def test_clip_to_seconds_series():
m = IndexMap()
stamp = 1234567890
k = pd.date_range(pd.to_datetime(stamp, unit='s'),
periods=10000,
freq='ns').to_series().astype(int)
assert len(m.clip_to_seconds(k).unique()) == 1
assert m.clip_to_seconds(k).unique()[0] == stamp
def test_update(mocker):
m = IndexMap()
keys = generate_keys(10000)
def hash_mock(k, salt=0):
seed = 123456
rs = np.random.RandomState(seed=seed + salt)
return pd.Series(rs.randint(0, len(k) * 10, size=len(k)), index=k)
with mocker.patch.object(m, 'hash_', side_effect=hash_mock):
m.update(keys)
assert len(m) == len(keys), "All keys not in mapping"
assert m._map.index.difference(keys).empty, "All keys not in mapping"
assert len(m._map.unique()) == len(keys), "Duplicate values in mapping"
# Can't have duplicate keys.
with pytest.raises(KeyError):
m.update(keys)
new_unique_keys = generate_keys(1000).difference(keys)
m.update(new_unique_keys)
assert len(
m) == len(keys) + len(new_unique_keys), "All keys not in mapping"
assert m._map.index.difference(
keys.union(new_unique_keys)).empty, "All keys not in mapping"
assert len(m._map.unique(
)) == len(keys) + len(new_unique_keys), "Duplicate values in mapping"
def test_convert_to_ten_digit_int():
m = IndexMap()
v = 1234567890
datetime_col = pd.date_range(pd.to_datetime(v, unit='s'), periods=10000, freq='ns').to_series()
int_col = pd.Series(v, index=range(10000))
float_col = pd.Series(1.1234567890, index=range(10000))
bad_col = pd.Series('a', index=range(10000))
assert m.convert_to_ten_digit_int(datetime_col).unique()[0] == v
assert m.convert_to_ten_digit_int(int_col).unique()[0] == 4072825790
assert m.convert_to_ten_digit_int(float_col).unique()[0] == v
with pytest.raises(RandomnessError):
m.convert_to_ten_digit_int(bad_col)
def test_spread_scalar():
m = IndexMap()
assert m.spread(1234567890) == 4072825790
def test_clip_to_seconds_scalar():
m = IndexMap()
k = pd.to_datetime("2010-01-25 06:25:31.123456789")
assert m.clip_to_seconds(k.value) == int(str(k.value)[:10])
def test_digit_series():
m = IndexMap()
k = pd.Series(123456789, index=range(10000))
for i in range(10):
assert len(m.digit(k, i).unique()) == 1
assert m.digit(k, i)[0] == 10 - (i + 1)
def test_digit_scalar():
m = IndexMap()
k = 123456789
for i in range(10):
assert m.digit(k, i) == 10 - (i + 1)
def map_size_and_hashed_values(request):
keys = generate_keys(*request.param)
m = IndexMap()
return m.map_size, m.hash_(keys)
def test_shift_series():
m = IndexMap()
s = pd.Series(1.1234567890, index=range(10000))
assert len(m.shift(s).unique()) == 1
assert m.shift(s).unique()[0] == 1234567890
def test_shift_scalar():
m = IndexMap()
assert m.shift(1.1234567890) == 1234567890
def test_spread_series():
m = IndexMap()
s = pd.Series(1234567890, index=range(10000))
assert len(m.spread(s).unique()) == 1
assert m.spread(s).unique()[0] == 4072825790