RDM: automated stratification of user datasets

[jdries]

• [ ] automatically run stratification script
• [ ] generate geoparquet file, if openEO requires separate file (TBD)

[santoshkaranam]

First version completed, need to be reviewed https://github.com/WorldCereal/worldcereal-referencedata/blob/main/harmonization/extraction_flag.py

Pending

• [x] 1. Review of the code
• [x] 2. Convert to docker image
• [x] 3. Host Service inside RDM

[santoshkaranam]

@cbutsko is currently working on converting script to service using FastAPI

[santoshkaranam]

@cbutsko has provided the first version of scripts to assign h3 index, extract samples and also mapping suggestions. I will test this and integrate

[santoshkaranam]

service deployed, need to trigger stratification based on criteria.

[santoshkaranam]

@ this is done. When user dataset is uploaded, h3 indexes are assigned and stratified to mark subset of samples as extract=1.

issue can be closed