Closed danscales closed 5 months ago
Attention: Patch coverage is 97.18310%
with 2 lines
in your changes are missing coverage. Please review.
Project coverage is 23.77%. Comparing base (
33cd4bc
) to head (413c210
). Report is 6 commits behind head on master.
Files | Patch % | Lines |
---|---|---|
...ats/gfwpro_dashboard/GfwProDashboardAnalysis.scala | 0.00% | 1 Missing :warning: |
...gfwpro_dashboard/GfwProDashboardRawDataGroup.scala | 96.00% | 1 Missing :warning: |
:exclamation: Your organization needs to install the Codecov GitHub app to enable full functionality.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
GTC-2734 Add JRC/SBTN, confidence levels, and total_ha to integrated alerts
For daily alerts, add new columns that divide the daily alerts into a 3x3 cross-product of (whole area, SBTN forest only, JRC forest only) x (lo, med, hi) alert confidence
Also, add a new 'total_ha' column which is the area of the location. This area will be used by the front-end to calculate measures such as "daily alerts per hectare".
Currently, I am calculating the entire location area in the usual per-pixel way with Geodesy.pixelArea. We were already calculating the area per pixel whenever isTreeCoverExtent30 was true (more than 30% tree cover). From a look at some input sets, it seems like this is usually a large part of the location areas. So, doing the area calculation for the entire geometry shouldn't be much more expensive.
However, I will make sure that the isTreeCoverExtent30 area is a large part of the whole area for most datasets, once this change is in. If it is not, then it may give better performance to calculate the total area of locations outside of the pixel function using an equal-area projection (for which I already have some working code).