def test_arxiv_categories_from_65017a_2_obsolete_category():
schema = load_schema('authors')
subschema = schema['properties']['arxiv_categories']
snippet = ('<datafield tag="650" ind1="1" ind2="7">'
' <subfield code="2">INSPIRE</subfield>'
' <subfield code="a">ATOM-PH</subfield>'
'</datafield>') # record/1010819
expected = [
'physics.atom-ph',
]
result = hepnames.do(create_record(snippet))
assert validate(result['arxiv_categories'], subschema) is None
assert expected == result['arxiv_categories']
expected = [
{
'2': 'arXiv',
'a': 'physics.atom-ph',
},
]
result = hepnames2marc.do(result)
assert expected == result['65017']
def test_inspire_categories_from_65017a_2_E():
schema = load_schema('authors')
subschema = schema['properties']['inspire_categories']
snippet = ('<datafield tag="650" ind1="1" ind2="7">'
' <subfield code="2">INSPIRE</subfield>'
' <subfield code="a">E</subfield>'
'</datafield>') # record/1019112
expected = [
{
'term': 'Experiment-HEP'
},
]
result = hepnames.do(create_record(snippet))
assert validate(result['inspire_categories'], subschema) is None
assert expected == result['inspire_categories']
expected = [
{
'2': 'INSPIRE',
'a': 'Experiment-HEP',
},
]
result = hepnames2marc.do(result)
assert expected == result['65017']
def test_other_names_from_400__triple_a():
schema = load_schema('authors')
subschema = schema['properties']['other_names']
snippet = ('<datafield tag="400" ind1=" " ind2=" ">'
' <subfield code="a">Yosef Cohen, Hadar</subfield>'
' <subfield code="a">Josef Cohen, Hadar</subfield>'
' <subfield code="a">Cohen, Hadar Josef</subfield>'
'</datafield>') # record/1292399
expected = [
'Yosef Cohen, Hadar',
'Josef Cohen, Hadar',
'Cohen, Hadar Josef',
]
result = hepnames.do(create_record(snippet))
assert validate(result['other_names'], subschema) is None
assert expected == result['other_names']
expected = [
{
'a': 'Yosef Cohen, Hadar'
},
{
'a': 'Josef Cohen, Hadar'
},
{
'a': 'Cohen, Hadar Josef'
},
]
result = hepnames2marc.do(result)
assert expected == result['400']
def test_positions_from_371__a():
schema = load_schema('authors')
subschema = schema['properties']['positions']
snippet = ('<datafield tag="371" ind1=" " ind2=" ">'
' <subfield code="a">Aachen, Tech. Hochsch.</subfield>'
'</datafield>') # record/997958
expected = [
{
'current': False,
'institution': {
'curated_relation': False,
'name': 'Aachen, Tech. Hochsch.',
},
},
]
result = hepnames.do(create_record(snippet))
assert validate(result['positions'], subschema) is None
assert expected == result['positions']
expected = [{'a': 'Aachen, Tech. Hochsch.'}]
result = hepnames2marc.do(result)
assert expected == result['371']
def test_ids_from_035__a_9_with_orcid():
schema = load_schema('authors')
subschema = schema['properties']['ids']
snippet = ('<datafield tag="035" ind1=" " ind2=" ">'
' <subfield code="9">ORCID</subfield>'
' <subfield code="a">0000-0001-6771-2174</subfield>'
'</datafield>')
expected = [
{
'schema': 'ORCID',
'value': '0000-0001-6771-2174',
},
]
result = hepnames.do(create_record(snippet))
assert validate(result['ids'], subschema) is None
assert expected == result['ids']
expected = [{
'9': 'ORCID',
'a': '0000-0001-6771-2174',
}]
result = hepnames2marc.do(result)
assert expected == result['035']
def test_ids_from_035__a_9_with_wikipedia():
schema = load_schema('authors')
subschema = schema['properties']['ids']
snippet = ('<datafield tag="035" ind1=" " ind2=" ">'
' <subfield code="9">Wikipedia</subfield>'
