SNAP CPython Adapter

Goals

• Let Python programmers write GPF operators using CPython and its scientific extension libraries numpy, scipy, pandas, etc.
• It should be transparent to users how operators are implemented, e.g. using the native Java GPF API, using the Python API or using the stand-alone tools adapter.

Background and strategic fit

Scientists prefer Python over Java.

Assumptions

• Users have Python 2.7, 3.3 or 3.4 installed.

Requirements

User interaction and design

A user's Python code may be loaded from the following locations:

• <module-jar-file>   the module JAR file
• ${user.home}/.snap/snappy/ext/*    the directories in directory snappy/ext in the SNAP user directory
• ${snap.snappy.ext}   from the directory paths given in the special system property snap.snappy.ext

A Python extension module must contain a plain text file META-INF/snappy.def which lists all Python extensions contained in the module. In this file, a line are either empty, is a comment line starting with the '#' character, or is a line declaring a Python extension:

The type of the extension is derived from the ExtClassName. If it ends with Op it is assumed to be a GPF operator, if it ends with Action it is assumed to be an action to be placed in the menu or in a toolbar. Each Python extension must follow this schema:

[<ext-rel-path>/]

<ext>.py   contains the Python extension ExtClassName  

<ext>-info.xml   contains the operator metadata such as processing parameter definitions (for GPF operators only)

Questions

Below is a list of questions to be addressed as a result of this requirements document:

Not Doing