Release v0.2.0

This release adds support for the input and output processing of the intrinsics in rag-intrinsics-lib.

We also include a number of bug fixes and API tweaks.

What's Changed

• Align chat completion format with vLLM by @frreiss in #16
• Add sanitization for Granite 3.2 and 3.3 by @kndtran in #13
• fix: Add init values for augmented hallucination object for error cases. by @kndtran in #25
• Updates to common IO processing for intrinsics by @frreiss in #23
• Add test of query rewrite by @frreiss in #27
• Align chat completion more closely with OpenAI APIs by @frreiss in #28
• Improve OpenAI compatibility by @frreiss in #29
• Update logit decoding to handle certainty model. by @frreiss in #31
• Add tests for context relevance by @frreiss in #32
• Update notebooks for manual testing by @frreiss in #37
• Add sentence splitting and stubs for sentence output processing by @frreiss in #39
• Sentence decoding for hallucinations by @frreiss in #40
• Updated processing for citations by @frreiss in #41
• Various fixes from testing by @frreiss in #43
• Updates for answer relevance intrinsics by @frreiss in #44
• Release prep by @frreiss in #45

New Contributors

• @kndtran made their first contribution in #13

Full Changelog: v0.1.0...v0.2.0