SelfQueryRetriever pass #1

app.py

@@ -1,28 +1,55 @@
 # https://towardsai.net/p/machine-learning/deploying-a-langchain-large-language-model-llm-with-streamlit-pinecone?amp=1
 
 """Python file to serve as the frontend"""
+import os
+import pinecone
 import streamlit as st
 from streamlit_chat import message
-
-import pinecone
-import os
+from langchain.llms import OpenAI
 from langchain.embeddings.openai import OpenAIEmbeddings
 from langchain.vectorstores import Pinecone
-
 from langchain.chains import ConversationChain
-from langchain.llms import OpenAI
-
+from langchain.retrievers.self_query.base import SelfQueryRetriever
+from langchain.chains.query_constructor.base import AttributeInfo
+
+metadata_field_info=[
+    AttributeInfo(
+        name="author",
+        description="The author of the excerpt", 
+        type="string or list[string]", 
+    ),
+    AttributeInfo(
+        name="chapter_number",
+        description="The chapter number of excerpt", 
+        type="integer", 
+    ),
+    AttributeInfo(
+        name="chapter_name",
+        description="The chapter name of the excerpt", 
+        type="string", 
+    ),
+]
+
+document_content_description = "Excerpt's from Reid Hoffman's book Impromptu"
 embeddings = OpenAIEmbeddings()
 
 pinecone.init(
  api_key=str(os.environ['PINECONE_API_KEY']), 
  environment=str(os.environ['PINECONE_ENV']))
-
 index_name = str(os.environ['PINECONE_INDEX_NAME'])
 
 def load_chain():
- docsearch = Pinecone.from_existing_index(index_name, embeddings)
- return docsearch
+    docsearch = Pinecone.from_existing_index(index_name, embeddings)
+    retriever = SelfQueryRetriever.from_llm(
+         llm, 
+         vectorstore, 
+         document_content_description, 
+         metadata_field_info, 
+         verbose=True)
+    return retriever
+    
+ # docsearch = Pinecone.from_existing_index(index_name, embeddings)
+ # return docsearch
 
 # def load_chain():
 #     """Logic for loading the chain you want to use should go here."""
@@ -51,7 +78,7 @@ def get_text():
 user_input = get_text()
 
 if user_input:
-    docs = chain.similarity_search(user_input)
+    docs = chain.get_relevant_documents(user_input)
     output = docs[0].page_content
 
     st.session_state.past.append(user_input)