Update main.py

main.py

@@ -26,9 +26,20 @@ app = FastAPI(title="E-Bikes Semantic Search API",
               description="API for finding similar e-bikes based on semantic search",
               version="1.0.0")
 
+def build_filter(pt: Optional[str], cat: Optional[str]) -> dict | None:
+    filt = {}
+    if pt:
+        filt["product_type"] = pt          # shorthand $eq
+    if cat:
+        filt["category"] = cat
+    return filt or None
+
 # Request and response models
 class SearchRequest(BaseModel):
     description: str
+    top_k: int = 3
+    product_type: Optional[str] = None   # "ebike" or "escooter"
+    category: Optional[str] = None       # e.g. "mountain"
 
 class BikeMatch(BaseModel):
     id: str
@@ -96,7 +107,7 @@ def load_ebikes_data(file_path="data.json"):
     try:
         with open(file_path, 'r') as f:
             data = json.load(f)
-        return data.get('ebikes', [])
+        return data.get('pogo-cycles-data', [])
     except Exception as e:
         print(f"Error loading e-bikes data: {e}")
         return []
@@ -114,8 +125,9 @@ def create_and_upload_embeddings(ebikes_data, encoder, pinecone_index):
         metadata.append({
             'id': bike['id'],
             'name': bike['name'],
-            'type': bike['type'],
-            'description': bike['description']
+            'type': bike['product_type'],
+            'description': bike['description'],
+            'category': bike['category']
         })
     
     # Create embeddings
@@ -168,7 +180,7 @@ async def health_check():
     return {"status": "healthy"}
 
 @app.post("/search", response_model=SearchResponse)
-async def search_ebikes(description: str):
+async def search_ebikes(description: str,filters:dict):
     """
     Search for e-bikes similar to the provided description
     
@@ -177,12 +189,14 @@ async def search_ebikes(description: str):
     try:
         # Create embedding for the query
         query_embedding = encoder.encode(description)[0]
-        
+        filter_payload = build_filter(filters.get("product_type"), filters.get("category"))
+
         # Query Pinecone
         results = pinecone_index.query(
             vector=query_embedding.tolist(),
             top_k=3,
-            include_metadata=True
+            include_metadata=True,
+            filter=filter_payload
         )
         
         # Parse results