def test_or_filter_multiple(self):
f1 = filters.Filter(dimension="dim1", value="val1")
f2 = filters.Filter(dimension="dim2", value="val2")
f3 = filters.Filter(dimension="dim3", value="val3")
filter = filters.Filter(type="or", fields=[f1, f2, f3])
actual = filters.Filter.build_filter(filter)
expected = {
"type":
"or",
"fields": [
{
"type": "selector",
"dimension": "dim1",
"value": "val1"
},
{
"type": "selector",
"dimension": "dim2",
"value": "val2"
},
{
"type": "selector",
"dimension": "dim3",
"value": "val3"
},
],
}
assert actual == expected
def test_search_filter(self):
# Without caseSensitive param - default:false
actual = filters.Filter.build_filter(
filters.Filter(type="search", dimension="dim", value='val'))
expected = {
'type': 'search',
'dimension': 'dim',
'query': {
'type': 'contains',
'caseSensitive': 'false',
'value': 'val'
}
}
assert actual == expected
# With caseSensitive param
actual = filters.Filter.build_filter(
filters.Filter(type="search",
dimension="dim",
value='val',
caseSensitive='true'))
expected = {
'type': 'search',
'dimension': 'dim',
'query': {
'type': 'contains',
'caseSensitive': 'true',
'value': 'val'
}
}
assert actual == expected
def test_nested_filtered_aggregator(self):
filter1 = filters.Filter(dimension='dim1', value='val')
filter2 = filters.Filter(dimension='dim2', value='val')
agg = aggregators.filtered(
filter1, aggregators.filtered(filter2,
aggregators.count('metric1')))
actual = aggregators.build_aggregators({'agg_name': agg})
# the innermost aggregation must have 'agg_name'
expected = [{
'type': 'filtered',
'aggregator': {
'type': 'filtered',
'aggregator': {
'fieldName': 'metric1',
'type': 'count',
'name': 'agg_name'
},
'filter': {
'dimension': 'dim2',
'value': 'val',
'type': 'selector'
}
},
'filter': {
'dimension': 'dim1',
'value': 'val',
'type': 'selector'
}
}]
assert expected == actual
def test_nested_filtered_aggregator(self):
filter1 = filters.Filter(dimension="dim1", value="val")
filter2 = filters.Filter(dimension="dim2", value="val")
agg = aggregators.filtered(
filter1, aggregators.filtered(filter2,
aggregators.count("metric1")))
actual = aggregators.build_aggregators({"agg_name": agg})
# the innermost aggregation must have 'agg_name'
expected = [{
"type": "filtered",
"aggregator": {
"type": "filtered",
"aggregator": {
"fieldName": "metric1",
"type": "count",
"name": "agg_name",
},
"filter": {
"dimension": "dim2",
"value": "val",
"type": "selector"
},
},
"filter": {
"dimension": "dim1",
"value": "val",
"type": "selector"
},
}]
assert expected == actual
def test_search_filter(self):
# Without caseSensitive param - default:false
actual = filters.Filter.build_filter(
filters.Filter(type="search", dimension="dim", value="val"))
expected = {
"type": "search",
"dimension": "dim",
"query": {
"type": "contains",
"caseSensitive": "false",
"value": "val"
},
}
assert actual == expected
# With caseSensitive param
actual = filters.Filter.build_filter(
filters.Filter(type="search",
dimension="dim",
value="val",
caseSensitive="true"))
expected = {
"type": "search",
"dimension": "dim",
"query": {
"type": "contains",
"caseSensitive": "true",
"value": "val"
},
}
assert actual == expected
def test_nested_not_or_filter(self):
f1 = filters.Filter(dimension='dim1', value='val1')
f2 = filters.Filter(dimension='dim2', value='val2')
actual = filters.Filter.build_filter(~(f1 | f2))
expected = {
'type': 'not',
'field': {'type': 'or',
'fields': [{'type': 'selector', 'dimension': 'dim1', 'value': 'val1'},
