init without binary files

Dockerfile

@@ -0,0 +1,11 @@
+FROM python:3.11
+RUN useradd -m -u 1000 user
+USER user
+ENV HOME=/home/user \
+    PATH=/home/user/.local/bin:$PATH
+WORKDIR $HOME/app
+COPY --chown=user . $HOME/app
+COPY ./requirements.txt ~/app/requirements.txt
+RUN pip install -r requirements.txt
+COPY . .
+CMD ["chainlit", "run", "app.py", "--port", "7860"]

README.md

@@ -1,10 +1,63 @@
 ---
-title: Chroma Qa Chat
-emoji: 🏢
-colorFrom: green
-colorTo: green
-sdk: docker
-pinned: false
+title: 'Chroma Q&A with Sources Element'
+tags: ['chroma', 'chainlit', 'qa']
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# Chroma Q&A with Sources Element
+
+This repository contains a Chainlit application that provides a question-answering service using documents stored in a Chroma vector store. It allows users to upload PDF documents, which are then chunked, embedded, and indexed for efficient retrieval. When a user asks a question, the application retrieves relevant document chunks and uses OpenAI's language model to generate an answer, citing the sources it used.
+
+## High-Level Description
+
+The `app.py` script performs the following functions:
+
+1. **PDF Processing (`process_pdfs`)**: Chunks PDF files into smaller text segments, creates embeddings for each chunk, and stores them in Chroma.
+2. **Document Indexing (`index`)**: Uses `SQLRecordManager` to track document writes into the vector store.
+3. **Question Answering (`on_message`)**: When a user asks a question, the application retrieves relevant document chunks and generates an answer using OpenAI's language model, providing the sources for transparency.
+
+## Quickstart
+
+### Prerequisites
+
+- Python 3.11 or higher
+- Chainlit installed
+- PDF documents to be indexed
+
+### Setup and Run
+
+1. **Install Dependencies:**
+
+Install the required Python packages specified in `requirements.txt`.
+
+```shell
+pip install -r requirements.txt
+```
+
+2. **Process PDFs:**
+
+Place your PDF documents in the `./pdfs` directory.
+
+3. **Run the Application:**
+
+Use the provided `Dockerfile` to build and run the application.
+
+```shell
+docker build -t chroma-qa-chat .
+docker run -p 7860:7860 chroma-qa-chat
+```
+
+   Access the application at `http://localhost:7860`.
+
+## Code Definitions
+
+- `process_pdfs`: Function that processes PDF files and indexes them into Chroma.
+- `on_chat_start`: Event handler that sets up the Chainlit session with the necessary components for question answering.
+- `on_message`: Event handler that processes user messages, retrieves relevant information, and sends back an answer.
+- `PostMessageHandler`: Callback handler that posts the sources of the retrieved documents as a Chainlit element.
+
+![Screenshot](./screenshot.png)
+
+## See Also
+
+For a visual guide on how to use this application, watch the video by [Chris Alexiuk](https://www.youtube.com/watch?v=9SBUStfCtmk&ab_channel=ChrisAlexiuk).
+

