Snappy GPT operators

[MarkWilliamMatthews]

Hi,

I’ve been looking into using snappy (the python SNAP API) to replace my graph processing tool (gpt) command line calls in my scripts.

However, when I am looking for operators e.g. Rad2Refl, these do not appear to be available in snappy (I am looking at http://step.esa.int/docs/v4.0/apidoc/engine/) in the gpt operators at org.esa.snap.core.gpf.common. I’d like to import these like done with the subset tool, e.g.:
NameOp = snappy.jpy.get_type(‘org.esa.snap.core.gpf.common.NameOp’)
where Name is the name of operator.

Am I correct in saying that all the operators available in the command line gpt tool are not all available with snappy? If false, could you point me to where they are sitting so I can import them? If true, are there plans to put more operators into snappy?

Thanks for the clarification.

[marpet]

Hi Mark
All operators from the command line are also available via snappy. But you should not instantiate the operator classes directly as long as there is no very good reason for it. (However I wonder why it didn’t work.)

It is much better to invoke the operators via GPF.

I just posted this one:

Compute zonal statistics from Sentinel-2 VI's

The statistics operator is called as any other operator from python. The same for the index operators. On the command line you can typegpt -h to get a list of all available operators. You can find examples in the following threads: snappy-where-to-start? Example script for multiple operations? If you invoke the statistics operator only with one product it should work as you expect it. regards Marco

Also helpful might be the links in:

About the python category

Put your question regarding the usage of snappy, the snap-python interface or anything else related to Python and SNAP into this category. This could be questions regarding the access of the SNAP API from Python or implementing an operator in Python, all questions like this are best suited here. Question regarding Sen2Cor should go into the sen2cor category. General documentation about snappy can be found in How to use the SNAP API from Python of the Developer Guide. There is also a page which…

Currently you have to call

 snappy.GPF.getDefaultInstance().getOperatorSpiRegistry().loadOperatorSpis()

once at the beginning of you script. In the next version this will not be necessary any more.

[MarkWilliamMatthews]

Hi All,

I have answered my own post by digging into the script shared in this previous post:

Python codes of GLCM for texture feature extraction

snappy_glcm.py (8.7 KB)

, but it’s not very obvious so I’ll spell it out here for future users:.

For those wanting to use GPT tools / operators in snappy this is the general procedure, in python:

import snappy
snappy.GPF.getDefaultInstance().getOperatorSpiRegistry().loadOperatorSpis() # Load all available operators
HashMap = snappy.jpy.get_type(‘java.util.HashMap’)
parameters = HashMap() #Specify the parameters, these can be seen from the command line using gpt -h operator
parameters.put(‘sensor’,‘OLCI’)
parameters.put(‘copyNonSpectralBands’,‘False’) # I had to use this option due to an error
refl_product = GPF.createProduct(‘Rad2Refl’, parameters, sub_product) # compute the product from input called sub_product

Hope that helps!

This can also be seen in the snappy_bandmaths.py example and here: [quote=“marpet, post:2, topic:2636”]
[/quote]

[MarkWilliamMatthews]

Thanks Marko, think I’ve figured it out now.