def test_validation_compare_title(self):
"""BibMatch comparison: compare title"""
original_record_instances = [
'Assault frequency and preformation probability'
]
matched_record_instances = [
'Assault frequency and preformation probability : The alpha emission process'
]
comparisons = get_paired_comparisons(original_record_instances,
matched_record_instances)
threshold = 0.9
matches_needed = 1
# This should fail
result, dummy = compare_fieldvalues_normal(comparisons, threshold,
matches_needed)
self.assertFalse(result)
# Title search however, takes separators into account
result, dummy = compare_fieldvalues_title(comparisons, threshold,
matches_needed)
self.assertTrue(result)
# Check longer titles
original_record_instances = ['Buffered Electropolishing \xe2\x80\x93 A New Way for ' \
'Achieving Extremely Smooth Surface Finish on Nb SRF ' \
'Cavities to be Used in Particle Accelerators']
matched_record_instances = ['Buffered Electropolishing: A New Way for Achieving ' \
'Extremely Smooth Surface Finish on Nb SRF Cavities ' \
'To be Used in Particle Accelerators']
comparisons = get_paired_comparisons(original_record_instances,
matched_record_instances)
result, dummy = compare_fieldvalues_title(comparisons, threshold,
matches_needed)
self.assertTrue(result)
def test_validation_get_paired_comparisons(self):
"""bibmatch - validation: check generated paired comparisons """
first_list = [1,2,3]
second_list = [4,5]
# Should return empty, as lists are not equal in length
self.assertFalse(get_paired_comparisons(first_list, second_list, False))
# Should return result, in un-ordered mode
result = [((1, 4), (1, 5)), ((2, 4), (2, 5)), ((3, 4), (3, 5))]
self.assertEqual(result, get_paired_comparisons(first_list, second_list))
def test_validation_get_paired_comparisons(self):
"""bibmatch - validation: check generated paired comparisons """
first_list = [1,2,3]
second_list = [4,5]
# Should return empty, as lists are not equal in length
self.assertFalse(get_paired_comparisons(first_list, second_list, False))
# Should return result, in un-ordered mode
result = [((1, 4), (1, 5)), ((2, 4), (2, 5)), ((3, 4), (3, 5))]
self.assertEqual(result, get_paired_comparisons(first_list, second_list))
def test_validation_compare_date(self):
"""BibMatch comparison: compare date"""
original_record_instances = ['2002-02']
matched_record_instances = ['2001']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
threshold = 1.0
matches_needed = 1
result, dummy = compare_fieldvalues_date(comparisons, threshold, matches_needed)
self.assertFalse(result)
original_record_instances = ['2001-02']
matched_record_instances = ['2001']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
result, dummy = compare_fieldvalues_date(comparisons, threshold, matches_needed)
self.assertTrue(result)
def test_validation_compare_identifiers(self):
"""BibMatch comparison: compare identifiers"""
original_record_instances = ['REP-NO-02123']
matched_record_instances = ['REPNO123']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
threshold = 1.0
matches_needed = 1
result, dummy = compare_fieldvalues_identifier(comparisons, threshold, matches_needed)
self.assertFalse(result)
original_record_instances = ['REP-NO-0123']
matched_record_instances = ['REPNO123']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
result, dummy = compare_fieldvalues_identifier(comparisons, threshold, matches_needed)
self.assertTrue(result)
def test_validation_compare_strings(self):
"""BibMatch comparison: compare strings"""
original_record_instances = ['This is some random text']
matched_record_instances = ['I have some random text about nothing']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
threshold = 0.8
matches_needed = 1
result, dummy = compare_fieldvalues_normal(comparisons, threshold, matches_needed)
self.assertFalse(result)
original_record_instances = ['This is some random text']
matched_record_instances = ['Is some random text']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
result, dummy = compare_fieldvalues_normal(comparisons, threshold, matches_needed)
self.assertTrue(result)
def test_validation_compare_authors(self):
"""BibMatch comparison: compare authors"""
original_record_instances = ['Brodsky, Stanley J.']
