new endpoint

process.py

@@ -223,7 +223,7 @@ async def improve_classification(
         response = await asyncio.get_event_loop().run_in_executor(
             None,
             lambda: client.chat.completions.create(
-                model="gpt-4",
+                model="gpt-3.5-turbo",
                 messages=[{"role": "user", "content": prompt}],
                 temperature=0,
                 max_tokens=300,

server.py

@@ -11,6 +11,7 @@ import os
 from dotenv import load_dotenv
 import pandas as pd
 from utils import validate_results
+from process import improve_classification
 
 # Load environment variables
 load_dotenv()
@@ -88,6 +89,21 @@ class ValidationResponse(BaseModel):
     misclassifications: Optional[List[Dict[str, Any]]] = None
     suggested_improvements: Optional[List[str]] = None
 
+class ImprovementRequest(BaseModel):
+    df: Dict[str, Any]  # JSON representation of the DataFrame
+    validation_report: str
+    text_columns: List[str]
+    categories: str
+    classifier_type: str
+    show_explanations: bool
+    file_path: str
+
+class ImprovementResponse(BaseModel):
+    improved_df: Dict[str, Any]  # JSON representation of the improved DataFrame
+    new_validation_report: str
+    success: bool
+    updated_categories: List[str]
+
 @app.get("/health", response_model=HealthResponse)
 async def health_check() -> HealthResponse:
     """Check the health status of the API"""
@@ -208,6 +224,37 @@ async def validate_classifications(validation_request: ValidationRequest) -> Val
     except Exception as e:
         raise HTTPException(status_code=500, detail=str(e))
 
+@app.post("/improve-classification", response_model=ImprovementResponse)
+async def improve_classification_endpoint(request: ImprovementRequest) -> ImprovementResponse:
+    """Improve classification based on validation report"""
+    try:
+        # Convert JSON DataFrame back to pandas DataFrame
+        df = pd.DataFrame.from_dict(request.df)
+
+        # Call the improve_classification function
+        improved_df, new_validation, success, updated_categories = await improve_classification(
+            df=df,
+            validation_report=request.validation_report,
+            text_columns=request.text_columns,
+            categories=request.categories,
+            classifier_type=request.classifier_type,
+            show_explanations=request.show_explanations,
+            file=request.file_path
+        )
+
+        # Convert improved DataFrame to JSON
+        improved_df_json = improved_df.to_dict() if improved_df is not None else None
+
+        return ImprovementResponse(
+            improved_df=improved_df_json,
+            new_validation_report=new_validation,
+            success=success,
+            updated_categories=updated_categories
+        )
+
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))
+
 if __name__ == "__main__":
     import uvicorn
     uvicorn.run("server:app", host="0.0.0.0", port=8000, reload=True) 

test_server.py

@@ -1,6 +1,7 @@
 import requests
 import json
 from typing import List, Dict, Any, Optional
+import pandas as pd
 
 BASE_URL: str = "http://localhost:8000"
 
@@ -123,6 +124,50 @@ def test_validate_classifications() -> None:
     )
     print("\nValidation results:")
     print(json.dumps(response.json(), indent=2))
+    return response.json()  # Return validation results for use in improve test
+
+def test_improve_classification() -> None:
+    """Test the improve-classification endpoint"""
+    # First get validation results
+    validation_results = test_validate_classifications()
+    
+    # Load emails from CSV file
+    import csv
+    
+    emails: List[Dict[str, str]] = []
+    with open("examples/emails.csv", "r", encoding="utf-8") as file:
+        reader = csv.DictReader(file)
+        for row in reader:
+            emails.append(row)
+    
+    # Create a DataFrame with the first 5 emails
+    df = pd.DataFrame(emails[:5])
+    
+    # Get current categories
+    categories_response: requests.Response = requests.post(
+        f"{BASE_URL}/suggest-categories",
+        json=[email["contenu"] for email in emails[:5]]
+    )
+    response_data: Dict[str, Any] = categories_response.json()
+    current_categories: str = ",".join(response_data["categories"])
+    
+    # Send improvement request
+    improvement_request: Dict[str, Any] = {
+        "df": df.to_dict(),
+        "validation_report": validation_results["validation_report"],
+        "text_columns": ["contenu"],
+        "categories": current_categories,
+        "classifier_type": "gpt35",
+        "show_explanations": True,
+        "file_path": "examples/emails.csv"
+    }
+    
+    response: requests.Response = requests.post(
+        f"{BASE_URL}/improve-classification",
+        json=improvement_request
+    )
+    print("\nImprovement results:")
+    print(json.dumps(response.json(), indent=2))
 
 if __name__ == "__main__":
     print("Testing FastAPI server endpoints...")
@@ -131,4 +176,5 @@ if __name__ == "__main__":
     test_classify_text()
     test_classify_batch()
     test_suggest_categories()
-    test_validate_classifications() 
+    test_validate_classifications()
+    test_improve_classification()