index samples by sample name for easier access.

[brentp]

No description provided.

[brentp]

SHA: c1af7f5

should be a separate pull request, but I can't see how to do that. It adds a genotypes() method so that one can get the un-encoded genotypes from the encoded ones.
This might work better if Record were a class rather than a named tuple, then we could add methods to record.

[jamescasbon]

A good case for an OrderedDict so that we can access by name, but iterating the values provides the order as in the file?
http://docs.python.org/dev/library/collections.html#collections.OrderedDict

[brentp]

Seems reasonable, especially since it's usually going to be a fairly small dictionary. I'll add this to the pull request later today.

[brentp]

although OrderedDicts are not in python2.6 collections

[jamescasbon]

Always with your backwards compatibility! Oh well, it was a nice to have, but version for 2.6 listed here:
http://stackoverflow.com/questions/1617078/ordereddict-for-older-versions-of-python

not sure its worth a dependency.

[jamescasbon]

Please continue issue at jamescasbon#1

[jamescasbon]

A quick note since everyone is watching this issue.

I have fixed this on my fork using OrderedDict from collections. If python<2.7 setup.py should install some code to do that.