datapipelab 0.1.8__py3-none-any.whl → 0.2.0__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.
@@ -3,16 +3,18 @@ from datapipelab.logger import logger
3
3
 
4
4
  class BigQueryAPIProcessorNode(TNode):
5
5
  def __init__(self, spark, tnode_config):
6
- from google.cloud import bigquery
7
- from google.oauth2 import service_account
8
6
  super().__init__(spark=spark)
9
7
  self.sql_query = tnode_config['options']['query']
10
8
  self.node_name = tnode_config['name']
11
9
  self.credentials_path = tnode_config['options']['credentials_path']
12
10
  self.return_as_spark_df = tnode_config['options']['return_as_spark_df']
13
11
  self.project_name = tnode_config['options']['project_name']
12
+ self.return_as_python_list = tnode_config['options'].get('return_as_python_list', False)
13
+ self.return_as_is = tnode_config['options'].get('return_as_is', False)
14
14
 
15
15
  def __sql_biqquery(self, sql_query):
16
+ from google.cloud import bigquery
17
+ from google.oauth2 import service_account
16
18
  credentials = service_account.Credentials.from_service_account_file(self.credentials_path)
17
19
  client = bigquery.Client(credentials=credentials, project=self.project_name)
18
20
 
@@ -20,12 +22,17 @@ class BigQueryAPIProcessorNode(TNode):
20
22
  query_job = client.query(sql_query)
21
23
 
22
24
  results = query_job.result()
23
- rows = [dict(row) for row in results]
24
25
  if self.return_as_spark_df:
26
+ rows = [dict(row) for row in results]
25
27
  self.node = self.spark.createDataFrame(rows)
28
+ elif self.return_as_python_list:
29
+ rows = [dict(row) for row in results]
30
+ self.node = rows
31
+ elif self.return_as_is:
32
+ self.node = results
26
33
  else:
27
34
  self.node = None
28
- logger.info(rows)
35
+ logger.info([dict(row) for row in results]) # TODO: Remove this
29
36
 
30
37
  def _process(self):
31
38
  self.__sql_biqquery(self.sql_query)
@@ -9,7 +9,8 @@ class TNode:
9
9
 
10
10
  # Source and Processor nodes
11
11
  def _createOrReplaceTempView(self):
12
- self.node.createOrReplaceTempView(self.node_name)
12
+ if self.node is not None:
13
+ self.node.createOrReplaceTempView(self.node_name)
13
14
 
14
15
  def run(self):
15
16
  return self._process()
@@ -1,4 +1,5 @@
1
1
  from datapipelab.app.node.processor.custom_node import CustomNode
2
+ from datapipelab.app.node.processor.shell_node import ShellProcessorNode
2
3
  from datapipelab.app.node.source.spark_node import SparkSourceNode
3
4
  from datapipelab.app.node.source.delta_node import DeltaSourceNode
4
5
  from datapipelab.app.node.processor.spark_node import SparkProcessorNode
@@ -50,6 +51,8 @@ class PipelineHandler:
50
51
  processor_df = BigQueryAPIProcessorNode(self.spark, tnode_config).run()
51
52
  if tnode_format == 'bigquery_spark':
52
53
  processor_df = BigQuerySparkProcessorNode(self.spark, tnode_config).run()
54
+ if tnode_format == 'shell':
55
+ processor_df = ShellProcessorNode(self.spark, tnode_config).run()
53
56
  return processor_df
54
57
 
