upgini 1.1.288__py3-none-any.whl → 1.1.288a0__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Potentially problematic release.


This version of upgini might be problematic. Click here for more details.

upgini/__about__.py CHANGED
@@ -1 +1 @@
1
- __version__ = "1.1.288"
1
+ __version__ = "1.1.288a0"
upgini/autofe/date.py CHANGED
@@ -21,20 +21,6 @@ class DateDiffMixin(BaseModel):
21
21
 
22
22
  return pd.to_datetime(x, unit=unit)
23
23
 
24
- def _convert_diff_to_unit(self, diff: Union[pd.Series, TimedeltaArray]) -> Union[pd.Series, TimedeltaArray]:
25
- if self.diff_unit == "D":
26
- if isinstance(diff, pd.Series) and diff.dtype == "object":
27
- return diff.apply(lambda x: None if isinstance(x, float) and np.isnan(x) else x.days)
28
- else:
29
- return diff / np.timedelta64(1, self.diff_unit)
30
- elif self.diff_unit == "Y":
31
- if isinstance(diff, TimedeltaArray):
32
- return (diff / 365 / 24 / 60 / 60 / 10**9).astype(int)
33
- else:
34
- return (diff / 365 / 24 / 60 / 60 / 10**9).dt.nanoseconds
35
- else:
36
- raise Exception(f"Unsupported difference unit: {self.diff_unit}")
37
-
38
24
 
39
25
  class DateDiff(PandasOperand, DateDiffMixin):
40
26
  name = "date_diff"
@@ -55,8 +41,7 @@ class DateDiff(PandasOperand, DateDiffMixin):
55
41
  def calculate_binary(self, left: pd.Series, right: pd.Series) -> pd.Series:
56
42
  left = self._convert_to_date(left, self.left_unit)
57
43
  right = self._convert_to_date(right, self.right_unit)
58
- diff = self._convert_diff_to_unit(left.dt.date - right.dt.date)
59
- return self.__replace_negative(diff)
44
+ return self.__replace_negative((left - right) / np.timedelta64(1, self.diff_unit))
60
45
 
61
46
  def __replace_negative(self, x: Union[pd.DataFrame, pd.Series]):
62
47
  x[x < 0] = None
@@ -122,7 +107,12 @@ class DateListDiff(PandasOperand, DateDiffMixin):
122
107
  return pd.Series(left - right.values).apply(lambda x: self._agg(self._diff(x)))
123
108
 
124
109
  def _diff(self, x: TimedeltaArray):
125
- x = self._convert_diff_to_unit(x)
110
+ if self.diff_unit == "Y":
111
+ x = (x / 365 / 24 / 60 / 60 / 10**9).astype(int)
112
+ elif self.diff_unit == "M":
113
+ raise Exception("Unsupported difference unit: Month")
114
+ else:
115
+ x = x / np.timedelta64(1, self.diff_unit)
126
116
  return x[x > 0]
127
117
 