app.py

@@ -0,0 +1,124 @@
+from typing import List
+from pathlib import Path
+from langchain_openai import ChatOpenAI, OpenAIEmbeddings
+from langchain.prompts import ChatPromptTemplate
+from langchain.schema import StrOutputParser
+from langchain_community.document_loaders import (
+    PyMuPDFLoader,
+)
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain.vectorstores.chroma import Chroma
+from langchain.indexes import SQLRecordManager, index
+from langchain.schema import Document
+from langchain.schema.runnable import Runnable, RunnablePassthrough, RunnableConfig
+from langchain.callbacks.base import BaseCallbackHandler
+
+import chainlit as cl
+
+chunk_size = 1024
+chunk_overlap = 50
+
+embeddings_model = OpenAIEmbeddings()
+
+PDF_STORAGE_PATH = "./pdfs"
+
+
+def process_pdfs(pdf_storage_path: str):
+    pdf_directory = Path(pdf_storage_path)
+    docs = []  # type: List[Document]
+    text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=100)
+
+    for pdf_path in pdf_directory.glob("*.pdf"):
+        loader = PyMuPDFLoader(str(pdf_path))
+        documents = loader.load()
+        docs += text_splitter.split_documents(documents)
+
+    doc_search = Chroma.from_documents(docs, embeddings_model)
+
+    namespace = "chromadb/my_documents"
+    record_manager = SQLRecordManager(
+        namespace, db_url="sqlite:///record_manager_cache.sql"
+    )
+    record_manager.create_schema()
+
+    index_result = index(
+        docs,
+        record_manager,
+        doc_search,
+        cleanup="incremental",
+        source_id_key="source",
+    )
+
+    print(f"Indexing stats: {index_result}")
+
+    return doc_search
+
+
+doc_search = process_pdfs(PDF_STORAGE_PATH)
+model = ChatOpenAI(model_name="gpt-4", streaming=True)
+
+
+@cl.on_chat_start
+async def on_chat_start():
+    template = """Answer the question based only on the following context:
+
+    {context}
+
+    Question: {question}
+    """
+    prompt = ChatPromptTemplate.from_template(template)
+
+    def format_docs(docs):
+        return "\n\n".join([d.page_content for d in docs])
+
+    retriever = doc_search.as_retriever()
+
+    runnable = (
+        {"context": retriever | format_docs, "question": RunnablePassthrough()}
+        | prompt
+        | model
+        | StrOutputParser()
+    )
+
+    cl.user_session.set("runnable", runnable)
+
+
+@cl.on_message
+async def on_message(message: cl.Message):
+    runnable = cl.user_session.get("runnable")  # type: Runnable
+    msg = cl.Message(content="")
+
+    class PostMessageHandler(BaseCallbackHandler):
+        """
+        Callback handler for handling the retriever and LLM processes.
+        Used to post the sources of the retrieved documents as a Chainlit element.
+        """
+
+        def __init__(self, msg: cl.Message):
+            BaseCallbackHandler.__init__(self)
+            self.msg = msg
+            self.sources = set()  # To store unique pairs
+
+        def on_retriever_end(self, documents, *, run_id, parent_run_id, **kwargs):
+            for d in documents:
+                source_page_pair = (d.metadata['source'], d.metadata['page'])
+                self.sources.add(source_page_pair)  # Add unique pairs to the set
+
+        def on_llm_end(self, response, *, run_id, parent_run_id, **kwargs):
+            if len(self.sources):
+                sources_text = "\n".join([f"{source}#page={page}" for source, page in self.sources])
+                self.msg.elements.append(
+                    cl.Text(name="Sources", content=sources_text, display="inline")
+                )
+
+    async with cl.Step(type="run", name="QA Assistant"):
+        async for chunk in runnable.astream(
+            message.content,
+            config=RunnableConfig(callbacks=[
+                cl.LangchainCallbackHandler(),
+                PostMessageHandler(msg)
+            ]),
+        ):
+            await msg.stream_token(chunk)
+
+    await msg.send()

chainlit.md

@@ -0,0 +1,14 @@
+# Welcome to Chainlit! 🚀🤖
+
+Hi there, Developer! 👋 We're excited to have you on board. Chainlit is a powerful tool designed to help you prototype, debug and share applications built on top of LLMs.
+
+## Useful Links 🔗
+
+- **Documentation:** Get started with our comprehensive [Chainlit Documentation](https://docs.chainlit.io) 📚
+- **Discord Community:** Join our friendly [Chainlit Discord](https://discord.gg/k73SQ3FyUh) to ask questions, share your projects, and connect with other developers! 💬
+
+We can't wait to see what you create with Chainlit! Happy coding! 💻😊
+
+## Welcome screen
+
+To modify the welcome screen, edit the `chainlit.md` file at the root of your project. If you do not want a welcome screen, just leave this file empty.

requirements.txt

@@ -0,0 +1,8 @@
+langchain
+langchain-community
+chainlit
+langchain_openai
+openai
+chromadb
+tiktoken
+pymupdf