Le programme des bourses d’excellence Eiffel: L’appel à candidature est ouvert jusqu’au 9 janvier 2020.

L’appel à candidature est ouvert jusqu’au 9 janvier 2020.

Le programme de bourses Eiffel est un outil développé par le Ministère de l’Europe et des Affaires étrangères afin de permettre aux établissements français d’enseignement supérieur d’attirer les meilleurs étudiants étrangers dans des formations diplômantes de niveau master et en doctorat.

Il permet de former les futurs décideurs étrangers, des secteurs privé et public, dans les domaines d’études prioritaires, et de stimuler les candidatures d’étudiants originaires de pays émergents âgés de 30 ans maximum pour le niveau master et de pays émergents et industrialisés âgés de 35 ans au plus pour le niveau doctorat.

Calendrier session 2020

  • Ouverture de l’appel à candidature : semaine du 21 octobre 2019
  • Date limite de réception des dossiers par Campus France : 9 janvier 2020
  • Annonce des résultats : semaine du 23 mars 2020 

Pour en savoir plus: https://www.campusfrance.org/fr/le-programme-de-bourses-d-excellence-eiffel

Face recognition using Python and OpenCV

Just playing around with deep learning and the folks at the office. It was retrieved 7785 face encodings from the users’ database, performed face detection, encoding and calculating the least Euclidean distance between face encodings, then returning the username of the detected face.

That script was developed with Python 3.6, OpenCV, and the face_recognition wrapper.

The code is available at: https://github.com/progerjkd/face_recognition

The TODO list includes receive the URL of photos via API and improves the affinity of detection.

Installing Python into your home directory

Sometimes we may need a newer Python version than the one installed into our station, and the Linux distribution does not provide the last release in its repositories. We could get rid the installed Python packages and make a new installation from the source code. However, various OS applications make use of the default Python version installed, thus there is a risk of breaking some applications or services of the system in the case of overriding the Python provided by the package management system used, yum or apt, for example.

Thus, we may install Python into the local user directory, and let the default Python installation intact.

Another reason to do this is to have multiple Python environments when developing or debugging old Python scripts.

The following commands were tested into an Ubuntu 16.04 amd64 docker container.

Install prerequisites for downloading and compiling Python from source:

# apt install xz-utils wget gcc make \
  libssl-dev libffi-dev zlib1g-dev

If you are running CentOS 7,  install the following packages:

# yum groupinstall -y "development tools"

# yum install -y \
  libffi-devel \
  zlib-devel \
  bzip2-devel \
  openssl-devel \
  ncurses-devel \
  sqlite-devel \
  readline-devel \
  tk-devel \
  gdbm-devel \
  db4-devel \
  libpcap-devel \
  xz-devel \

Create a folder for the new Python installation, download, and extract the source code. In this example, I used the hidden directory $HOME/.python.

mkdir $HOME/.python
cd $HOME/.python
wget https://www.python.org/ftp/python/3.7.0/Python-3.7.0.tar.xz
tar Jxvf Python-3.7.0.tar.xz
cd Python-3.7.0

Compile and install Python.

./configure --prefix=$HOME/.python --enable-optimizations
make install

Next, we need to edit $HOME/.profile , and $HOME/.bashrc files to add the Python’s bin directory to $PATH.

export PATH="$HOME/.python/bin:$PATH"
export PYTHONPATH="$HOME/.python"

Then we can reload the current session by doing:

source $HOME/.profile

Now we have multiple python installations on our system. If we type ‘python’ and press TAB twice we can see:

roger@9e65d2dd5598:~$ python
python python2.7 python3-config python3.5m python3.7-config python3.7m-config
python2 python3 python3.5 python3.7 python3.7m python3m

So, to run the new python REPL, we must enter:

roger@9e65d2dd5598:~$ python3.7
Python 3.7.0 (default, Sep 26 2018, 22:06:04)
[GCC 5.4.0 20160609] on linux
Type "help", "copyright", "credits" or "license" for more information.

Pip also is referred by its version:

roger@9e65d2dd5598:~$ pip3.7

pip3.7 <command> [options]

When developing executable python scripts for this environment, the bash shebang used must be like:

#!/usr/bin/env python3.7

Delete the Python source code and compilation directory:

rm -fr $HOME/.python/Python-3.7.0*


This article is adapted from the solution presented in http://thelazylog.com/install-python-as-local-user-on-linux/.