deepset-ai / haystack

:mag: AI orchestration framework to build customizable, production-ready LLM applications. Connect components (models, vector DBs, file converters) to pipelines or agents that can interact with your data. With advanced retrieval methods, it's best suited for building RAG, question answering, semantic search or conversational agent chatbots.
https://haystack.deepset.ai
Apache License 2.0
16.93k stars 1.85k forks source link

ci: Slightly rework docs release process #8363

Closed silvanocerza closed 1 week ago

silvanocerza commented 2 weeks ago

Fixes #8331.

This changes some stuff about the documentation release process.

The release_docs.py used to create a new unstable version from the existing one and then promote the existing one to stable. If we were releasing 2.9.x we would create 2.10-unstable on Readme based on 2.9-unstable and change 2.9-unstable to 2.9.

That file has been split into create_unstable_docs.py and promote_unstable_docs.py.

create_unstable_docs.py only creates a new unstable docs from the existing unstable.

promote_unstable_docs.py changes an unstable version to stable.

The common parts have been moved to readme_api.py.


The minor_version_release.yml workflow has been changed to just create the unstable docs for the version being released calling create_unstable_docs.py.

The new promote_unstable_docs.yml workflow runs whenever the first patch release of a new version is being released and promote that version unstable to stable. If I'm releasing 2.9.0 it will change the docs on Readme from 2.9-unstable to 2.9. If we're releasing 2.9.1 it won't run.