{'type': 'selector', 'dimension': 'dim2', 'value': 'val2'}]}
}
assert actual == expected
def test_and_filter(self):
f1 = filters.Filter(dimension='dim1', value='val1')
f2 = filters.Filter(dimension='dim2', value='val2')
actual = filters.Filter.build_filter(f1 & f2)
expected = {
'type': 'and',
'fields': [
{'type': 'selector', 'dimension': 'dim1', 'value': 'val1'},
{'type': 'selector', 'dimension': 'dim2', 'value': 'val2'}
]
}
assert actual == expected
def test_nested_mix_filter(self):
f1 = filters.Filter(dimension='dim1', value='val1')
f2 = filters.Filter(dimension='dim2', value='val2')
f3 = filters.Filter(dimension='dim3', value='val3')
f4 = filters.Filter(dimension='dim4', value='val4')
f5 = filters.Filter(dimension='dim5', value='val5')
f6 = filters.Filter(dimension='dim6', value='val6')
f7 = filters.Filter(dimension='dim7', value='val7')
f8 = filters.Filter(dimension='dim8', value='val8')
actual = filters.Filter.build_filter(f1 & ~f2 & f3 & (f4 | ~f5 | f6 | (f7 & ~f8)))
expected = {
'fields': [{'dimension': 'dim1', 'type': 'selector', 'value': 'val1'},
{'field': {'dimension': 'dim2', 'type': 'selector', 'value': 'val2'},
'type': 'not'},
{'dimension': 'dim3', 'type': 'selector', 'value': 'val3'},
{'fields': [{'dimension': 'dim4', 'type': 'selector', 'value': 'val4'},
{'field': {'dimension': 'dim5', 'type': 'selector',
'value': 'val5'},
'type': 'not'},
{'dimension': 'dim6', 'type': 'selector', 'value': 'val6'},
{'fields': [
{'dimension': 'dim7', 'type': 'selector', 'value': 'val7'},
{'field': {'dimension': 'dim8', 'type': 'selector',
'value': 'val8'},
'type': 'not'}],
'type': 'and'}],
'type': 'or'}],
'type': 'and'
}
assert actual == expected
def test_and_filter_multiple(self):
f1 = filters.Filter(dimension='dim1', value='val1')
f2 = filters.Filter(dimension='dim2', value='val2')
f3 = filters.Filter(dimension='dim3', value='val3')
filter = filters.Filter(type='and', fields=[f1, f2, f3])
actual = filters.Filter.build_filter(filter)
expected = {
'type': 'and',
'fields': [
{'type': 'selector', 'dimension': 'dim1', 'value': 'val1'},
{'type': 'selector', 'dimension': 'dim2', 'value': 'val2'},
{'type': 'selector', 'dimension': 'dim3', 'value': 'val3'}
]
}
assert actual == expected
def test_in_filter(self):
actual = filters.Filter.build_filter(
filters.Filter(type='in', dimension='dim',
values=['val1', 'val2', 'val3']))
expected = {'type': 'in', 'dimension': 'dim',
'values': ['val1', 'val2', 'val3']}
assert actual == expected
def test_filtered_aggregator(self):
filter_ = filters.Filter(dimension='dim', value='val')
aggs = [aggregators.count('metric1'),
aggregators.longsum('metric2'),
aggregators.doublesum('metric3'),
aggregators.doublemin('metric4'),
aggregators.doublemax('metric5'),
aggregators.hyperunique('metric6'),
aggregators.cardinality('dim1'),
aggregators.cardinality(['dim1', 'dim2'], by_row=True),
aggregators.thetasketch('dim1'),
aggregators.thetasketch('metric7'),
aggregators.thetasketch('metric8', isinputthetasketch=True, size=8192)
]
for agg in aggs:
expected = {
'type': 'filtered',
'filter': {
'type': 'selector',
'dimension': 'dim',
'value': 'val'
},
'aggregator': agg
}
actual = aggregators.filtered(filter_, agg)
assert actual == expected
def test_filtered_aggregator(self):
filter_ = filters.Filter(dimension="dim", value="val")
aggs = [
aggregators.count("metric1"),
aggregators.longsum("metric2"),
aggregators.doublesum("metric3"),
aggregators.doublemin("metric4"),
aggregators.doublemax("metric5"),
