We were at La Cantine on May 21th 2012 in Paris for the " Replay session".

La Cantine is a coworking space where hackers, artists, students and so on can meet and work. It also organises some meetings and conferences about digital culture, computer science, ...

On May 21th 2012, it was a dev day about Python. "Would you like to have more PyCon?" is a french wordplay where PyCon sounds like Picon, a french "apéritif" which traditionally accompanies beer. A good thing because the meeting began at 6:30 PM! Presentations and demonstrations were about some Python projects presented at PyCon 2012 in Santa Clara (California) last March. The original pycon presentations are accessible on

PDB Introduction

By Gael Pasgrimaud (@gawel_).

pdb is the well-known Python debugger. Gael showed us how to easily use this almost-mandatory tool when you develop in Python. As with the gdb debugger, you can stop the execution at a breakpoint, walk up the stack, print the value of local variables or modify temporarily some local variables.

The best way to define a breakpoint in your source code, it's to write:

import pdb; pdb.set_trace()

Insert that where you would like pdb to stop. Then, you can step trough the code with s, c or n commands. See help for more information. Following, the help command in pdb command-line interpreter:

(Pdb) help

Documented commands (type help <topic>):
EOF    bt         cont      enable  jump  pp       run      unt
a      c          continue  exit    l     q        s        until
alias  cl         d         h       list  quit     step     up
args   clear      debug     help    n     r        tbreak   w
b      commands   disable   ignore  next  restart  u        whatis
break  condition  down      j       p     return   unalias  where

Miscellaneous help topics:
exec  pdb

It is also possible to invoke the module pdb when you run a Python script such as:

$> python -m pdb


By Alexis Metereau (@ametaireau).

Pyramid is an open source Python web framework from Pylons Project. It concentrates on providing fast, high-quality solutions to the fundamental problems of creating a web application:

  • the mapping of URLs to code ;
  • templating ;
  • security and serving static assets.

The framework allows to choose different approaches according the simplicity//feature tradeoff that the programmer need. Alexis, from the French team of Services Mozilla, is working with it on a daily basis and seemed happy to use it. He told us that he uses Pyramid more as web Python library than a web framework.


By Benoit Chesneau (@benoitc).

Circus is a process watcher and runner. Python scripts, via an API, or command-line interface can be used to manage and monitor multiple processes.

A very useful web application, called circushttpd, provides a way to monitor and manage Circus through the web. Circus uses zeromq, a well-known tool used at Logilab.

matplotlib demo

This session was a well prepared and funny live demonstration by Julien Tayon of matplotlib, the Python 2D plotting library . He showed us some quick and easy stuff.

For instance, how to plot a sinus with a few code lines with matplotlib and NumPy:

import numpy as np
import matplotlib.pyplot as plt

fig = plt.figure()
ax = fig.add_subplot(111)

# A simple sinus.
ax.plot(np.sin(np.arange(-10., 10., 0.05)))

which gives:

You can make some fancier plots such as:

# A sinus and a fancy Cardioid.
a = np.arange(-5., 5., 0.1)
ax_sin = fig.add_subplot(211)
ax_sin.plot(np.sin(a), '^-r', lw=1.5)
ax_sin.set_title("A sinus")

# Cardioid.
ax_cardio = fig.add_subplot(212)
x = 0.5 * (2. * np.cos(a) - np.cos(2 * a))
y = 0.5 * (2. * np.sin(a) - np.sin(2 * a))
ax_cardio.plot(x, y, '-og')
ax_cardio.set_xlabel(r"$\frac{1}{2} (2 \cos{t} - \cos{2t})$", fontsize=16)

where you can type some LaTeX equations as X label for instance.

The force of this plotting library is the gallery of several examples with piece of code. See the matplotlib gallery.

Using Python for robotics

Dimitri Merejkowsky reviewed how Python can be used to control and program Aldebaran's humanoid robot NAO.

Wrap up

Unfortunately, Olivier Grisel who was supposed to make three interesting presentations was not there. He was supposed to present :

  • A demo about injecting arbitrary code and monitoring Python process with Pyrasite.
  • Another demo about Interactive Data analysis with Pandas and the new IPython NoteBook.
  • Wrap up : Distributed computation on cluster related project: IPython.parallel, picloud and Storm + Umbrella

Thanks to La Cantine and the different organisers for this friendly dev day.

blog entry of