' <subfield code="a">Guido_Tonelli</subfield>'
'</datafield>') # record/985898/export/xme
expected = [
{
'schema': 'WIKIPEDIA',
'value': 'Guido_Tonelli',
},
]
result = hepnames.do(create_record(snippet))
assert validate(result['ids'], subschema) is None
assert expected == result['ids']
expected = [
{
'9': 'WIKIPEDIA',
'a': 'Guido_Tonelli',
},
]
result = hepnames2marc.do(result)
assert expected == result['035']
def test_ids_from_035__a_9_with_scopus():
schema = load_schema('authors')
subschema = schema['properties']['ids']
snippet = ('<datafield tag="035" ind1=" " ind2=" ">'
' <subfield code="9">SCOPUS</subfield>'
' <subfield code="a">7103280792</subfield>'
'</datafield>') # record/1017182
expected = [
{
'schema': 'SCOPUS',
'value': '7103280792',
},
]
result = hepnames.do(create_record(snippet))
assert validate(result['ids'], subschema) is None
assert expected == result['ids']
expected = [
{
'9': 'SCOPUS',
'a': '7103280792',
},
]
result = hepnames2marc.do(result)
assert expected == result['035']
def test_stub_from_980__a_useful():
schema = load_schema('authors')
subschema = schema['properties']['stub']
snippet = ('<datafield tag="980" ind1=" " ind2=" ">'
' <subfield code="a">USEFUL</subfield>'
'</datafield>') # record/1222902
expected = False
result = hepnames.do(create_record(snippet))
assert validate(result['stub'], subschema) is None
assert expected == result['stub']
expected = [
{
'a': 'USEFUL'
},
{
'a': 'HEPNAMES'
},
]
result = hepnames2marc.do(result)
assert expected == result['980']
def test_ids_from_035__a_9_with_viaf():
schema = load_schema('authors')
subschema = schema['properties']['ids']
snippet = ('<datafield tag="035" ind1=" " ind2=" ">'
' <subfield code="9">VIAF</subfield>'
' <subfield code="a">34517183</subfield>'
'</datafield>') # record/1008109
expected = [
{
'schema': 'VIAF',
'value': '34517183',
},
]
result = hepnames.do(create_record(snippet))
assert validate(result['ids'], subschema) is None
assert expected == result['ids']
expected = [
{
'9': 'VIAF',
'a': '34517183',
},
]
result = hepnames2marc.do(result)
assert expected == result['035']
def test_ids_from_035__a_9_with_researcherid():
schema = load_schema('authors')
subschema = schema['properties']['ids']
snippet = ('<datafield tag="035" ind1=" " ind2=" ">'
' <subfield code="9">RESEARCHERID</subfield>'
' <subfield code="a">B-4717-2008</subfield>'
'</datafield>') # record/1051026
expected = [
{
'schema': 'RESEARCHERID',
'value': 'B-4717-2008',
},
]
result = hepnames.do(create_record(snippet))
assert validate(result['ids'], subschema) is None
assert expected == result['ids']
expected = [
{
'9': 'RESEARCHERID',
'a': 'B-4717-2008',
},
]
result = hepnames2marc.do(result)
assert expected == result['035']
def test_ids_from_035__a_9_with_googlescholar():
schema = load_schema('authors')
subschema = schema['properties']['ids']
snippet = ('<datafield tag="035" ind1=" " ind2=" ">'
' <subfield code="9">GoogleScholar</subfield>'
' <subfield code="a">Tnl-9KoAAAAJ</subfield>'
'</datafield>') # record/1467553
expected = [
{
'schema': 'GOOGLESCHOLAR',
'value': 'Tnl-9KoAAAAJ',
},
]
result = hepnames.do(create_record(snippet))
assert validate(result['ids'], subschema) is None
assert expected == result['ids']
expected = [
{
'9': 'GOOGLESCHOLAR',
'a': 'Tnl-9KoAAAAJ',
},
]
result = hepnames2marc.do(result)
assert expected == result['035']
def test_ids_from_035__a_with_bai():
schema = load_schema('authors')
