def test_str_add_ser(self):
edadd = "The last name is: " + self.ed_ecommerce(
)["customer_last_name"]
pdadd = "The last name is: " + self.pd_ecommerce(
)["customer_last_name"]
assert_pandas_eland_series_equal(pdadd, edadd)
def test_ser_add_str(self):
edadd = self.ed_ecommerce(
)["customer_first_name"] + " is the first name."
pdadd = self.pd_ecommerce(
)["customer_first_name"] + " is the first name."
assert_pandas_eland_series_equal(pdadd, edadd)
def test_bad_str_add_ser(self):
# TODO encode special characters better
# Elasticsearch accepts this, but it will cause problems
edadd = " *" + self.ed_ecommerce()["customer_last_name"]
pdadd = " *" + self.pd_ecommerce()["customer_last_name"]
assert_pandas_eland_series_equal(pdadd, edadd)
def test_ser_add_str_add_ser(self):
pdadd = (self.pd_ecommerce()["customer_first_name"] + " " +
self.pd_ecommerce()["customer_last_name"])
edadd = (self.ed_ecommerce()["customer_first_name"] + " " +
self.ed_ecommerce()["customer_last_name"])
assert_pandas_eland_series_equal(pdadd, edadd)
def test_flights_filter_index_items(self, items):
ed_flights_small = self.ed_flights_small()["FlightDelayType"]
pd_flights_small = self.pd_flights_small()["FlightDelayType"]
ed_ser = ed_flights_small.filter(items=items, axis=0)
pd_ser = pd_flights_small.filter(items=items, axis=0)
assert_pandas_eland_series_equal(pd_ser, ed_ser)
def test_ecommerce_series_simple_series_addition(self):
pd_df = self.pd_ecommerce().head(100)
ed_df = self.ed_ecommerce().head(100)
pd_series = pd_df["taxful_total_price"] + pd_df["total_quantity"]
ed_series = ed_df["taxful_total_price"] + ed_df["total_quantity"]
assert_pandas_eland_series_equal(pd_series, ed_series, check_less_precise=True)
def test_getitem_one_argument(self):
ed_flights = self.ed_flights().head(89)
pd_flights = self.pd_flights().head(89)
ed_flights_OriginAirportID = ed_flights.OriginAirportID
pd_flights_OriginAirportID = pd_flights.OriginAirportID
assert_pandas_eland_series_equal(pd_flights_OriginAirportID,
ed_flights_OriginAirportID)
def test_getitem_one_attribute(self):
ed_flights = self.ed_flights().head(103)
pd_flights = self.pd_flights().head(103)
ed_flights_OriginAirportID = ed_flights["OriginAirportID"]
pd_flights_OriginAirportID = pd_flights["OriginAirportID"]
assert_pandas_eland_series_equal(pd_flights_OriginAirportID,
ed_flights_OriginAirportID)
def test_head_tail(self):
pd_s = self.pd_flights()["Carrier"]
ed_s = ed.Series(ES_TEST_CLIENT, FLIGHTS_INDEX_NAME, "Carrier")
pd_s_head = pd_s.head(10)
ed_s_head = ed_s.head(10)
assert_pandas_eland_series_equal(pd_s_head, ed_s_head)
pd_s_tail = pd_s.tail(10)
ed_s_tail = ed_s.tail(10)
assert_pandas_eland_series_equal(pd_s_tail, ed_s_tail)
def test_ecommerce_series_simple_arithmetics(self):
pd_df = self.pd_ecommerce().head(100)
ed_df = self.ed_ecommerce().head(100)
pd_series = (pd_df["taxful_total_price"] + 5 +
pd_df["total_quantity"] / pd_df["taxless_total_price"] -
pd_df["total_unique_products"] * 10.0 +
pd_df["total_quantity"])
ed_series = (ed_df["taxful_total_price"] + 5 +
ed_df["total_quantity"] / ed_df["taxless_total_price"] -
ed_df["total_unique_products"] * 10.0 +
ed_df["total_quantity"])
assert_pandas_eland_series_equal(pd_series,
ed_series,
check_less_precise=True)
def test_getitem_multiple_calls(self):
ed_flights = self.ed_flights().head(89)
pd_flights = self.pd_flights().head(89)
ed_col0 = ed_flights[
["DestCityName", "DestCountry", "DestLocation", "DestRegion"]
]
try:
ed_col1 = ed_col0["Carrier"]
except KeyError:
pass
pd_col1 = pd_flights["DestCountry"]
ed_col1 = ed_col0["DestCountry"]
assert_pandas_eland_series_equal(pd_col1, ed_col1)
def test_all_formats(self):
index_name = self.time_index_name
ed_df = ed.DataFrame(ES_TEST_CLIENT, index_name)
for format_name in self.time_formats.keys():
times = [
pd.to_datetime(
datetime.strptime(dt, "%Y-%m-%dT%H:%M:%S.%f%z").strftime(
self.time_formats[format_name]),
format=self.time_formats[format_name],
) for dt in self.times
]
ed_series = ed_df[format_name]
pd_series = pd.Series(
times,
index=[str(i) for i in range(len(self.times))],
name=format_name)
assert_pandas_eland_series_equal(pd_series, ed_series)
def test_simple_lat_lon(self):
"""
Note on nested object order - this can change when
note this could be a bug in ES...
