New Freebie: Local-RAG

Author: ctavolazzi

freebies/local-rag/README.md

@@ -12,25 +12,27 @@ This is a Retrieval-Augmented Generation (RAG) chatbot application using Langcha
 ## Setup
 
 1. Clone this repository:
-   ```
-   git clone https://github.com/yourusername/rag-chatbot.git
-   cd rag-chatbot
-   ```
+```
+git clone https://github.com/yourusername/rag-chatbot.git
+cd rag-chatbot
+```
 
 2. Run the setup script:
-   ```
-   python setup.py
-   ```
-   This will download necessary files, install dependencies, and set up Ollama.
+
+```python
+python setup.py
+```
+This will download necessary files, install dependencies, and set up Ollama.
 
 3. Activate the virtual environment:
-   - On Windows: `.\venv\Scripts\activate`
-   - On macOS/Linux: `source venv/bin/activate`
+- On Windows: `.\venv\Scripts\activate`
+- On macOS/Linux: `source venv/bin/activate`
 
 4. Run the application:
-   ```
-   python app.py
-   ```
+
+```python
+python app.py
+```
 
 5. Open your web browser and go to `http://localhost:7860` to interact with the chatbot.
 
@@ -59,14 +61,11 @@ vector_store = get_vector_store()
 add_documents_to_store(vector_store, new_chunks)
 ```
 
-## Troubleshooting
-
+### Troubleshooting
 If you encounter any issues, please check the application logs or file an issue on the GitHub repository.
 
-## Contributing
-
+### Contributing
 Contributions are welcome! Please feel free to submit a Pull Request.
 
-## License
-
-This project is licensed under the MIT License.
+### License
+This project is licensed under the GPL-3 license.

freebies/local-rag/app.py

@@ -1,13 +1,15 @@
-"""
-Main application script for the RAG Chatbot.
-"""
-
 import gradio as gr
 import logging
-from chatbot import chat_ollama
-from data_loader import get_document_chunks
-from vector_store import get_vector_store, VectorStoreError
+from scrapegraphai.graphs import SmartScraperGraph
+from langchain.document_loaders import UnstructuredMarkdownLoader
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain.vectorstores import Chroma
+from langchain.embeddings import OllamaEmbeddings
+from langchain.schema import StrOutputParser
+from langchain.schema.runnable import RunnablePassthrough
 from config import CONFIG
+from vector_store import VectorStoreError, get_vector_store
+from chatbot import chat_ollama
 
 # Set up logging
 logging.basicConfig(
@@ -16,6 +18,80 @@
 )
 logger = logging.getLogger(__name__)
 
+def scrape_with_scrapegraphai(url):
+    """Scrape content using ScrapeGraphAI library."""
+    graph_config = {
+        "llm": {
+            "model": CONFIG['OLLAMA_MODEL'],
+            "temperature": 0,
+            "format": "json",  
+            "base_url": CONFIG['OLLAMA_URL'],  
+        },
+        "embeddings": {
+            "model": "ollama/nomic-embed-text",
+            "base_url": CONFIG['OLLAMA_URL'],  
+        },
+        "verbose": True,
+    }
+    smart_scraper_graph = SmartScraperGraph(
+        prompt="Extract all the text content",
+        source=url,
+        config=graph_config
+    )
+    result = smart_scraper_graph.run()
+    markdown_path = "scraped_content.md"
+    with open(markdown_path, "w", encoding="utf-8") as file:
+        for item in result['content']:
+            file.write(item + "\n")
+    
+    return markdown_path
+
+def ingest_markdown(markdown_path):
+    """Ingest the markdown content into the vector store."""
+    loader = UnstructuredMarkdownLoader(markdown_path)
+    docs = loader.load()
+
+    chunk_size = CONFIG['CHUNK_SIZE']
+    chunk_overlap = CONFIG['CHUNK_OVERLAP']
+    text_splitter = RecursiveCharacterTextSplitter(chunk_size=chunk_size, chunk_overlap=chunk_overlap)
+    splits = text_splitter.split_documents(docs)
+
+    embeddings = OllamaEmbeddings(base_url=CONFIG['OLLAMA_URL'], model=CONFIG['OLLAMA_MODEL'])
+    vectorstore = Chroma.from_documents(documents=splits, embedding=embeddings, persist_directory=CONFIG['CHROMA_PATH'])
+    retriever = vectorstore.as_retriever()
+
+    prompt_template = "Your local prompt template here"
+
+    def format_docs(docs):
+        return "\n\n".join(doc.page_content for doc in docs)
+
+    rag_chain = (
+        {"context": retriever | format_docs, "question": RunnablePassthrough()}
+        | prompt_template
+        | chat_ollama  # Using local Ollama model for processing
+        | StrOutputParser()
+    )
+    
+    return vectorstore, rag_chain
+
+def ingest_url(url):
+    """Ingest content from the URL into the vector store."""
+    try:
+        markdown_path = scrape_with_scrapegraphai(url)
+        vectorstore, rag_chain = ingest_markdown(markdown_path)
+        return "Content ingested successfully!", vectorstore, rag_chain
+    except Exception as e:
+        logger.error(f"Error ingesting content from {url}: {str(e)}")
+        return f"Failed to ingest content from {url}: {str(e)}", None, None
+
+def query_vectorstore(query, rag_chain):
+    """Query the vector store with a given query."""
+    if rag_chain:
+        response = rag_chain.invoke(query)
