Borb

This release is a small feature release.
- Small fix in `TextAnnotation`.
- Small fix(es) in tests related to redaction.
- All tests that produce a PDF now check the validity of that PDF using an external validator.
- The default profile of this validator checks against the latest version of the PDF spec. `borb` was built to adhere to iso32000.
Most notable difference is the embedding of the standard 14 fonts. I made a separate profile that checks only against the iso32000 spec.
- `CheckBox` should be debugged. The validator marks it as a bad PDF.
- For PDF A1/b to really match the spec, the fonts (even standard 14) need to be embedded.
- You can now specify `multiplied_inter_column_spacing` in `MultiColumnLayout`.
- `x or Decimal(y)` evaluates to `Decimal(y)` when `x == Decimal(0)`. This was unintended. I only wanted it to evaluate to `Decimal(y)` when `x is None`. I fixed all unwanted occurrences. This mostly affected the `LayoutElement` classes.
- Add `Alignment` to easy imports
- Add `LoremIpsum` class to quickly generate dummy text (using a markov chain)

This release is a small feature release.
- a strong effort has been made to annotate all asserts
- all public methods are documented
- fixed a small bug in `Page.apply_redact_annotations`
- fixed a ton of `mypy` warnings

This release is a small feature release.
- rounded corners are available on `LayoutElement` and all its children
- all properties of `LayoutElement` are now available on all its children
- with the exception of `FormField`
- with the exception of nonsensical properties
- e.g.: `font_size` on `Shape`
- `PDFToJPG` can now locate fonts on Windows and Mac
- `OCRImageRenderEventListener` has been made more resilient to bad input

This release is a small feature release.
- rounded corners are available on `LayoutElement` and all its children
- all properties of `LayoutElement` are now available on all its children
- with the exception of `FormField`
- with the exception of nonsensical properties
- e.g.: `font_size` on `Shape`
- `PDFToJPG` can now locate fonts on Windows and Mac
- `OCRImageRenderEventListener` has been made more resilient to bad input

This release is a small bugfix release.
By adding al lot of imports to `borb/pdf/__init__.py` suddenly all imports for `Chart` were imported.
This included `matplotlib.pyplot`. This import is only needed when working with `Chart`.
I changed the code in `Chart` to only declare the type, but not do the import (as it didn't really need to in the first place).

This release is a small feature release:

- Test have been added corresponding to issues on GitHub.
- More documentation has been added in general. No undocumented public methods!
- Unsplash API has been added, to ensure you can build rapid prototypes for Documents by just specifying keywords for images, rather than having to look for the perfect `Image`
- Small bugfix w.r.t. annotation names
- Small improvement in writing a PDF to bytes, inline array objects no longer have trailing newline character
- Small bugfix to fix layout of `FormField` objects using `FlexibleWidthColumnTable`
- Added some imports to `borb/pdf/__init__.py` to make it easier to import `borb` objects in general
- Added the EURion symbol to the `LineArtFactory`
- Added methods to rotate `Shape` and `DisjointShape`