def test_do_not_add_bibtex_citation(self):
converter = OrcidConverter(
record=self.inspire_record,
url_pattern='http://inspirehep.net/record/{recid}',
)
xml_root = converter.get_xml()
top_level_tags = [
etree.QName(node).localname for node in xml_root.getchildren()
]
assert 'citation' not in top_level_tags
def calculate_hash_for_record(record):
"""Generate hash for an ORCID-serialised HEP record
Args:
record (dict): HEP record
Returns:
string: hash of the record
"""
orcid_rec = OrcidConverter(record, app.config['LEGACY_RECORD_URL_PATTERN'])
return hash_xml_element(orcid_rec.get_xml())
def test_format_thesis(app, api_client):
response = api_client.get('/literature/1395663')
assert response.status_code == 200
phdthesis = json.loads(response.data)
expected = xml_parse("""
<work:work xmlns:common="http://www.orcid.org/ns/common" xmlns:work="http://www.orcid.org/ns/work">
<work:title>
<common:title>MAGIC $\\gamma$-ray observations of distant AGN and a study of source variability and the extragalactic background light using FERMI and air Cherenkov telescopes</common:title>
</work:title>
<work:type>dissertation</work:type>
<work:url>http://inspirehep.net/record/1395663</work:url>
<work:contributors>
<work:contributor>
<work:credit-name>Mankuzhiyil, Nijil</work:credit-name>
<work:contributor-attributes>
<work:contributor-sequence>first</work:contributor-sequence>
<work:contributor-role>author</work:contributor-role>
</work:contributor-attributes>
</work:contributor>
</work:contributors>
</work:work>
""")
result = OrcidConverter(
phdthesis['metadata'],
url_pattern='http://inspirehep.net/record/{recid}',
).get_xml()
assert valid_against_schema(result)
assert xml_compare(expected, result)
assert hash_xml_element(expected) == hash_xml_element(result)
def test_format_book(app, api_client):
response = api_client.get('/literature/736770')
assert response.status_code == 200
book = json.loads(response.data)
expected = xml_parse("""
<work:work xmlns:common="http://www.orcid.org/ns/common" xmlns:work="http://www.orcid.org/ns/work">
<work:title>
<common:title>Differential geometry and Lie groups for physicists</common:title>
</work:title>
<work:type>book</work:type>
<common:publication-date>
<common:year>2011</common:year>
<common:month>03</common:month>
<common:day>03</common:day>
</common:publication-date>
<common:external-ids>
<common:external-id>
<common:external-id-type>isbn</common:external-id-type>
<common:external-id-value>9780521187961</common:external-id-value>
</common:external-id>
<common:external-id>
<common:external-id-type>isbn</common:external-id-type>
<common:external-id-value>9780521845076</common:external-id-value>
</common:external-id>
<common:external-id>
<common:external-id-type>isbn</common:external-id-type>
<common:external-id-value>9780511242960</common:external-id-value>
</common:external-id>
</common:external-ids>
<work:url>http://inspirehep.net/record/736770</work:url>
<work:contributors>
<work:contributor>
<work:credit-name>Fecko, M.</work:credit-name>
<work:contributor-attributes>
<work:contributor-sequence>first</work:contributor-sequence>
<work:contributor-role>author</work:contributor-role>
</work:contributor-attributes>
</work:contributor>
</work:contributors>
</work:work>
""")
result = OrcidConverter(
book['metadata'],
url_pattern='http://inspirehep.net/record/{recid}',
).get_xml()
assert valid_against_schema(result)
assert xml_compare(expected, result)
assert _OrcidHasher(
mock.Mock())._hash_xml_element(expected) == _OrcidHasher(
mock.Mock())._hash_xml_element(result)
def test_format_proceedings(app, api_client):
response = api_client.get('/literature/701585')
assert response.status_code == 200
proceedings = json.loads(response.data)
expected = xml_parse("""
<work:work xmlns:common="http://www.orcid.org/ns/common" xmlns:work="http://www.orcid.org/ns/work">
<work:title>
<common:title>HERA and the LHC: A Workshop on the implications of HERA for LHC physics: Proceedings Part A</common:title>
</work:title>
<work:type>edited-book</work:type>
<common:publication-date>
<common:year>2005</common:year>
</common:publication-date>
<common:external-ids>
<common:external-id>
<common:external-id-type>arxiv</common:external-id-type>
<common:external-id-value>hep-ph/0601012</common:external-id-value>
