obo-parser.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""
A constant-space parser for the GeneOntology OBO v1.2 format

Modified from work by Uli Koehler: http://techoverflow.net/blog/2013/11/18/a-geneontology-obo-v1.2-parser-in-python/
"""
from __future__ import with_statement
from collections import defaultdict
import types

import couchdb
couch = couchdb.Server()
db = couch['doid']

__author__    = "Uli Koehler"
__copyright__ = "Copyright 2013 Uli Koehler"
__license__   = "Apache v2.0"

def processGOTerm(goTerm):
    """
    In an object representing a GO term, replace single-element lists with
    their only member.
    Returns the modified object as a dictionary.
    """
    ret = dict(goTerm) #Input is a defaultdict, might express unexpected behaviour
    for key, value in ret.iteritems():
        if len(value) == 1:
            ret[key] = value[0]
    return ret

def parseGOOBO(filename):
    """
    Parses a Gene Ontology dump in OBO v1.2 format.
    Yields each 
    Keyword arguments:
        filename: The filename to read
    """
    with open(filename, "r") as infile:
        currentGOTerm = None
        for line in infile:
            line = line.strip()
            if not line: continue #Skip empty
            if line == "[Term]":
                if currentGOTerm: yield processGOTerm(currentGOTerm)
                currentGOTerm = defaultdict(list)
            elif line == "[Typedef]":
                #Skip [Typedef sections]
                currentGOTerm = None
            else: #Not [Term]
                #Only process if we're inside a [Term] environment
                if currentGOTerm is None: continue
                key, sep, val = line.partition(":")
                currentGOTerm[key].append(val.strip())
        #Add last term
        if currentGOTerm is not None:
            yield processGOTerm(currentGOTerm)

if __name__ == "__main__":
    """Print out the number of GO objects in the given GO OBO file"""
    import argparse
    parser = argparse.ArgumentParser()
    parser.add_argument('infile', help='The input file in GO OBO v1.2 format.')
    args = parser.parse_args()
    #Iterate over GO terms
    termCounter = 0
    newDocs = []
    for goTerm in parseGOOBO(args.infile):
	newDoc = goTerm
	newDoc['_id'] = newDoc['id']
	if 'is_a' in newDoc:
	    newDoc['parents'] = []
	    if isinstance(newDoc['is_a'], basestring):
		parent = newDoc['is_a']
                newDoc['parents'].append(parent.split('!')[0].strip())
	print newDoc
        newDocs.append(newDoc)
        termCounter += 1
    print "Found %d GO terms" % termCounter
    for doc in db.update(newDocs):
        print repr(doc)