Parser for retrieving PubMed metadata

pubmed.py

"""
Lookup PMIDS and parse the returned metadata into the BibJSON format.  

See: http://www.ncbi.nlm.nih.gov/books/NBK25499/#chapter4.ESearch

"""

from lxml import etree
import urllib
from pprint import pprint

class BibJSON(object):
	def __init__(self, pmid):
		self.pmid = pmid
		handle = urllib.urlopen('http://eutils.ncbi.nlm.nih.gov/entrez/eutils/esummary.fcgi?db=pubmed&id=%s' % pmid)
		self.doc = etree.parse(handle)
		self.dsum = self.doc.xpath('/eSummaryResult/DocSum')[0]

	def _x(self, path, text=True):
		if text:
			return self.dsum.xpath('%s/text()' % path)[0]
		else:
			return self.dsum.xpath(path)

	def _name(self, name):
		try:
			return self._x('Item[@Name="%s"][1]' % name)
		except IndexError:
			return ''

	def identifiers(self):
		out = [
			{'type': 'pmid',
			'value': 'pmid:%s' % self.pmid}
		]
		d = {'type': 'doi',
		  'value': "doi:%s" % self._name('DOI')}
		out.append(d)
		d = {}
		for k in ['ISSN', 'ESSN']:
			v = self._name(k)
			if v != '':
				if k == 'ESSN':
					k = 'eissn'
				out.append({'type': k.lower(), 'value': v})

		#Look at article ids section
		aids = self.doc.xpath('/eSummaryResult/DocSum/Item[@Name="ArticleIds"]')[0]
		for aid in aids:
			if aid.get('Name') == 'pmc':
				out.append({'type': 'pmcid', 'value': aid.text})
		#Try pubmed central id
		#pmc = self.xpath
		return out

	def authors(self):
		#author => [{'lastname': 'Frogner', 'name': 'BK Frogner', 'firstname': 'BK'}]
		out = []
		author_list = self.doc.xpath('/eSummaryResult/DocSum/Item[@Name="AuthorList"]')[0]
		for auth in author_list:
			out.append({'name': auth.text})
		return out

	def type(self):
		pub = self.doc.xpath('/eSummaryResult/DocSum/Item/Item[@Name="PubType"]')[0].text
		if pub == 'Journal Article':
			return 'article'
		return pub

	def parse(self):
		def _name(name):
			try:
				return self._x('Item[@Name="%s"][1]' % name)
			except IndexError:
				return ''
		b = {}
		b['type'] = self.type()
		b['title'] = _name('Title')
		b['year'] = _name('PubDate')[:4]
		b['volume'] = _name('Volume')
		b['issue'] = _name('Issue')
		b['pages'] = _name('Pages')
		if b['pages'] != '':
			chunks = b['pages'].split('-')
			b['start_page'] = chunks[0]
			b['end_page'] = chunks[1]
		b['journal'] = {'name': _name('FullJournalName'),
						'shortcode': _name('Source')}
		b['_pubstatus'] = _name('PubStatus')

		b['identifiers'] = self.identifiers()
		b['link'] = []
		for bid in b['identifiers']:
			if bid['type'] == 'pmcid':
				b['link'].append({'anchor': 'Full text from PubMed Central',
								  'url': 'http://www.ncbi.nlm.nih.gov/pmc/articles/%s' % bid['value']})
		if b['link'] == []:
			del b['link']

		b['author'] = self.authors()

		return b

def to_bibjson(pmid):
	d = BibJSON(pmid).parse()
	return d

to_check = [
'6869634',
'9285281',
'8996046'
]


from bibjsontools import to_openurl
for pmid in to_check:
	print pmid
	bib = to_bibjson(pmid)
	pprint(bib)