tab complete lurking groupby methods

[cpcloud]

Would be nice to have tab completion for groupby methods that are "lazily determined" (not sure if that is the correct terminology)

see related issue #4500

see this SO question for an example (also linked to in #4500 )

[jreback]

shouldn't this function definition go in core/base.py and then can be run on individual class/modules as needed?
@cpcloud @jtratner (this is bottom of core/frame.py and others now

def install_ipython_completers():  # pragma: no cover
    """Register the DataFrame type with IPython's tab completion machinery, so
    that it knows about accessing column names as attributes."""
    from IPython.utils.generics import complete_object

    @complete_object.when_type(DataFrame)
    def complete_dataframe(obj, prev_completions):
        return prev_completions + [c for c in obj.columns
                                   if isinstance(c, compat.string_types) and compat.isidentifier(c)]


# Importing IPython brings in about 200 modules, so we want to avoid it unless
# we're in IPython (when those modules are loaded anyway).
if "IPython" in sys.modules:  # pragma: no cover
    try:
        install_ipython_completers()
    except Exception:
        pass

[cpcloud]

fyi these would have to be everything on the current whitelist plus whatever attributes there are (which i think is automatically done by ipython)

[jtratner]

Couldn't we just define dir instead? Any reason not to do that?

[jtratner]

Er __dir__ rather

[cpcloud]

yep looks like that's the way to do it