use optional dependencies

[thorinaboenke]

Context
enable the use of optional dependencies, corresponding issue in the Service:
ait-detectmate/DetectMateService#179

Currently all dependencies are imported whether you need it or not. This should change. We can use optional.dependencies in pyproject.toml to make "features or extras" that can be installed if needed.

https://pydevtools.com/handbook/explanation/what-are-optional-dependencies-and-dependency-groups/

Alternatives

Can you achieve the same result doing it in an alternative way? Is the alternative considerable?

Has the feature been requested before?

Please provide a link to the issue.

If the feature request is approved, would you be willing to submit a PR?

Yes / No (Help can be provided if you need assistance submitting a PR)

[viktorbeck98]

Good idea. In fact, these dependencies are only necessary for the LLM-parser:

"openai>=2.26.0",
"tenacity>=9.1.4",
"scipy>=1.17.1",
"scikit-learn>=1.8.0",
"tiktoken>=0.12.0",
"pandas>=2.3.2",

And these are necessary for the Dataframe persistency backends which are just nice helper classes for someone experimenting with log data (not production):

"pandas>=2.3.2",
"polars>=1.40.1",

We could define groups like "prod" (default), and "dev" or maybe define setup procedures for specific features, when they are activated but not sure how that works.

[thorinaboenke]

I would not mix the optional dependencies up with dev and prod labels, since it's more like features that users can opt into
Otherwise updated the dependency groups into llm and dataframes @viktorbeck98