PUT my_index/doc/1
{
"location": {
"lat": "50.033333",
"lon": "8.570556"
}
}
GET my_index/_search
"_source": {
"location": {
"lat": "50.033333",
"lon": "8.570556"
}
}
GET my_index/_search
{
"_source": "location"
}
"_source": {
"location": {
"lon": "8.570556",
"lat": "50.033333"
}
}
Hence we store the pandas df source json as 'lon', 'lat'
"""
pd_dest_location = self.pd_flights()["DestLocation"].head(1)
ed_dest_location = self.ed_flights()["DestLocation"].head(1)
assert_pandas_eland_series_equal(
pd_dest_location, ed_dest_location, check_exact=False, rtol=2
)
def test_ecommerce_series_basic_rarithmetics(self):
pd_df = self.pd_ecommerce().head(10)
ed_df = self.ed_ecommerce().head(10)
ops = [
"__radd__",
"__rtruediv__",
"__rfloordiv__",
"__rpow__",
"__rmod__",
"__rmul__",
"__rsub__",
"radd",
"rtruediv",
"rfloordiv",
"rpow",
"rmod",
"rmul",
"rsub",
]
for op in ops:
pd_series = getattr(pd_df["taxful_total_price"], op)(
pd_df["total_quantity"]
)
ed_series = getattr(ed_df["taxful_total_price"], op)(
ed_df["total_quantity"]
)
assert_pandas_eland_series_equal(
pd_series, ed_series, check_less_precise=True
)
pd_series = getattr(pd_df["taxful_total_price"], op)(3.141)
ed_series = getattr(ed_df["taxful_total_price"], op)(3.141)
assert_pandas_eland_series_equal(
pd_series, ed_series, check_less_precise=True
)
pd_series = getattr(pd_df["taxful_total_price"], op)(np.float32(2.879))
ed_series = getattr(ed_df["taxful_total_price"], op)(np.float32(2.879))
assert_pandas_eland_series_equal(
pd_series, ed_series, check_less_precise=True
)
pd_series = getattr(pd_df["taxful_total_price"], op)(int(6))
ed_series = getattr(ed_df["taxful_total_price"], op)(int(6))
assert_pandas_eland_series_equal(
pd_series, ed_series, check_less_precise=True
)
def test_ecommerce_series_basic_arithmetics(self):
pd_df = self.pd_ecommerce().head(100)
ed_df = self.ed_ecommerce().head(100)
ops = [
"__add__",
"__truediv__",
"__floordiv__",
"__pow__",
"__mod__",
"__mul__",
"__sub__",
"add",
"truediv",
"floordiv",
"pow",
"mod",
"mul",
"sub",
]
for op in ops:
pd_series = getattr(pd_df["taxful_total_price"], op)(
pd_df["total_quantity"]
)
ed_series = getattr(ed_df["taxful_total_price"], op)(
ed_df["total_quantity"]
)
assert_pandas_eland_series_equal(
pd_series, ed_series, check_less_precise=True
)
pd_series = getattr(pd_df["taxful_total_price"], op)(10.56)
ed_series = getattr(ed_df["taxful_total_price"], op)(10.56)
assert_pandas_eland_series_equal(
pd_series, ed_series, check_less_precise=True
)
pd_series = getattr(pd_df["taxful_total_price"], op)(np.float32(1.879))
ed_series = getattr(ed_df["taxful_total_price"], op)(np.float32(1.879))
assert_pandas_eland_series_equal(
pd_series, ed_series, check_less_precise=True
)
pd_series = getattr(pd_df["taxful_total_price"], op)(int(8))
ed_series = getattr(ed_df["taxful_total_price"], op)(int(8))
assert_pandas_eland_series_equal(
pd_series, ed_series, check_less_precise=True
)
def test_name(self):
# deep copy pandas DataFrame as .name alters this reference frame
pd_series = self.pd_flights()["Carrier"].copy(deep=True)
ed_series = ed.Series(ES_TEST_CLIENT, FLIGHTS_INDEX_NAME, "Carrier")
assert_pandas_eland_series_equal(pd_series, ed_series)
