giosystem workshop¶

This is a short workshop that covers the various aspects of the giosystem software package.

Session #0 - Tools of the trade¶

In this session we introduce some software tools used when developing giosystem components.

virtualenv¶

http://virtualenv.readthedocs.org/en/latest/

Virtualenv is a Python package that allows creating isolated Python environments in any directory. These environments can be controlled independently of the system wide Python installation and can be installed anywhere inside the user’s home directory. This means that we can:

install Python packages without requiring root access;
install different versions of the same package without conflicts;
install as many packages as a project needs without polluting the system wide Python installation with packages that are only used by a single project;
create a virtualenv for each project, which keeps each projects libraries stable and isolated

In short, a virtualenv is a custom Python interpreter, located in a directory chosen by the user, where it is possible to install project-specific packages.

Installation¶

sudo apt-get install python-virtualenv

Operation¶

In order to use a virtualenv, we must:

create it
activate it

Lets say we are working on project ~/dev/myproject and wish to use a virtualenv.

Virtual environments can be created wherever the user wishes, but usually we create them as a sub-directory under the project that will be using the virtualenv. They can have any name, but it is common practice to name them venv.

mkdir -p ~/dev/myproject && cd ~/dev/myproject
virtualenv venv

Now that the virtualenv has been created, we need to activate it, so that it can be used. There are two ways to activate and use a virtualenv:

When working in a terminal, we can source the activate script. The terminal’s command prompt then changes, including the name of the virtualenv to provide a visual cue that the virtualenv is active.
```
geo2@geo2:~/dev/myproject$ source venv/bin/activate
(venv)geo2@geo2:~/dev/myproject$ # the virtualenv has been activated
```
After activating a virtualenv, the python command will now refer to a new Python interpreter, located inside the virtualenv’s directory. We can verify this by issuing the which command:
```
(venv)geo2@geo2:~/dev/myproject$ which python
~/dev/myproject/venv/bin/python
```
When we no longer wish to use the virtualenv, we can issue the deactivate command and it will be deactivated. The terminal’s prompt returns to its normal state to indicate that the virtualenv is no longer active.
```
(venv)geo2@geo2:~/dev/myproject$ deactivate
geo2@geo2:~/dev/myproject$ # the virtualenv is no longer active
```
When running scripts and applications, we can directly use the Python interpreter that was created by the virtualenv instead of using the default system Python interpreter. For example, if we had a myscript.py file to run, we’d execute it like this:
```
# instead of calling python myscript.py
venv/bin/python myscript.py
```
In this case, the virtualenv is only used for executing the script, as soon as execution is terminated, so is the virtualenv.

Since a virtualenv is just a directory with custom (but isolated) Python interpreter and packages, uninstalling it is just a matter of deleting the virtualenv directory.

rm -rf ~/dev/myproject/venv

pip¶

http://www.pip-installer.org/en/latest/

Pip is an installer of Python packages. It provides some very convenient functionality:

download and install packages from:
- the Python Package Index (http://pypi.python.org)
- source code repositories
install and uninstall packages
install packages from a list of requirements

Installation¶

Pip can be installed system wide, for using with the default Python interpreter. It can also be used inside a virtualenv. In this case, the virtualenv command takes care of installing pip for us when creating the virtualenv, so there is no need to do anything.

sudo apt-get install python-pip

Operation¶

Using pip to install packages is straightforward. Lets try installing the requests package, which is available on pypi:

# pip install <package_name or url>
(venv) pip install requests

Pip will download and install the packages. There are many other options available, such as specifying the version number for a package, separating the downloading and installing steps, etc.

Uninstalling a package is just as simple:

(venv) pip uninstall requests

For inspectings which packages are currently installed we can use the freeze command:

(venv) pip freeze

Pip has the ability to install packages from a requirements file. A requirements file is just a normal text file where each package is defined in a new line. This means we can store the output of pip freeze into a file and then use this file in another location to install the same packages.

(venv) pip install requests
(venv) pip freeze > requirements.txt
(venv) pip uninstall requests
(venv) pip install -r requirements.txt

This makes requirements files a nice thing to have in a source code repository, as they can be used to install dependencies.

django¶

https://www.djangoproject.com/

Django is a web framework. It is used to create applications that operate over the Internet, such as websites and the like.

Django documentation is really good, so the best way to learn it is going through the tutorial at:

https://docs.djangoproject.com/en/1.6/intro/tutorial01/

Generally speaking, django is composed of the following components:

models - Store data and application logic. Models are saved in a database and are manipulated using an object-relational mapper. This means that Django takes care of writing the actual SQL code to fetch data from the database while we can operate on them as regular objects and methods;
views - Retrieve data from models and process it to produce the requested information. Views are normal Python functions that usually fetch model objects, do something with them and then return a result as a form of a rendered template.
urlconfs - Do the matching between URLs and the view functions. Urlconfs take care of matching specific URLs with the respective view functions that fetch the actual data from the models.
templates - Present the resulting information to the user. Templates use the returned data from views and present it to the user, with nicely formatted fonts, colors, etc.

Installation¶

Operation¶

Other libraries¶

requests
pyxb

#Session #1 - giosystem_settings_django application #:::::::::::::::::::::::::::::::::::::::::::::::::: # #*to be written* # #Session #2 - giosystemcore library #:::::::::::::::::::::::::::::::::: # #*to be written* # #Session #3 - giosystem-processing components #:::::::::::::::::::::::::::::::::::::::::::: # #*to be written* # #Session #4 - giosystem-ecflow integration #::::::::::::::::::::::::::::::::::::::::: # #*to be written* # #Session #5 - giosystem ordering service #::::::::::::::::::::::::::::::::::::::: # #*to be written*