def test_match_simple(self):
df = pd.DataFrame({
"matching_specificity":
["flu", "ebv", "GAD", "PPI", "GAD", "PPI", "ebv"],
"p_val": [0.01, 0.00001, 0.000001, 0.01, 0.01, 0.0000001, 0.1],
"odds_ratio": [0.51, 0.5, 0, 0, 0, 0, 0],
"a": ["yes", "no", "no", "no", "no", "no", "no"],
"b": ["no", "yes", "no", "no", "no", "no", "no"]
})
filter_params = {
"type": OperationType.TOP_N,
"value": {
"type": DataType.COLUMN,
"name": "odds_ratio",
},
"number": 3
}
matcher = CriteriaMatcher()
result = matcher.match(filter_params, df)
self.assertTrue(
np.array_equal(
result,
np.array([True, True, False, False, False, False, True])))
def test_match(self):
df = pd.DataFrame({
"matching_specificity":
["flu", "ebv", "GAD", "PPI", "GAD", "PPI", "ebv"],
"p_val": [0.01, 0.00001, 0.000001, 0.01, 0.01, 0.0000001, 0.1],
"odds_ratio": [0.51, 0, 0, 0, 0, 0, 0],
"a": ["yes", "no", "no", "no", "no", "no", "no"],
"b": ["no", "yes", "no", "no", "no", "no", "no"]
})
filter_params = {
"type":
BooleanType.OR,
"operands": [
{
"type":
BooleanType.AND,
"operands": [
{
"type": OperationType.IN,
"allowed_values": ["GAD", "PPI"],
"value": {
"type": DataType.COLUMN,
"name": "matching_specificity"
}
},
{
"type": OperationType.LESS_THAN,
"threshold": 0.001,
"value": {
"type": DataType.COLUMN,
"name": "p_val"
}
},
]
},
{
"type":
BooleanType.AND,
"operands": [
{
"type": OperationType.IN,
"allowed_values": ["yes"],
"value": {
"type": DataType.COLUMN,
"name": "a"
}
},
{
"type": OperationType.GREATER_THAN,
"threshold": 0.5,
"value": {
"type": DataType.COLUMN,
"name": "odds_ratio"
}
},
]
},
]
}
matcher = CriteriaMatcher()
result = matcher.match(filter_params, df)
self.assertTrue(
np.array_equal(
result,
np.array([True, False, True, False, False, True, False])))
def get_matching_indices(dataset: RepertoireDataset, criteria):
metadata = pd.DataFrame(dataset.get_metadata(None))
matches = CriteriaMatcher().match(criteria, metadata)
indices = np.where(matches)[0]
return indices