Configure Python to use the new SNAP-Python (esa_snappy) interface (SNAP version 10+)

Access to the esa_snappy plugin

The esa_snappy plugin has been implemented as a dedicated SNAP module. As for the other SNAP modules, the corresponding source code is available from Github .

Installation of the esa_snappy plugin

The esa_snappy plugin is one of the ‘internal’ SNAP plugins which are automatically installed during the SNAP installation. Thus, no further user action is required.

Configuration of esa_snappy from the Command Line

Open a command line window (shell, terminal window on Unixes, cmd on Windows) at the bin folder of the SNAP installation directory. Now type

Unix/MacOS:
$ ./snappy-conf <python-exe>

Windows:

$ snappy-conf <python-exe>

This will generate the Python module esa_snappy configured for the current SNAP installation and your Python interpreter <python-exe> into the .snap/snap-pythondirectory of the user home directory. The parameter <python-exe> must be the full path to the Python interpreter executable which you want to use with SNAP (supported versions are 2.7, 3.3 to 3.10, we recommend to use a fairly recent version).  If you want the esa_snappy module to be placed somewhere else use:

Unix/MacOS:
$ ./snappy-conf <python-exe> <esa_snappy-dir>

Windows:

$ snappy-conf <python-exe> <esa_snappy-dir>

When you see as output something like below, the configuration was successful:

Known issues and pitfalls

Configuration fails with empty stack trace

This issue has been observed while trying to configure esa_snappy on Windows as described above. The output of snappy-conf may look like this:

Configuring ESA SNAP-Python interface...

Configuration failed!

Error: Python configuration error

Full stack trace:

The reason for this could be that you have permanent environment variables PYTHONHOME and PYTHONPATH set and pointing to a Python installation different to the one which you are trying to use with esa_snappy. If not ultimately needed, we recommend not to use PYTHONHOME and PYTHONPATH.

Configuration failed with exit code 30

This error has been occasionally observed during the esa_snappy installation and Python
configuration on Linux. Here, the environment variable LD_LIBRARY_PATH is likely not set
correctly, and thus the shared library for the JVM cannot be found. This can be solved by
performing the following steps:

• `locate libjvm.so`

• Output is, say, `/path/to/libjvm.so`

• `export LD_LIBRARY_PATH=/path/to/:${LD_LIBRARY_PATH}`

• Re-try the installation

Cannot open shared object

After a successful esa_snappy installation and Python configuration, you might get an error
similar to

`ImportError: libjvm.so: cannot open shared object file: No such file or directory`

when doing `import jpy` within your Python script. Again, the shared library for the JVM cannot be found. This might happen if the LD_LIBRARY_PATH has previously been set correctly, but was changed or not set permanently. In this case, set the LD_LIBRARY_PATH as described above and restart your Python script.

Configuration of esa_snappy during SNAP installation

During the SNAP installation, a dedicated screen for the Python configuration will appear. Follow the instructions given there.

Configuration of esa_snappy from GUI

The option to configure esa_snappy from a graphical interface in SNAP Desktop will be provided in a future SNAP release.

Testing esa_snappy

To test esa_snappy,

$ cd <esa_snappy-dir>
$ <python-exe>     (start your Python interpreter)

Then try the following code:

from esa_snappy import ProductIO
p = ProductIO.readProduct('esa_snappy/testdata/MER_FRS_L1B_SUBSET.dim')
list(p.getBandNames())

This approach only works if the current working directory is in the <esa_snappy-dir>. To generally make use of snappy you need to do one of the following configuration steps.

Usage of esa_snappy from Python

To effectively use the SNAP Python API from Python, the esa_snappy module must be detectable by your Python interpreter. There are a number of ways to achieve this. 

• To make esa_snappy permanently accessible, you could install it into your Python installation. On the command line (shell, terminal window on Unixes, cmd on Windows), type

$ cd <esa-snappy-dir>/esa_snappy
$ <python-exe> setup.py installThis might require root privileges on Unix systems and results in a esa_snappy folder in usr/local/lib/python/dist-packages/

• If you encounter any problems with this approach, you can also try to copy the <esa-snappy-dir>/esa_snappy directory directly into the site-packages directory of your Python installation.

• Or you could also temporarily or permanently set your PYTHONPATH environment variable:

export PYTHONPATH=$PYTHONPATH:<esa-snappy-dir>   (on Unix OS)
set PYTHONPATH=%PYTHONPATH%;<esa-snappy-dir>    (on Windows OS)

• Finally, you could also append <esa-snappy-dir> to the sys.path variable in your Python code before importing esa_snappy:

import sys
sys.path.append('<esa-snappy-dir>') # or sys.path.insert(1, '<esa-snappy-dir>')
import esa_snappy

• In case you seek for a generic solution without needing to set <snappy-dir> you may automatically find snappy through the 'USERPROFILE' environment variable. Note that this solution requires esa_snappy to be located at the value of 'USERPROFILE':

import os
esa_snappy_envar = 'USERPROFILE'
envs = os.environ
if not esa_snappy_envar in envs.keys():
	raise Exception('Can’t find esa_snappy')
else:
	esa_snappy_dir = os.path.join(envs.get(esa_snappy_envar), '.snap', 'snap-python')
sys.path.append(esa_snappy_dir)
import snappy

Change the Memory Settings

Within <esa-snappy-dir>/esa_snappy a file named esa_snappy.ini is located. Here you can change how much memory esa_snappy can use.

Change the line from

# java_max_mem: 8G

to e.g.

java_max_mem: 10G

This means that esa_snappy can use 10GB of your RAM. A recommended value is 70%-80% of the available RAM in your system.