Future tutorial possibilities

[keflavich]

Here are some tutorial ideas / suggestions we have in mind. If they're here and unchecked, they haven't been scheduled. Once they're checked, they should be scheduled on the google calendar.

• May 5, 2014; Day 1: Intro, Units & Coordinates in astropy Adam Ginsburg
• pyCUDA and ffts by Marco Tazzari
• pyMultiNest by Johannes Buchner
• astroquery by Julien Woillez and Adam Ginsburg
• Python IFU tools by Bernd Husemann
• May 12, 2014 Numpy by Julian Taylor
• Package Management and Installation (anaconda, pip, etc.)
• Debugging interactively (and testing)
• Making nice matplotlib plots by Bernd Husemann
• Wrapping fortran & C code
• How to optimize (profile and improve speed) of python code
• Working with tables (astropy, maybe pandas)
• ADS-related tools
• Glue
• Testing (py.test, nose, etc.)
• Parallelization. joblib, multiprocessing, mpi4py, some notes on parallelism

[keflavich]

Package management and installation

[keflavich]

Getting started with ipython

Debugging interactively (without the mess)

• for packages
• for scripts

[keflavich]

How to write fast code: numpy arrays

• basic numpy
• advanced array stuff? numba?

[keflavich]

3D plotting (incorporating into PDFs)

[mtazzari]

Hi,
a collaborator of mine, Frederik Beaujean (LMU, C2PAP), offered to come and present us

Fast FFT computation on CPU with FFTW and the pyfftw wrapper.

[keflavich]

From Leonardo:

• Recommended python distribution and interactive development environment
• Developing some useful code on a git-like repository and packaging in a way that can be downloaded/installed as module by other people
• Computation efficient code:
  • how best to use python with my existing C/C++ or Fortran codes
  • how to develop and link small compute efficient functions (in C/C++/F) and link them to python as a module
  • how to use hardware accelerators with pyCUDA and pyOpenCL (does it make sense to do this?)
  • parallel computing in python multiprocessing and/or mpi4py (or other options)

We are looking for speakers for many of these

[keflavich]

Eric:

• Optimizing code; profiling (related: C/Fortran wrapping) - Julian
• parallelism in general
• Automatically generating bibliographies: ADS, ADS-labs, etc.

Joe:

• working with tables. astropy tables; maybe pandas (or maybe not)
• SQL database querying