aggregators.hyperunique("metric6"),
aggregators.cardinality("dim1"),
aggregators.cardinality(["dim1", "dim2"], by_row=True),
aggregators.thetasketch("dim1"),
aggregators.thetasketch("metric7"),
aggregators.thetasketch("metric8",
isinputthetasketch=True,
size=8192),
]
for agg in aggs:
expected = {
"type": "filtered",
"filter": {
"type": "selector",
"dimension": "dim",
"value": "val"
},
"aggregator": agg,
}
actual = aggregators.filtered(filter_, agg)
assert actual == expected
def test_selector_filter_extraction_fn(self):
extraction_fn = dimensions.RegexExtraction('([a-b])')
f = filters.Filter(dimension='dim', value='v',
extraction_function=extraction_fn)
actual = filters.Filter.build_filter(f)
expected = {'type': 'selector', 'dimension': 'dim', 'value': 'v',
'extractionFn': {'type': 'regex', 'expr': '([a-b])'}}
assert actual == expected
def test_extraction_filter(self):
extraction_fn = dimensions.PartialExtraction('([a-b])')
f = filters.Filter(type='extraction', dimension='dim', value='v',
extraction_function=extraction_fn)
actual = filters.Filter.build_filter(f)
expected = {'type': 'extraction', 'dimension': 'dim', 'value': 'v',
'extractionFn': {'type': 'partial', 'expr': '([a-b])'}}
assert actual == expected
def test_not_filter(self):
f = filters.Filter(dimension='dim', value='val')
actual = filters.Filter.build_filter(~f)
expected = {
'type': 'not',
'field': {'type': 'selector', 'dimension': 'dim', 'value': 'val'}
}
assert actual == expected
def test_javascript_filter(self):
actual = filters.Filter.build_filter(
filters.Filter(type="javascript",
dimension="dim",
function="function(x){return true}"))
expected = {
"type": "javascript",
"dimension": "dim",
"function": "function(x){return true}",
}
assert actual == expected
def test_in_filter(self):
actual = filters.Filter.build_filter(
filters.Filter(type="in",
dimension="dim",
values=["val1", "val2", "val3"]))
expected = {
"type": "in",
"dimension": "dim",
"values": ["val1", "val2", "val3"],
}
assert actual == expected
def test_columnComparison_filter(self):
actual = filters.Filter.build_filter(
filters.Filter(type='columnComparison', dimensions=[
'dim1',
dimensions.DimensionSpec('dim2', 'dim2')
]))
expected = {'type': 'columnComparison', 'dimensions': [
'dim1',
{'type': 'default', 'dimension': 'dim2', 'outputName': 'dim2'}
]}
assert actual == expected
def test_not_filter(self):
f = ~filters.Filter(dimension='dim', value='val')
actual = filters.Filter.build_filter(f)
# Call `build_filter` twice to make sure it does not
# change the passed filter object argument `f`.
actual = filters.Filter.build_filter(f)
expected = {
'type': 'not',
'field': {'type': 'selector', 'dimension': 'dim', 'value': 'val'}
}
assert actual == expected
def test_javascript_filter(self):
actual = filters.Filter.build_filter(
filters.Filter(type='javascript',
dimension='dim',
function='function(x){return true}'))
expected = {
'type': 'javascript',
'dimension': 'dim',
'function': 'function(x){return true}'
}
assert actual == expected
def test_or_filter(self):
f1 = filters.Filter(dimension="dim1", value="val1")
f2 = filters.Filter(dimension="dim2", value="val2")
actual = filters.Filter.build_filter(f1 | f2)
expected = {
"type":
"or",
"fields": [
{
"type": "selector",
"dimension": "dim1",
"value": "val1"
},
{
"type": "selector",
"dimension": "dim2",
"value": "val2"
},
],
}
assert actual == expected
def test_not_filter(self):
f = ~filters.Filter(dimension="dim", value="val")
actual = filters.Filter.build_filter(f)
