Why is Chardet a dependency?

[dhensen]

Because of strict licensing regulations at our company we can not use pyexcel, because chardet has the LGPL license.
Upon inspection I see that nowhere the chardet library is imported. When I clone this project, switch to v0.7.0 branch and remove the chardet depdendency everywhere, I can run a succesfull test suite.

Can chardet be removed or is there a reason that chardet has to remain as a dependency of pyexcel?

[dhensen]

When searching for "chardet" in your whole organization I find a single occurence of chardet actually being used in python code:

https://github.com/pyexcel/pyexcel-io/blob/c7039124b2ab2d069f9955e38d10ac790c3c5389/pyexcel_io/readers/csvz.py#L46

note that this in pyexcel-io package

Did you maybe forget to move chardet to the requirements.txt for pyexcel-io? it's not in there.

[dhensen]

@chfw if you have some spare time it would be really appreciated if you could take a look at this one

[dhensen]

@chfw I think I know what happened:
In #247 I think you were adding github workflows to run your tests, I think your pipeline then failed complaining about not having chardet.

I think you then added chardet to pyexcel/pyexcel, but this code is only used by pyexcel/pyexcel-io. I think it should have been added as a requirement for that package and not toplevel on pyexcel/pyexcel.

Our problem is that we (currently I work for a corporate fin. comp) can not use LGPL packages, so we can not use chardet.

To fix this:

1. move dependency from pyexcel/pyexcel to pyexcel/pyexcel-io
2. mark chardet for pyexcel/pyexcel-io as extras package
3. try/except around import chardet and supply a basic encoding for csvz/tsvz zipfile encode guessing. I can do a PR for this