Scripts for XML -> Python object -> YAML conversion (re #129)

[mbollmann]

This PR adds a script that converts the authoritative XML to YAML format, in preparation for using these files as a data source in the static rewrite of the site.

• anthology.py contains classes that read in the XML and represent it as simple Python objects. They might potentially be useful for further checks or conversions as well.

• xml_to_yaml.py outputs one YAML file per XML file with the paper infos, and additionally compiles an author index in a separate YAML file. This is done to facilitate the generation of the static pages, and might be extended with other auxiliary files in the future. Conflation of name variants could, in principle, also be done at this point, depending on how exactly we decide to handle them ( Authors being stored under multiple spellings #86 ).

I will continue to add to and/or modify these scripts during the static rewrite, according to the needs of the page generation.

[mjpost]

Wow, this is really cool!

Are you familiar with the argparse module? It's builtin to Python and is quite nice. Just wondering—maybe you have a preference or reason for docopt.

I think you can just merge it unless you wanted comments.

[mbollmann]

I've used argparse for a long time, but have recently come to prefer docopt for all but very complex CLI interfaces. The main reasons are that it's easier to write, self-documenting, and an implementation exists for most programming languages, so you can apply it everywhere. It's not a strong preference though, argparse is still pretty good. :)