<common:external-id-url>http://arxiv.org/abs/hep-ph/0601012</common:external-id-url>
<common:external-id-relationship>self</common:external-id-relationship>
</common:external-id>
</common:external-ids>
<work:url>http://inspirehep.net/record/701585</work:url>
<work:contributors>
<work:contributor>
<work:credit-name>De Roeck, A.</work:credit-name>
<work:contributor-attributes>
<work:contributor-sequence>first</work:contributor-sequence>
<work:contributor-role>editor</work:contributor-role>
</work:contributor-attributes>
</work:contributor>
<work:contributor>
<work:credit-name>Jung, H.</work:credit-name>
<work:contributor-attributes>
<work:contributor-sequence>additional</work:contributor-sequence>
<work:contributor-role>editor</work:contributor-role>
</work:contributor-attributes>
</work:contributor>
</work:contributors>
</work:work>
""")
result = OrcidConverter(
proceedings['metadata'],
url_pattern='http://inspirehep.net/record/{recid}',
).get_xml()
assert valid_against_schema(result)
assert xml_compare(expected, result)
assert _OrcidHasher(
mock.Mock())._hash_xml_element(expected) == _OrcidHasher(
mock.Mock())._hash_xml_element(result)
def test_format_thesis_with_author_orcid(app, api_client):
response = api_client.get('/literature/1395663')
assert response.status_code == 200
phdthesis = json.loads(response.data)
phdthesis['metadata']['authors'][0]['ids'] = [{
'schema':
'ORCID',
'value':
'0000-0002-1825-0097',
}]
phdthesis['metadata']['authors'][0]['emails'] = ['*****@*****.**']
expected = xml_parse("""
<work:work xmlns:common="http://www.orcid.org/ns/common" xmlns:work="http://www.orcid.org/ns/work">
<work:title>
<common:title>MAGIC $\\gamma$-ray observations of distant AGN and a study of source variability and the extragalactic background light using FERMI and air Cherenkov telescopes</common:title>
</work:title>
<work:type>dissertation</work:type>
<work:url>http://inspirehep.net/record/1395663</work:url>
<work:contributors>
<work:contributor>
<common:contributor-orcid>
<common:uri>http://orcid.org/0000-0002-1825-0097</common:uri>
<common:path>0000-0002-1825-0097</common:path>
<common:host>orcid.org</common:host>
</common:contributor-orcid>
<work:credit-name>Mankuzhiyil, Nijil</work:credit-name>
<work:contributor-email>[email protected]</work:contributor-email>
<work:contributor-attributes>
<work:contributor-sequence>first</work:contributor-sequence>
<work:contributor-role>author</work:contributor-role>
</work:contributor-attributes>
</work:contributor>
</work:contributors>
</work:work>
""")
result = OrcidConverter(
phdthesis['metadata'],
url_pattern='http://inspirehep.net/record/{recid}',
).get_xml()
assert valid_against_schema(result)
assert xml_compare(expected, result)
assert _OrcidHasher(
mock.Mock())._hash_xml_element(expected) == _OrcidHasher(
mock.Mock())._hash_xml_element(result)
def test_format_article(app, api_client):
response = api_client.get('/literature/4328')
assert response.status_code == 200
article = json.loads(response.data)
expected = xml_parse("""
<work:work xmlns:common="http://www.orcid.org/ns/common" xmlns:work="http://www.orcid.org/ns/work">
<work:title>
<common:title>Partial Symmetries of Weak Interactions</common:title>
</work:title>
<work:journal-title>Nucl.Phys.</work:journal-title>
<work:type>journal-article</work:type>
<common:publication-date>
<common:year>1961</common:year>
</common:publication-date>
<common:external-ids>
<common:external-id>
<common:external-id-type>doi</common:external-id-type>
<common:external-id-value>10.1016/0029-5582(61)90469-2</common:external-id-value>
<common:external-id-url>http://dx.doi.org/10.1016/0029-5582(61)90469-2</common:external-id-url>
<common:external-id-relationship>self</common:external-id-relationship>
</common:external-id>
</common:external-ids>
<work:url>http://inspirehep.net/record/4328</work:url>
<work:contributors>
<work:contributor>
<work:credit-name>Glashow, S.L.</work:credit-name>
<work:contributor-attributes>
<work:contributor-sequence>first</work:contributor-sequence>
<work:contributor-role>author</work:contributor-role>
</work:contributor-attributes>
</work:contributor>
</work:contributors>
</work:work>
""")
result = OrcidConverter(
article['metadata'],
url_pattern='http://inspirehep.net/record/{recid}',
