Update app.py

app.py

@@ -1,6 +1,6 @@
 from fastapi import FastAPI, HTTPException
 from fastapi.middleware.cors import CORSMiddleware
-from pydantic import BaseModel, validator
+from pydantic import BaseModel, field_validator, ValidationInfo
 from transformers import AutoTokenizer, AutoModelForSequenceClassification
 import torch
 import logging
@@ -32,23 +32,25 @@ app.add_middleware(
     allow_headers=["*"],
 )
 
-# ---------------- Models ----------------
+# ---------------- Pydantic Models ----------------
 class ScanRequest(BaseModel):
     text: str
-    # Accept both scan_type and scanType
     scan_type: Optional[str] = None
     scanType: Optional[str] = None
     userId: Optional[str] = None
     
-    @validator('scan_type', 'scanType', pre=True, always=True)
-    def determine_scan_type(cls, v, values, field):
-        if field.name == 'scanType' and v:
-            # Map scanType to scan_type for internal use
-            values['scan_type'] = v
+    @field_validator('scanType')
+    @classmethod
+    def map_scantype_to_scan_type(cls, v: Optional[str], info: ValidationInfo) -> Optional[str]:
+        """Mapper to ensure backward compatibility with old 'scanType' parameter name."""
+        if v is not None:
+            # Map the old 'scanType' field value to the new 'scan_type' field
+            info.data['scan_type'] = v
         return v
-    
+
     def get_scan_type(self) -> str:
-        """Get the scan type, defaulting to 'basic' if not provided"""
+        """Get the scan type, defaulting to 'basic' if not provided."""
+        # scan_type takes precedence as it's the canonical field name
         return self.scan_type or "basic"
 
 class ScanResponse(BaseModel):
@@ -58,7 +60,7 @@ class ScanResponse(BaseModel):
     credits: Optional[dict] = None
     test_mode: bool = False
 
-# ---------------- AI Detector ----------------
+# ---------------- AI Detector Core ----------------
 MODEL_NAME = "openai-community/roberta-large-openai-detector"
 
 class AIDetector:
@@ -93,7 +95,7 @@ class AIDetector:
         logger.info("Model loaded successfully.")
 
     def predict(self, text: str, max_length: int = 512) -> dict:
-        """Return both human and AI probabilities with debugging info"""
+        """Return both human and AI probabilities."""
         if self.model is None:
             self.load_model()
         
@@ -116,9 +118,7 @@ class AIDetector:
             ai_prob = float(probs[0][1].item())     # Class 1
             
             # Debug logging
-            logger.debug(f"Raw probabilities: {probs}")
-            logger.debug(f"Class 0 (Human): {human_prob:.4f}")
-            logger.debug(f"Class 1 (AI): {ai_prob:.4f}")
+            logger.debug(f"Class 0 (Human): {human_prob:.4f}, Class 1 (AI): {ai_prob:.4f}")
             
             # Verify probabilities sum to ~1.0
             total = human_prob + ai_prob
@@ -133,9 +133,9 @@ class AIDetector:
 
 detector = AIDetector()
 
-# ---------------- ChatGPT Pattern Detection ----------------
+# ---------------- Pattern Detection ----------------
 def detect_chatgpt_patterns(text: str) -> bool:
-    """Return True if ChatGPT patterns are detected"""
+    """Return True if ChatGPT patterns are detected."""
     patterns = [
         "as an ai language model",
         "i am an ai model",
@@ -237,10 +237,10 @@ def compute_overall_score(sections: List[dict], confidence_threshold: float = 0.
         "confident_sections": len(confident_sections)
     }
 
-# ---------------- Endpoints ----------------
+# ---------------- API Endpoints ----------------
 @app.on_event("startup")
 async def startup():
-    """Initialize the model on startup"""
+    """Initialize the model on startup."""
     logger.info("Starting Detextly AI Detector API...")
     try:
         detector.load_model()
@@ -257,22 +257,21 @@ async def root():
         "device": str(detector.device),
         "version": "2.1.0",
         "features": ["basic_scan", "highlight_scan", "chatgpt_pattern_detection"],
-        "endpoints": ["POST /api/scan", "GET /health", "GET /debug/test"]
+        "note": "Accepts both 'scan_type' and 'scanType' parameters"
     }
 
 @app.get("/health")
 async def health():
-    health_status = {
+    return {
         "status": "healthy",
         "model_loaded": detector.model is not None,
         "model": MODEL_NAME,
         "timestamp": time.time()
     }
-    return health_status
 
 @app.get("/debug/test")
 async def debug_test():
-    """Test endpoint to verify model is working correctly"""
+    """Test endpoint to verify model is working correctly."""
     test_texts = [
         "I went to the store yesterday to buy groceries.",
         "As an AI language model, I don't have personal experiences.",
@@ -299,7 +298,7 @@ async def debug_test():
 
 @app.post("/api/scan", response_model=ScanResponse)
 async def scan_text(request: ScanRequest):
-    """Main scanning endpoint"""
+    """Main scanning endpoint."""
     start_time = time.time()
     
     try:
@@ -307,7 +306,7 @@ async def scan_text(request: ScanRequest):
         if not request.text or len(request.text.strip()) < 10:
             raise HTTPException(status_code=400, detail="Text must be at least 10 characters long.")
         
-        # Get scan type (handles both scan_type and scanType)
+        # Get scan type (handles both scan_type and scanType via the validator)
         scan_type = request.get_scan_type()
         logger.info(f"Scan request: type={scan_type}, userId={request.userId}, text_length={len(request.text)}")
         
@@ -398,7 +397,7 @@ async def scan_text(request: ScanRequest):
 
 @app.get("/api/credits")
 async def get_credits(userId: Optional[str] = None):
-    """Get credits information (for compatibility with worker)"""
+    """Get credits information (for compatibility with worker)."""
     return {
         "basic": 5,
         "highlight": 1,