Divide dependencies into categories about data-describe HOT 3 CLOSED

setup.py is using extras_requires for optional dependencies:

https://github.com/brianray/data-describe/blob/ba1bc43d3cd610563b85d12567ec5676f1a056b2/setup.py#L39-L44

Is this the recommended pattern or is there a different pattern to more explicitly define categories of dependencies? How should requirements.txt be organized?

Within the module imports, is the recommended pattern to update a variable on ImportError? Should there be one try block for each optional dependency, or one per application?

try:
    import pandas
except ImportError as e:
    _PANDAS_INSTALLED = False

from data-describe.

The pattern in requirements.txt looks like a good start. Though module imports need more attention.

Yes the try-except block above is recommended. I think it's better to have one try block for each optional dependency. For example, we can have a module named compat.py and put all the try-except blocks there and finally other modules would import variables like _PANDAS_INSTALLED from compat.py. Within geo module, we check geo module related dependencies.

from data-describe.

Resolved by #109

from data-describe.