Changelog

All notable changes to the bigdata-research-tools package will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[0.15.1] - 2025-06-12

Changed

• Fix bug in labeler logic which affected Colab

[0.15.0] - 2025-06-11

Added

• Logic to add a LLM generated motivation explaining why a company exposed to a theme
• Logic to construct a portfolio of a certain size, with balancing (by sector/region/other) and weighting (by thematic score/other score) capabilities
• Logic to build queries with custom batching and non-company entities
• Workflow usage metrics

[0.14.0] - 2025-05-30

Changed

• Chunk numbers now match the Chunk index from the Vector DB
• Add dependencies like bigdata logo and IPython to run 2 workflows seamlessly
• Cleanup of Miners and Screeners logic
• Update Miners docs to use newly refactored code
• Remove any tech debt/unused code from the library

[0.13.0] - 2025-04-25

Added

• Add prompt which better integrates analyst focus into mindmapper

Changed

• Updated narrative miners cookbook to use another source (as we no longer have CNBC)
• Fix excel logo
• Update cookbooks section of documentation to explain why to look at the cookbook(s)
• Change ordering of docs
• Fix the labelling prompt to be the same as in thematic screener notebook
• Set the default temperature for all LLM calls to 0

[0.12.0] - 2025-03-24

Added

• Add screener class for the Narrative Executive Factor:

  • screeners.ExecutiveNarrativeFactor

• Add labeler for company screening:

  • labeler.screener_labeler

• Add screener search:

  • search.screener_search

• Add a themes module with logic to generate sub-themes from a main one:

  • themes.generate_theme_tree
  • themes.ThemeTree class.

• Resources folder moved inside the package source data.

• API Reference updated:

  • Screeners
  • Search
  • Themes

[0.11.0] - 2025-02-27

Added

• Add narrative miner classes to track narratives in transcipts, filings, news (miners folder)
• Add capability to choose which LLM to run prompts with (llm folder)
• Add supporting functionality for hybrid searches and labelling (labeler and search folders)
• Add logic to export the structured dataset of labelled chunks to excel workbook (excel.py)
• Cookbook Miners
• API Reference API Reference