# Call `build_filter` twice to make sure it does not
# change the passed filter object argument `f`.
actual = filters.Filter.build_filter(f)
expected = {
"type": "not",
"field": {
"type": "selector",
"dimension": "dim",
"value": "val"
},
}
assert actual == expected
def test_selector_filter_extraction_fn(self):
extraction_fn = dimensions.RegexExtraction("([a-b])")
f = filters.Filter(dimension="dim",
value="v",
extraction_function=extraction_fn)
actual = filters.Filter.build_filter(f)
expected = {
"type": "selector",
"dimension": "dim",
"value": "v",
"extractionFn": {
"type": "regex",
"expr": "([a-b])"
},
}
assert actual == expected
def test_extraction_filter(self):
extraction_fn = dimensions.PartialExtraction("([a-b])")
f = filters.Filter(
type="extraction",
dimension="dim",
value="v",
extraction_function=extraction_fn,
)
actual = filters.Filter.build_filter(f)
expected = {
"type": "extraction",
"dimension": "dim",
"value": "v",
"extractionFn": {
"type": "partial",
"expr": "([a-b])"
},
}
assert actual == expected
def test_columnComparison_filter(self):
actual = filters.Filter.build_filter(
filters.Filter(
type="columnComparison",
dimensions=["dim1",
dimensions.DimensionSpec("dim2", "dim2")],
))
expected = {
"type":
"columnComparison",
"dimensions": [
"dim1",
{
"type": "default",
"dimension": "dim2",
"outputName": "dim2"
},
],
}
assert actual == expected
def druid_simple_groupby(dimensions, filter_list=[], filter_type="and",
datasource=settings.DRUID_SPRAYDAY_DATASOURCE):
"""
Inputs:
dimensions => list of dimensions to group by
filter_list => list of list of things to filter with e.g.
filter_list=[['target_area_id', operator.ne, 1],
['sprayable', operator.eq, "true"],
['dimension', operator, "value"]])
filter_type => type of Druid filter to perform
"""
query = PyDruid(get_druid_broker_url(), 'druid/v2')
params = dict(
datasource=datasource,
granularity='all',
intervals=settings.DRUID_INTERVAL,
limit_spec={
"type": "default",
"limit": 50000,
}
)
params['dimensions'] = dimensions
if filter_list:
fields = []
for this_filter in filter_list:
compare_dim = filters.Dimension(this_filter[0])
comparison_operator = this_filter[1] # e.g. operator.eq
compare_dim_value = this_filter[2]
fields.append(comparison_operator(compare_dim, compare_dim_value))
params['filter'] = filters.Filter(
type=filter_type,
fields=fields
)
try:
request = query.groupby(**params)
except OSError:
pass
else:
return request.result
return []
def get_druid_data(dimensions=None, filter_list=[], filter_type="and",
order_by=["target_area_name"],
datasource=settings.DRUID_SPRAYDAY_DATASOURCE):
"""
Runs a query against Druid, returns data with metrics
Inputs:
dimensions => list of dimensions to group by
filter_list => list of list of things to filter with e.g.