subschema = schema['properties']['ids']
snippet = ('<datafield tag="035" ind1=" " ind2=" ">'
' <subfield code="a">Jian.Long.Han.1</subfield>'
'</datafield>') # record/1464894
expected = [
{
'schema': 'INSPIRE BAI',
'value': 'Jian.Long.Han.1',
},
]
result = hepnames.do(create_record(snippet))
assert validate(result['ids'], subschema) is None
assert expected == result['ids']
expected = [
{
'9': 'BAI',
'a': 'Jian.Long.Han.1',
},
]
result = hepnames2marc.do(result)
assert expected == result['035']
def test_ids_from_035__a_9_with_slac():
schema = load_schema('authors')
subschema = schema['properties']['ids']
snippet = ('<datafield tag="035" ind1=" " ind2=" ">'
' <subfield code="9">SLAC</subfield>'
' <subfield code="a">SLAC-218626</subfield>'
'</datafield>') # record/1028379
expected = [
{
'schema': 'SLAC',
'value': 'SLAC-218626',
},
]
result = hepnames.do(create_record(snippet))
assert validate(result['ids'], subschema) is None
assert expected == result['ids']
expected = [
{
'9': 'SLAC',
'a': 'SLAC-218626',
},
]
result = hepnames2marc.do(result)
assert expected == result['035']
def test_private_notes_from_595__double_a():
schema = load_schema('authors')
subschema = schema['properties']['_private_notes']
snippet = (
'<datafield tag="595" ind1=" " ind2=" ">'
' <subfield code="a">"I want to hide my personal information on REDACTED" 7/2017</subfield>'
' <subfield code="a">REDACTED</subfield>'
'</datafield>') # record/1279232
expected = [
{
'value':
'"I want to hide my personal information on REDACTED" 7/2017'
},
{
'value': 'REDACTED'
},
]
result = hepnames.do(create_record(snippet))
assert validate(result['_private_notes'], subschema) is None
assert expected == result['_private_notes']
expected = [
{
'a': '"I want to hide my personal information on REDACTED" 7/2017'
},
{
'a': 'REDACTED'
},
]
result = hepnames2marc.do(result)
assert expected == result['595']
def test_name_from_100__a_g_q():
schema = load_schema('authors')
subschema = schema['properties']['name']
snippet = ('<datafield tag="100" ind1=" " ind2=" ">'
' <subfield code="a">Abarbanel, Henry D.I.</subfield>'
' <subfield code="q">Henry D.I. Abarbanel</subfield>'
' <subfield code="g">ACTIVE</subfield>'
'</datafield>') # record/1019100
expected = {
'value': 'Abarbanel, Henry D.I.',
'preferred_name': 'Henry D.I. Abarbanel',
}
result = hepnames.do(create_record(snippet))
assert validate(result['name'], subschema) is None
assert expected == result['name']
expected = {
'a': 'Abarbanel, Henry D.I.',
'q': 'Henry D.I. Abarbanel',
'g': 'active',
}
result = hepnames2marc.do(result)
assert expected == result['100']
def test_ids_from_970__a():
schema = load_schema('authors')
subschema = schema['properties']['ids']
snippet = ('<datafield tag="970" ind1=" " ind2=" ">'
' <subfield code="a">HEPNAMES-646482</subfield>'
'</datafield>') # record/1498151
expected = [
{
'schema': 'SPIRES',
'value': 'HEPNAMES-646482',
},
]
result = hepnames.do(create_record(snippet))
assert validate(result['ids'], subschema) is None
assert expected == result['ids']
expected = [
{
'a': 'HEPNAMES-646482'
},
]
result = hepnames2marc.do(result)
assert expected == result['970']
def test_name_from_100__g_q_populates_value_from_preferred_name():
schema = load_schema('authors')
subschema = schema['properties']['name']
snippet = ('<datafield tag="100" ind1=" " ind2=" ">'
' <subfield code="g">ACTIVE</subfield>'
' <subfield code="q">Vyacheslav I. Yukalova</subfield>'
'</datafield>') # record/1259075
