def test_get_histograms(spark_co):
pytest.age["data"]["name"] = "b'age'"
pytest.company["data"]["name"] = "b'company'"
pytest.eyesColor["data"]["name"] = "b'eyeColor'"
pytest.gender["data"]["name"] = "b'gender'"
pytest.isActive["data"]["name"] = "b'isActive'"
pytest.latitude["data"]["name"] = "b'latitude'"
pytest.longitude["data"]["name"] = "b'longitude'"
pytest.transaction["data"]["name"] = "b'transaction'"
pytest.latitude_longitude["data"]["name"] = "b'latitude:longitude'"
pytest.latitude_longitude["data"]["bins:name"] = "unit_func"
spark = spark_co
spark_df = spark.createDataFrame(pytest.test_df)
spark_filler = SparkHistogrammar(
features=[
"date",
"isActive",
"age",
"eyeColor",
"gender",
"company",
"latitude",
"longitude",
["isActive", "age"],
["latitude", "longitude"],
"transaction",
],
bin_specs={
"transaction": {
'num': 100,
'low': -2000,
'high': 2000
},
"longitude": {
"bin_width": 5.0,
"bin_offset": 0.0
},
"latitude": {
"bin_width": 5.0,
"bin_offset": 0.0
},
},
read_key="input",
store_key="output",
)
# test get_histograms() function call
current_hists = spark_filler.get_histograms(spark_df)
# current_hists = make_histograms(spark_df, features, bin_specs)
assert current_hists["age"].toJson() == pytest.age
assert current_hists["company"].toJson() == pytest.company
assert current_hists["eyeColor"].toJson() == pytest.eyesColor
assert current_hists["gender"].toJson() == pytest.gender
assert current_hists["latitude"].toJson() == pytest.latitude
assert current_hists["longitude"].toJson() == pytest.longitude
assert current_hists["transaction"].toJson() == pytest.transaction
def test_get_histograms_timestamp(spark_co):
from pyspark.sql.functions import to_timestamp
spark = spark_co
data_date = [
"2018-12-10 00:00:00", "2018-12-10 00:00:00", "2018-12-10 00:00:00",
"2018-12-10 00:00:00", "2018-12-10 00:00:00", "2018-12-17 00:00:00",
"2018-12-17 00:00:00", "2018-12-17 00:00:00", "2018-12-17 00:00:00",
"2018-12-19 00:00:00"
]
df = pd.DataFrame(data_date, columns=['dt'])
sdf = spark.createDataFrame(df).withColumn(
"dt", to_timestamp("dt", "yyyy-MM-dd HH:mm:ss"))
expected = {
'data': {
'binWidth': 2592000000000000.0,
'bins': {
'108': 9.0,
'109': 1.0
},
'bins:type': 'Count',
'entries': 10.0,
'name': "b'dt'",
'nanflow': 0.0,
'nanflow:type': 'Count',
'origin': 1.2625632e+18
},
'type': 'SparselyBin',
'version': '1.0'
}
filler = SparkHistogrammar(features=['dt'])
current_hists = filler.get_histograms(sdf)
assert current_hists['dt'].toJson() == expected
def test_get_histograms(spark_co):
pytest.age["data"]["name"] = "b'age'"
pytest.company["data"]["name"] = "b'company'"
pytest.eyesColor["data"]["name"] = "b'eyeColor'"
pytest.gender["data"]["name"] = "b'gender'"
pytest.isActive["data"]["name"] = "b'isActive'"
pytest.latitude["data"]["name"] = "b'latitude'"
pytest.longitude["data"]["name"] = "b'longitude'"
pytest.latitude_longitude["data"]["name"] = "b'latitude:longitude'"
pytest.latitude_longitude["data"]["bins:name"] = "unit_func"
spark = spark_co
spark_df = spark.createDataFrame(pytest.test_df)
spark_filler = SparkHistogrammar(features=[
'date', 'isActive', 'age', 'eyeColor', 'gender', 'company', 'latitude',
'longitude', ['isActive', 'age'], ['latitude', 'longitude']
],
bin_specs={
'longitude': {
'bin_width': 5.0,
'bin_offset': 0.0
},
'latitude': {
'bin_width': 5.0,
'bin_offset': 0.0
}
},
read_key='input',
store_key='output')
# test get_histograms() function call
current_hists = spark_filler.get_histograms(spark_df)
# current_hists = make_histograms(spark_df, features, bin_specs)
assert current_hists['age'].toJson() == pytest.age
assert current_hists['company'].toJson() == pytest.company
assert current_hists['eyeColor'].toJson() == pytest.eyesColor
assert current_hists['gender'].toJson() == pytest.gender
assert current_hists['latitude'].toJson() == pytest.latitude
assert current_hists['longitude'].toJson() == pytest.longitude
def test_get_histograms_timestamp(spark_co):
from pyspark.sql.functions import to_timestamp
spark = spark_co
data_date = [
"2018-12-10 00:00:00",
"2018-12-10 00:00:00",
"2018-12-10 00:00:00",
"2018-12-10 00:00:00",
"2018-12-10 00:00:00",
"2018-12-17 00:00:00",
"2018-12-17 00:00:00",
"2018-12-17 00:00:00",
"2018-12-17 00:00:00",
"2018-12-19 00:00:00",
]
df = pd.DataFrame(data_date, columns=["dt"])
sdf = spark.createDataFrame(df).withColumn(
"dt", to_timestamp("dt", "yyyy-MM-dd HH:mm:ss"))
expected = {
"data": {
"binWidth": 2592000000000000.0,
"bins": {
"108": 9.0,
"109": 1.0
},
"bins:type": "Count",
"entries": 10.0,
"name": "b'dt'",
"nanflow": 0.0,
"nanflow:type": "Count",
"origin": 1.2625632e18,
},
"type": "SparselyBin",
"version": "1.0",
}
filler = SparkHistogrammar(features=["dt"])
current_hists = filler.get_histograms(sdf)
assert current_hists["dt"].toJson() == expected