* Add changelog workflow to github actions, remove old script
This basically takes the exact script, and turns it into a workflow. The
only difference is that the new workflow doesn't detect a release build
like the script did (releases will be a separate workflow).
* remove old config file for changelog generator (it's in workflow now)
* whitespace change
* remove ableist language
* Add initial config to test github actions
* remove other CI configs so PR doesn't get bogged down
* add poetry
* skip failing test, update behave output
* add multiple python versions
* add test cache
* try to use matrix for python version
* try to cache pip installation
* add other OSes
* whitespace changes
* skip failing test on windows
* add pip cache
* take out pip cache
* clean up, add linting steps
* add release branch to testing
* clean up
* add initial a11y config for docs site
* take out junit for now
* add junit annotations
* try different junit reporter
* add junit annotations into same workflow
* linter fix
* make junit step always run
* more junit
* clean up
* disable a11y exit code for now (always pass)
* put back travis and circle so we can temporarily run everything in tandem
* fix for issues with secret tokens
* Improve GitHub PR Template
* suggest running commands with `poetry run ...`
pyflakes here is given only certain subdirectories because `.` will try and run across all packages installed in the virtual environment, if the virtual environment is within the project's root folder (which is my common practice).
