PyPDF2 2.0: Cleanup plan

[MartinThoma]

The PyPDF2 2.X versions should have a clean interface. Here is a list of things I want to do - please comment / add suggestions!

1. ✔️ Distribution Package: Remove Tests, Resources, and Scripts from the package (discussion)
2. ✔️ PEP8 Names: Use PEP8-compliant function/method/variable names. In many cases that means using snake_case instead of camelCase. There will be a PyPDF2 1.X version that has both + deprecation warnings so that people can make the switch gradually. I would also rename Tests to tests.
3. Explicit Exports: PyPDF2 will explicitly export some modules / classes / functions. Those are meant for users of PyPDF2 to be used. Everything else can be changed without deprecation warnings (discussion: Which parts of PyPDF2 should be part of the public API?)
4. Dropping methods in favor of magic methods: I have a couple of ideas for a modern interface to PyPDF2 (links). Those can be introduced additionally and don't cause any technical issues having them side-by-side. However, for users it's simpler when the usage of PyPDF2 looks consistent in examples. For this reason I would favor deprecating the non-magic method functions.
5. ✔️ Dropping overwriteWarnings (done): warnings.showwarning = _showwarning cases issues, see PyPDF2 should not overwrite warnings.formatwarning. #67 - merging Custom implementation of warnings.formatwarning removed #225
6. ✔️ Black code formatting
7. ✔️ strict=False by default (done): see Should strict=False be the default? #783 (comment)
8. ✔️ Private methods / functions: Some methods, like the rectangle accessors mentioned in Unbound & undocumented instance methods in pdf.py (getRectangle, setRectangle, deleteRectangle, createRectangleAccessor) #517 , should have a leading underscore to mark them as private.
9. ✔️ Remove merger.OutlinesObject (done) : Removing PyPDF2.merger.OutlinesObject as nobody seems to use it

PRs that I want to merge with 2.0:

• ENH: Add decrypt support for V5 and AES-128, AES-256 (R5 only) #749 : Better encryption support

[MartinThoma]

Is the PyPDF2.merger.OutlinesObject used by anybody? Can somebody show a small script how it's used?

[sekrause]

I think auto formatting code with something like https://github.com/psf/black is generally a good idea.

However it's usually not a good idea to simply reformat an existing project because it would break all pull requests, so maybe take a look at https://github.com/akaihola/darker which only auto-formats code lines that have changed in a commit (and maybe only start doing that until all old PRs are either merged or rejected).

[MartinThoma]

I'd also deprecate things like RectangleObject.getLowerRight in favor of RectangleObject.lowerRight (using the properties instead of Java-style getter/setter methods - although things don't change under the hood)