Update app/services/embedding_service.py

app/services/embedding_service.py

@@ -1,6 +1,6 @@
-from typing import List
 import logging
-from openai import OpenAI
+from typing import List
+import google.generativeai as genai
 from app.config import settings
 
 logger = logging.getLogger(__name__)
@@ -8,61 +8,49 @@ logger = logging.getLogger(__name__)
 class EmbeddingService:
     def __init__(self):
         try:
-            self.client = OpenAI(api_key=settings.OPENAI_API_KEY)
-            self.model_name = settings.EMBEDDING_MODEL or "llama-text-embed-v2"
-            self.dimension = int(settings.PINECONE_DIMENSION)
-            logger.info(f"🔹 Using embedding model: {self.model_name}")
+            genai.configure(api_key=settings.GEMINI_API_KEY)
+            self.model_name = "models/embedding-001"  # Gemini text embedding model
+            logger.info(f"🔹 Using Gemini embedding model: {self.model_name}")
         except Exception as e:
-            logger.error(f"Error initializing embedding service: {e}")
+            logger.error(f"Error initializing Gemini embedding service: {e}")
             raise
 
     async def embed_text(self, text: str) -> List[float]:
-        """Generate embeddings for a single text using the Llama model."""
+        """Generate text embeddings using Gemini API"""
         try:
-            response = self.client.embeddings.create(
+            response = genai.embed_content(
                 model=self.model_name,
-                input=text
+                content=text
             )
-            embedding = response.data[0].embedding
+            embedding = response["embedding"]
 
-            # Ensure correct dimensionality
-            if len(embedding) < self.dimension:
-                embedding += [0.0] * (self.dimension - len(embedding))
-            elif len(embedding) > self.dimension:
-                embedding = embedding[:self.dimension]
+            # Ensure vector dimension matches Pinecone index (1024)
+            if len(embedding) < settings.PINECONE_DIMENSION:
+                padding = [0.0] * (settings.PINECONE_DIMENSION - len(embedding))
+                embedding.extend(padding)
+            elif len(embedding) > settings.PINECONE_DIMENSION:
+                embedding = embedding[:settings.PINECONE_DIMENSION]
 
             return embedding
         except Exception as e:
-            logger.error(f"Error generating embedding: {e}")
+            logger.error(f"Error generating Gemini embedding: {e}")
             raise
 
     async def embed_batch(self, texts: List[str]) -> List[List[float]]:
-        """Generate embeddings for a batch of texts."""
+        """Generate batch embeddings"""
         try:
-            response = self.client.embeddings.create(
-                model=self.model_name,
-                input=texts
-            )
-            embeddings = [d.embedding for d in response.data]
-
-            # Pad/truncate each embedding
-            fixed_embeddings = []
-            for emb in embeddings:
-                if len(emb) < self.dimension:
-                    emb += [0.0] * (self.dimension - len(emb))
-                elif len(emb) > self.dimension:
-                    emb = emb[:self.dimension]
-                fixed_embeddings.append(emb)
-
-            return fixed_embeddings
+            embeddings = []
+            for text in texts:
+                response = genai.embed_content(
+                    model=self.model_name,
+                    content=text
+                )
+                embeddings.append(response["embedding"])
+            return embeddings
         except Exception as e:
             logger.error(f"Error generating batch embeddings: {e}")
             raise
 
-    async def encode_product(self, product) -> List[float]:
-        """Combine product info for embedding."""
-        text = f"{product.title or ''} {product.brand or ''} {product.material or ''} {product.color or ''} {' '.join(product.categories) if product.categories else ''}"
-        return await self.embed_text(text)
 
 # Global instance
 embedding_service = EmbeddingService()