expected = {
'preferred_name': 'Vyacheslav I. Yukalova',
'value': 'Yukalova, Vyacheslav I.',
}
result = hepnames.do(create_record(snippet))
assert validate(result['name'], subschema) is None
assert expected == result['name']
expected = {
'a': 'Yukalova, Vyacheslav I.',
'g': 'active',
'q': 'Vyacheslav I. Yukalova',
}
result = hepnames2marc.do(result)
assert expected == result['100']
def test_deleted_from_980__c():
schema = load_schema('authors')
subschema = schema['properties']['deleted']
snippet = ('<datafield tag="980" ind1=" " ind2=" ">'
' <subfield code="c">DELETED</subfield>'
'</datafield>') # record/1511071
expected = True
result = hepnames.do(create_record(snippet))
assert validate(result['deleted'], subschema) is None
assert expected == result['deleted']
expected = [
{
'c': 'DELETED'
},
{
'a': 'HEPNAMES'
},
]
result = hepnames2marc.do(result)
assert expected == result['980']
def test_title_from_100__a_c_q_discards_default_title():
schema = load_schema('authors')
subschema = schema['properties']['name']
snippet = ('<datafield tag="100" ind1=" " ind2=" ">'
' <subfield code="a">Joosten, Sylvester Johannes</subfield>'
' <subfield code="c">title (e.g. Sir)</subfield>'
' <subfield code="q">Sylvester Johannes Joosten</subfield>'
'</datafield>') # record/1270441
expected = {
'preferred_name': 'Sylvester Johannes Joosten',
'value': 'Joosten, Sylvester Johannes',
}
result = hepnames.do(create_record(snippet))
assert validate(result['name'], subschema) is None
assert expected == result['name']
expected = {
'a': 'Joosten, Sylvester Johannes',
'q': 'Sylvester Johannes Joosten',
}
result = hepnames2marc.do(result)
assert expected == result['100']
def test_ids_from_035__a_9_with_desy():
schema = load_schema('authors')
subschema = schema['properties']['ids']
snippet = ('<datafield tag="035" ind1=" " ind2=" ">'
' <subfield code="a">DESY-1001805</subfield>'
' <subfield code="9">DESY</subfield>'
'</datafield>') # record/993224/export/xme
expected = [
{
'schema': 'DESY',
'value': 'DESY-1001805',
},
]
result = hepnames.do(create_record(snippet))
assert validate(result['ids'], subschema) is None
assert expected == result['ids']
expected = [
{
'9': 'DESY',
'a': 'DESY-1001805',
},
]
result = hepnames2marc.do(result)
assert expected == result['035']
def test_private_notes_from_595__a_9():
schema = load_schema('authors')
subschema = schema['properties']['_private_notes']
snippet = (
'<datafield tag="595" ind1=" " ind2=" ">'
' <subfield code="a">Author prefers Alexandrov, A.S.</subfield>'
' <subfield code="9">SPIRES-HIDDEN</subfield>'
'</datafield>') # record/1050484
expected = [{
'source': 'SPIRES-HIDDEN',
'value': 'Author prefers Alexandrov, A.S.',
}]
result = hepnames.do(create_record(snippet))
assert validate(result['_private_notes'], subschema) is None
assert expected == result['_private_notes']
expected = [{
'9': 'SPIRES-HIDDEN',
'a': 'Author prefers Alexandrov, A.S.',
}]
result = hepnames2marc.do(result)
assert expected == result['595']
def test_acquisition_source_from_541__a_b_c_d_e_converts_dates_to_datetimes():
schema = load_schema('authors')
subschema = schema['properties']['acquisition_source']
snippet = ('<datafield tag="541" ind1=" " ind2=" ">'
' <subfield code="a">inspire:uid:51852</subfield>'
' <subfield code="b">[email protected]</subfield>'
' <subfield code="c">submission</subfield>'
' <subfield code="d">2016-05-24</subfield>'
' <subfield code="e">805819</subfield>'
