I am planning to build a chatbot that works on specific types of e-commerce data sets. To start with I have a large data set with product pricing, stock status, review count, promotions, etc. This information is scraped from e-commerce websites. The data is updated periodically in a Postgres database. I've already connected metabase with the database and built a model from the product data table.
I built a custom open ai model that generates SQL queries from natural language text for this type of data set. How can I pass the SQL to the metabase API and get the visualization from metabase as a public link inside my chatbot?
Questions:
- Is it possible on the open-source version?
- Any references or tips for the implementation?