def test_xml_parsing_endnote():
xp = EndnoteXMLParser(ENDNOTE_DATA)
documents = xp.process()
assert documents is not None
assert len(documents['authors']) == 3
assert len(documents['journals']) == 1
assert len(documents['keywords']) == 3
def create_index(self, xmldata):
xp = EndnoteXMLParser(xmldata)
documents = xp.process()
# Make the index directories
if not os.path.exists(self.index_dir):
os.mkdir(self.index_dir)
if not os.path.exists(self.authors_dir):
os.mkdir(self.authors_dir)
if not os.path.exists(self.keywords_dir):
os.mkdir(self.keywords_dir)
if not os.path.exists(self.journals_dir):
os.mkdir(self.journals_dir)
self.index_authors = create_in(self.authors_dir, self.schema_authors)
self.index_keywords = create_in(self.keywords_dir, self.schema_keywords)
self.index_journals = create_in(self.journals_dir, self.schema_journals)
# Authors
writer = self.index_authors.writer()
for a in documents["authors"]:
writer.add_document(author=smart_unicode(a))
writer.commit()
# Keywords
writer = self.index_keywords.writer()
for k in documents["keywords"]:
writer.add_document(keyword=smart_unicode(k))
writer.commit()
# Journals
writer = self.index_journals.writer()
for j in documents["journals"]:
writer.add_document(journal=smart_unicode(j))
writer.commit()
