(Redmine issue: #3361,#3443)

The account registration process needed more work to get it into stable and usefull state:
* Changed validator for user login from email to string containing restricted set of characters.
* Renamed field 'Login (eppn)' to just 'Login'. It made no sense to be used that was for example with Basic or certificate based authentication service.
* Implemented sendind of confirmation emails after successfull registration to system admins and to the user that performed the registration. The emails are generated using Jinja2 templates.
* Implemented sending of confirmation emails after successfull account activation to the user and as BCC to system admins.
* Added missing rollback() calls to base View implementations to resolve SQLAlchemy errors after unsuccessfull database operation.

(Redmine issue: #3734)

Added valid maxDiff variable name from unit test framework to list of valid names in Pylint to globally ignore this warning.

(Redmine issue: #3443,#1017)

The mentat-informant.py module is a Python replacement for previous mentat-briefer.py module. It is designed to provide periodical statistical overviews via email. The email reports are localized using babel package and can be translated into any language. Additionally, the reporting engine uses Jinja2 templating system for message rendering, so it enables very easy customization. (Redmine issue: #3377)

E:422,34: Module 'traceback' has no 'TracebackException' member (no-member)

(Redmine issue: #3443,#3387)

(Redmine issue: #3443)

Fixed all pylint issues in mentat.geoip module. (Redmine issue: #3443)

Removed all reported errors for library code reported by pylint, with additional fixes of minor problems. (Redmine issue: #3443)

Added configuration file fo pylint tool for inspecting python modules (files in ./lib subdirectory).

(Redmine issue: #3443)

Greatly improved code quality using pylint and pyflakes. All files now rank above 9 points with custom .pylintrc-bin file. Following exceptions were added to default pylint configuration:

  * line-too-long - I do not want to have calls to logging service spanning over multiple lines and distracting from code flow. I want them to be only single line, no matter how long, because it is not important.
  * bad-whitespace - I like to add additional whitespaces in places, where they should not be, mostly to align related items vertically, which in my opinion improves readability.
  * logging-format-interpolation - In Python3 I prefer string.format() to string % (args).

Additionally the pattern for module name validation was altered, because of our executable file naming policy (added suppot for '-' and '.' characters in name).

(Redmine issue: #3443)