - - name: Common Legal Forms - regex: "\\b(SPRL|SA|SCS|SNC|SCRI|SCA|SRL|SC)\\b" - score: 0.9 - context: ["société", "entreprise"] + - name: Credit Card with spaces + 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 bancaire", "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", "bancaire"] + - name: BICRecognizer + supported_language: fr + supported_entity: BE_BIC + patterns: + - name: BIC/SWIFT Code + regex: "\\b([A-Z]{4}[A-Z]{2}[A-Z0-9]{2}([A-Z0-9]{3})?)\\b" + score: 1.0 + context: ["bic", "swift"] + + # --- NUMÉROS D'IDENTIFICATION (FRANCE) --- + - 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"] + - name: FrenchSIRENSIRETRecognizer + supported_language: fr + supported_entity: FR_SIREN_SIRET + patterns: + - {name: SIRET Pattern, regex: "\\b[0-9]{3}[\\s]?[0-9]{3}[\\s]?[0-9]{3}[\\s]?[0-9]{5}\\b", score: 0.9} + - {name: SIREN Pattern, regex: "\\b[0-9]{3}[\\s]?[0-9]{3}[\\s]?[0-9]{3}\\b", score: 0.85} + context: ["siren", "siret"] + - name: FrenchPassportRecognizer + supported_language: fr + supported_entity: FR_PASSPORT + patterns: [{name: French Passport, regex: "\\b[0-9]{2}[A-Z]{2}[0-9]{5}\\b", score: 1.0}] + context: ["passeport"] + - name: FrenchIDCardRecognizer + supported_language: fr + supported_entity: FR_ID_CARD_NUMBER + patterns: [{name: French CNI, regex: "\\b[0-9]{12}\\b", score: 0.8}] + context: ["cni", "carte nationale d'identité"] + - name: FrenchLicensePlateRecognizer + supported_language: fr + supported_entity: FR_LICENSE_PLATE + patterns: [{name: French License Plate, regex: "\\b([A-Z]{2}-\\d{3}-[A-Z]{2}|\\d{1,4}\\s[A-Z]{2,3}\\s\\d{2,3})\\b", score: 0.8}] + context: ["plaque", "immatriculation"] + + # --- NUMÉROS D'IDENTIFICATION (BELGIQUE) --- + - name: BelgianNRNRecognizer + supported_language: fr + supported_entity: BE_NATIONAL_REGISTER_NUMBER + 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", "nrn", "niss"] + - name: BelgianEnterpriseRecognizer + supported_language: fr + supported_entity: BE_ENTERPRISE_NUMBER + patterns: [{name: BE Enterprise Number, regex: "\\bBE\\s?0\\d{3}[\\.\\s]?\\d{3}[\\.\\s]?\\d{3}\\b", score: 0.95}] + context: ["numéro d'entreprise", "btw", "tva", "BCE", "KBO"] + - name: BelgianIDCardRecognizer + supported_language: fr + supported_entity: BE_ID_CARD_NUMBER + patterns: [{name: "Belgian ID Card Number", regex: "\\b\\d{3}-\\d{7}-\\d{2}\\b", score: 1.0}] + context: ["carte d'identité", "eid"] + - name: BelgianLicensePlateRecognizer + supported_language: fr + supported_entity: BE_LICENSE_PLATE + patterns: [{name: "Belgian License Plate", regex: "\\b(?:[1-9]-\\w{3}-\\d{3}|\\w{3}-\\d{3})\\b", score: 0.8}] + context: ["plaque", "immatriculation"] + - name: BelgianPassportRecognizer + supported_language: fr + supported_entity: BE_PASSPORT + patterns: [{name: Belgian Passport, regex: "\\b[A-Z]{2}\\d{6}\\b", score: 1.0}] + context: ["passeport"] # 3. LISTE D'EXCLUSION (ALLOW LIST) # ===================================================================== @@ -163,41 +214,45 @@ anonymizer_config: LOCATION: replace ORGANIZATION: replace DATE_TIME: replace + CUSTOM_DATE: replace CREDIT_CARD_NUMBER: replace EMAIL_ADDRESS: replace - IP_ADDRESS: replace PHONE_NUMBER: replace - URL: replace IBAN: replace MONEY: replace - BE_NATIONAL_REGISTER_NUMBER: replace - BE_ENTERPRISE_NUMBER: replace - FR_SOCIAL_SECURITY_NUMBER: replace - FR_SIREN_SIRET: replace - BE_ID_CARD_NUMBER: replace - BE_LICENSE_PLATE: replace BE_BIC: replace JOB_TITLE: replace - LEGAL_FORM: replace + BE_NATIONAL_REGISTER_NUMBER: replace + BE_ENTERPRISE_NUMBER: replace + BE_ID_CARD_NUMBER: replace + BE_LICENSE_PLATE: replace + BE_PASSPORT: replace + FR_SOCIAL_SECURITY_NUMBER: replace + FR_SIREN_SIRET: replace + FR_PASSPORT: replace + FR_ID_CARD_NUMBER: replace + FR_LICENSE_PLATE: replace replacements: PERSON: "" LOCATION: "" ORGANIZATION: "" DATE_TIME: "" + CUSTOM_DATE: "" CREDIT_CARD_NUMBER: "" EMAIL_ADDRESS: "" - IBAN: "" - IP_ADDRESS: "" PHONE_NUMBER: "" - URL: "" + IBAN: "" MONEY: "" - BE_NATIONAL_REGISTER_NUMBER: "" - BE_ENTERPRISE_NUMBER: "" - FR_SOCIAL_SECURITY_NUMBER: "" - FR_SIREN_SIRET: "" - BE_ID_CARD_NUMBER: "" - BE_LICENSE_PLATE: "" BE_BIC: "" JOB_TITLE: "" - LEGAL_FORM: "" + BE_NATIONAL_REGISTER_NUMBER: "" + BE_ENTERPRISE_NUMBER: "" + BE_ID_CARD_NUMBER: "" + BE_LICENSE_PLATE: "" + BE_PASSPORT: "" + FR_SOCIAL_SECURITY_NUMBER: "" + FR_SIREN_SIRET: "" + FR_PASSPORT: "" + FR_ID_CARD_NUMBER: "" + FR_LICENSE_PLATE: ""