Create app.py

app.py

@@ -0,0 +1,58 @@
+!pip install langchain
+!pip install langchain pypdf
+!pip install openai==0.28
+!pip install chromadb
+!pip install tiktoken
+
+import json
+import openai
+import numpy as np
+import getpass
+import os
+
+from langchain.llms import OpenAI
+from langchain.vectorstores import Chroma
+from langchain_community.document_loaders import TextLoader
+from langchain.text_splitter import RecursiveCharacterTextSplitter, CharacterTextSplitter
+from langchain.embeddings.openai import OpenAIEmbeddings
+from langchain.chains import ConversationalRetrievalChain
+
+documents = []
+loader = TextLoader("sentences.txt")
+documents.extend(loader.load())
+
+text_splitter = RecursiveCharacterTextSplitter(
+    chunk_size = 1000,
+    chunk_overlap = 150
+)
+
+# Recursive Splitting the whole text of emails into chunks
+splits = text_splitter.split_documents(documents)
+
+# Creating the Embeddings from the splits we created
+embedding = OpenAIEmbeddings(openai_api_key='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')
+
+# Storing the Embeddings into ChromaDB
+persist_directory = 'docs/chroma/'
+vectordb = Chroma.from_documents(
+    documents=splits[0:500],
+    embedding=embedding,
+    persist_directory=persist_directory
+)
+
+retriever = vectordb.as_retriever(search_type="similarity", search_kwargs={"k":2})
+qa = ConversationalRetrievalChain.from_llm(OpenAI(temperature=0), retriever)
+
+def respond(message, history):
+    chat_history = []
+    print(message)
+    print(chat_history)
+    # Getting the response from QA langchain 
+    response = qa({"question": message, "chat_history": chat_history})
+
+    # Append user messages and responses to chat history
+    chat_history.append((message, response['answer']))
+    print(chat_history)
+    return response['answer']
+
+gr.ChatInterface(respond).launch(debug=True)