Use area definition names to check sunlight coverage #67
Conversation
Codecov Report
@@ Coverage Diff @@
## master #67 +/- ##
==========================================
+ Coverage 92.13% 92.25% +0.11%
==========================================
Files 10 10
Lines 1449 1523 +74
==========================================
+ Hits 1335 1405 +70
- Misses 114 118 +4
Continue to review full report at Codecov.
|
| prod = scn[product] | ||
| except KeyError: | ||
| try: | ||
| prod = scn[list(scn.keys())[0]] |
There was a problem hiding this comment.
What happens if some of the datasets aren't complete (like 3a or 3b on avhrr ?)
There was a problem hiding this comment.
I guess then the user could move the plugin to be run after resampling.
|
I've now tested this wih live data using an named area definition, and directly saving to satellite projection (area set to |
|
The coverage calculation doesn't work for projections where the area has space in it, like the full disk |
|
Ok, found a solution and it seems to work. I'll check the status in the morning. |
|
Yes, works nicely. |
|
The ZMQ and crash handling commits should have been in a different branch -> reverted. |
This PR changes the behaviour of
check_sunlight_coverage()so that it will use the name of area definitions instead of individual composite/channel areas read from resampled scene attributes. This way the filtering can be done before loading and resampling. If the data are not to be projected (area defined asnullin the config), the area is read from the dataset attributes and the plugin needs to be after the data is loaded withload_composites.Closes pytroll/pyresample#228
flake8 trollflow2