filter_list=[['target_area_id', operator.ne, 1],
['sprayable', operator.eq, "true"],
['dimension', operator, "value"]])
filter_type => type of Druid filter to perform,
order_by => field(s) to order the data by
"""
query = PyDruid(get_druid_broker_url(), 'druid/v2')
params = dict(
datasource=datasource,
granularity='all',
intervals=settings.DRUID_INTERVAL,
aggregations={
'num_not_sprayable': aggregators.filtered(
filters.Filter(
type='and',
fields=[filters.Dimension('sprayable') == 'false']
),
aggregators.longsum('count')
),
'num_not_sprayed': aggregators.filtered(
filters.Filter(
type='and',
fields=[filters.Dimension('sprayable') == 'true',
filters.Dimension('sprayed') ==
settings.MSPRAY_WAS_NOT_SPRAYED_VALUE]
),
aggregators.longsum('count')
),
'num_sprayed': aggregators.filtered(
filters.Dimension('sprayed') ==
settings.MSPRAY_WAS_SPRAYED_VALUE,
aggregators.longsum('count')
),
'num_new': aggregators.filtered(
filters.Dimension('is_new') == 'true',
aggregators.longsum('count')
),
'num_new_no_duplicates': aggregators.filtered(
filters.Filter(
type='and',
fields=[filters.Dimension('is_duplicate') == 'false',
filters.Dimension('is_new') == 'true']
),
aggregators.longsum('count')
),
'num_duplicate': aggregators.filtered(
filters.Dimension('is_duplicate') == 'true',
aggregators.longsum('count')
),
'num_sprayed_no_duplicates': aggregators.filtered(
filters.Filter(
type='and',
fields=[filters.Dimension('is_duplicate') == 'false',
filters.Dimension('sprayed') ==
settings.MSPRAY_WAS_SPRAYED_VALUE]
),
aggregators.longsum('count')
),
'num_not_sprayed_no_duplicates': aggregators.filtered(
filters.Filter(
type='and',
fields=[filters.Dimension('is_duplicate') == 'false',
filters.Dimension('sprayable') == 'true',
filters.Dimension('sprayed') ==
settings.MSPRAY_WAS_NOT_SPRAYED_VALUE]
),
aggregators.longsum('count')
),
'num_sprayed_duplicates': aggregators.filtered(
filters.Filter(
type='and',
fields=[filters.Dimension('is_duplicate') == 'true',
filters.Dimension('sprayable') == 'true',
filters.Dimension('sprayed') ==
settings.MSPRAY_WAS_SPRAYED_VALUE]
),
aggregators.longsum('count')
),
'num_not_sprayable_no_duplicates': aggregators.filtered(
filters.Filter(
type='and',
fields=[filters.Dimension('is_duplicate') == 'false',
filters.Dimension('sprayable') == 'false']
),
aggregators.longsum('count')
),
'num_refused': aggregators.filtered(
filters.Filter(
type='and',
fields=[filters.Dimension('is_duplicate') == 'false',
filters.Dimension('is_refused') == 'true',
filters.Dimension('sprayed') ==
settings.MSPRAY_WAS_NOT_SPRAYED_VALUE]
),
aggregators.longsum('count')
),
},
post_aggregations={
'num_found': Field('num_sprayed_no_duplicates') +
Field('num_sprayed_duplicates') +
Field('num_not_sprayed_no_duplicates')
},
limit_spec={
"type": "default",
"limit": 50000,
"columns": order_by
}
)
if filter_list:
fields = []
for this_filter in filter_list:
compare_dim = filters.Dimension(this_filter[0])
comparison_operator = this_filter[1] # e.g. operator.eq
compare_dim_value = this_filter[2]
fields.append(comparison_operator(compare_dim, compare_dim_value))
params['filter'] = filters.Filter(
type=filter_type,
fields=fields
)
if dimensions is None:
params['dimensions'] = ['target_area_id', 'target_area_name',
'target_area_structures']
else:
params['dimensions'] = dimensions
try:
request = query.groupby(**params)
except OSError:
return []
else:
return request.result
def test_invalid_filter(self):
with pytest.raises(NotImplementedError):
filters.Filter(type='invalid', dimension='dim', value='val')
def test_selector_filter(self):
actual = filters.Filter.build_filter(
filters.Filter(dimension='dim', value='val'))
expected = {'type': 'selector', 'dimension': 'dim', 'value': 'val'}
assert actual == expected
def test_like_filter(self):
actual = filters.Filter.build_filter(
filters.Filter(type="like", dimension="dim", pattern="%val%"))
expected = {'type': 'like', 'dimension': 'dim', 'pattern': '%val%'}
assert actual == expected
