On 2020/04/25 21:16, Mikolaj Kucharski wrote:
> Hi,
>
> I did some quick testing of some random code from GitHub, which uses
> pandas Python module and it failed with:
>
> holders = _pd.read_html(url)
> File "/usr/local/lib/python3.7/site-packages/pandas/io/html.py", line 1100,
> in read_html
> displayed_only=displayed_only,
> File "/usr/local/lib/python3.7/site-packages/pandas/io/html.py", line 891,
> in _parse
> parser = _parser_dispatch(flav)
> File "/usr/local/lib/python3.7/site-packages/pandas/io/html.py", line 848,
> in _parser_dispatch
> raise ImportError("lxml not found, please install it")
> ImportError: lxml not found, please install it
>
> Above `_pd` is pandas. After adding py3-lxml package problems goes
> away. I can provide a patch and add a run dep. Does it make sense?
>
> --
> Regards,
> Mikolaj
>
there's a whole section on this in doc/source/getting_started/install.rst
explaining the 4 different options of combinations of libraries that you
need for parsing html.. looks like the port is taking the approach of
setting deps for the basics and leaving the user to pick any extras.
since there's 30+ optional dependencies this seems a reasonable approach
but maybe it's worth adding a note somewhere about this.
