Software Carpentry workshop for MAT at QMUL, Feb 2013¶

Post-workshop links¶

• Assessed exercise - How we marked the exercise
• Download links for workshop material

Dates/Where¶

• 6th-7th February 2013 -- MAT lab

Planned schedule¶

Day 1 - 9.30am to 3pm

• Day 1 - Morning
  • Introduction - material - 5 mins - Chris
  • Shell - material - 45 mins - Steve
  • Short coffee break
  • Python - material - 90 mins - Luis

• Day 1 - Lunch 12 - 12.30

• Day 1 - Afternoon
  • Version control - material - 90 mins - Chris
  • Data management - material - conversation points in area outside lab - 60 mins - Steve

Day 2 - 10am to 4pm

• Day 2 - Morning
  • Python modules and NumPy - material in preparation - 60 mins - Luis
  • Coffee
  • Audio files and more NumPy - material in preparation based on this from Codasign - 80 mins - Luis

• Day 2 - Lunch - 12.30 to 1.15

• Day 2 - Afternoon
  • Testing - material - 90 mins - Chris
  • Set the assessment exercise; coffee
  • What We Know - material - 50 mins - Chris

• Post-workshop assessment exercise
  • Register for soundsoftware site No, the exercise will use BitBucket as well (though we tell them about the SoundSoftware site)
  • Take the exercise from the last bit of the afternoon (based on ISMIR, but not quite the same) and complete some aspect of it.
  • Do this within version control on soundsoftware - add one of us as project member

Actual schedule¶

The timings we actually ended up having.

Day 1 - 9.30am to 3pm

• Wait for people to arrive - 15 mins
• Short intro - Chris - 5 mins
• Shell - Steve - 60 mins
• Coffee - 15 mins
• Intro to Python - Luis - 50 mins
• Lunch - 40 mins (12.00 - 12.40)
• Version control - Chris - 80 mins
• Data - Steve - 60 mins (inc. words from Pat at the start)

Day 2 - 10am to 4pm

• Wait for people to arrive - 15 mins
• Python functions - Luis - 95 mins
• Coffee - 10 mins
• Audio in Python - Luis - 45 mins
• Lunch - 60 mins (12.45 - 13.45) (inc. words from Pat at the end)
• Testing - Chris - 60 mins
• Coffee - 20 mins
• Preview assessed exercise - 5 mins
• What We Know - Chris - 50 mins

MAT lab computer setup¶

Computer lacks:

• Login for Luis (and Steve?) fixed
• hg and EasyMercurial fixed
• Matplotlib fixed
• iPython fixed
• Nose fixed
• pip unsure
• scikits.audiolab unsure
• Emacs with a window (as opposed to in terminal) - Aquamacs? use textwrangler

Computer does have:

• A text editor (TextWrangler)
• Python 2.7 system install
• NumPy

Detailed plan for schedule¶

Day 1¶

9am (?) to 2pm.

Day 1 is rather short because of the MAT lab being booked by someone else from 2pm.

Morning

• Intro to Shell (SJW)
  • Directories: pwd, cd, mkdir, ls
  • Files: cp, mv, rm, cat, editors
  • scripts #!/bin/bash
  • access control: chmod
  • Finding stuff: grep, find, locate, which
    • Regular expressions
    • Shell wildcards
  • Background tasks: &, fg, bg
  • Disk usage: du, df
  • Tying things together > < |
  • commands history, tab autocompletion
  • variables
  • doing stuff somewhere else: ssh
  • doing serious stuff: sudo

• More possibilities:
  • dot
  • wget
  • rsync
  • mounting network (home) directories
  • working on octave

• Version Control (CC) (exercise over lunch?)
  • Creating repositories on SSW

Afternoon

• Data Management (SJW) - Should content be integrated into discussion of shell / be before VC ?
  • Rationale
  • File formats - open vs. proprietary, lossy vs. lossless
    • Open formats should remain accessible in the future
    • The actual output from a lossy data format may vary depending upon how it's decoded
    • Converting files - ImageMagick convert, mplayer (to wav) + lame/flac/ogg, ffmpeg
  • Backing up
  • Archiving
    • "standard" compression tools good on text-like files especially if areas of content are similar
    • special compression needed for complex binary data which varies a lot e.g. FLAC
  • Publishing
  • Documenting
  • Available facilities in C4DM (relevance for MAT ?)

Day 2¶

10am to 4pm.

Morning

• Python (LF)
  • Variables
    • strings, integers, floats, lists, sets and dictionaries
    • By reference vs. by value
  • Control-of-flow
  • Libraries and modules - pip, easy_install
  • OOP, objects and classes ?

• Python for research
  • NumPy
    • Matplotlib
    • SciPy
  • Python for Audio/Media: audioop, ossaudiodev, wave, aifc, sunau, PyMedia, pygame (sndarray, mixer, mixer.music), libpd (?), Python Audio Tools, scikits.audiolab, Medussa

Afternoon

• Software engineering
  • feedback cycles
  • Unit Tests and TDD
    • nose

Note: We need to have some sort of evaluation so that we can give attendees a mark (some exercise to be submitted via version control a week later?)

Outcomes¶

Note some of these are C4DM-specific for the C4DM version of the workshop

• Students to have their own version control repository on soundsoftware.ac.uk and to know how to use it (command-line and easyHg)
• Removal of fear of the command-line
• Knowledge of basic UNIX shell commands
• Knowledge of basic Python
• Knowledge of (some) Python libraries for media research
• Understanding of the need for data management and how to manage data
• Understand facilities available at C4DM
• Know where to go for more info
  • Software Carpentry - software-carpentry.org
  • Version control / project management - soundsoftware.ac.uk
  • Python - python.org
  • Data management - SoDaMaT
• Improved practices for software engineering, souce code management, data management