ENC: better pandas typesetting in ipython nbconvert --to latex

[lmarti]

Currently, when IPython notebooks are converted to latex, pandas DataFrames are appear as raw (verbatim) text. One simple solution is to include a _repl_latex_ function. I have tried this solution and works. See ipython/ipython#8261

[shoyer]

This is a nice idea!

Could you please add a basic test? Could be something as simple as assert ds._repl_latex_() == ds.to_latex(), just so we know this doesn't get broken in the future.

Also, this will need a release note in the what's new:
http://pandas-docs.github.io/pandas-docs-travis/contributing.html#documenting-your-code

[shoyer]

can you add a comment to note that this is for ipython? adding a reference to this GH issue would probably also be a good idea.

[jorisvandenbossche]

Nice idea indeed, I was only wondering if there is a way to make this configurable? Maybe on the nbconvert side?
As I can imagine you don't always want a full tex table, eg when writing a tutorial where you want to have it look like console input/output

[lmarti]

Well, yes... I was thinking of a simple fix... but it would be nice to be able to pass options in the way that you can pass them no to to_latex. Maybe store those options in a singleton and reuse them? Is there any similar situation in pandas?

[shoyer]

We already have pd.set_options and pd.option_context: http://pandas.pydata.org/pandas-docs/version/0.16.0/options.html

So that would be a pretty easy way to add this customization if you're interested. But (from my perspective), that shouldn't be a requirement for this PR.

[jreback]

@lmarti can you update with a test and options as described above

[lmarti]

@jreback yes! in the next couple of days. Sorry for the delay.

[jreback]

@lmarti can you add a test

[jreback]

@lmarti can you add a basic tests for this?

[jreback]

@lmarti pls update according to the comments

[jreback]

@lmarti can you update according to comments

[lmarti]

I have sent a new PR (#11399) that contains all the requested modifications. Thanks for the wait.