Progress Report for Week 4 (June 23 – June 29)
1. What did I get done this week?
- Implemented STAC to GeoCroissant converter in Python with validation and full metadata preservation.
- Handled
references
,item_assets
, anddistribution
fields for improved dataset interoperability. - Introduced custom fields like
geocr:visualizations
andgeocr:summaries
to retain extra STAC context. - Ensured compliance with Croissant and GeoCroissant specs including checksum, encoding format, and record structure.
- Developed a visual map overlay using Folium for AGB dataset with raster, GPKG tile boundaries, and training data points.
- Validated and exported the final
croissant.json
with support for visualization and external STAC item linking.
2. Plan for Next Week (June 30 – July 6):
- Generate example conversions from Hugging Face datasets into valid GeoCroissant format.
- Integrate Hugging Face dataset download links (
hf.co/datasets/...
) asaccessURL
in thedistribution
field. - Validate all generated metadata with
mlcroissant
to ensure compliance. - Review viewer compatibility for GeoCroissant metadata on Hugging Face (
dataset-viewer
behavior).
3. Am I blocked on anything?
- No, I am not currently blocked on anything.
Links to Work Done:
- Project Wiki Page: GSoC 2025 AI‐ready Dataset Metadata as a Service using ZOO‐Project · HarshShinde0/ZOO-AI-DATASET-MAAS Wiki · GitHub
- Public Repository: ZOO-AI-DATASET-MAAS/STAC to GeoCroissant at main · HarshShinde0/ZOO-AI-DATASET-MAAS · GitHub
Best Regards,
Harsh Shinde