WIP: Fix to model summary methods

[pstjohn]

This fix resolves #267, in which models that do not conform to the
standard boundary reaction format would cause the model_summary method
to fail. Now the summary method will print the ID of the first
metabolite and raise a warning if a boundary reaction exchanges more
than one species.

[coveralls]

Coverage increased (+0.01%) to 84.328% when pulling c4f2dc3 on pstjohn:summary_fix into 9b279ff on opencobra:master.

[pstjohn]

Good catch -- I'll work on that.
You're right, splitting coupled boundary reactions shouldn't be too tough either

[cdiener]

Awesome. If you need help with anything you know where to find me :)

[pstjohn]

While I'm at it we might as well get to #266, #238, and #240.

How would we feel about exporting table formatting to a third party? It might make the code a bit more extendable / maintainable. tabulate looks promising, but I'm certainly happy to look at others if you have suggestions. (Or if we want to avoid additional dependencies)

[cdiener]

I agree. This kind of formatting usually has hundreds of edge cases one has to consider ("what happens if the name of the reaction is super long", etc.). I like how it's done now with converting everything to a pandas data frame and then writing some formatting for that data frame. tabulate looks good, we just have to check how it deals with long strings and how it plays with pandas (for instance the reaction strings). In terms of dependencies I think everything that does not have compiled code is fine.

[hredestig]

pardon the premature merge, the bug fix was already good, please open new pr for the additional issues you mentioned. Anyway, tabulate looks like good option, prettytable is another but tabulate directly takes pandas df so maybe the better option.

[pstjohn]

Yikes, as @cdiener mentioned there was still an issue with assigning input and output fluxes..
I'll try to send in another PR soon that fixes the current master then.

[hredestig]

ouch, yes mea culpa :( This also reminds me I think it would be good to start doing merges to devel first, and only upon release to master..

[cdiener]

I think there is no problem here. The bug I reported initially was already fixed here and the rest can be in another pull req. I think bug fix pull requests are fine in master, otherwise master would just be a glorified tag for the latest release :)

[pstjohn]

Sounds great. Regardless, it shouldn't be too long until I have a newer implementation ready.
Tabulate seems to be working great, and should make maintenance for these methods a lot easier going forward.