'</datafield>') # record/982806
expected = {
'datetime': '2016-05-24T00:00:00',
'email': '*****@*****.**',
'internal_uid': 51852,
'method': 'submitter',
'submission_number': '805819',
}
result = hepnames.do(create_record(snippet))
assert validate(result['acquisition_source'], subschema) is None
assert expected == result['acquisition_source']
expected = {
'b': '*****@*****.**',
'c': 'submission',
'd': '2016-05-24T00:00:00',
'e': '805819',
}
result = hepnames2marc.do(result)
assert expected == result['541']
def test_old_single_email_from_371__a():
schema = load_schema('authors')
subschema = schema['properties']['positions']
snippet = (
'<datafield tag="371" ind1=" " ind2=" ">'
' <subfield code="a">IMSc, Chennai</subfield>'
' <subfield code="o">[email protected]</subfield>'
' <subfield code="r">PD</subfield>'
' <subfield code="s">2012</subfield>'
' <subfield code="t">2013</subfield>'
'</datafield>'
) # record/1060782
expected = [
{
"current": False,
"old_emails": [
"*****@*****.**"
],
"end_date": "2013",
"rank": "POSTDOC",
"institution": {
"name": "IMSc, Chennai",
"curated_relation": False
},
"_rank": "PD",
"start_date": "2012"
}
]
result = hepnames.do(create_record(snippet))
assert validate(result['positions'], subschema) is None
assert expected == result['positions']
expected = [
{
"a": "IMSc, Chennai",
"o": [
"*****@*****.**"
],
"s": "2012",
"r": "PD",
"t": "2013"
}
]
result = hepnames2marc.do(result)
assert expected == result['371']
def test_ids_from_double_035__a_9():
schema = load_schema('authors')
subschema = schema['properties']['ids']
snippet = (
'<record>'
' <datafield tag="035" ind1=" " ind2=" ">'
' <subfield code="a">INSPIRE-00134135</subfield>'
' <subfield code="9">INSPIRE</subfield>'
' </datafield>'
' <datafield tag="035" ind1=" " ind2=" ">'
' <subfield code="a">H.Vogel.1</subfield>'
' <subfield code="9">BAI</subfield>'
' </datafield>'
'</record>'
) # record/984519
expected = [
{
'schema': 'INSPIRE ID',
'value': 'INSPIRE-00134135',
},
{
'schema': 'INSPIRE BAI',
'value': 'H.Vogel.1',
},
]
result = hepnames.do(create_record(snippet))
assert validate(result['ids'], subschema) is None
assert expected == result['ids']
expected = [
{
'a': 'INSPIRE-00134135',
'9': 'INSPIRE'
},
{
'a': 'H.Vogel.1',
'9': 'BAI'
}
]
result = hepnames2marc.do(result)
for el in expected:
assert el in result['035']
for el in result['035']:
assert el in expected
def test_experiments(mock_get_record_ref, mock_get_recid_from_ref, test_name,
xml_snippet, expected_json, expected_marc):
mock_get_record_ref.side_effect = \
lambda x, *_: x and 'mocked_recid_%s' % x
mock_get_recid_from_ref.side_effect = \
lambda x, *_: x and int(x.rsplit('_')[-1])
if not xml_snippet.strip().startswith('<record>'):
xml_snippet = '<record>%s</record>' % xml_snippet
json_data = hepnames.do(create_record(xml_snippet))
json_experiments = json_data['experiments']
marc_experiments = hepnames2marc.do(json_data)['693']
assert marc_experiments == expected_marc
assert json_experiments == expected_json
def test_advisors_from_701__a_g_i():
schema = load_schema('authors')
subschema = schema['properties']['advisors']
snippet = (
'<datafield tag="701" ind1=" " ind2=" ">'
' <subfield code="a">Rivelles, Victor O.</subfield>'
' <subfield code="g">PhD</subfield>'
' <subfield code="i">INSPIRE-00120420</subfield>'
' <subfield code="x">991627</subfield>'
' <subfield code="y">1</subfield>'
'</datafield>'
) # record/1474091/export/xme
expected = [
{
