def _convert_wos_record(record, ns):
"""
Takes an XML tree of a single WoS record and returns a dictionary
representing the record's information.
Args:
record: An XML tree, where the root is the 'REC' tag
ns: A dictionary that contains the key 'ns' whose value is the null XML namespace.
Returns:
A dictionary representing the record's information
with the following keys and value types:
wosid: unicode
title: unicode
journal: unicode
issue: unicode
volume: unicode
pubdate: int
institutions: {int: (unicode, [unicode])}
authors: [(unicode, [int])]
citcount: int
"""
r = dict()
r['wosid'] = xpath_str(record, "ns:UID/text()", ns)
r['title'] = xpath_str(record, "ns:static_data/ns:summary/ns:titles/ns:title[@type='item']/text()",ns)
r['journal'] = xpath_str(record, "ns:static_data/ns:summary/ns:titles/ns:title[@type='source']/text()", ns)
pubinfo = record.xpath("ns:static_data/ns:summary/ns:pub_info", namespaces=ns)[0]
(r['issue'], r['volume'], pubdate) = (pubinfo.attrib.get('issue'), pubinfo.attrib.get('vol'), pubinfo.attrib.get('sortdate'))
if pubdate:
m = _date_re.match(pubdate)
r['pubdate'] = int(m.group('yr') + m.group('mon') + m.group('day'))
r['institutions'] = {}
num_institutions = int(record.xpath("ns:static_data/ns:fullrecord_metadata/ns:addresses", namespaces=ns)[0].attrib['count'])
for institution_tag in record.xpath("ns:static_data/ns:fullrecord_metadata/ns:addresses/ns:address_name/ns:address_spec", namespaces=ns):
index = int(institution_tag.attrib['addr_no'])
address = xpath_str(institution_tag, "ns:full_address/text()", ns)
organizations = xpath_strs(institution_tag, "ns:organizations/ns:organization/text()", ns)
r['institutions'][index] = (address, organizations)
r['authors'] = []
num_authors = int(record.xpath("ns:static_data/ns:summary/ns:names", namespaces=ns)[0].attrib['count'])
for i in range(1, num_authors + 1):
author_tag = record.xpath("ns:static_data/ns:summary/ns:names/ns:name[@seq_no='%d']" % i, namespaces=ns)[0]
author_name = xpath_str(author_tag, "ns:wos_standard/text()", ns)
if author_name == None: continue
affiliation_indices = map(int, author_tag.attrib['addr_no'].split(' ')) if 'addr_no' in author_tag.attrib else None
r['authors'].append((author_name, affiliation_indices))
cittag = record.xpath("ns:dynamic_data/ns:citation_related/ns:tc_list/ns:silo_tc[@coll_id='WOS']/@local_count", namespaces=ns)
if cittag:
r['citcount'] = int(cittag[0])
return r
def _convert_wos_record(record, ns):
"""
Takes an XML tree of a single WoS record and returns a dictionary
representing the record's information.
Args:
record: An XML tree, where the root is the 'REC' tag
ns: A dictionary that contains the key 'ns' whose value is the null XML namespace.