).get_xml()
assert valid_against_schema(result)
assert xml_compare(expected, result)
assert _OrcidHasher(
mock.Mock())._hash_xml_element(expected) == _OrcidHasher(
mock.Mock())._hash_xml_element(result)
def test_format_conference_paper(app, api_client):
response = api_client.get('/literature/524480')
assert response.status_code == 200
inproceedings = json.loads(response.data)
expected = xml_parse("""
<work:work xmlns:common="http://www.orcid.org/ns/common" xmlns:work="http://www.orcid.org/ns/work">
<work:title>
<common:title>CMB anisotropies: A Decadal survey</common:title>
</work:title>
<work:journal-title>4th RESCEU International Symposium on Birth and Evolution of the Universe</work:journal-title>
<work:type>conference-paper</work:type>
<common:external-ids>
<common:external-id>
<common:external-id-type>arxiv</common:external-id-type>
<common:external-id-value>astro-ph/0002520</common:external-id-value>
<common:external-id-url>http://arxiv.org/abs/astro-ph/0002520</common:external-id-url>
<common:external-id-relationship>self</common:external-id-relationship>
</common:external-id>
</common:external-ids>
<work:url>http://inspirehep.net/record/524480</work:url>
<work:contributors>
<work:contributor>
<work:credit-name>Hu, Wayne</work:credit-name>
<work:contributor-attributes>
<work:contributor-sequence>first</work:contributor-sequence>
<work:contributor-role>author</work:contributor-role>
</work:contributor-attributes>
</work:contributor>
</work:contributors>
</work:work>
""")
result = OrcidConverter(
inproceedings['metadata'],
url_pattern='http://inspirehep.net/record/{recid}',
).get_xml()
assert valid_against_schema(result)
assert xml_compare(expected, result)
assert _OrcidHasher(
mock.Mock())._hash_xml_element(expected) == _OrcidHasher(
mock.Mock())._hash_xml_element(result)
def test_format_book_chapter(app, api_client):
response = api_client.get('/literature/1375491')
assert response.status_code == 200
inbook = json.loads(response.data)
expected = xml_parse("""
<work:work xmlns:common="http://www.orcid.org/ns/common" xmlns:work="http://www.orcid.org/ns/work">
<work:title>
<common:title>Supersymmetry</common:title>
</work:title>
<work:type>book-chapter</work:type>
<common:publication-date>
<common:year>2015</common:year>
</common:publication-date>
<common:external-ids>
<common:external-id>
<common:external-id-type>doi</common:external-id-type>
<common:external-id-value>10.1007/978-3-319-15001-7_10</common:external-id-value>
<common:external-id-url>http://dx.doi.org/10.1007/978-3-319-15001-7_10</common:external-id-url>
<common:external-id-relationship>self</common:external-id-relationship>
</common:external-id>
<common:external-id>
<common:external-id-type>arxiv</common:external-id-type>
<common:external-id-value>1506.03091</common:external-id-value>
<common:external-id-url>http://arxiv.org/abs/1506.03091</common:external-id-url>
<common:external-id-relationship>self</common:external-id-relationship>
</common:external-id>
</common:external-ids>
<work:url>http://inspirehep.net/record/1375491</work:url>
<work:contributors>
<work:contributor>
<work:credit-name>Bechtle, Philip</work:credit-name>
<work:contributor-attributes>
<work:contributor-sequence>first</work:contributor-sequence>
<work:contributor-role>author</work:contributor-role>
</work:contributor-attributes>
</work:contributor>
<work:contributor>
<work:credit-name>Plehn, Tilman</work:credit-name>
<work:contributor-attributes>
<work:contributor-sequence>additional</work:contributor-sequence>
<work:contributor-role>author</work:contributor-role>
</work:contributor-attributes>
</work:contributor>
<work:contributor>
<work:credit-name>Sander, Christian</work:credit-name>
<work:contributor-attributes>
<work:contributor-sequence>additional</work:contributor-sequence>
<work:contributor-role>author</work:contributor-role>
</work:contributor-attributes>
</work:contributor>
</work:contributors>
</work:work>
""")
result = OrcidConverter(
inbook['metadata'],
url_pattern='http://inspirehep.net/record/{recid}',
).get_xml()
assert valid_against_schema(result)
assert xml_compare(expected, result)
assert hash_xml_element(expected) == hash_xml_element(result)
def push_record_with_orcid(recid, orcid, oauth_token):
"""Push record to ORCID with a specific ORCID ID.
