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

Update app.py

Browse Source
This commit is contained in:
Nacim
2025-06-23 16:34:12 +02:00
committed by GitHub
parent 37f3b95298
commit 2b9236bbe1
1 changed files with 27 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files

51
app.py
View File

@@ -5,7 +5,7 @@ from flask import Flask, request, jsonify, make_response
from presidio_analyzer import AnalyzerEngine, RecognizerRegistry, PatternRecognizer, Pattern
from presidio_analyzer.nlp_engine import NlpEngineProvider
# On importe les recognizers prédéfinis qu'on veut pouvoir utiliser
# On importe les classes des détecteurs prédéfinis que l'on veut pouvoir utiliser depuis le YAML
from presidio_analyzer.predefined_recognizers import (
CreditCardRecognizer, CryptoRecognizer, DateRecognizer, IpRecognizer,
MedicalLicenseRecognizer, UrlRecognizer, SpacyRecognizer
@@ -19,7 +19,6 @@ logger = logging.getLogger(__name__)
app = Flask(__name__)
# --- Dictionnaire pour mapper les noms du YAML aux classes Python ---
# C'est ce qui nous permet de lire la liste 'recognizer_registry' du YAML
PREDEFINED_RECOGNIZERS_MAP = {
"SpacyRecognizer": SpacyRecognizer,
"CreditCardRecognizer": CreditCardRecognizer,
@@ -30,7 +29,6 @@ PREDEFINED_RECOGNIZERS_MAP = {
"UrlRecognizer": UrlRecognizer,
}
# --- Initialisation Globale de l'Analyseur ---
analyzer = None
try:
@@ -50,23 +48,14 @@ try:
# 3. Créer le registre de recognizers EN SUIVANT LE YAML
logger.info("Creating and populating recognizer registry from config file...")
registry = RecognizerRegistry()
supported_languages = config.get("supported_languages", ["en"])
# === DÉBUT DE LA CORRECTION MAJEURE ===
# A) Charger les recognizers PRÉDÉFINIS listés dans le YAML
supported_languages = config.get("supported_languages", ["en"])
for recognizer_name in config.get("recognizer_registry", []):
if recognizer_name in PREDEFINED_RECOGNIZERS_MAP:
recognizer_class = PREDEFINED_RECOGNIZERS_MAP[recognizer_name]
# On passe les langues supportées à chaque recognizer qu'on instancie
registry.add_recognizer(recognizer_class(supported_languages=supported_languages))
logger.info(f"Loaded predefined recognizer: {recognizer_name}")
# B) Charger les recognizers PERSONNALISÉS définis dans le YAML
custom_recognizers_conf = config.get("recognizers", [])
for recognizer_conf in custom_recognizers_conf:
# Étape A: On pré-construit tous les détecteurs personnalisés ("custom") définis dans la section 'recognizers'
custom_recognizers = {}
for recognizer_conf in config.get("recognizers", []):
patterns = [Pattern(name=p['name'], regex=p['regex'], score=p['score']) for p in recognizer_conf['patterns']]
# On s'assure de ne pas recréer un recognizer prédéfini mais bien un custom
custom_recognizer = PatternRecognizer(
supported_entity=recognizer_conf['entity_name'],
name=recognizer_conf['name'],
@@ -74,8 +63,26 @@ try:
patterns=patterns,
context=recognizer_conf.get('context')
)
registry.add_recognizer(custom_recognizer)
logger.info(f"Loaded custom recognizer from YAML: {custom_recognizer.name}")
custom_recognizers[recognizer_conf['name']] = custom_recognizer
# Étape B: On parcourt la liste 'recognizer_registry' pour activer les détecteurs demandés
for recognizer_name in config.get("recognizer_registry", []):
# Cas 1: Le détecteur est dans notre liste de détecteurs personnalisés
if recognizer_name in custom_recognizers:
registry.add_recognizer(custom_recognizers[recognizer_name])
logger.info(f"Loaded custom recognizer from registry list: {recognizer_name}")
# Cas 2: Le détecteur est un détecteur prédéfini connu
elif recognizer_name in PREDEFINED_RECOGNIZERS_MAP:
recognizer_class = PREDEFINED_RECOGNIZERS_MAP[recognizer_name]
# On crée une instance pour chaque langue supportée (en, fr)
for lang in supported_languages:
# CORRECTION : On utilise le mot-clé au singulier 'supported_language'
instance = recognizer_class(supported_language=lang)
registry.add_recognizer(instance)
logger.info(f"Loaded predefined recognizer '{recognizer_name}' for languages: {supported_languages}")
else:
logger.warning(f"Recognizer '{recognizer_name}' from registry list was not found in custom or predefined lists.")
# === FIN DE LA CORRECTION MAJEURE ===
@@ -86,7 +93,6 @@ try:
registry=registry,
supported_languages=supported_languages
)
# L'allow list est chargée automatiquement par l'AnalyzerEngine
analyzer.set_allow_list(config.get("allow_list", []))
logger.info("--- Presidio Analyzer Service Ready ---")
@@ -95,6 +101,7 @@ except Exception as e:
logger.exception("FATAL: Error during AnalyzerEngine initialization.")
analyzer = None
# Le reste du fichier Flask reste identique...
@app.route('/analyze', methods=['POST'])
def analyze_text():
if not analyzer:
@@ -103,17 +110,13 @@ def analyze_text():
try:
data = request.get_json(force=True)
text_to_analyze = data.get("text", "")
# Utiliser la première langue supportée comme langue par défaut si non fournie
default_lang = analyzer.supported_languages[0] if analyzer.supported_languages else "en"
language = data.get("language", default_lang)
if not text_to_analyze:
return jsonify({"error": "text field is missing or empty"}), 400
results = analyzer.analyze(
text=text_to_analyze,
language=language
)
results = analyzer.analyze(text=text_to_analyze, language=language)
response_data = [res.to_dict() for res in results]
return make_response(jsonify(response_data), 200)