Returns:
A dictionary representing the record's information
with the following keys and value types:
wosid: unicode
title: unicode
journal: unicode
issue: unicode
volume: unicode
pubdate: int
institutions: {int: (unicode, [unicode])}
authors: [(unicode, [int])]
citcount: int
"""
r = dict()
r['wosid'] = xpath_str(record, "ns:UID/text()", ns)
r['title'] = xpath_str(
record,
"ns:static_data/ns:summary/ns:titles/ns:title[@type='item']/text()",
ns)
r['journal'] = xpath_str(
record,
"ns:static_data/ns:summary/ns:titles/ns:title[@type='source']/text()",
ns)
pubinfo = record.xpath("ns:static_data/ns:summary/ns:pub_info",
namespaces=ns)[0]
(r['issue'], r['volume'], pubdate) = (pubinfo.attrib.get('issue'),
pubinfo.attrib.get('vol'),
pubinfo.attrib.get('sortdate'))
if pubdate:
m = _date_re.match(pubdate)
r['pubdate'] = int(m.group('yr') + m.group('mon') + m.group('day'))
r['institutions'] = {}
num_institutions = int(
record.xpath("ns:static_data/ns:fullrecord_metadata/ns:addresses",
namespaces=ns)[0].attrib['count'])
for institution_tag in record.xpath(
"ns:static_data/ns:fullrecord_metadata/ns:addresses/ns:address_name/ns:address_spec",
namespaces=ns):
index = int(institution_tag.attrib['addr_no'])
address = xpath_str(institution_tag, "ns:full_address/text()", ns)
organizations = xpath_strs(institution_tag,
"ns:organizations/ns:organization/text()",
ns)
r['institutions'][index] = (address, organizations)
r['authors'] = []
num_authors = int(
record.xpath("ns:static_data/ns:summary/ns:names",
namespaces=ns)[0].attrib['count'])
for i in range(1, num_authors + 1):
author_tag = record.xpath(
"ns:static_data/ns:summary/ns:names/ns:name[@seq_no='%d']" % i,
namespaces=ns)[0]
author_name = xpath_str(author_tag, "ns:wos_standard/text()", ns)
if author_name == None: continue
affiliation_indices = map(int, author_tag.attrib['addr_no'].split(
' ')) if 'addr_no' in author_tag.attrib else None
r['authors'].append((author_name, affiliation_indices))
cittag = record.xpath(
"ns:dynamic_data/ns:citation_related/ns:tc_list/ns:silo_tc[@coll_id='WOS']/@local_count",
namespaces=ns)
if cittag:
r['citcount'] = int(cittag[0])
return r
def _article_to_pubmed_ref(article):
'''Convert PubMed XML data about an article into a ref (dictionary containing the article data).
The returned dictionary will contain this:
{
"pmid": a string containing the article's PMID
"authors": a list of tuples (string, integer), where the first element is the author's name and the second is the author's affiliation if known (otherwise it's None)
"institutions": a dictionary, where the key is an integer and value is a list of strings contanining the institutional hierarchy
"title": the article's title
"pubdate": an integer of the form 19850726 (i.e., 1985/07/26)
"year": the publication year as integer
"journal": a string
"grantagencies": a list of strings
"pubtypes": a list of strings specifying the publication types as per PubMed
"meshterms": a nested list of strings specifying the MeSH terms as per PubMed
}'''
r = {}
r['pmid'] = xpath_str(
article,
'PubmedData/ArticleIdList/ArticleId[@IdType=\'pubmed\']/text()')
institutions = {}
authors = []
for author in article.xpath('MedlineCitation/Article/AuthorList/Author'):
lastname = xpath_str(author, 'LastName/text()')
initials = xpath_str(author, 'Initials/text()')
if lastname and initials:
name = lastname + u' ' + initials
else:
continue
institution_address = xpath_str(author, 'Affiliation/text()')
institution_index = len(
institutions) + 1 if institution_address else None
if institution_address:
institutions[institution_index] = (institution_address, None)
authors.append((name, institution_index))
r['authors'] = authors
r['institutions'] = institutions
r['title'] = xpath_str(article,
'MedlineCitation/Article/ArticleTitle/text()')
pubdate_str = u''
pubdate_elem = article.xpath(
'PubmedData/History/PubMedPubDate[@PubStatus="pubmed"]')[0]
pubdate_yr = xpath_str(pubdate_elem, 'Year/text()')
if pubdate_yr:
pubdate_str += pubdate_yr
pubdate_mon = xpath_str(pubdate_elem, 'Month/text()')
if pubdate_mon:
pubdate_str += '%02d' % int(pubdate_mon)
pubdate_day = xpath_str(pubdate_elem, 'Day/text()')
if pubdate_day:
pubdate_str += '%02d' % int(pubdate_day)
else:
pubdate_str += '00'
else:
pubdate_str += '0000'
r['pubdate'] = int(pubdate_str) if pubdate_str else None
r['year'] = pubdate_yr
r['journal'] = xpath_str(
article, 'MedlineCitation/MedlineJournalInfo/MedlineTA/text()')
r['grantagencies'] = xpath_strs(
article,
'MedlineCitation/Article/GrantList[last()]/Grant/Agency/text()')
r['pubtypes'] = xpath_strs(
article,
'MedlineCitation/Article/PublicationTypeList/PublicationType/text()')
allterms = []
for meshheading in article.xpath(
'MedlineCitation/MeshHeadingList/MeshHeading'):
terms = xpath_strs(meshheading,
'DescriptorName/text() | QualifierName/text()')
allterms.append(terms)
r['meshterms'] = allterms
return r
def _article_to_pubmed_ref(article):
'''Convert PubMed XML data about an article into a ref (dictionary containing the article data).
The returned dictionary will contain this:
{
"pmid": a string containing the article's PMID
"authors": a list of tuples (string, integer), where the first element is the author's name and the second is the author's affiliation if known (otherwise it's None)
"institutions": a dictionary, where the key is an integer and value is a list of strings contanining the institutional hierarchy
"title": the article's title
"pubdate": an integer of the form 19850726 (i.e., 1985/07/26)
"year": the publication year as integer
"journal": a string
"grantagencies": a list of strings
"pubtypes": a list of strings specifying the publication types as per PubMed
"meshterms": a nested list of strings specifying the MeSH terms as per PubMed
}'''
r = {}
r['pmid'] = xpath_str(article, 'PubmedData/ArticleIdList/ArticleId[@IdType=\'pubmed\']/text()')
institutions = {}
authors = []
for author in article.xpath('MedlineCitation/Article/AuthorList/Author'):
lastname = xpath_str(author, 'LastName/text()')
initials = xpath_str(author, 'Initials/text()')
if lastname and initials:
name = lastname + u' ' + initials
else:
continue
institution_address = xpath_str(author, 'Affiliation/text()')
institution_index = len(institutions) + 1 if institution_address else None
if institution_address:
institutions[institution_index] = (institution_address, None)
authors.append((name, institution_index))
r['authors'] = authors
r['institutions'] = institutions
r['title'] = xpath_str(article, 'MedlineCitation/Article/ArticleTitle/text()')
pubdate_str = u''
pubdate_elem = article.xpath('PubmedData/History/PubMedPubDate[@PubStatus="pubmed"]')[0]
pubdate_yr = xpath_str(pubdate_elem, 'Year/text()')
if pubdate_yr:
pubdate_str += pubdate_yr
pubdate_mon = xpath_str(pubdate_elem, 'Month/text()')
if pubdate_mon:
pubdate_str += '%02d' % int(pubdate_mon)
pubdate_day = xpath_str(pubdate_elem, 'Day/text()')
if pubdate_day:
pubdate_str += '%02d' % int(pubdate_day)
else:
pubdate_str += '00'
else:
pubdate_str += '0000'
r['pubdate'] = int(pubdate_str) if pubdate_str else None
r['year'] = pubdate_yr
r['journal'] = xpath_str(article, 'MedlineCitation/MedlineJournalInfo/MedlineTA/text()')
r['grantagencies'] = xpath_strs(article, 'MedlineCitation/Article/GrantList[last()]/Grant/Agency/text()')
r['pubtypes'] = xpath_strs(article, 'MedlineCitation/Article/PublicationTypeList/PublicationType/text()')
allterms = []
for meshheading in article.xpath('MedlineCitation/MeshHeadingList/MeshHeading'):
terms = xpath_strs(meshheading, 'DescriptorName/text() | QualifierName/text()')
allterms.append(terms)
r['meshterms'] = allterms
return r