Exoskeleton

Breaking Changes:
* `__assign_labels_to_version` is now `assign_labels_to_uuid`
* `__assign_labels_to_master` is now `__assign_labels_to_master`
* Moved `utils.determine_file_extension()` into the `userprovided` project. Now it has to be called via `userprovided.url.determine_file_extension()`.
* Renamed database table `blocklist`to `blockList`: could be a breaking change depending on your system's settings.

New features:
* `version_uuids_by_label` gets the new option `processed_only`: if that is set to `True` it returns only UUIDs from files / content / ... that has been processed. Otherwise the returned set might include queue items with that label.
* New function `remove_labels_from_uuid` allows to remove one or more labels from an UUID / version.

Breaking Changes:
* Add a [blocklist](documentation/blocklist.md "How to use the blocklist"): requires changes to the database.
* Moved function `utils.get_file_hash()` into the userprovided library.
* Harmonized all functions that either return labels associated with ids or that return ids associated with labels to return the datatype `set`. If no associations are found the return value is an empty set.
* Issue 17 / Bugfix: The duplicate prevention did not work as expected for already processed files / pages. Had to extend the table `fileVersions` and the stored procedures.

New features:
* Adding to queue returns the UUID for the new queue item: The functions `add_page_to_pdf`, `add_save_page_code`, and `add_file_download` return the UUID *if* the task was added. In case the task was not added (for example because it is already in the queue or the host is in the blocklist) the return value is None. This functionality is needed for automatic tests, but it might also be useful to create some bots.
* Introduce new functions to organize files with labels:
+ `filemaster_labels_by_id`: Get a list of label names (not id numbers!) attached to a specific filemaster entry.
+ `get_filemaster_id`: Get the id of the filemaster entry associated with a specific version identified by its UUID.
+ `all_labels_by_uuid`: Get a set of ALL label names (not id numbers!) attached to a specific version of a file AND its filemaster entry.

Breaking Changes:
* There is a new parameter `mail_behavior` that contains the already existing settings `send_start_msg`, `send_finish_msg`, and `milestone_num`.
* The prefix `mail_` was removed in settings as all mail related settings are within the `mail_settings` dictionary. `mail_admin` was renamed to `recipient`.
* Move `utils.convert_to_set()` into the `userprovided` library.

New Features / Improvements:
* New function `version_uuids_by_label`: returns all UUIDs which have this label attached.
* New function `version_labels_by_uuid`: return all labels attached to a specific version / UUID, but not the filemaster entry.
* The functionality to send mails has been externalized into the [`bote` library](https://github.com/RuedigerVoigt/bote "Homepage of the bote project"). As it has the same maintainer, compatibility is ensured.
* It is now possible to send mail with another server than localhost.
* Signal compliance with [PEP 561](https://www.python.org/dev/peps/pep-0561/): If you type-check code that imports this package, tools like mypy now know that `userprovided` has type-hints and extend their checks to calls of these functions.
* The automatic tests now also cover Windows and MacOS.

Breaking Changes:
* Changed the database structure especially regarding labels. (Reason: issue 12)
* Several class parameters were joined into dictionaries.

New Features / Improvements:
* Added an extensive [documentation](documentation/README.md).
* Restore lost database connection after timeout.
* Add Ability to store any web page as PDF file using headless Chrome.
* Prettify HTML before storing it.
* The content header is used to determine the file extension.
* Build process improved: with GitHub-Actions syntax checks and unit-tests are run for each build in Python 3.6, 3.7 and 3.8.

Other:
* Input checks were moved to the sister package [userprovided](https://github.com/RuedigerVoigt/userprovided)

* Bugfix in job_get_current_url()

* Add ability to create jobs: In essence a convenient way to store the current URL while parsing a large set of results. In case the program is terminated, now it is easy to pick up at the same page.