MAINT: Use grouped constants instead of literals

[MartinThoma]

This allows us to document what they are good for and to distinguish
literals of the same name but of different contexts,
e.g. /Type in Pages or /Type in Page

The constants are in PyPDF2/pdf_attributes.py

[MartinThoma]

@MasterOdin What do you think about this type of change (it's not done jet; at the moment it's more an example how I imagine it).

I see it like this:

• Linking help texts / context: Having attributes instead of literals allows us to use the editor/IDE to jump to an explanation of the literal. I see this as a massive advantage
• Readability: We need to be careful that things are still readable. So we shouldn't end up with Java-esque lines where every name is super explicit, but you have a hard time grasping the complete line of code as it's simply too long. That is a disadvantage of this approach, but it could be mitigated by abbreviating the names when importing
• Runtime: I hope that it doesn't really make a difference for typical workflows. However, we still don't have any performance tests.

[MasterOdin]

I really like this approach as it makes it harder to have "text misspelling bugs" and as you note it makes it easier to then have reference to relevant documentation on these attributes.

For readability, I think that what you have (alias the long class name to an abbreviation) renders this point largely moot, and that the gain in documentation and not having "magic strings" outweighs this heavily.

Performance is going to be so neglibly affected (if at all really) that I would not worry about it.

[codecov-commenter]

Codecov Report

Merging #745 (74094ea) into main (87aafd6) will increase coverage by 0.95%.
The diff coverage is 86.28%.

@@            Coverage Diff             @@
##             main     #745      +/-   ##
==========================================
+ Coverage   69.64%   70.59%   +0.95%     
==========================================
  Files           9       10       +1     
  Lines        3317     3425     +108     
  Branches      783      798      +15     
==========================================
+ Hits         2310     2418     +108     
  Misses        763      763              
  Partials      244      244              

Impacted Files	Coverage Δ
PyPDF2/pagerange.py	98.11% <ø> (ø)
PyPDF2/filters.py	58.67% <66.66%> (+0.97%)	⬆️
PyPDF2/generic.py	66.85% <70.37%> (+0.04%)	⬆️
PyPDF2/pdf.py	72.42% <82.45%> (+0.03%)	⬆️
PyPDF2/merger.py	72.56% <83.33%> (+0.09%)	⬆️
PyPDF2/__init__.py	100.00% <100.00%> (ø)
PyPDF2/constants.py	100.00% <100.00%> (ø)
PyPDF2/xmp.py	43.09% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 87aafd6...74094ea. Read the comment docs.