Should we use 'raise from'?

[MartinThoma]

I'm currently looking at several issues Flake8 points out. One group is this:

./PyPDF2/_reader.py:354:17: R100 raise in except handler without from
./PyPDF2/_reader.py:1000:21: R100 raise in except handler without from
./PyPDF2/_reader.py:1336:17: R100 raise in except handler without from
./PyPDF2/generic.py:591:17: R100 raise in except handler without from
./PyPDF2/generic.py:1914:13: R100 raise in except handler without from

Basically when we capture an exception and raise a new one, we could use raise from (introduced with PEP 3134 in Python 3.0).

I like using raise from in my applications as it gives me a better idea where an issue came from.

However, for users of PyPDF2 I'm not so sure if that will help in any way. When I use PyPDF2 I'm typically only interested in the PDF that caused the issue, not the exact path how it caused an issue.

What do you think?