Create app.py

app.py

@@ -0,0 +1,57 @@
+import streamlit as st
+import os
+from langchain.chains import RetrievalQA, ConversationalRetrievalChain
+from langchain.chat_models import ChatOpenAI
+from langchain.document_loaders import TextLoader
+from langchain.text_splitter import CharacterTextSplitter
+from langchain.embeddings import OpenAIEmbeddings
+from langchain.vectorstores import FAISS
+from langchain.memory import ConversationBufferMemory
+import re
+
+# Initialize the Streamlit app
+st.title('Document-Based Q&A System')
+
+# API Key input securely
+api_key = st.text_input("Enter your OpenAI API key:", type="password")
+if api_key:
+    os.environ["OPENAI_API_KEY"] = api_key
+    st.success("API Key has been set!")
+
+# File uploader
+uploaded_file = st.file_uploader("Upload your document", type=['txt'])
+if uploaded_file is not None:
+    # Read and process the document
+    text_data = uploaded_file.getvalue().decode("utf-8")
+    text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
+    data = text_splitter.split_documents(text_data)
+
+    # Create vector store
+    embeddings = OpenAIEmbeddings()
+    vectorstore = FAISS.from_documents(data, embedding=embeddings)
+
+    # Create conversation chain
+    llm = ChatOpenAI(temperature=0.3, model_name="gpt-4-turbo")
+    memory = ConversationBufferMemory(
+        memory_key='chat_history', return_messages=True, output_key='answer')
+    conversation_chain = ConversationalRetrievalChain.from_llm(
+        llm=llm,
+        chain_type="stuff",
+        retriever=vectorstore.as_retriever(),
+        memory=memory,
+        return_source_documents=True
+    )
+
+    # Question input
+    query = st.text_input("Ask a question about the document:")
+    if query:
+        result = conversation_chain({"question": query})
+        answer = result["answer"]
+        st.write("Answer:", answer)
+
+        # Optionally display source text snippets
+        if st.checkbox("Show source text snippets"):
+            st.write("Source documents:")
+            for i in result["source_documents"]:
+                res = re.search(r'^[^\n]*', i.page_content)
+                st.write(i.page_content[res.span()[0]:res.span()[1]])