matched_record_instances = ['Brodsky, S.J.', 'Not, M E']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
threshold = 0.8
matches_needed = 1
result, dummy = compare_fieldvalues_authorname(comparisons, threshold, matches_needed)
self.assertTrue(result)
original_record_instances = ['Brodsky, J.']
matched_record_instances = ['Brodsky, S.J.']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
result, dummy = compare_fieldvalues_authorname(comparisons, threshold, matches_needed)
self.assertFalse(result)
def test_validation_compare_identifiers(self):
"""BibMatch comparison: compare identifiers"""
original_record_instances = ['REP-NO-02123']
matched_record_instances = ['REPNO123']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
threshold = 1.0
matches_needed = 1
result, dummy = compare_fieldvalues_identifier(comparisons, threshold, matches_needed)
self.assertFalse(result)
original_record_instances = ['REP-NO-0123']
matched_record_instances = ['REPNO123']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
result, dummy = compare_fieldvalues_identifier(comparisons, threshold, matches_needed)
self.assertTrue(result)
def test_validation_compare_strings(self):
"""BibMatch comparison: compare strings"""
original_record_instances = ['This is some random text']
matched_record_instances = ['I have some random text about nothing']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
threshold = 0.8
matches_needed = 1
result, dummy = compare_fieldvalues_normal(comparisons, threshold, matches_needed)
self.assertFalse(result)
original_record_instances = ['This is some random text']
matched_record_instances = ['Is some random text']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
result, dummy = compare_fieldvalues_normal(comparisons, threshold, matches_needed)
self.assertTrue(result)
def test_validation_compare_authors(self):
"""BibMatch comparison: compare authors"""
original_record_instances = ['Brodsky, Stanley J.']
matched_record_instances = ['Brodsky, S.J.', 'Not, M E']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
threshold = 0.85
matches_needed = 1
result, dummy = compare_fieldvalues_authorname(comparisons, threshold, matches_needed)
self.assertTrue(result)
original_record_instances = ['Brodsky, J.']
matched_record_instances = ['Brodsky, S.J.']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
result, dummy = compare_fieldvalues_authorname(comparisons, threshold, matches_needed)
self.assertFalse(result)
def test_validation_compare_date(self):
"""BibMatch comparison: compare date"""
original_record_instances = ['2002-02']
matched_record_instances = ['2001']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
threshold = 1.0
matches_needed = 1
result, dummy = compare_fieldvalues_date(comparisons, threshold, matches_needed)
self.assertFalse(result)
original_record_instances = ['2001-02']
matched_record_instances = ['2001']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
result, dummy = compare_fieldvalues_date(comparisons, threshold, matches_needed)
self.assertTrue(result)
def test_validation_compare_title(self):
"""BibMatch comparison: compare title"""
original_record_instances = ['Assault frequency and preformation probability']
matched_record_instances = ['Assault frequency and preformation probability : The alpha emission process']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
threshold = 0.9
matches_needed = 1
# This should fail
result, dummy = compare_fieldvalues_normal(comparisons, threshold, matches_needed)
self.assertFalse(result)
# Title search however, takes separators into account
result, dummy = compare_fieldvalues_title(comparisons, threshold, matches_needed)
self.assertTrue(result)
# Check longer titles
original_record_instances = ['Buffered Electropolishing \xe2\x80\x93 A New Way for ' \
'Achieving Extremely Smooth Surface Finish on Nb SRF ' \
'Cavities to be Used in Particle Accelerators']
matched_record_instances = ['Buffered Electropolishing: A New Way for Achieving ' \
'Extremely Smooth Surface Finish on Nb SRF Cavities ' \
'To be Used in Particle Accelerators']
comparisons = get_paired_comparisons(original_record_instances, matched_record_instances)
result, dummy = compare_fieldvalues_title(comparisons, threshold, matches_needed)
self.assertTrue(result)