+        return response
+    else:
+        return "Vector store not initialized."
+
 def main():
     """Initialize and run the RAG Chatbot application."""
     try:
@@ -27,20 +103,30 @@ def main():
             logger.info("Loading existing vector database...")
             get_vector_store()
 
-        gradio_interface = gr.ChatInterface(
-            chat_ollama,
-            chatbot=gr.Chatbot(),
-            textbox=gr.Textbox(placeholder="Example: Who is Alice?", container=False, scale=7),
-            title="The Ollama RAG Chatbot",
-            description=f"Ask the {CONFIG['OLLAMA_MODEL']} chatbot a question!",
-            theme='gradio/base',
-            retry_btn=None,
-            undo_btn="Delete Previous",
-            clear_btn="Clear",
-        )
+        with gr.Blocks() as gradio_interface:
+            gr.Markdown("## The Ollama RAG Chatbot")
+            with gr.Row():
+                with gr.Column():
+                    url_input = gr.Textbox(label="Enter URL to Ingest", placeholder="Example: https://example.com/article")
+                    ingest_button = gr.Button("Ingest URL")
+                    ingestion_status = gr.Textbox(label="Ingestion Status", interactive=False)
+                    chatbot = gr.Chatbot()
+                    user_input = gr.Textbox(placeholder="Example: Who is Alice?", container=False, scale=7)
+                    send_button = gr.Button("Send")
+                    
+                    def ingest_callback(url):
+                        status, vectorstore, rag_chain = ingest_url(url)
+                        return status, vectorstore, rag_chain
+                    
+                    def query_callback(user_input, rag_chain):
+                        response = query_vectorstore(user_input, rag_chain)
+                        return response
+                    
+                    ingest_button.click(ingest_callback, inputs=url_input, outputs=[ingestion_status, chatbot, None])
+                    send_button.click(query_callback, inputs=user_input, outputs=chatbot)
 
         logger.info("Starting Gradio interface...")
-        gradio_interface.launch()
+        gradio_interface.launch(server_name="0.0.0.0", server_port=7860, inbrowser=True)
     except VectorStoreError as e:
         logger.error(f"Vector store error: {str(e)}")
         print(f"An error occurred with the vector store: {str(e)}")
@@ -49,4 +135,4 @@ def main():
         print(f"An unexpected error occurred: {str(e)}")
 
 if __name__ == "__main__":
-    main()
+    main()

freebies/local-rag/chatbot.py

@@ -75,4 +75,4 @@ def chat_ollama(message, history):
         return "I'm sorry, but I'm having trouble generating a response right now."
     except Exception as e:
         logger.error(f"Unexpected error in chat_ollama: {str(e)}")
-        return "I apologize, but I'm experiencing an unexpected issue. Please try again later."
+        return "I apologize, but I'm experiencing an unexpected issue. Please try again later."

freebies/local-rag/config.py

@@ -33,4 +33,4 @@
 
     # Application settings
     'DEBUG': os.getenv('DEBUG', 'False').lower() == 'true',
-}
+}

freebies/local-rag/data_loader.py

@@ -63,4 +63,4 @@ def add_new_document(file_path):
         return chunks
     except Exception as e:
         logger.error(f"Error adding new document {file_path}: {str(e)}")
-        raise
+        raise

freebies/local-rag/llm.py

@@ -21,4 +21,4 @@ def get_ollama():
         return ollama
     except Exception as e:
         logger.error(f"Error initializing Ollama: {str(e)}")
-        raise OllamaError(f"Failed to initialize Ollama: {str(e)}")
+        raise OllamaError(f"Failed to initialize Ollama: {str(e)}")

freebies/local-rag/setup.py

@@ -1,14 +1,14 @@
 """
 Setup script for the RAG Chatbot application.
-This script downloads necessary files from the NovaSystem repository and sets up the application.
+This script downloads necessary files from the specified repository and sets up the application.
 """
 
 import os
 import subprocess
 import requests
 
-REPO_OWNER = 'ctavolazzi'
-REPO_NAME = 'NovaSystem'
+REPO_OWNER = 'yourusername'
+REPO_NAME = 'rag-chatbot'
 BRANCH = 'main'
 FOLDER_PATH = 'freebies/local-rag'
 FILES_TO_DOWNLOAD = [
@@ -18,7 +18,8 @@
     'vector_store.py',
     'llm.py',
     'chatbot.py',
-    'requirements.txt'
+    'requirements.txt',
+    'README.md'
 ]
 
 def download_file(file_name):
@@ -83,4 +84,4 @@ def main():
     print("2. Run the application: python app.py")
 
 if __name__ == "__main__":
-    main()
+    main()

freebies/local-rag/vector_store.py

@@ -57,4 +57,4 @@ def add_documents_to_store(vector_store, new_chunks):
         logger.info(f"Added {len(new_chunks)} new chunks to the vector store")
     except Exception as e:
         logger.error(f"Error adding new documents to vector store: {str(e)}")
-        raise VectorStoreError(f"Failed to add new documents to vector store: {str(e)}")
+        raise VectorStoreError(f"Failed to add new documents to vector store: {str(e)}")