Skip to content

feat(data): add sampling and transform utilities#2

Merged
Goldokpa merged 2 commits into
mainfrom
feature/data-pipeline-improvements
Mar 23, 2026
Merged

feat(data): add sampling and transform utilities#2
Goldokpa merged 2 commits into
mainfrom
feature/data-pipeline-improvements

Conversation

@Goldokpa
Copy link
Copy Markdown
Member

Summary

  • Add data sampling utilities (stratified split, balanced sampling, k-fold CV)
  • Add custom satellite imagery transforms (augmentations, normalization, cloud masking)

Changes

  • src/climatevision/data/sampling.py — Sampling strategies for training data
  • src/climatevision/data/transforms.py — Custom transforms for satellite imagery

Test plan

  • Run unit tests for sampling functions
  • Verify transforms work with existing dataset
  • Test k-fold cross-validation split

🤖 Generated with Claude Code

Francis Umo added 2 commits March 23, 2026 19:19
Add sampling strategies for training data:
- Stratified train/val/test split
- Balanced class sampling
- Weighted sampler for imbalanced data
- K-fold cross-validation splits
- Random subset selection for debugging
Add augmentation transforms for satellite data:
- Geometric: flip, rotate, compose
- Photometric: brightness, contrast, noise
- Satellite-specific: cloud masking, normalization
- Pre-built training/validation transform pipelines
@Goldokpa Goldokpa merged commit 391ab1e into main Mar 23, 2026
Goldokpa added a commit that referenced this pull request Mar 23, 2026
feat(data): add sampling and transform utilities

- Add stratified split, balanced sampling, k-fold CV
- Add custom satellite imagery transforms
- Add cloud masking and normalization utilities
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant