Data backend for Contact Maps

[camtauxe]

This PR is a first step towards implementing the contact map widget and addresses the server/database's handling of contact map data.

It includes:

• Rough contact maps for the two structure files in the test.00 project (named hic.tsv in each one).
• The script used to generate these maps from the structure files: bin/structure2hic.pl
• Code in bin/db_pop to load these contact maps into a table in the database.
• A new route in the gtkserver.py to retrieve contact records as JSON.

When the server is running, you can use the following command to query the contact records and view them as a tsv again: curl -sL 'localhost:8000/data/contactmap/1/contacts' | perl -MJSON -0777 -E 'say join "\t", @{$_}{qw(x y value)} for @{ decode_json(<>)->{contacts} }'

NOTE: I wasn't exactly sure how to incorporate the contact map data into the project.json, since it had placeholders for both "hic" data and "md-contact-map" data. But since neither was actually implemented yet, I went ahead and just picked one.

[camtauxe]

( this is associated with issue #1 btw )

[dhrogers]

Looks good.

The md-contact-map is the correct place to put these, as these are files generated by us, and that aren't strict 'hic' format. At the moment, md-contact-map format is loosely defined as a csv-ish file. We should nail this down.

In general, I'm placing data into the following structure: source/ (anything that's raw data) 0/ (dataset ID. A dataset is a group of data. There can be as many of these as we'd like) file.ext file.ext

So I'll approve this request, and then if you could update things so that the md-contact-map files are in the correct place, that would be great. The schema doesn't require this organization, it's just easier to keep track of things in the project if they are semented like this.