Create app.py

app.py

@@ -0,0 +1,95 @@
+# Import required modules from 'langchain' for document processing, embeddings, Q&A, etc.
+from langchain.document_loaders import PyPDFLoader
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain.vectorstores import Chroma
+from langchain.embeddings import OpenAIEmbeddings
+from langchain.chat_models import ChatOpenAI
+from langchain.chains import RetrievalQA
+
+# Importing Streamlit for creating the web app, and other necessary modules for file handling.
+import streamlit as st
+import tempfile
+import os
+
+# Import a handler for streaming outputs.
+from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
+
+# Set the title of the Streamlit web application.
+st.title("ChatPDF")
+# Create a horizontal line for better visual separation in the app.
+st.write("---")
+
+# Provide an input box for users to enter their OpenAI API key.
+openai_key = st.text_input('Enter OPEN_AI_API_KEY', type="password")
+
+# Provide a file upload widget to let users upload their PDF files.
+uploaded_file = st.file_uploader("Upload your PDF file!", type=['pdf'])
+# Another visual separation after the file uploader.
+st.write("---")
+
+# Define a function that converts the uploaded PDF into a document format.
+def pdf_to_document(uploaded_file):
+    # Create a temporary directory to store the uploaded PDF file temporarily.
+    temp_dir = tempfile.TemporaryDirectory()
+    # Join the directory path with the uploaded file name to get the complete path.
+    temp_filepath = os.path.join(temp_dir.name, uploaded_file.name)
+    
+    # Write the content of the uploaded file into the temporary file path.
+    with open(temp_filepath, "wb") as f:
+        f.write(uploaded_file.getvalue())
+    
+    # Use PyPDFLoader to read and split the PDF into individual pages.
+    loader = PyPDFLoader(temp_filepath)
+    pages = loader.load_and_split()
+    return pages
+
+# Check if a file has been uploaded by the user.
+if uploaded_file is not None:
+    # Convert the uploaded PDF into a document format.
+    pages = pdf_to_document(uploaded_file)
+
+    # Initialize a text splitter to break the document into smaller chunks.
+    text_splitter = RecursiveCharacterTextSplitter(
+        # Define parameters for the splitter: chunk size, overlap, etc.
+        chunk_size = 300,
+        chunk_overlap  = 20,
+        length_function = len
+    )
+    # Split the document pages into chunks.
+    texts = text_splitter.split_documents(pages)
+
+    # Initialize the OpenAIEmbeddings model for creating embeddings from texts using the provided API key.
+    embeddings_model = OpenAIEmbeddings(openai_api_key=openai_key)
+
+    # Load the textual chunks into Chroma after creating embeddings.
+    db = Chroma.from_documents(texts, embeddings_model)
+
+    # Define a custom handler to stream outputs to the Streamlit app.
+    from langchain.callbacks.base import BaseCallbackHandler
+    class StreamHandler(BaseCallbackHandler):
+        def __init__(self, container, initial_text=""):
+            self.container = container
+            self.text=initial_text
+        def on_llm_new_token(self, token: str, **kwargs) -> None:
+            self.text+=token
+            self.container.markdown(self.text)
+
+    # Display a header for the question section of the web app.
+    st.header("Ask the PDF a question!")
+    # Provide an input box for users to type in their questions.
+    question = st.text_input('Type your question')
+
+    # Check if the user has clicked on the 'Ask' button.
+    if st.button('Ask'):
+        # Show a spinner animation while processing the user's question.
+        with st.spinner('Processing...'):
+            # Create a space to display the answer.
+            chat_box = st.empty()
+            # Initialize a handler to stream outputs.
+            stream_hander = StreamHandler(chat_box)
+            # Initialize the ChatOpenAI model for Q&A tasks with streaming enabled.
+            llm = ChatOpenAI(model_name="gpt-3.5-turbo", temperature=0, openai_api_key=openai_key, streaming=True, callbacks=[stream_hander])
+            # Create a RetrievalQA chain that uses the ChatOpenAI model and Chroma retriever to answer the question.
+            qa_chain = RetrievalQA.from_chain_type(llm, retriever=db.as_retriever())
+            # Fetch the answer to the user's question.
+            qa_chain({"query": question})