feature/run_large_tracksets - Githubissues

nismod / open-gira

Open-data Global Infrastructure Risk/Resilience Analysis

https://nismod.github.io/open-gira/

MIT License

12 stars 3 forks source link

feature/run_large_tracksets #150

Closed thomas-fred closed 9 months ago

thomas-fred commented 10 months ago

Performance improvements for running large (synthetic) storm track sets:

Faster inverse geodesic solve (haversine)
Cache JSON file reads
Faster storm-country intersection
Filter out disruption supply_factor above 0.95 (much smaller files)
Store wind speeds as 16 bit integers on disk (smaller and (much!) faster than zlib compression)

Should have their own PRs, but there you go:

Add IRIS tracksets
Add grid exposure calculation (and rename exposure outputs to more appropriate disruption)
Spatial aggregation rules for disruption and exposure outputs