kotti_user_manual

A User Manual for Administrators and Users of Kotti CMS websites

Setup

Make a virtualenv and prepare to use Kotti and selenium:

mkvirtualenv kotti_user_manual
wget https://raw.github.com/Pylons/Kotti/master/requirements.txt
wget https://raw.github.com/Pylons/Kotti/master/app.ini
pip install -U selenium
pip install -r requirements.txt

pip install Kotti

pip install kotti_docs_theme

pip install BlockDiag

PIL is already installed by Kotti requirements, and BlockDiag needs it. Check that you have freetype2 on your system and that the installation of PIL sees it.

Visit http://dejavu-fonts.org/wiki/Download and download into the /scripts directory a set of fonts to use. We should probably try to use a given set. So far, the set used is:

wget http://sourceforge.net/projects/dejavu/files/dejavu/2.33/dejavu-fonts-ttf-2.33.tar.bz2

After exploding the fonts, you should have the ttf files in:

scripts/dejavu-fonts-ttf-2.33/ttf/

Run Kotti from the kotti_user_manual virtualenv directory:

pserve app.ini

In a Second Terminal

Download the Selenium Java server and run it. On Mac OS X, after downloading and clicking the tarball, it will explode and run. Or run with java yourself, with, for example:

java -jar selenium-server-standalone-2.28.0.jar

In a Third Terminal

For using Chrome, see the list of drivers at http://code.google.com/p/chromedriver/downloads/list, then download the chromedriver to the scripts directory. You do not have to download a driver for using the Firefox webdriver.

Then run the script from the kotti_user_manual/scripts/ directory:

python kotti_capture.py

This script saves screen capture images made with Selenium in the docs/images/ directory, and stores a chromedriver.log locally.

Then run the script:

python kotti_blockdiag.py

This script uses the BlockDiag python package to generate PNG images for use in the docs. It is possible to embed the BlockDiag diagram definitions in the rst docs as directives. However, so far at least, to take advantage of ttf fonts as this script does, we need to generate the PNG files on our own machines that have PIL installed with freetype2 installed.

Working with the kotti_capture.py Script

In docs/ .rst files, add images as you normally would for Sphinx, according to its documentation. The entries will have the form:

``.. Image:: ../images/add_menu.png``

You need to add a matching function, e.g. add_menu(), which must be made through study of existing functions and experimentation. It is very much a system where success is found by trial and error, combining attempts in the Python code to use Selenium to find elements and do the screen captures with element inspection using Browser tools to find ids, class names, and other targets for Selenium.

Write functions that are dedicated to producing a given image, using whatever steps are needed in python-selenium operations. It can be necessary to perform several find, text and key entry, and click operations to achieve the desired state before capturing the image. See existing functions for the pattern, paying attention to the order of steps. Little things matter.

Working with the kotti_blockdiag.py Script

Add image directives for diagrams in the same way as you would for screen captures, e.g. a diagram PNG image might be:

``.. Image:: ../images/anonymous_vs_authenticated.png``

Add a corresponding BlockDiag diagram definition in scripts/blockdiag, e.g.

scripts/blockdiag/anonymous_vs_authenticated.diag

See existing examples and the BlockDiag docs, and experiment.

Overall Routine

• Change / Run kotti_capture.py when something about Kotti changes, and fresh screen captures are needed, or when something is added or removed.
• Likewise, run kotti_blockdiag.py as needed when existing diagrams are modified or new ones added.
• Update docs files as needed for image directives from either kotti_capture.py or kotti_blockdiag.py.
• Update text of docs as needed.
• Rerun docs/make html.
• Push to github.com/Kotti/kotti_user_manual.
• Log in to readthedocs.org and manually click the Build button.