Args:
recid (string): HEP record to push
orcid (string): ORCID identifier to push onto
oauth_token (string): ORCID user OAUTH token
Returns:
Tuple[string, string]: a tuple with two elements:
- the put-code of the inserted item,
- and the new hash of the ORCID record
"""
inspire_record = get_db_record('lit', recid)
orcid_cache = OrcidCache(orcid, recid)
putcode = orcid_cache.read_work_putcode()
if not orcid_cache.has_work_content_changed(inspire_record):
LOGGER.info(
'Hash unchanged: not pushing #%s as not a meaningful update'
' for orcid:', recid, orcid)
return putcode
# TODO Move bibtex serialization into the the converter
try:
bibtex = bibtex_v1.serialize(recid, inspire_record)
except Exception:
bibtex = None
LOGGER.error(
'Pushing record #%s without BibTex, as fetching it failed!', recid)
# TODO Split orcid API client networking part from our business logic.
# This TODO is also a requirement for proper app monitoring.
orcid_api = _get_api()
orcid_xml = OrcidConverter(inspire_record,
app.config['LEGACY_RECORD_URL_PATTERN'],
put_code=putcode,
bibtex_citation=bibtex).get_xml()
lock_name = 'orcid:{}'.format(orcid)
# It's an update: PUT.
if putcode:
LOGGER.info(
"Pushing record #%s with put-code %s onto %s.",
recid,
putcode,
orcid,
)
with log_time_context('Pushing updated record', LOGGER), \
distributed_lock(lock_name, blocking=True):
orcid_api.update_record(
orcid_id=orcid,
token=oauth_token,
request_type='work',
data=orcid_xml,
put_code=putcode,
content_type='application/orcid+xml',
)
# It's a new record: POST.
else:
LOGGER.info(
"No put-code found, pushing new record #%s to ORCID %s.",
recid,
orcid,
)
try:
with log_time_context('Pushing new record', LOGGER), \
distributed_lock(lock_name, blocking=True):
putcode = orcid_api.add_record(
orcid_id=orcid,
token=oauth_token,
request_type='work',
data=orcid_xml,
content_type='application/orcid+xml',
)
except Exception as exc:
if DuplicatedExternalIdentifiersError.match(exc):
recache_author_putcodes(orcid, oauth_token)
putcode = orcid_cache.read_work_putcode()
if not putcode:
raise PutcodeNotFoundInCacheException(
'Putcode not found in cache for recid {} after having'
' recached all putcodes for the orcid {}'.format(
recid, orcid))
return push_record_with_orcid(recid, orcid, oauth_token)
else:
raise
if not putcode:
raise ValueError('Empty putcode not allowed')
LOGGER.info("Push of %s onto %s completed with put-code %s.", recid, orcid,
putcode)
orcid_cache.write_work_putcode(putcode, inspire_record)
return putcode
def push_record_with_orcid(recid,
orcid,
oauth_token,
put_code=None,
old_hash=None):
"""Push record to ORCID with a specific ORCID ID.
Args:
recid (string): HEP record to push
orcid (string): ORCID identifier to push onto
oauth_token (string): ORCID user OAUTH token
put_code (Union[string, NoneType]): put-code to push record onto,
if None will push as a new record
old_hash (Optional[string]): previous hash of the record
Returns:
Tuple[string, string]: a tuple with two elements:
- the put-code of the inserted item,
- and the new hash of the ORCID record
"""
record = get_db_record('lit', recid)
new_hash = calculate_hash_for_record(record)
if new_hash == old_hash:
LOGGER.info(
'Hash unchanged: not pushing #%s as not a meaningful update',
recid)
return put_code, new_hash
try:
bibtex = bibtex_v1.serialize(recid, record)
except Exception:
bibtex = None
LOGGER.error(
'Pushing record #%s without BibTex, as fetching it failed!', recid)
orcid_api = _get_api()
orcid_xml = OrcidConverter(record,
app.config['LEGACY_RECORD_URL_PATTERN'],
put_code=put_code,
bibtex_citation=bibtex).get_xml()
lock_name = 'orcid:{}'.format(orcid)
if put_code:
LOGGER.info(
"Pushing record #%s with put-code %s onto %s.",
recid,
put_code,
orcid,
)
with log_time_context('Pushing updated record', LOGGER), \
distributed_lock(lock_name, blocking=True):
orcid_api.update_record(
orcid_id=orcid,
token=oauth_token,
request_type='work',
data=orcid_xml,
put_code=put_code,
content_type='application/orcid+xml',
)
else:
LOGGER.info(
"No put-code found, pushing new record #%s to ORCID %s.",
recid,
orcid,
)
with log_time_context('Pushing new record', LOGGER), \
distributed_lock(lock_name, blocking=True):
put_code = orcid_api.add_record(
orcid_id=orcid,
token=oauth_token,
request_type='work',
data=orcid_xml,
content_type='application/orcid+xml',
)
LOGGER.info("Push of %s onto %s completed with put-code %s.", recid, orcid,
put_code)
return put_code, new_hash