A chatbot to talk to AWS Blogs.
git clone https://github.com/praveenc/bedrock-blog-buddycd bedrock-blog-buddy
pip install -r requirements.txtcd app
# Ensure you're logged into aws using CLI.
streamlit run BlogBuddy.pyOnce the app is launched, select Embedding Model, Text Generation Model (llm) and AWS Region and click save.
Navigate to Refresh Data page and click on Refresh Feed, this seeds your local DB.
You can now click on Chat to start your session.
We use LanceDB as our vector store. Vectors and metadata are stored locally.
To avoid scraping RSS feeds multiple times, we cache scraped html data to disk and log the scraping activity to DuckDB locally.
Blog posts are indexed from the below AWS RSS feeds.