upgini 1.1.296a3521.dev9__py3-none-any.whl → 1.1.296a3521.dev10__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__ = "1.1.296a3521.dev9"
1
+ __version__ = "1.1.296a3521.dev10"
upgini/autofe/feature.py CHANGED
@@ -146,7 +146,7 @@ class Feature:
146
146
  elif shorten and not self.op.is_unary:
147
147
  components = ["f_autofe", self.get_op_display_name()]
148
148
  else:
149
- components = ["f_" + "_f_".join(self.get_columns())] + [
149
+ components = ["f_" + "_f_".join(self.get_columns(**kwargs))] + [
150
150
  "autofe",
151
151
  self.get_op_display_name(),
152
152
  ]
@@ -59,9 +59,35 @@ class DataSourcePublisher:
59
59
  features_for_embeddings: Optional[List[str]] = DEFAULT_GENERATE_EMBEDDINGS,
60
60
  data_table_id_to_replace: Optional[str] = None,
61
61
  keep_features: Optional[List[str]] = None,
62
+ date_features: Optional[List[str]] = None,
63
+ date_vector_features: Optional[List[str]] = None,
62
64
  _force_generation=False,
63
65
  _silent=False,
64
66
  ) -> str:
67
+ """Register new ADS
68
+
69
+ Parameters
70
+ ----------
71
+ data_table_uri - str - table name in format {project_id}.{datasource_name}.{table_name}
72
+
73
+ search_keys - dict with column names as keys and SearchKey as value
74
+
75
+ update_frequency - str - (Monthly, Weekly, Daily, Annually, Quarterly)
76
+
77
+ exclude_from_autofe_generation - optional list of features that should be excluded from AutoFE
78
+
79
+ secondary_search_keys - optional dict of secondary search keys
80
+
81
+ sort_column - optional str - name of unique column that could be used for sort
82
+
83
+ date_format - optional str - format of date if it is present in search keys
84
+
85
+ ...
86
+
87
+ data_table_id_to_replace - optional str - id of registered ADS that should be replaced by new table
88
+
89
+ keep_features - optional list - features that should not be removed from ADS (even if they are personal)
90
+ """
65
91
  trace_id = str(uuid.uuid4())
66
92
 
67
93
  with MDC(trace_id=trace_id):
@@ -124,6 +150,14 @@ class DataSourcePublisher:
124
150
  request["excludeFromGeneration"] = exclude_from_autofe_generation
125
151
  if keep_features is not None:
126
152
  request["keepFeatures"] = keep_features
153
+ if date_features is not None:
154
+ if date_format is None:
155
+ raise ValidationError("date_format should be presented if you use date features")
156
+ request["dateFeatures"] = date_features
157
+ if date_vector_features is not None:
158
+ if date_format is None:
159
+ raise ValidationError("date_format should be presented if you use date vector features")
160
+ request["dateVectorFeatures"] = date_vector_features
127
161
  self.logger.info(f"Start registering data table {request}")
128
162
 
129
163
  task_id = self._rest_client.register_ads(request, trace_id)
@@ -181,6 +215,9 @@ class DataSourcePublisher:
181
215
  msg = f"Data table successfully registered with id: {data_table_id}"
182
216
  self.logger.info(msg)
183
217
  print(msg)
218
+ if "warnings" in status_response and status_response["warnings"]:
219
+ self.logger.warning(status_response["warnings"])
220
+ print(status_response["warnings"])
184
221
  return data_table_id
185
222
  except KeyboardInterrupt:
186
223
  if task_id is not None:
@@ -2870,7 +2870,7 @@ class FeaturesEnricher(TransformerMixin):
2870
2870
  self.logger.info(f"Dates interval is ({min_date}, {max_date})")
2871
2871
 
2872
2872
  except Exception:
2873
- self.logger.exception("Failed to log debug information")
2873
+ self.logger.warning("Failed to log debug information", exc_info=True)
2874
2874
 
