Merge remote-tracking branch 'origin/main'

.env.example

@@ -2,4 +2,4 @@ API_URI=
 API_KEY=
 API_TEAM=
 GAME_FILES_DIR=/project_absolute_path/game_files
-GROQ_API_KEY=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
+GOOGLE_API_KEY=

app.py

@@ -1,14 +1,12 @@
-from dto.requests import GameStartRequestDTO
+import logging
-from services.extractor import run_extraction_chain
-from services.julius_baer_api_client import JuliusBaerApiClient
-from validation.from_passport import FromPassport
 
-from services.player import Player
-from utils.parsers import process_passport
 from flask import Flask
 
-import logging
 import config
+from dto.requests import GameStartRequestDTO
+from services.extractor import extract_profile
+from services.julius_baer_api_client import JuliusBaerApiClient
+from services.player import Player
 
 app = Flask(__name__)
 logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - [%(module)s] - %(message)s')
@@ -18,37 +16,15 @@ logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - [%
 def hello_world():
     return 'Hello World!'
 
+
 if __name__ == '__main__':
     jb_client = JuliusBaerApiClient()
     game_start_request = GameStartRequestDTO(player_name=config.API_TEAM)
     res = jb_client.start_game(game_start_request)
 
-
+    result = extract_profile(res.client_data)
-    passport_data = res.client_data.get("passport")
-
-    prompt_template = (
-        "Extract the following information from the provided passport text.\n"
-        "Return only JSON matching this format:\n{format_instructions}\n\n"
-        "Pay special attention to the passport number and signature.\n"
-        "Passport text:\n{processed_text}"
-    )
-
-    result = run_extraction_chain(
-        raw_file_data=passport_data,
-        file_processor=process_passport,
-        pydantic_model=FromPassport,
-        prompt_template=prompt_template,
-    )
-
-    print(result)
 
     player = Player()
     player.play_on_separate_thread()
 
     app.run()
-
-    # res.session_id
-    # UUID('fde19363-a3d5-432e-8b87-54a6dd54f0dd')
-    # second test UUID('e3d58302-400a-4bc6-9772-ae50de43c9f4')
-    # UUID('f8b2a0a6-d4e0-45e6-900f-8ecb3c28f993')
-    # UUID('f8b2a0a6-d4e0-45e6-900f-8ecb3c28f993')

resources/decision_log2.csv

@@ -8,3 +8,21 @@ a482b28b-6f4c-4f2d-a167-42c071a92470,Accept
 2d7d0293-291c-46ec-91fb-75351be347f8,Accept
 3006006d-1060-4b53-afd3-b702a8fc2358,Accept
 f4a42e3e-75a8-43dc-92fe-e38fe23b1d82,Accept
+a5e06a84-2b05-47d1-8149-649d7a9e8bb6,Accept
+36adf081-fef6-4696-81da-e6ba73a6c8a0,Accept
+154b3c9d-a2e0-4d40-bb39-f3a249b26bc2,Accept
+71fdff4e-466a-40c8-a944-4958be13f974,Accept
+7b20c9c6-1bd6-4675-9e46-9b9829e50252,Accept
+1591ebcd-d0c2-44c7-b130-72f9a15c8a35,Accept
+10fd6524-a2a0-4ecf-953e-6c758f7147dc,Accept
+44418b3d-e2cd-4105-a599-5165c00c4971,Accept
+06dcf9d6-3ec4-451c-9bf4-75cb9c8061ee,Accept
+d1d0eb32-9f99-422c-a404-606b4d5c3a10,Accept
+3efa6b4e-8a4b-43d8-b570-0d02dd28b5ee,Accept
+42d0b5e4-03ab-4199-a74b-2f3fbe68680a,Accept
+1ad14242-15a9-4142-bc1f-c3fdd1165021,Accept
+e11b296a-5278-49de-a6b1-01aa31a508c4,Accept
+ad6b6980-7b2e-4af9-bda0-2ec004574211,Accept
+ac62e33d-6645-4360-8e14-21bfb0b6902a,Accept
+d5789f9c-a0f4-4663-9c6e-0d416bbbffb8,Accept
+25a429d6-bd8a-45d2-af1d-a0b8e5ec2e72,Accept

services/extractor.py

@@ -1,20 +1,65 @@
 import base64
 import binascii
-from typing import Callable, Type
+from typing import Callable, Type, Any, TypeVar
 from langchain_core.runnables import Runnable
 from langchain_core.prompts import ChatPromptTemplate
 from langchain_core.output_parsers import PydanticOutputParser
 from langchain_google_genai import ChatGoogleGenerativeAI
+from pydantic import BaseModel
+
+from utils.parsers import process_profile, process_passport
+from validation.from_passport import FromPassport
+from validation.from_profile import FromProfile
 
 
-def run_extraction_chain(
+def extract_passport(client_data: dict[str, Any]):
+    passport_data = client_data.get("passport")
+
+    prompt_template = (
+        "Extract the following information from the provided passport text.\n"
+        "Return only JSON matching this format:\n{format_instructions}\n\n"
+        "Pay special attention to the passport number\n"
+        "Passport text:\n{processed_text}"
+    )
+
+    result = __run_extraction_chain(
+        raw_file_data=passport_data,
+        file_processor=process_passport,
+        pydantic_model=FromPassport,
+        prompt_template=prompt_template,
+    )
+
+    return result
+
+
+def extract_profile(client_data: dict[str, Any]):
+    passport_data = client_data.get("profile")
+
+    prompt_template = (
+        "Extract the following information from the provided text.\n"
+        "Return only JSON matching this format:\n{format_instructions}\n\n"
+        "Pay special attention to the passport number and signature.\n"
+        "Passport text:\n{processed_text}"
+    )
+
+    result = __run_extraction_chain(
+        raw_file_data=passport_data,
+        file_processor=process_profile,
+        pydantic_model=FromProfile,
+        prompt_template=prompt_template,
+    )
+
+    return result
+
+ModelType = TypeVar("ModelType", bound=BaseModel)
+def __run_extraction_chain(
     *,
     raw_file_data: str,
     file_processor: Callable[[str], str],
-    pydantic_model: Type,
+    pydantic_model: type[ModelType],
     prompt_template: str,
     model_name: str = "gemini-2.0-flash"
-):
+) -> ModelType:
     """
     Traite un fichier encodé en base64, applique un parser OCR, génère un prompt, envoie à un modèle LLM, et retourne le résultat parsé.