vanna Releases

Jan 17, 2024

Improved SQL parsing from code blocks and fixed transaction handling in Postgres.

Jan 17, 2024

Added Flask app integration to notebooks and updated documentation.

Jan 17, 2024

This release introduces an experimental integrated Flask application, allowing users to launch a UI within Jupyter notebooks or Colab environments without needing to clone a repository.

Jan 16, 2024

This update adds Marqo as a new vector storage option and includes necessary formatting improvements for notebooks.

Jan 8, 2024

This release adds support for the Mistral LLM, expanding the framework's compatibility with different AI models.

Dec 22, 2023

This update focuses on documentation and usability, featuring new starter notebooks, OpenAI API refinements, and the ability to use the Vanna-hosted vector database with any LLM.

Dec 17, 2023

This release contains a bug fix to ensure the correct message is displayed by correcting the string usage from 'Question' to 'question'.

Nov 20, 2023

This release introduces dynamic URL handling for SQLite connections and adds schema support to Snowflake connections.

Oct 30, 2023

This release includes improvements to the visualization capabilities by updating the Plotly figures used in the vn.ask method.

Sep 28, 2023

This update refactors the codebase to fully support GPT-4 and ensures function synchronization between local and remote environments.

Sep 22, 2023

This release resolves a connectivity issue that prevented users from connecting to BigQuery when using Google Colab.

Sep 13, 2023

This release adds more configuration options for the engine and warehouse settings to provide greater flexibility.

Aug 31, 2023

This release focuses on repository restructuring, introduces local context storage using ChromaDB, and refactors the DDL parameter to SQL. It also includes various import fixes and formatting improvements.

Aug 31, 2023

An experimental feature was introduced to generate generic training plans automatically using the database information schema.

Aug 4, 2023

Aug 4, 2023

The library now supports Postgres and BigQuery connectors, along with specific training capabilities for Postgres databases.

Aug 3, 2023

This update improves documentation, adds validation for model types, and introduces optional dependencies and tox for better development workflows.

Aug 2, 2023

A 'training plan' feature was implemented to support the training workflow functionality.

Jul 29, 2023

Documentation was updated to rename datasets to 'models' for clarity, and the training process was made generic to handle various use cases.

Jul 28, 2023