Open cfusting opened 7 years ago
Let's explore dealing directly with the hdf files. Here is an example of exploring and hdf file and getting the data.
from pyhdf.SD import SD, SDC
import numpy as np
f = SD("tests/MOD13A3.A2014274.h20v02.006.2015291193621.hdf", SDC.READ)
for i in f.datasets():
print i
data = f.select("1 km monthly NDVI")
print data.get()
data.attributes()
See also this post for further ideas. I'm not convinced as to whether we should put back the masked data directly into the original hdf file or if we should begin to construct some "master" hdf file that will hold all the data we wish to use in our model. Either way if dealing directly with the hdf files works out I'd like to go ahead and have you refactor the existing methods to do so as well.
Have a look at extract_lst_data.py, extract_lst.sh, extract_lst_job.sh for an example of how to do this.