import xml.etree.ElementTree as ET
from xml.etree.ElementTree import XML
from StringIO import StringIO
import urllib, urllib2, gzip, re

query = "HIV"
searchUrl = 'http://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi'
fetchUrl = 'ftp://ftp-private.ncbi.nlm.nih.gov/pubchem/Bioassay/CSV/Description/%s.descr.xml.gz'
params = urllib.urlencode({'db':'pcassay', 'retmode':'xml', 'term':query})
req = urllib2.Request(searchUrl, params)
root = XML(urllib2.urlopen(req).read())

## get the assay ID's
ids = [x.text for x in root.findall('IdList/Id')]

def addns(path, ns):
    path = re.sub('/', '/{%s}' % (ns), path)
    path = '{%s}%s' % (ns,path)
    return path

## get the assay documents
def getProteinID(aid):
    req = urllib2.Request(fetchUrl % (aid))
    resp = urllib2.urlopen(req).read()
    root = XML(gzip.GzipFile('', 'r', 0, StringIO(resp)).read())
    path = addns('PC-AssaySubmit/PC-AssaySubmit_assay/PC-AssaySubmit_assay_descr/PC-AssayDescription/PC-AssayDescription_target/PC-AssayTargetInfo/PC-AssayTargetInfo_mol-id', 'http://www.ncbi.nlm.nih.gov')
    pid = root.find(path)
    if pid is not None: return pid.text 

## get the protein ID's
pids = [getProteinID(x) for x in ids]
print pids