2875
2875
  def __handle_index_search_keys(self, df: pd.DataFrame, search_keys: Dict[str, SearchKey]) -> pd.DataFrame:
2876
2876
  index_names = df.index.names if df.index.names != [None] else [DEFAULT_INDEX]
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.3
2
2
  Name: upgini
3
- Version: 1.1.296a3521.dev9
3
+ Version: 1.1.296a3521.dev10
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/
@@ -1,9 +1,9 @@
1
- upgini/__about__.py,sha256=2F_HOCFdMZavmlP6CydWnRXZDOzDzF1tIkugqbcE9-g,34
1
+ upgini/__about__.py,sha256=JKvS85WNogY-2YIe8YIQ4nIWCAaaPwBgVr2LGrQzI7g,35
2
2
  upgini/__init__.py,sha256=ObEtjFkIssl83qeKNMLpIQygfwK8TzztwiI43YTsAP0,353
3
3
  upgini/ads.py,sha256=nvuRxRx5MHDMgPr9SiU-fsqRdFaBv8p4_v1oqiysKpc,2714
4
4
  upgini/dataset.py,sha256=7TLVVhGtjgx_9yaiaIUK3kZSe_R9wg5dY0d4F5qCGM4,45636
5
5
  upgini/errors.py,sha256=2b_Wbo0OYhLUbrZqdLIx5jBnAsiD1Mcenh-VjR4HCTw,950
6
- upgini/features_enricher.py,sha256=rLy6BwhL94VkRCp8W4RxJ0lBb7sZqTOsz-bUVI8nXZU,177530
6
+ upgini/features_enricher.py,sha256=HQFLw3VyEsZfAt4xFnIYOnp3fzQSHAsyHzIm0gTJpOI,177543
7
7
  upgini/http.py,sha256=bp6jWl422Icy3AhHMdCcJv5NjExE45gSMmzMTPJjPuk,42600
8
8
  upgini/lazy_import.py,sha256=EwoM0msNGbSmWBhGbrLDny1DSnOlvTxCjmMKPxYlDms,610
9
9
  upgini/metadata.py,sha256=qDAIO7NLSSQp_XiXCv3U4XJTLO0KH3YuQ8lvCLYPqzs,9781
@@ -17,13 +17,13 @@ upgini/autofe/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
17
17
  upgini/autofe/all_operands.py,sha256=XbvgX2IU4aee9rJZ--d5MdmrfKhON_emle5-RU1qlEY,2506
18
18
  upgini/autofe/binary.py,sha256=8FXPJxN7fnC5wphO0Dp1tQCa0lFMSDGQGvBMkSIVAcE,4155
19
19
  upgini/autofe/date.py,sha256=8zYVhjl7jVS4xt-IjCgk9px2LHnACX2YlMlmDELlRTc,7943
20
- upgini/autofe/feature.py,sha256=f4j3n0WlXmIacvnYgd-dcnBFlzCBkGao-qyTplH2X3c,13492
20
+ upgini/autofe/feature.py,sha256=nV1oJCT65nsAYFfXYCXYNYXZPaZVpqKJ09iCMTldooc,13500
21
21
  upgini/autofe/groupby.py,sha256=4WjDzQxqpZxB79Ih4ihMMI5GDxaFqiH6ZelfV82ClT4,3091
22
22
  upgini/autofe/operand.py,sha256=MKEsl3zxpWzRDpTkE0sNJxTu62U20sWOvEKhPjUWS6s,2915
23
23
  upgini/autofe/unary.py,sha256=ZWjLd-CUkNt_PpM8YuWLLipW1v_RdBlsl4JxXIVo9aM,3652
24
24
  upgini/autofe/vector.py,sha256=dLxfAstJs-gw_OQ1xxoxcM6pVzORlV0HVzdzt7cLXVQ,606
25
25
  upgini/data_source/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
26
- upgini/data_source/data_source_publisher.py,sha256=6paupnciqagACnSzjGSkA2a5i-c9ETvZheLqBYOJemk,17810
26
+ upgini/data_source/data_source_publisher.py,sha256=1cQZrK630VztwGGDp41ec9gqIeUtkefaqSSQEitVWiM,19581
27
27
  upgini/mdc/__init__.py,sha256=aM08nIWFc2gWdWUa3_IuEnNND0cQPkBGnYpRMnfFN8k,1019
28
28
  upgini/mdc/context.py,sha256=3u1B-jXt7tXEvNcV3qmR9SDCseudnY7KYsLclBdwVLk,1405
29
29
  upgini/normalizer/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -57,7 +57,7 @@ upgini/utils/sklearn_ext.py,sha256=13jQS_k7v0aUtudXV6nGUEWjttPQzAW9AFYL5wgEz9k,4
57
57
  upgini/utils/target_utils.py,sha256=Y96_PJ5cC-WsEbeqg20v9uqywDQobLoTb-xoP7S3o4E,7807
58
58
  upgini/utils/track_info.py,sha256=G5Lu1xxakg2_TQjKZk4b5SvrHsATTXNVV3NbvWtT8k8,5663
59
59
  upgini/utils/warning_counter.py,sha256=dIWBB4dI5XRRJZudvIlqlIYKEiwLLPcXarsZuYRt338,227
60
- upgini-1.1.296a3521.dev9.dist-info/METADATA,sha256=rPLhUPU3flV46eJUOaAVuDlezi56zAhzub_IRvH1_L0,48161
61
- upgini-1.1.296a3521.dev9.dist-info/WHEEL,sha256=zEMcRr9Kr03x1ozGwg5v9NQBKn3kndp6LSoSlVg-jhU,87
62
- upgini-1.1.296a3521.dev9.dist-info/licenses/LICENSE,sha256=5RRzgvdJUu3BUDfv4bzVU6FqKgwHlIay63pPCSmSgzw,1514
63
- upgini-1.1.296a3521.dev9.dist-info/RECORD,,
60
+ upgini-1.1.296a3521.dev10.dist-info/METADATA,sha256=RZTVD4L0dwwznJDxfpxjmCTZGzSQjE8lX1fw5CjP9ZA,48162
61
+ upgini-1.1.296a3521.dev10.dist-info/WHEEL,sha256=zEMcRr9Kr03x1ozGwg5v9NQBKn3kndp6LSoSlVg-jhU,87
62
+ upgini-1.1.296a3521.dev10.dist-info/licenses/LICENSE,sha256=5RRzgvdJUu3BUDfv4bzVU6FqKgwHlIay63pPCSmSgzw,1514
63
+ upgini-1.1.296a3521.dev10.dist-info/RECORD,,