Auto Update Feluda and Operator package version numbers #440
Labels
enhancement
New feature or request
level:feature
priority:high
python
Pull requests that update Python code
There seems to be some confusion on the requirements of what auto update means, so I thought i'll clarify the intent and give some examples.
With the new refactor we have a lot of packages to manage. The main one being the feluda package and then for every operator there would be one package. Soon we will have > 10 operators and they all have individual version numbers.
What we should have is that anytime a new release is created, the appropriate .toml files should be updated with new version numbers. The versioning uses semantic versioning which follow the syntax MAJOR.MINOR.PATCH
One proposed way to identify which number to increment is to use conventional commits. All commit messages in feluda follow convetions from conventional commits. we use the prefixes like "feat", "chore", "docs", "fix" to denote the type of change. Those usually map to the kind of increment to do on version number. For example feat is usually a new feature and can be a MINOR version whereas fix would be a PATCH. One of the convention is to add the scope of your commit in parentheses, in which case your prefix becomes "feat(image_vec_rep_resnet)", this indicates that a new feature has been added to the image_vec_rep_resnet operator and hence only its version should be updated.
This is a suggestion, there could be other approaches as well.
As part of this issue, I'd appreciate if first we could collect examples of projects like ours and see how they are managing versions. After comparing multiple approaches, we can pick one.
The text was updated successfully, but these errors were encountered: