def __init__(self):

super(NLMParser, self).__init__()

def get_references(self, xml):

references = []

for reference in xml.getElementsByTagName("ref"):

plain_text = None

ref_type = reference.getElementsByTagName('citation')[0].getAttribute('publication-type').encode('utf-8')

label = get_value_in_tag(reference, "label").strip('.')

authors = []

for author in reference.getElementsByTagName("name"):

given_name = get_value_in_tag(author, "given-names")

surname = get_value_in_tag(author, "surname")

if given_name:

name = "%s, %s" % (surname, given_name)

else:

name = surname

if name.strip().split() == []:

name = get_value_in_tag(author, "string-name")

authors.append(name)

doi_tag = reference.getElementsByTagName("pub-id")

doi = ""

for tag in doi_tag:

if tag.getAttribute("pub-id-type") == "doi":

doi = xml_to_text(tag)

issue = get_value_in_tag(reference, "issue")

page = get_value_in_tag(reference, "fpage")

page_last = get_value_in_tag(reference, "lpage")

title = get_value_in_tag(reference, "source")

volume = get_value_in_tag(reference, "volume")

year = get_value_in_tag(reference, "year")

ext_link = format_arxiv_id(super(NLMParser, self).get_ref_link(reference, "arxiv"))

if ref_type != 'journal':

plain_text = get_value_in_tag(reference, "mixed-citation")

references.append((label, authors, doi, issue, page, page_last, title, volume, year, ext_link, plain_text))

self.references = references

def get_record(self, f_path, publisher=None, collection=None, logger=None):

xml = super(NLMParser, self).get_article(f_path)

rec = {}

title = super(NLMParser, self).get_title(xml)

if title:

record_add_field(rec, '245', subfields=[('a', title)])

record_add_field(rec, '260', subfields=[('c', time.strftime('%Y-%m-%d'))])

journal, issn, volume, issue, first_page, last_page, year, doi = super(NLMParser, self).get_publication_information(xml)

journal = "PTEP" ## Let's override the journal information

if logger:

logger.info("Creating record: %s %s" % (join(f_path, pardir), doi))

if doi:

record_add_field(rec, '024', ind1='7', subfields=[('a', doi), ('2', 'DOI')])

authors = super(NLMParser, self).get_authors(xml)

first_author = True

for author in authors:

if author.get('surname'):

subfields = [('a', '%s, %s' % (author.get('surname'), author.get('given_name') or author.get('initials', '')))]

else:

subfields = [('a', '%s' % (author.get('name', '')))]

if 'orcid' in author:

subfields.append(('j', author['orcid']))

if 'affiliation' in author:

for aff in author["affiliation"]:

subfields.append(('v', aff))

if author.get('email'):

subfields.append(('m', author['email']))

if first_author:

record_add_field(rec, '100', subfields=subfields)

first_author = False

else:

record_add_field(rec, '700', subfields=subfields)

abstract = super(NLMParser, self).get_abstract(xml)

if abstract:

record_add_field(rec, '520', subfields=[('a', abstract), ('9', publisher)])

record_add_field(rec, '540', subfields=[('a', 'CC-BY-3.0'), ('u', 'http://creativecommons.org/licenses/by/3.0/')])

copyright = super(NLMParser, self).get_copyright(xml)

if copyright:

record_add_field(rec, '542', subfields=[('f', copyright)])

keywords = super(NLMParser, self).get_keywords(xml)

if keywords['pacs']:

for keyword in keywords['pacs']:

record_add_field(rec, '084', ind1='1', subfields=[('a', keyword), ('9', 'PACS')])

## Oxford is giving us bad keywords. Better ignore them.

#if keywords['other']:

#for keyword in keywords['other']:

#record_add_field(rec, '653', ind1='1', subfields=[('a', keyword), ('9', 'author')])

record_add_field(rec, '773', subfields=[('p', journal), ('v', volume), ('n', issue), ('c', '%s-%s' % (first_page, last_page)), ('y', year)])

self.get_references(xml)

for label, authors, doi, issue, page, page_last, title, volume, year, ext_link, plain_text in self.references:

subfields = []

if doi:

subfields.append(('a', doi))

for author in authors:

subfields.append(('h', author))

if issue:

subfields.append(('n', issue))

if label:

subfields.append(('o', label))

if year:

subfields.append(('y', year))

if ext_link:

subfields.append(('r', ext_link))

# should we be strict about it?

if title and volume and year and page:

subfields.append(('s', '%s %s (%s) %s' % (title, volume, year, page)))

elif not plain_text:

subfields.append(('m', ('%s %s %s %s' % (title, volume, year, page))))

if plain_text:

subfields.append(('m', plain_text))

if subfields:

record_add_field(rec, '999', ind1='C', ind2='5', subfields=subfields)

f_path_pdf = f_path[:-(len('.xml'))] + '.pdf'

f_path_pdfa = join(dirname(f_path), 'archival_pdfs', basename(f_path)[:-len('.xml')] + '-hires.pdf')

record_add_field(rec, 'FFT', subfields=[('a', f_path), ('n', 'main')])

record_add_field(rec, 'FFT', subfields=[('a', f_path_pdf), ('n', 'main')])

record_add_field(rec, 'FFT', subfields=[('a', f_path_pdfa), ('n', 'main'), ('f', '.pdf;pdfa')])

extra_subfields = []

if collection:

extra_subfields.append(('a', collection))

if publisher:

extra_subfields.append(('b', publisher))

record_add_field(rec, '980', subfields=extra_subfields)