Which parts of PyPDF2 should be part of the public API?

[MartinThoma]

I want to define clearly what parts of PyPDF2 are meant to be used via __all__.

The following is clearly part of that (and should be importable from PyPDF2 directly):

__version__
PageRange
PdfFileMerger
PdfFileReader
PdfFileWriter

Where I'm less sure:

• the pdf module overall, e.g. the ContentStream
• parse_filename_page_ranges

Most likely internal:

• filters.py
• generic.py
• xmp.py