2008 Winter Project Week:PythonSupport

Return to 2008_Winter_Project_Week

[1]

Python module

Key Investigators

• Mario Negri Institute: Luca Antiga
• UCSD: Bryan Smith

Objective

Python has a great potential in Slicer3. Its uses span:

• Python command line modules (advantages: easy to write, access to Python libraries, don't need a compiler to write new modules)
• matplotlib and scipy Python modules (they provide a matlab-like processing environment for images directly within Slicer).
• Python as Slicer's scripting language (for automating tasks, programmatically access modules - this is Dan's new baby! - enabling staged processing)

Approach, Plan

In the project week, we plan to

• solve building issues, check stability on different platforms
• threading issues: modifying the MRML scene from Python command line modules in a thread-safe way
• stabilize calling modules from Python, write example on staged processing
• provide a Python equivalent to ScriptedModules
• define what aspects will end up in the next release
• propose a working set of Python command line modules (interact with Sonia's breakout session on Tuesday?) to be included in the next release
• write documentation on how to write Python command line modules

Progress

January 2008 Project Week

• Matplotlib with a Cairo backend is almost there
• Python command line modules now work correctly (handling arguments multiple=true has been fixed)
• A set of working Python command line modules has been committed
• Scripted modules have been extended to Python, and a sample scripted module has been written
• Meta-modules are now possible; the first meta-modules will be written soon

June 2007 Project Week

These are the results of the MIT project week (largely Dan's work)

• Implemented Python Modules
  • Add search for Python in ModuleFactory
    • Can put .py files in the CommandLine Module path
  • Execute Python within Slicer
    • Full access to VTK, vtkTEEM, vtkITK, etc.
  • If the Python module provides an XML description, and an Execute function, will be added as a CommandLine Module
• ToDo
  • Check-in (needs USE_PYTHON guards)
  • Progress reporting
  • Simplify the API
  • Automate command line parsing using XML description

Later on, after the meeting

• command line parsing has been completed
• Python scripting and Python modules have been made independent of Numpy
• Python command line modules are now run in the main thread (allows interaction with MRML scene)
• a set of Python command line modules has been written