Update app.py

app.py

@@ -12,8 +12,8 @@ from langchain_core.prompts import ChatPromptTemplate
 import os
 from dotenv import load_dotenv
 from helper import SYSTEM_PROMPT
-# from langchain_google_genai import GoogleGenerativeAIEmbeddings
-from langchain.embeddings import HuggingFaceEmbeddings # open source free embedding
+from langchain_google_genai import GoogleGenerativeAIEmbeddings
+# from langchain.embeddings import HuggingFaceEmbeddings # open source free embedding
 load_dotenv()
 
 
@@ -37,17 +37,17 @@ class PDFQAProcessor:
             
     question_answer_chain = create_stuff_documents_chain(llm, prompt)
 
-    EMBEDDING_MODEL = "intfloat/e5-large-v2"
+    # EMBEDDING_MODEL = "intfloat/e5-large-v2"
     
-    embeddings = HuggingFaceEmbeddings(
-        model_name=EMBEDDING_MODEL,
-        model_kwargs={'device': 'cpu'},
-        encode_kwargs={'normalize_embeddings': True}
-    )
-
-    # embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
-    CHUNK_SIZE = 700
-    CHUNK_OVERLAP = 150
+    # embeddings = HuggingFaceEmbeddings(
+    #     model_name=EMBEDDING_MODEL,
+    #     model_kwargs={'device': 'cpu'},
+    #     encode_kwargs={'normalize_embeddings': True}
+    # )
+
+    embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
+    CHUNK_SIZE = 400
+    CHUNK_OVERLAP = 50
     text_splitter = RecursiveCharacterTextSplitter(chunk_size=CHUNK_SIZE,chunk_overlap = CHUNK_OVERLAP)
     # persist_directory="./chroma_db"
 
@@ -81,7 +81,7 @@ class PDFQAProcessor:
                 splits,
                 self.embeddings
             )
-            self.retriever = self.vectorstore.as_retriever(search_kwargs={"k": 10})
+            self.retriever = self.vectorstore.as_retriever(search_kwargs={"k": 18})
             return "PDFs processed successfully! Ask your questions now."
         
         except Exception as e: