Fixes for nested labels

[kozlov721]

• Changed return type of BaseDataset.get_tasks to a dictionary mapping task names to task types (before it was just a list of combined task names)
• Changed return type of BaseDataset.get_classes from a tuple of all classes and a dictionary mapping task names to classes to just the later. The list of all classes doesn't make semantic sense with the new tasks, and also it was rarely even used anywhere
• Removed the created_at timestamp from parquet files - wasn't used anywhere
• Fixed task_name and task_type values for annotations with nested sub-detections
  • Names of the sub-detections are now part of the task_name instead of task_type
  • {"car/driver": "boundingbox"} instead of {"car": "driver/boundingbox"}
• Added additional overloads to methods loading parquet files
• Simplified some tests

[codecov]

Codecov Report

Attention: Patch coverage is 96.90722% with 3 lines in your changes missing coverage. Please review.

Project coverage is 95.86%. Comparing base (2fbec83) to head (37fed81).
Report is 1 commits behind head on main.

✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
luxonis_ml/data/datasets/luxonis_dataset.py	88.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #224      +/-   ##
==========================================
- Coverage   95.93%   95.86%   -0.07%     
==========================================
  Files          88       89       +1     
  Lines        4744     4715      -29     
==========================================
- Hits         4551     4520      -31     
- Misses        193      195       +2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[klemen1999]

LGTM, left one comment regarding tests