[DRAFT] Add RandomSizedBBoxSafeCrop augmentation for bbox-safe training crops.

[bw4sz]

Crop a random window that contains all boxes, with configurable context around detections, then resize to a fixed size. Wire bbox context through box and mask datasets, document usage, and add an example figure.

Description

Related Issue(s)

AI-Assisted Development

• I used AI tools (e.g., GitHub Copilot, ChatGPT, etc.) in developing this PR
• I understand all the code I'm submitting
• I have reviewed and validated all AI-generated code

AI tools used (if applicable):

[codecov]

Codecov Report

❌ Patch coverage is 75.60976% with 30 lines in your changes missing coverage. Please review.
✅ Project coverage is 86.33%. Comparing base (d2851ec) to head (3f61b6e).
⚠️ Report is 20 commits behind head on main.

Files with missing lines	Patch %	Lines
src/deepforest/augmentations.py	74.57%	30 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1420      +/-   ##
==========================================
- Coverage   86.61%   86.33%   -0.28%     
==========================================
  Files          26       28       +2     
  Lines        3736     4187     +451     
==========================================
+ Hits         3236     3615     +379     
- Misses        500      572      +72     

Flag	Coverage Δ
unittests	86.33% <75.60%> (-0.28%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

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

🚀 New features to boost your workflow:

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