def test_extractFeatures_keyerror():
# Test when some groundtruth are not in the local caseversion file
cvs = finddup.loadLocalCaseversions('tests/data/small_274_0_key_error.json')
gt = finddup.loadGroundTruth('tests/data/groundtruth-274.csv', cvs['objects'])
features = finddup.extractFeatures(cvs, gt['ids'])
assert(features.shape[0] == len(gt['perdictions']))
def test_calcDiffs():
cvs = finddup.loadLocalCaseversions('tests/data/small_274_0.json')
comb_it = finddup.genAllCombinations(cvs)
selected_pairs = [next(comb_it) for i in range(2)]
diffs = filters.calcDiffs(cvs, selected_pairs)
assert (len(selected_pairs) == len(diffs))
assert (type(diffs[0]) == type([]))
assert (type(diffs[0][0]) == type(""))
def test_extractFeatures_default():
cvs = finddup.loadLocalCaseversions('tests/data/small_274_0.json')
#gt = finddup.loadGroundTruth('tests/data/groundtruth-274.csv')
comb = [x for x in finddup.genAllCombinations(cvs)] #FIXME: remove this dependency
features = finddup.extractFeatures(cvs, comb)
print(len(cvs['objects']))
pairs_count = len(cvs['objects']) * (len(cvs['objects']) - 1) / 2
assert(features.shape[0] == pairs_count)
def test_loadGroundTruth_filter_by_csv():
cvs = finddup.loadLocalCaseversions('tests/data/small_274_0_key_error.json')
groundtruth = finddup.loadGroundTruth('tests/data/groundtruth-274.csv', cvs['objects'])
expected = {
"perdictions": [
"merge",
"none"
],
"ids": [
{
"lhs_id": "210201",
"rhs_id": "210202"
}, # 210521 does not exist in csv
{
"lhs_id": "210201",
"rhs_id": "211079"
}
]
}
assert(groundtruth == expected)
def test_extractFeatures_select():
cvs = finddup.loadLocalCaseversions('tests/data/small_274_0.json')
gt = finddup.loadGroundTruth('tests/data/groundtruth-274.csv')
features = finddup.extractFeatures(cvs, gt['ids'])
assert(features.shape[0] == len(gt['perdictions']))
