https://github.com/basedosdados/chatbot - v0.6.2

Changes since v0.6.1: - Fixed a circular import issue by moving prompts into a dedicated prompts module.

- Python
Published by vrtornisiello 11 months ago

https://github.com/basedosdados/chatbot - v0.6.1

Changes since v0.6.0: - Metadata Formatters: Now returns full table IDs to avoid name collisions.

- Python
Published by vrtornisiello 11 months ago

https://github.com/basedosdados/chatbot - v0.6.1

Changes since v0.6.0: - Metadata Formatters: Now returns full table IDs to avoid name collisions. - SQL Agent: Agent now responds in the same language as the user’s question.

- Python
Published by vrtornisiello 11 months ago

https://github.com/basedosdados/chatbot - v0.6.0

Changes since v0.5.3:

• Router Agent: Now stores the data retrieved for each chat turn and passes the appropriate data and context to the VizAgent based on the conversation history.
• Visualization Agent: – Now stateless between runs, responsible solely for generating a Python script using pandas and plotly to visualize the retrieved data.
• Assistant Protocols: – Fixed issues in the Assistant and AsyncAssistant protocols.

- Python
Published by vrtornisiello 12 months ago

https://github.com/basedosdados/chatbot - v0.5.3

Changes since v0.5.2: - SQLAgent System Prompt Update: The new prompt introduces a clearer strategy for analyzing and correcting generated SQL queries that return no results, which often indicates an incorrect filter in a WHERE clause. This should help the agent make more reliable self-corrections.

- Python
Published by vrtornisiello 12 months ago

https://github.com/basedosdados/chatbot - v0.5.2

Changes since v0.5.1:

• Added response streaming capabilities to all agents and assistants.

- Python
Published by vrtornisiello about 1 year ago

https://github.com/basedosdados/chatbot - v0.5.2

Changes since v0.5.1: - Added response streaming capabilities to all agents and assistants.

- Python
Published by vrtornisiello about 1 year ago

https://github.com/basedosdados/chatbot - v0.5.1

Changes since v0.5.0: - Added a query rewriting step to the SQLAgent, enabling improved semantic search results when using vector stores. - Added the rewrite_query parameter to both the assistants and SQLAgent invocation methods. It defaults to False.

- Python
Published by vrtornisiello about 1 year ago

https://github.com/basedosdados/chatbot - v0.5.0

Changes since v0.4.5: - Simplified the Assistant api: - Removed log messages. - Removed error handling. - Replaced the Database protocol with the BaseContextProvider interface. - Users must implement their own context providers for supplying context to the SQLAgent. - Created the BasePromptFormatter interface. - Users must implement their own prompt formatters for custom prompt formatting. - Updated dataset selection system prompt. - Updated README file with documentation. - Created LICENSE file.

- Python
Published by vrtornisiello about 1 year ago

https://github.com/basedosdados/chatbot - v0.4.5

Changes since v0.4.4: - Removed the ModelFactory factory, along with the whole models module. - Refactored assistants to accept langchain chat model objects instead of model URIs.

- Python
Published by vrtornisiello about 1 year ago

https://github.com/basedosdados/chatbot - v0.4.4

Changes since v0.4.3: - Disabled logging by default when the package is imported as a library.

- Python
Published by vrtornisiello about 1 year ago

https://github.com/basedosdados/chatbot - v0.4.3

Changes since v0.4.2: - Added a decision node to the RouterAgent, placed immediately after invoking the SQLAgent, to determine whether to call the VizAgent or return the response directly to the user. - Fixed bugs in the Assistant's _format_response() method. - Updated dependency files.

- Python
Published by vrtornisiello about 1 year ago

https://github.com/basedosdados/chatbot - v0.4.2

Changes since v0.4.1:

Fixed BaseBigQueryTool - it now correctly accepts any object that implements the Database protocol. Removed ModelURI, so that any OpenAI or Google model can be passed to ModelFactory.

- Python
Published by vrtornisiello about 1 year ago

https://github.com/basedosdados/chatbot - v0.4.1

Changes since v0.4.0:

• Fixed a bug on the loguru_logging module.
• Constrained the ChartData structured output schema for better validation.
• Enhanced error handling when invoking the VizAgent.

- Python
Published by vrtornisiello about 1 year ago

https://github.com/basedosdados/chatbot - v0.4.0

Changes since v0.3.0: - Refactored SQLAssistant and SQLVizAssistant: - Reverted previous changes - checkpointers and vector stores must now be passed explicitly to the constructor. - Assistants can no longer be used as context managers. - Removed the storage module: - Users are now responsible for creating and passing vector stores directly to the assistants.

- Python
Published by vrtornisiello about 1 year ago

https://github.com/basedosdados/chatbot - v0.3.0

Changes since v0.2.1: - Refactored SQLAssistant and SQLVizAssistant: - Created a sync and an async version of each assistant. - The assistants can now be used as context managers. - If not being used as context managers, users must call the setup() and shutdown() methods. - Chroma vector store is now being created from an HTTP Chroma client.

- Python
Published by vrtornisiello over 1 year ago

https://github.com/basedosdados/chatbot - v0.2.1

Changes since v0.2.0: - Pinned grpcio package version to the latest stable release to avoid errors related to chromadb/opentelemetry/protobuf. Check issue https://github.com/basedosdados/chatbot/issues/6 for more information.

- Python
Published by vrtornisiello over 1 year ago

https://github.com/basedosdados/chatbot - v0.2.0

Changes since v0.1.0: - Replaced BigQueryAssistant with the new SQLVizAssistant, which is decoupled from BigQuery database. - Simplified assistant inputs and outputs. - Added a CI workflow. - Restructured the repo directories, to make it a python package.

- Python
Published by vrtornisiello over 1 year ago

https://github.com/basedosdados/chatbot - v0.1.0

First official release of the chatbot package, featuring LLM-powered agents for querying and visualizing databases.

- Python
Published by vrtornisiello over 1 year ago