BUG: Escape < and > in info _repr_html_

[TomAugspurger]

Closes #7105

Before: http://nbviewer.ipython.org/gist/TomAugspurger/4c83a646ffb93d5ee978
After: http://nbviewer.ipython.org/gist/TomAugspurger/81bf15086864ca090bb6

After conosle:

In [28]: df
 Out[5]: 
<class 'pandas.core.frame.DataFrame'>
Int64Index: 2 entries, 0 to 1
Data columns (total 2 columns):
0    2 non-null float64
1    2 non-null float64
dtypes: float64(2)

FYI, I don't know HTML, but I think this is mostly right.

[jreback]

this going to be very slow I think, maybe do this where the class is actually put their (if you know its html?).

[jreback]

never mind, this is just info repr which is almost always short...

[TomAugspurger]

I'm doing a bit of profiling right now. I was hoping that adding the limit of 1 to .replace would only read the str until it replaces, and I think it might:

In [11]: s = 'abc' + 'S' + ('d' * 1000000)

In [12]: %timeit s.replace('S', 's')
10000 loops, best of 3: 145 µs per loop

In [13]: %timeit s.replace('S', 's', 1)
10000 loops, best of 3: 85.2 µs per loop

I'll have a bit more soon.

[jreback]

rebase on master and you can merge this in when you are ready, I put up the display changes section: 8a8ead6

[TomAugspurger]

Perf change vs. master

Doesn't look like much. The x axis is the size of an NxN DataFrame of floats in log base 2.

[TomAugspurger]

@jreback done. Will just wait on Travis.

[TomAugspurger]

Merged via 2bb2a8d