Releases: nicofretti/DataGenFlow
Releases · nicofretti/DataGenFlow
release-v1.2.0
What's Changed
- Model/Embed UI: added the ability for users to create, set up, and test various models and embeddings directly in the application.
- New 'Seeders' block: introduced a new block type for directly loading and chunking content from Markdown files.
- Q&A generation template: uncluded a new template for direct generation of Q&A pairs from a Markdown file.
- Dockerized application: the application is now fully runnable on Docker for simplified setup and deployment.
- Improved User Experience:
- Pipeline status and error alignment
- Validation of the seeds before running a pipeline
- Models can be selected directly in the blocks instead of write them everytime, them are taken from the config
- New official home: the GitHub Pages site has been replaced by the new official website: datagenflow.com (a big contribution by @kevinagyeman)
- Bug fixes and stability
Config of models/embeddings avaiable also in the blocks
Improvement of the generation phase
Issues and PR:
- 🧩 Fix: mobile/tablet responsive by @kevinagyeman in #22
- 🧩 Fix: add change log + website menu page + cleanup + fix pytest by @nicofretti in #25
- 🧩 fix: swagger api reference not visible by @giuseppeambrosio97 in #21
- 🚀 Feat: docker by @giuseppeambrosio97 in #28
- 🚀 Feat: improve the feedbacks with toast instead of banners by @NicolaLC in #29
- 🚀 Feat: seeders blocks + launch validation by @nicofretti in #27
- 🚀 Feat: support multiple models + add configuration UI by @nicofretti in #32
- 📐 Refactor: best practice file + dismiss website + agent code helper by @nicofretti in #36
- 🧩 Fix: bug provieders in llm/embeddings models by @nicofretti and @kevinagyeman in #41
New Contributors
- @kevinagyeman made their first contribution in #22
- @giuseppeambrosio97 made their first contribution in #21
- @NicolaLC made their first contribution in #29
Full Changelog: release-v1.1.0...release-v1.2.0
v1.1.0
What's Changed
- Integrated LiteLLM for robust multi-provider stability, structured output, and intelligent fallbacks.
- Improved the Pipeline Editor with enhanced block visualization and new auto-layout.
- Introduced three new Pipeline Templates (JSON Generation, Text Classification, Q&A Generation).
- Refactored UI blocks to support Jinja2 templating for dynamic prompts.
- Launched the new website with feature guides and documentation.
- Resolved various bugs and stability issues.
New Contributors
- @Ujjansh05 made their first contribution in #13
Full Changelog: release...release-v1.1.0
v1.0.0
Base working version of DataGenFlow
Full Changelog: https://github.com/nicofretti/DataGenFlow/commits/release