55
58
  def write_sink_node(self, tnode_config, t_df):
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: datapipelab
3
- Version: 0.1.8
3
+ Version: 0.2.0
4
4
  Summary: A data pipeline library with connectors, sources, processors, and sinks.
5
5
  Requires-Dist: json5
6
6
  Requires-Dist: loguru
@@ -3,14 +3,14 @@ datapipelab/engine.py,sha256=3QRsedRYNov6xIDOZ1tukinFE-SKv39Fn3sNCnD3L6g,442
3
3
  datapipelab/logger.py,sha256=Ugv0A4TfD3JWCWXNWu0lURcnfAEyuVrK3IrvVVgcHBo,864
4
4
  datapipelab/pipeline.py,sha256=dw9D9KM_hztt9g_YzqoNgQBRyCYR92cRZwrU5duP_Pg,1464
5
5
  datapipelab/pipeline_config.py,sha256=2bFAJepViE7rT7CaRANZU07aeQpOYcZ954ISujm9pXA,3816
6
- datapipelab/pipeline_handler.py,sha256=lQv6HwwdgZDQvICgABtWiuvZQ9jG9cJjy8s_7qLZr9s,3871
6
+ datapipelab/pipeline_handler.py,sha256=oWhpnzMQJp_5Dhy2_EiGbq8W0twsIPLpnU3OOJkYxT8,4058
7
7
  datapipelab/app/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
8
8
  datapipelab/app/connector_node/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
9
9
  datapipelab/app/node/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
10
10
  datapipelab/app/node/custom_node.py,sha256=VvjwkECTobRhO_fYKUrJCd117B5MoR9P6UKYZfRLhV4,1017
11
- datapipelab/app/node/tnode.py,sha256=npHG4fFZty5JZ3F_okO9xml-BRhu4DkrZuNE6oaLbvw,446
11
+ datapipelab/app/node/tnode.py,sha256=-2hnQkIuLwEy7xVTig54TByO7L2l7UujolXMQL0CQJA,484
12
12
  datapipelab/app/node/processor/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
13
- datapipelab/app/node/processor/bigquery_api_node.py,sha256=g0LSHofMH9xxb6TW_5kSb-oY495E4lY1FZspO_CEkHw,1299
13
+ datapipelab/app/node/processor/bigquery_api_node.py,sha256=fLs8g1R8iRIZIJL66wbXONmStbhR4Q8Mw3QsZgp7MpI,1710
14
14
  datapipelab/app/node/processor/bigquery_spark_node.py,sha256=S9kIYW0RE5b0RjniKFFBTzA3Tx4_plFdkFQXzhl1xTY,1039
15
15
  datapipelab/app/node/processor/custom_node.py,sha256=1nqbJEhNiMP1rmN9ufpUuKO1IkuI2BEM5auW4JceGMA,933
16
16
  datapipelab/app/node/processor/shell_node.py,sha256=s3dKgfEqbpUIEiwORERgvp7FNDE5JkFHBo7EnJYBPnA,669
@@ -25,7 +25,7 @@ datapipelab/app/node/source/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMp
25
25
  datapipelab/app/node/source/delta_node.py,sha256=gg7SfuKBAAfjk6OX2jNrot9XX61HoBe3us3D8O-dscE,529
26
26
  datapipelab/app/node/source/hive_node.py,sha256=h_AMCnnmon7uLRIGsaHAPWEReD3VaWZXnz9r0TpLGNM,478
27
27
  datapipelab/app/node/source/spark_node.py,sha256=S_x2atRFPDnXmhCUtcmaLc4BDFd2H4uQq6wnEJb7Uug,480
28
- datapipelab-0.1.8.dist-info/METADATA,sha256=JhNtfMCVgxF4ZTiCcdYQNnzfFV7lq5Xkw3GEDlR5lz4,220
29
- datapipelab-0.1.8.dist-info/WHEEL,sha256=zaaOINJESkSfm_4HQVc5ssNzHCPXhJm0kEUakpsEHaU,91
30
- datapipelab-0.1.8.dist-info/top_level.txt,sha256=HgeBjHvXorKzvNqU5BNPutoI771HtiqVit9_-0Zyrb4,12
31
- datapipelab-0.1.8.dist-info/RECORD,,
28
+ datapipelab-0.2.0.dist-info/METADATA,sha256=_vL7fMDr9xkCojabJuY6-RUAQoyle8LxH1aRdeWVe1o,220
29
+ datapipelab-0.2.0.dist-info/WHEEL,sha256=zaaOINJESkSfm_4HQVc5ssNzHCPXhJm0kEUakpsEHaU,91
30
+ datapipelab-0.2.0.dist-info/top_level.txt,sha256=HgeBjHvXorKzvNqU5BNPutoI771HtiqVit9_-0Zyrb4,12
31
+ datapipelab-0.2.0.dist-info/RECORD,,