assert ed_series.name == pd_series.name
pd_series.name = "renamed1"
ed_series.name = "renamed1"
assert_pandas_eland_series_equal(pd_series, ed_series)
assert ed_series.name == pd_series.name
pd_series.name = "renamed2"
ed_series.name = "renamed2"
assert_pandas_eland_series_equal(pd_series, ed_series)
assert ed_series.name == pd_series.name
def test_rename(self):
pd_carrier = self.pd_flights()["Carrier"]
ed_carrier = ed.Series(ES_TEST_CLIENT, FLIGHTS_INDEX_NAME, "Carrier")
assert_pandas_eland_series_equal(pd_carrier, ed_carrier)
pd_renamed = pd_carrier.rename("renamed")
ed_renamed = ed_carrier.rename("renamed")
print(pd_renamed)
print(ed_renamed)
print(ed_renamed.info_es())
assert_pandas_eland_series_equal(pd_renamed, ed_renamed)
pd_renamed2 = pd_renamed.rename("renamed2")
ed_renamed2 = ed_renamed.rename("renamed2")
print(ed_renamed2.info_es())
assert "renamed2" == ed_renamed2.name
assert_pandas_eland_series_equal(pd_renamed2, ed_renamed2)
def test_supported_series_dtypes_rops(self):
pd_df = self.pd_ecommerce().head(100)
ed_df = self.ed_ecommerce().head(100)
# Test some specific operations that are and aren't supported
numeric_ops = [
"__radd__",
"__rtruediv__",
"__rfloordiv__",
"__rpow__",
"__rmod__",
"__rmul__",
"__rsub__",
]
non_string_numeric_ops = [
"__radd__",
"__rtruediv__",
"__rfloordiv__",
"__rpow__",
"__rmod__",
"__rsub__",
]
# __rmul__ is supported for int * str in pandas
# float op float
for op in numeric_ops:
pd_series = getattr(pd_df["taxful_total_price"], op)(
pd_df["taxless_total_price"]
)
ed_series = getattr(ed_df["taxful_total_price"], op)(
ed_df["taxless_total_price"]
)
assert_pandas_eland_series_equal(
pd_series, ed_series, check_less_precise=True
)
# int op float
for op in numeric_ops:
pd_series = getattr(pd_df["total_quantity"], op)(
pd_df["taxless_total_price"]
)
ed_series = getattr(ed_df["total_quantity"], op)(
ed_df["taxless_total_price"]
)
assert_pandas_eland_series_equal(
pd_series, ed_series, check_less_precise=True
)
# float op int
for op in numeric_ops:
pd_series = getattr(pd_df["taxful_total_price"], op)(
pd_df["total_quantity"]
)
ed_series = getattr(ed_df["taxful_total_price"], op)(
ed_df["total_quantity"]
)
assert_pandas_eland_series_equal(
pd_series, ed_series, check_less_precise=True
)
# str op int (throws)
for op in non_string_numeric_ops:
with pytest.raises(TypeError):
pd_series = getattr(pd_df["currency"], op)(pd_df["total_quantity"])
with pytest.raises(TypeError):
ed_series = getattr(ed_df["currency"], op)(ed_df["total_quantity"])
with pytest.raises(TypeError):
pd_series = getattr(pd_df["currency"], op)(10.0)
with pytest.raises(TypeError):
ed_series = getattr(ed_df["currency"], op)(10.0)
# int op str (throws)
for op in non_string_numeric_ops:
with pytest.raises(TypeError):
pd_series = getattr(pd_df["total_quantity"], op)(pd_df["currency"])
with pytest.raises(TypeError):
ed_series = getattr(ed_df["total_quantity"], op)(ed_df["currency"])
def test_sample(self):
ed_s = ed.Series(ES_TEST_CLIENT, FLIGHTS_INDEX_NAME, "Carrier")
pd_s = self.build_from_index(ed_s.sample(n=10, random_state=self.SEED))
ed_s_sample = ed_s.sample(n=10, random_state=self.SEED)
assert_pandas_eland_series_equal(pd_s, ed_s_sample)
