Update server.py

server.py

@@ -3,6 +3,7 @@ import shutil
 import zipfile
 import logging
 import urllib.parse
+import unicodedata
 from fastapi import FastAPI, File, UploadFile, Form, HTTPException
 from fastapi.middleware.cors import CORSMiddleware
 from fastapi.staticfiles import StaticFiles
@@ -25,14 +26,25 @@ DATASET_REPO = "aniketkumar1106/orbit-data"
 IMAGE_DIR = "Productimages"
 DB_TARGET = "orbiitt.db" 
 
+def normalize_filename(name):
+    """
+    Cleans filenames by removing smart quotes, normalizing unicode, 
+    and stripping spaces to ensure the URL matches the file on disk.
+    """
+    # 1. Replace Unicode smart quotes/dashes with standard ASCII
+    name = name.replace('’', "'").replace('‘', "'").replace('“', '"').replace('”', '"').replace('–', '-')
+    # 2. Convert to ASCII (removes accents/unsupported symbols)
+    name = unicodedata.normalize('NFKD', name).encode('ascii', 'ignore').decode('ascii')
+    # 3. Final cleanup
+    return name.strip()
+
 def setup_environment():
     token = os.environ.get("HF_TOKEN")
-    # Clean old data to prevent version mismatch
     if os.path.exists(IMAGE_DIR): shutil.rmtree(IMAGE_DIR)
     os.makedirs(IMAGE_DIR)
 
     try:
-        logger.info(f"Downloading dataset from {DATASET_REPO}...")
+        logger.info("Syncing Dataset from Hugging Face...")
         snapshot_download(repo_id=DATASET_REPO, repo_type="dataset", token=token, local_dir=".")
 
         if os.path.exists("orbiitt_db.zip"):
@@ -42,26 +54,19 @@ def setup_environment():
             for root, _, files in os.walk("temp_extract"):
                 for f in files:
                     src = os.path.join(root, f)
-                    # 1. DB LOCATOR: Handles nested or renamed DB files
                     if f.lower() in ["orbiitt_db", "orbiitt.db"] or (f.endswith(".db") and not f.startswith(".")):
                         shutil.copy(src, f"./{DB_TARGET}")
-                    # 2. IMAGE NORMALIZER: Fixes encoding and moves to static folder
                     elif f.lower().endswith(('.png', '.jpg', '.jpeg', '.webp')):
-                        clean_name = urllib.parse.unquote(f)
+                        # Normalize the file name before saving it to disk
+                        clean_name = normalize_filename(urllib.parse.unquote(f))
                         dest = os.path.join(IMAGE_DIR, clean_name)
                         shutil.copy(src, dest)
             shutil.rmtree("temp_extract")
-        
-        # Root fallback for database
-        if os.path.exists("orbiitt_db") and not os.path.exists(DB_TARGET):
-            shutil.move("orbiitt_db", DB_TARGET)
-
-        logger.info(f"READY: Images={len(os.listdir(IMAGE_DIR))} DB={os.path.exists(DB_TARGET)}")
+        logger.info(f"Environment Ready: {len(os.listdir(IMAGE_DIR))} images stored.")
     except Exception as e:
-        logger.error(f"FATAL SETUP ERROR: {e}")
+        logger.error(f"Setup Error: {e}")
 
 setup_environment()
-
 app.mount("/Productimages", StaticFiles(directory=IMAGE_DIR), name="Productimages")
 
 engine = None
@@ -69,18 +74,12 @@ try:
     from orbiitt_engine import OrbiittEngine
     if os.path.exists(DB_TARGET):
         engine = OrbiittEngine()
-        logger.info("SIGLIP ENGINE ONLINE")
 except Exception as e:
-    logger.error(f"ENGINE ERROR: {e}")
-
-@app.get("/")
-def health():
-    imgs = os.listdir(IMAGE_DIR) if os.path.exists(IMAGE_DIR) else []
-    return {"engine": engine is not None, "db": os.path.exists(DB_TARGET), "img_count": len(imgs)}
+    logger.error(f"AI Engine Failed to Start: {e}")
 
 @app.post("/search")
 async def search(text: str = Form(None), weight: float = Form(0.5), file: UploadFile = File(None)):
-    if not engine: raise HTTPException(503, detail="AI Engine Loading")
+    if not engine: raise HTTPException(503, detail="AI Engine Offline")
     t_path = None
     try:
         if file:
@@ -89,15 +88,29 @@ async def search(text: str = Form(None), weight: float = Form(0.5), file: Upload
         
         results = engine.search(text_query=text, image_file=t_path, text_weight=weight)
         
-        # 3. PATH SANITIZER: Converts internal paths to public browser URLs
+        all_files = os.listdir(IMAGE_DIR)
+        valid_results = []
+
         for r in results:
             raw_path = r.get('url') or r.get('path') or ""
-            filename = os.path.basename(raw_path)
-            r['url'] = f"Productimages/{urllib.parse.quote(filename)}"
+            # Normalize the database name to match the cleaned names on disk
+            fname = normalize_filename(os.path.basename(raw_path))
+            
+            if fname in all_files:
+                r['url'] = f"Productimages/{urllib.parse.quote(fname)}"
+                valid_results.append(r)
+            else:
+                # Fallback: check for substring matches (handles truncation errors)
+                for disk_file in all_files:
+                    if fname[:30] in disk_file: # Compare first 30 chars
+                        r['url'] = f"Productimages/{urllib.parse.quote(disk_file)}"
+                        valid_results.append(r)
+                        break
                 
-        return {"results": results}
+        return {"results": valid_results}
     except Exception as e:
-        return {"results": [], "error": str(e)}
+        logger.error(f"Search Process Error: {e}")
+        return {"results": []}
     finally:
         if t_path and os.path.exists(t_path): os.remove(t_path)