Fix redundant media downloads & improve loader initialization performance

[JSabadin]

Purpose

Bug Fix

• Description:
  When a dataset created on one device was pushed to the cloud and then pulled on a new device, media files were always re-downloaded—even if they already existed in the local media folder. This didn’t corrupt any data, but it caused unnecessary loading delays.
• Solution:
  Before pulling from the cloud, verify whether each media file already exists in local_dir/<dataset_name>/media/ and skip downloads for files that are present.

Performance Improvement

Loader startup and cloud-pull operations are now much faster, reducing overall load times.

Specification

None / not applicable

Dependencies & Potential Impact

None / not applicable

Deployment Plan

None / not applicable

Testing & Validation

None / not applicable

[codecov]

Codecov Report

Attention: Patch coverage is 93.75000% with 2 lines in your changes missing coverage. Please review.

Project coverage is 95.54%. Comparing base (45ad4ed) to head (b0ce13b).
Report is 2 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	92.00%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff            @@
##           main     #325       +/-   ##
=========================================
+ Coverage      0   95.54%   +95.54%     
=========================================
  Files         0      102      +102     
  Lines         0     6067     +6067     
=========================================
+ Hits          0     5797     +5797     
- Misses        0      270      +270     

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