def pubmed(gi,ids,query):
"""
Get the pubmed articles listed by *ids
"""
_ids=",".join(ids)
for id in ids:
handle = efetch(db="pubmed",id=id,retmode='xml',rettype='xml',retmax=MAX_RETURN)
try:
#print handle.read()
results = eread(handle)
for citation in results:
#runtime().debug(citation.keys())
citation = citation['MedlineCitation']
pmid = citation['PMID']
article = citation['Article']
title = article['ArticleTitle']
journal = article['Journal']['Title']
try:
date = citation['DateCompleted'] if citation.has_key('DateCompleted') else citation['DateCreated']
year = date['Year']
month = date['Month']
day = date['Day']
datetime = "%s-%s-%s" % (year,month,day)
except:
datetime = '0000-00-00'
runtime().debug("Parsed pmid:%s" % id)
yield Citation(gi, pmid, title, journal, datetime, query)
except:
runtime().debug("Failure fetching pmid:%s" % id)
continue
finally:
handle.close()
def publications(gi):
"""
Grab the list of publications for a given protein
"""
gi_names = list(names(gi))
if len(gi_names) == 0:
return
#runtime().debug("NAMES",gi_names)
term = " OR ".join(['"%s"' %n for n in gi_names])
runtime().debug(term)
handle = esearch(db="pubmed",
term=term,
retmax=MAX_RETURN,
mindate="2000")
try:
results = eread(handle)
finally:
handle.close()
results[COUNT] = int(results[COUNT])
runtime().debug("Found",results[COUNT])
if results[COUNT]>0:
runtime().debug(gi,term,"\t",results[COUNT], "\n\t",results[IDLIST])
return list(pubmed(gi,set(results[IDLIST]),term))