128
118
  def _agg(self, x):
@@ -320,7 +320,7 @@ def cross_validate(
320
320
  shuffle = cv.shuffle
321
321
  else:
322
322
  shuffle = False
323
- if hasattr(cv, "random_state"):
323
+ if hasattr(cv, "random_state") and shuffle:
324
324
  random_state = cv.random_state
325
325
  else:
326
326
  random_state = None
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.3
2
2
  Name: upgini
3
- Version: 1.1.288
3
+ Version: 1.1.288a0
4
4
  Summary: Intelligent data search & enrichment for Machine Learning
5
5
  Project-URL: Bug Reports, https://github.com/upgini/upgini/issues
6
6
  Project-URL: Homepage, https://upgini.com/
@@ -1,4 +1,4 @@
1
- upgini/__about__.py,sha256=dc9KjdGN41zFpIykuStCL0GQPyjThIUdODgFyKcKALg,24
1
+ upgini/__about__.py,sha256=3Qkh5WTdySU-oJImISkkJO-aROpU4gchsmPuaaEbmuU,26
2
2
  upgini/__init__.py,sha256=asENHgEVHQBIkV-e_0IhE_ZWqkCG6398U3ZLrNzAH6k,407
3
3
  upgini/ads.py,sha256=nvuRxRx5MHDMgPr9SiU-fsqRdFaBv8p4_v1oqiysKpc,2714
4
4
  upgini/dataset.py,sha256=7TLVVhGtjgx_9yaiaIUK3kZSe_R9wg5dY0d4F5qCGM4,45636
@@ -15,7 +15,7 @@ upgini/ads_management/ads_manager.py,sha256=igVbN2jz80Umb2BUJixmJVj-zx8unoKpecVo
15
15
  upgini/autofe/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
16
16
  upgini/autofe/all_operands.py,sha256=7UyvmmqGSqQu4kDgoFwQRKY__b9xKDk3Fpp2-H8A7AA,2399
17
17
  upgini/autofe/binary.py,sha256=441BRuqMsxlxuw4c8rMZB6h5EpRdVMk-bVa03U7T5Hg,3973
18
- upgini/autofe/date.py,sha256=qzk0NT332Q0vR1eRwTuNiMSrGE3ulh6Ic3QLBZqSdvw,7284
18
+ upgini/autofe/date.py,sha256=w0C2n261Uzd9sEk3s7QdDrXLZBWv6Vv7EBuv0W1g-LU,6738
19
19
  upgini/autofe/feature.py,sha256=_V9B74B3ue7eAYXSOt9JKhVC9klkAKks22MwnBRye_w,12487
20
20
  upgini/autofe/groupby.py,sha256=4WjDzQxqpZxB79Ih4ihMMI5GDxaFqiH6ZelfV82ClT4,3091
21
21
  upgini/autofe/operand.py,sha256=JjEVT1U3kY9NDjUPMdoki7Oa8hMDG0-_h_NklVjIFyc,2882
@@ -52,11 +52,11 @@ upgini/utils/ip_utils.py,sha256=Zf3F2cnQmOCH09QLQHetpjMFu1PnD0cTmDymn0SnSy8,1672
52
52
  upgini/utils/phone_utils.py,sha256=JNSkF8G6mgsN8Czy11pamaJdsY6rBINEMpi7jbVt_RA,408
53
53
  upgini/utils/postal_code_utils.py,sha256=_8CR9tBqsPptQsmMUvnrCAmBaMIQSWH3JfJ4ly3x_zs,409
54
54
  upgini/utils/progress_bar.py,sha256=N-Sfdah2Hg8lXP_fV9EfUTXz_PyRt4lo9fAHoUDOoLc,1550
55
- upgini/utils/sklearn_ext.py,sha256=N-eJrfAJxYpDPc85sKQyMFcIeL9Ug2lwlqDyS4jFOdE,44499
55
+ upgini/utils/sklearn_ext.py,sha256=13jQS_k7v0aUtudXV6nGUEWjttPQzAW9AFYL5wgEz9k,44511
56
56
  upgini/utils/target_utils.py,sha256=Y96_PJ5cC-WsEbeqg20v9uqywDQobLoTb-xoP7S3o4E,7807
57
57
  upgini/utils/track_info.py,sha256=G5Lu1xxakg2_TQjKZk4b5SvrHsATTXNVV3NbvWtT8k8,5663
58
58
  upgini/utils/warning_counter.py,sha256=dIWBB4dI5XRRJZudvIlqlIYKEiwLLPcXarsZuYRt338,227
59
- upgini-1.1.288.dist-info/METADATA,sha256=lUjVrD2DHFZAwdi5_HpALW7wbFRkKhu3C9rqpQwr6_k,48117
60
- upgini-1.1.288.dist-info/WHEEL,sha256=zEMcRr9Kr03x1ozGwg5v9NQBKn3kndp6LSoSlVg-jhU,87
61
- upgini-1.1.288.dist-info/licenses/LICENSE,sha256=5RRzgvdJUu3BUDfv4bzVU6FqKgwHlIay63pPCSmSgzw,1514
62
- upgini-1.1.288.dist-info/RECORD,,
59
+ upgini-1.1.288a0.dist-info/METADATA,sha256=E4zY2U029vSoJLOe5NFkIxJ5_Loj342ORFXKIYI8BjY,48119
60
+ upgini-1.1.288a0.dist-info/WHEEL,sha256=zEMcRr9Kr03x1ozGwg5v9NQBKn3kndp6LSoSlVg-jhU,87
61
+ upgini-1.1.288a0.dist-info/licenses/LICENSE,sha256=5RRzgvdJUu3BUDfv4bzVU6FqKgwHlIay63pPCSmSgzw,1514
62
+ upgini-1.1.288a0.dist-info/RECORD,,