'name': 'Rivelles, Victor O.',
'degree_type': 'PhD',
'ids': [
{
'schema': 'INSPIRE ID',
'value': 'INSPIRE-00120420'
}
],
'record': {
'$ref': 'http://localhost:5000/api/authors/991627',
},
'curated_relation': True
},
]
result = hepnames.do(create_record(snippet))
assert validate(result['advisors'], subschema) is None
assert expected == result['advisors']
expected = [
{
'a': 'Rivelles, Victor O.',
'g': 'PhD',
'i': 'INSPIRE-00120420',
},
]
result = hepnames2marc.do(result)
assert expected == result['701']
def test_ids_from_double_035__a_9_with_kaken():
schema = load_schema('authors')
subschema = schema['properties']['ids']
snippet = (
'<record>'
' <datafield tag="035" ind1=" " ind2=" ">'
' <subfield code="9">BAI</subfield>'
' <subfield code="a">Toshio.Suzuki.2</subfield>'
' </datafield>'
' <datafield tag="035" ind1=" " ind2=" ">'
' <subfield code="9">KAKEN</subfield>'
' <subfield code="a">70139070</subfield>'
' </datafield>'
'</record>'
) # record/1474271/export/xme
expected = [
{
'schema': 'INSPIRE BAI',
'value': 'Toshio.Suzuki.2',
},
{
'schema': 'KAKEN',
'value': 'KAKEN-70139070',
},
]
result = hepnames.do(create_record(snippet))
assert validate(result['ids'], subschema) is None
assert expected == result['ids']
expected = [
{
'9': 'BAI',
'a': 'Toshio.Suzuki.2',
},
{
'9': 'KAKEN',
'a': 'KAKEN-70139070',
},
]
result = hepnames2marc.do(result)
assert expected == result['035']
def test_new_record_from_970__d():
schema = load_schema('authors')
subschema = schema['properties']['new_record']
snippet = ('<datafield tag="970" ind1=" " ind2=" ">'
' <subfield code="d">1039458</subfield>'
'</datafield>') # record/1271254
expected = {'$ref': 'http://localhost:5000/api/authors/1039458'}
result = hepnames.do(create_record(snippet))
assert validate(result['new_record'], subschema) is None
assert expected == result['new_record']
expected = {'d': 1039458}
result = hepnames2marc.do(result)
assert expected == result['970']
def test_legacy_creation_date_from_961__double_x_does_not_raise():
schema = load_schema('authors')
subschema = schema['properties']['legacy_creation_date']
snippet = ('<datafield tag="961" ind1=" " ind2=" ">'
' <subfield code="x">2006-04-21</subfield>'
' <subfield code="x">1996-09-01</subfield>'
'</datafield>') # record/982164
expected = '1996-09-01'
result = hepnames.do(create_record(snippet))
assert validate(result['legacy_creation_date'], subschema) is None
assert expected == result['legacy_creation_date']
expected = {'x': '1996-09-01'}
result = hepnames2marc.do(result)
assert expected == result['961']
def test_positions_from_371__a_r_t_z():
schema = load_schema('authors')
subschema = schema['properties']['positions']
snippet = (
'<datafield tag="371" ind1=" " ind2=" ">'
' <subfield code="a">San Luis Potosi U.</subfield>'
' <subfield code="r">Master</subfield>'
' <subfield code="t">2007</subfield>'
' <subfield code="z">903830</subfield>'
'</datafield>'
) # record/1037568
expected = [
{
'_rank': 'Master',
'current': False,
'end_date': '2007',
'institution': {
'curated_relation': True,
'name': 'San Luis Potosi U.',
'record': {
'$ref': 'http://localhost:5000/api/institutions/903830',
},
},
'rank': 'MASTER',
},
]
result = hepnames.do(create_record(snippet))
assert validate(result['positions'], subschema) is None
assert expected == result['positions']
expected = [
{
'a': 'San Luis Potosi U.',
'r': 'MAS',
't': '2007',
},
]
result = hepnames2marc.do(result)
assert expected == result['371']