v0.2.0

What's Changed

• feat: Update training to accomodate antiSMASH and BGCFlow schema by @matinnuhamunada in #8

Full Changelog: v0.1.4...v0.2.0

v0.1.4

What's Changed

• feat: added streamlit interface by @matinnuhamunada in #7
• fix: update table schema

Full Changelog: v0.1.3...v0.1.4

v0.1.3

What's Changed

• dev 0.1.3 by @matinnuhamunada in #5

• Further improvements on training dataset, including antiSMASH glossary and abbreviations

Full Changelog: v0.1.2...v0.1.3

v0.1.2

What's Changed

• Feature improvement by @matinnuhamunada in #4

• Change default LLMs - llama3 for Ollama and gpt-4o for OpenAI
• Improve training: add question-sql pair dataset

Full Changelog: v0.1.1...v0.1.2

v0.1.1

What's Changed

• 0.1.0 - First attempt at creating chatBGC by @matinnuhamunada in #1
• feat: query duckdb dataset from BGCFlow using Ollama and ChromaDB
• feat: add option to use OpenAI API key to access latest model such as GPT-4
• feat: add training step by @matinnuhamunada in #2
• feat: add dataset to train - DDL, antismash documentation, BGCFlow table description, and question-sql pair

New Contributors

• @matinnuhamunada made their first contribution in #1

Full Changelog: https://github.com/NBChub/chatBGC/commits/v0.1.1