v2.0.5

What's Changed

• update comparison operator by @ArslanSaleem in #953
• build: fix ci for windows by @gventuri in #1005
• fix: add number result to memory as string by @chengwaikoo in #997
• fix: remove logger refs from SmartDataframe by @PavelAgurov in #985
• chore(code_manager): filter code for redeclaration of pd.DataFrame fr… by @ArslanSaleem in #1003
• fix: import polars by @gventuri in #1009

New Contributors

• @chengwaikoo made their first contribution in #997

Full Changelog: v2.0.3...v2.0.5

v2.0.3

What's Changed

• fix(llm): Google Gemini signature and memory usage by @mspronesti in #980

Full Changelog: v2.0.2...v2.0.3

v2.0.2

What's Changed

• fix: restore support for Google Gemini API by @gventuri

Full Changelog: v2.0.1...v2.0.2

v2.0.1

What's Changed

• fix: update domain name by @ArslanSaleem in #970
• fix(agent): langchain LLM instantiation by @mspronesti in #977

Full Changelog: v2.0...v2.0.1

v2.0

PandasAI v2.0 is here! The Panda has never been so smart! 🚀

After months of dedicated effort and collaboration, we're super excited to unveil PandasAI v2.0! This release marks a significant leap forward, bringing the capabilities of our favorite panda to unprecedented levels.

What's New in v2.0?
✅ Training: you can now tailor PandasAI to your needs with custom instructions and processes, ensuring more deterministic responses that align perfectly with your company's workflows
✅ Field descriptions: eliminate ambiguity by providing clear descriptions for any table field, boosting accuracy
✅ Agent description: you can now define how the agent should act during the interactions with the user
✅ Improved multi-turn conversations: up to 70% more accurate context retrieval in multi-turn conversations
✅ Google BigQuery connector
✅ BambooLLM support (currently in beta, will be released soon)
✅ Jinja2 template engine: better template maintainability
✅ Refined feature set: based on community feedback, we’ve remove less utilized features, focusing on what truly matters to users

Full Changelog: v1.5.21...v2.0

v1.5.21

What's Changed

• fix: error for missing langchain_core package installing with pip

Full Changelog: v1.5.20...v1.5.21

v1.5.20

What's Changed

• chore: add Flask openai example by @aloha-fim in #941
• chore(llms): upgrade to langchain API >= 0.1 by @mspronesti in #957
• fix(llms): langchain response for chat models by @mspronesti in #960
• feat(pandasai): support pydantic v2 by @mspronesti in #963

New Contributors

• @aloha-fim made their first contribution in #941

Full Changelog: v1.5.19...v1.5.20

v1.5.19

What's Changed

• feat: add support for modin by @mspronesti in #907
• docs: update Google Colab Badge by @HemantSachdeva in #914
• docs: Rectify code examples by adding missing statements by @HemantSachdeva in #915

New Contributors

• @HemantSachdeva made their first contribution in #914

Full Changelog: v1.5.18...v1.5.19

v1.5.18

What's Changed

• docs: update examples.md by @PavelAgurov in #887
• refactor: TypeVar for IResponseParser (#889) by @nautics889 in #890
• fix: logging chart saving only if code contains chart by @Lorenzobattistela in #897
• feat: add df summarization shortcut by @dudesparsh in #901
• fix: badge for "Open in Colab" by @devashishmamgain in #903
• feat: add support Google Gemini API in LLMs by @dudesparsh in #902

New Contributors

• @PavelAgurov made their first contribution in #887
• @dudesparsh made their first contribution in #901
• @devashishmamgain made their first contribution in #903

Full Changelog: v1.5.17...v1.5.18

v1.5.17

What's Changed

• fix(code manager): parsing of called functions by @mspronesti in #883
• fix open charts return on format plot by @lhlong in #881
• feat(project): add Makefile, re-lint project, restructure tests by @mspronesti in #884

New Contributors

• @lhlong made their first contribution in #881

Full Changelog: v1.5.16...v1.5.17