nasa-nccs-hpda / evhr

Enhanced Very-high Resolution (EVHR)
4 stars 0 forks source link

Some issues running EVHR on ADAPT (maybe because of duplicated entries) #16

Open jordancaraballo opened 12 months ago

jordancaraballo commented 12 months ago
cat /explore/nobackup/projects/3sl/data/VHR/_lists/ETZ-Tappan-Per-Year/ETZ-Tappan_M1BS_scenes_noDups_2017.csv | grep M1BS_R

Went through the input file you guys are using for EVHR yesterday. Looks like there are a lot of duplicated scene 1 ntfs in the X1BS subdir. I saw this before and those seem to cause EVHR problems. Anything with the RC in the filename should not be in the input file

Ex:

/css/nga/WV02/1B/2017/297/WV02_103001007374BB00_X1BS_500934118100_08/WV02_20171024115548_103001007374BB00_17OCT24115548-M1BS_R3C1-500934118100_08_P001.ntf /css/nga/WV02/1B/2017/297/WV02_103001007374BB00_X1BS_500934118100_08/WV02_20171024115548_103001007374BB00_17OCT24115548-M1BS_R1C1-500934118100_08_P001.ntf /css/nga/WV02/1B/2017/295/WV02_10300100735A9400_X1BS_500934119020_06/WV02_20171022112808_10300100735A9400_17OCT22112808-M1BS_R1C1-500934119020_06_P001.ntf /css/nga/WV02/1B/2017/295/WV02_10300100735A9400_X1BS_500934119020_06/WV02_20171022112808_10300100735A9400_17OCT22112808-M1BS_R2C1-500934119020_06_P001.ntf

You might want to check with Maggie and see if you can get rid of any of these and see if EVHR runs.

Of course that is all WV02 and your original problem was with some WV03 strips, I am still not sure what is going on there. Still taking a look.

mwooten3 commented 11 months ago

I actually do not think that these files cause problems inherently. It is possible for them to be part of a duplicate error, but whenever I have actually looked at these files, it seems the RxCy notation is just another way of dividing a strip into scenes. If you pull these specific example scenes into QGIS, they are not duplicates and form different parts of the same strip. dg_mosaic will not usually fail with these input files as long as their .xml files are unique. I am not sure why some scenes are given this naming convention, but exemplifies why we can't just use filename's catalogIDs and P00X suffix to detect duplicates (though this usually suffices).