[MRG] Handle torch tensor in DomainAwareDataset

[tom-yneuro]

Adding support for flexible data types (numpy arrays and torch tensors) and a new return type (DeepDADataset) inside DomainAwareDataset.

Enhanced pack Method Functionality:

• Implemented logic in the pack method to handle data concatenation and return the appropriate format based on the return_type. Added checks for torch availability and raised appropriate errors if required dependencies are missing.
• Added support for returning data as a DeepDADataset object, which is compatible with PyTorch's Dataset interface and includes domain-specific information.

Testing Enhancements:

• Added a new test in test_deep_dadataset.py to validate the functionality of the pack method with different return_type values ("array", "tensor", "DeepDADataset"). Ensures that the output matches the expected types and formats.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 96.50%. Comparing base (53a494b) to head (e71a980).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #337      +/-   ##
==========================================
+ Coverage   96.18%   96.50%   +0.32%     
==========================================
  Files          63       51      -12     
  Lines        7020     6068     -952     
==========================================
- Hits         6752     5856     -896     
+ Misses        268      212      -56     

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[tom-yneuro]

I noticed that the behavior of return_X_y isn't obvious when set to False (i.e it returns a Bunch object or a DomainAwareDataset depending on the file)
Here I proposed to change this argument to return_type, in my case allowing several types
Maybe this approach can be generalized to datasets fetch with return_X_y argument
(Subquestion in my case : should I remove the return_X_y argument or deprecate it, for now it's deprecated)

[tgnassou]

LGTM!