Hey,
Now we have reports of code coverage in codecov.io. Code coverage is basically percentage of statements executed during tests comparing to all statements in a package.
Right now 73% of pywkibot library is covered but only 10% of scripts are being tested. For more detailed report (per file) see this.
Note that Appveyor tests are not included in the report yet.

Some steps that can make pywikibot test coverage better:
* Write tests for scripts
* site.py, page.py and pagegenerators.py have noticeable number of missing statements. We can write tests for them.
* Try coverage suggestions

Let's talk about how we can have better tests!
Best