dependencies (html5lib)

[DaveBackus]

I suspect this is a naive suggestion, but here's the issue. I'm teaching a course to newbies (data + python) that uses the datareader a lot. We started with Anaconda, which includes pandas, then installed datareader with conda.

It didn't work because we didn't have html5lib installed. Which was easy to fix once we figured it out. But would it make sense to add a check? The code already gives a message referring to html5lib, but it's buried in the traceback. There are probably similar dependencies elsewhere that might get similar treatment.

The context here is that this is a terrific tool and I can see it used by lots of people, including many who are not experienced coders.

Specifics. This refers to the Yahoo options module.

import pandas_datareader.yahoo.options as yho

ticker = 'amzn'
otk = yho.Options(ticker)

cols = [0, 1, 2, 7]
ops = otk.get_all_data()[cols]

[femtotrader]

lxml is an optional dependency for Pandas DataReader but is "required" for Yahoo Options
see https://github.com/pydata/pandas-datareader/blob/master/pandas_datareader/yahoo/options.py#L665

Maybe we should

• either improve setup.py to be more explicit about it (extras_require might help - see https://pythonhosted.org/setuptools/setuptools.html#declaring-extras-optional-features-with-their-own-dependencies)
• or add lxml to dependencies

We have to weight the pros and cons of each solution

[bashtage]

This is now a requirement.