Update app.py

app.py

@@ -1,15 +1,17 @@
+# app.py
 import io
 import uvicorn
 import numpy as np
 import cv2
+import re
 from fastapi import FastAPI, UploadFile, File, HTTPException
 from bson import ObjectId
 from motor.motor_asyncio import AsyncIOMotorClient, AsyncIOMotorGridFSBucket
 from rapidocr_onnxruntime import RapidOCR
 
-# ------------------------------
+# --------------------------------------------------
 # CONFIG
-# ------------------------------
+# --------------------------------------------------
 MONGO_URI = (
     "mongodb+srv://harilogicgo_db_user:jFhyDM4oA4dklUsp"
     "@api-logs.i7rqf9p.mongodb.net/ocr_fastapi"
@@ -17,33 +19,37 @@ MONGO_URI = (
 )
 
 DB_NAME = "ocr_fastapi"
-BUCKET_NAME = "ocr_images"
+BUCKET_NAME = "fs"  # <-- FIXED (use default GridFS bucket)
 
 app = FastAPI()
 
-# DB INIT
+# MongoDB Init
 _client = AsyncIOMotorClient(MONGO_URI)
 _db = _client[DB_NAME]
 
 def gridfs():
+    # Always use default "fs" bucket (fs.files + fs.chunks)
     return AsyncIOMotorGridFSBucket(_db, bucket_name=BUCKET_NAME)
 
 ocr_engine = RapidOCR()
 OCR_RESULTS = {}
 
-# ------------------------------
+# --------------------------------------------------
 @app.get("/health")
 async def health():
     return {"status": "ok"}
 
-# ------------------------------
+# --------------------------------------------------
 @app.post("/upload")
 async def upload_image(file: UploadFile = File(...)):
+    """
+    Uploads image to MongoDB GridFS
+    """
     try:
         data = await file.read()
-
         fs = gridfs()
 
+        # Upload to GridFS
         oid = await fs.upload_from_stream(
             file.filename,
             data,
@@ -52,24 +58,39 @@ async def upload_image(file: UploadFile = File(...)):
 
         print("📌 Uploaded ID:", oid)
 
+        # Debug: Confirm stored in fs.files
+        stored = await _db[f"{BUCKET_NAME}.files"].count_documents({"_id": oid})
+        print("📦 Stored in DB:", stored)
+
         return {"image_id": str(oid)}
 
     except Exception as e:
-        raise HTTPException(status_code=500, detail=str(e))
+        raise HTTPException(status_code=500, detail=f"Upload failed: {str(e)}")
 
-# ------------------------------
+# --------------------------------------------------
 async def load_image_from_gridfs(image_id: str):
+    """
+    Loads the image bytes from GridFS bucket
+    """
     try:
-        print("🔍 Fetching ID:", image_id)
+        print("🔍 Fetching from GridFS:", image_id)
 
         oid = ObjectId(image_id)
         fs = gridfs()
 
+        # Debug: check existence
+        file_exists = await _db[f"{BUCKET_NAME}.files"].count_documents({"_id": oid})
+        print("📂 Exists in fs.files:", file_exists)
+
+        if file_exists == 0:
+            raise HTTPException(status_code=404, detail="Image not found in GridFS")
+
+        # Read file
         stream = await fs.open_download_stream(oid)
         data = await stream.read()
         await stream.close()
 
-        print("✅ Found image in GridFS")
+        print("✅ Loaded image bytes from GridFS")
 
         return data
 
@@ -77,26 +98,32 @@ async def load_image_from_gridfs(image_id: str):
         print("❌ GridFS error:", e)
         raise HTTPException(status_code=404, detail="Image not found")
 
-
-# ------------------------------
+# --------------------------------------------------
 @app.post("/generate/{image_id}")
 async def generate(image_id: str):
+    """
+    Reads image → performs OCR → returns extracted text
+    """
     raw_bytes = await load_image_from_gridfs(image_id)
 
+    # Decode image
     img_array = np.frombuffer(raw_bytes, np.uint8)
     img = cv2.imdecode(img_array, cv2.IMREAD_COLOR)
 
     if img is None:
         raise HTTPException(status_code=400, detail="Unable to decode image")
 
+    # Perform OCR
     result, times = ocr_engine(img)
 
     if not result:
         raise HTTPException(status_code=500, detail="OCR returned empty result")
 
+    # Extract plain text
     extracted = [text for box, text, score in result]
     full_text = "\n".join(extracted)
 
+    # Cache in memory for view endpoint
     OCR_RESULTS[image_id] = {
         "text": full_text,
         "details": result,
@@ -105,13 +132,17 @@ async def generate(image_id: str):
 
     return {"image_id": image_id, "text": full_text}
 
-# ------------------------------
+# --------------------------------------------------
 @app.get("/view/{image_id}")
 async def view_details(image_id: str):
+    """
+    Returns cached OCR result
+    """
     if image_id not in OCR_RESULTS:
-        raise HTTPException(status_code=404, detail="No OCR result found")
+        raise HTTPException(status_code=404, detail="No OCR result found for this ID")
     return OCR_RESULTS[image_id]
 
 
+# --------------------------------------------------
 if __name__ == "__main__":
-    uvicorn.run("app:app", host="0.0.0.0", port=8000)
+    uvicorn.run("app:app", host="0.0.0.0", port=7860)