Tabular datamodule (Custom dataset from DataFrame, CSV, or Parquet)

[manuelkonrad]

📝 Description

Hi, I submitted a similar PR about six months ago (PR #2403). It has not been reviewed yet, and in the meantime, Anomalib has gone from v1 to v2. Therefore, I decided to refactor the feature according to the new structure and submit it as a new PR.

• This PR adds the Tabular datamodule which is instantiated directly from a pandas DataFrame. It is an alternative to the Folder datamodule for custom datasets where the labels are not encoded in the directory structure. Useful for situations where labels are refined regularly or for sub-sampling large datasets without copying or moving files.
• The datamodule also includes a from_file constructor which loads the data from a tabular file supported by pandas.

✨ Changes

Select what type of change your PR is:

• 🐞 Bug fix (non-breaking change which fixes an issue)
• 🔨 Refactor (non-breaking change which refactors the code base)
• 🚀 New feature (non-breaking change which adds functionality)
• 💥 Breaking change (fix or feature that would cause existing functionality to not work as expected)
• 📚 Documentation update
• 🔒 Security update

✅ Checklist

Before you submit your pull request, please make sure you have completed the following steps:

• 📋 I have summarized my changes in the CHANGELOG and followed the guidelines for my type of change (skip for minor changes, documentation updates, and test enhancements).
• 📚 I have made the necessary updates to the documentation (if applicable).
• 🧪 I have written tests that support my changes and prove that my fix is effective or my feature works (if applicable).

For more information about code review checklists, see the Code Review Checklist.

[samet-akcay]

Thanks for creating this updated PR, and for your patience. We've been a bit side tracked by some other tasks.

@rajeshgangireddy and @ashwinvaidya17 can you prioritise this PR review please?

[rajeshgangireddy]

Hi @manuelkonrad
Great PR. Thank you.
I am yet to try it out, but so far it LGTM.

If it's not too much, could you also add a simple example notebook under examples/notebooks/100_datamodules

[rajeshgangireddy]

Minor comments.

[manuelkonrad]

Hi @rajeshgangireddy, thanks a lot for your helpful review!

I implemented the proposed changes and also added an example notebook.

[rajeshgangireddy]

Hi @manuelkonrad ,
It looks like our semgrep check is failing (unrelated to this PR).
I will first need to fix that and then merge your PR.

[samet-akcay]

Great PR! Thanks a lot for your efforts, @manuelkonrad!

I only have a single ask to have an example in the docstring to show from_file() method. Otherwise, looking great!

[manuelkonrad]

Great PR! Thanks a lot for your efforts, @manuelkonrad!

I only have a single ask to have an example in the docstring to show from_file() method. Otherwise, looking great!

Thanks for the feedback, @samet-akcay! I added the example.