datapipelab 0.1.6__py3-none-any.whl → 0.1.8__py3-none-any.whl

datapipelab/app/node/processor/bigquery_api_node.py

@@ -12,7 +12,7 @@ class BigQueryAPIProcessorNode(TNode):
         self.return_as_spark_df = tnode_config['options']['return_as_spark_df']
         self.project_name = tnode_config['options']['project_name']
 
-    def __sql_query(self, sql_query):
+    def __sql_biqquery(self, sql_query):
         credentials = service_account.Credentials.from_service_account_file(self.credentials_path)
         client = bigquery.Client(credentials=credentials, project=self.project_name)
 
@@ -28,6 +28,6 @@ class BigQueryAPIProcessorNode(TNode):
         logger.info(rows)
 
     def _process(self):
-        self.__sql_query(self.sql_query)
+        self.__sql_biqquery(self.sql_query)
         self._createOrReplaceTempView()
         return self.node

datapipelab/app/node/processor/bigquery_spark_node.py

@@ -1,30 +1,22 @@
 from datapipelab.app.node.tnode import TNode
 from datapipelab.logger import logger
 
+
 class BigQuerySparkProcessorNode(TNode):
     def __init__(self, spark, tnode_config):
         super().__init__(spark=spark)
         self.sql_query = tnode_config['options']['query']
         self.node_name = tnode_config['name']
-        self.credentials_path = tnode_config['options']['materialization_dataset'] # materializationDataset
-        self.return_as_spark_df = tnode_config['options']['parent_project'] # parentProject
-
-    def __sql_query(self, sql_query):
-        credentials = service_account.Credentials.from_service_account_file(self.credentials_path)
-        client = bigquery.Client(credentials=credentials, project=self.project_name)
-
-        # run the job
-        query_job = client.query(sql_query)
+        self.materialization_dataset = tnode_config['options']['materialization_dataset']  # materializationDataset
+        self.parent_project = tnode_config['options']['parent_project']  # parentProject
 
-        results = query_job.result()
-        rows = [dict(row) for row in results]
-        if self.return_as_spark_df:
-            self.node = self.spark.createDataFrame(rows)
-        else:
-            self.node = None
-        logger.info(rows)
+    def __sql_query(self):
+        self.node = self.spark.read.format("bigquery").option("materializationDataset",
+                                                              self.materialization_dataset).option("query",
+                                                                                                   self.sql_query).option(
+            "parentProject", self.parent_project).load()
 
     def _process(self):
-        self.__sql_query(self.sql_query)
+        self.__sql_query()
         self._createOrReplaceTempView()
         return self.node

datapipelab/pipeline_handler.py

@@ -7,6 +7,7 @@ from datapipelab.app.node.sink.csv_node import CSVSinkNode
 from datapipelab.app.node.sink.pandas_csv_node import PandasCSVSinkNode
 from datapipelab.app.node.sink.teams_notification_node import TeamsNotificationSinkNode
 from datapipelab.app.node.processor.bigquery_spark_node import BigQuerySparkProcessorNode
+from datapipelab.app.node.processor.bigquery_api_node import BigQueryAPIProcessorNode
 
 
 class PipelineHandler:
@@ -46,9 +47,9 @@ class PipelineHandler:
         if tnode_format == 'query':
             processor_df = SparkProcessorNode(self.spark, tnode_config).run()
         if tnode_format == 'bigquery_api':
-            processor_df = None
+            processor_df = BigQueryAPIProcessorNode(self.spark, tnode_config).run()
         if tnode_format == 'bigquery_spark':
-            processor_df = datapipelab.app.node.processor.bigquery_spark.BigQuerySparkProcessorNode(self.spark, tnode_config).run()
+            processor_df = BigQuerySparkProcessorNode(self.spark, tnode_config).run()
         return processor_df
 
     def write_sink_node(self, tnode_config, t_df):

{datapipelab-0.1.6.dist-info → datapipelab-0.1.8.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: datapipelab
-Version: 0.1.6
+Version: 0.1.8
 Summary: A data pipeline library with connectors, sources, processors, and sinks.
 Requires-Dist: json5
 Requires-Dist: loguru

{datapipelab-0.1.6.dist-info → datapipelab-0.1.8.dist-info}/RECORD

@@ -3,15 +3,15 @@ datapipelab/engine.py,sha256=3QRsedRYNov6xIDOZ1tukinFE-SKv39Fn3sNCnD3L6g,442
 datapipelab/logger.py,sha256=Ugv0A4TfD3JWCWXNWu0lURcnfAEyuVrK3IrvVVgcHBo,864
 datapipelab/pipeline.py,sha256=dw9D9KM_hztt9g_YzqoNgQBRyCYR92cRZwrU5duP_Pg,1464
 datapipelab/pipeline_config.py,sha256=2bFAJepViE7rT7CaRANZU07aeQpOYcZ954ISujm9pXA,3816
-datapipelab/pipeline_handler.py,sha256=HCyvyW3Yx2QYWQ-D-ArMDZdGtMRsJ348ZHSrqbRYwd0,3779
+datapipelab/pipeline_handler.py,sha256=lQv6HwwdgZDQvICgABtWiuvZQ9jG9cJjy8s_7qLZr9s,3871
 datapipelab/app/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 datapipelab/app/connector_node/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 datapipelab/app/node/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 datapipelab/app/node/custom_node.py,sha256=VvjwkECTobRhO_fYKUrJCd117B5MoR9P6UKYZfRLhV4,1017
 datapipelab/app/node/tnode.py,sha256=npHG4fFZty5JZ3F_okO9xml-BRhu4DkrZuNE6oaLbvw,446
 datapipelab/app/node/processor/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-datapipelab/app/node/processor/bigquery_api_node.py,sha256=2kF6hgYOUi7te-aXXr-cTZfkvitBDEPJAHT8KG-i8fE,1293
-datapipelab/app/node/processor/bigquery_spark_node.py,sha256=t8JJqMWTZwBuJUKV7-l72ZLdLVoHHSFJHFNovAY-2nc,1179
+datapipelab/app/node/processor/bigquery_api_node.py,sha256=g0LSHofMH9xxb6TW_5kSb-oY495E4lY1FZspO_CEkHw,1299
+datapipelab/app/node/processor/bigquery_spark_node.py,sha256=S9kIYW0RE5b0RjniKFFBTzA3Tx4_plFdkFQXzhl1xTY,1039
 datapipelab/app/node/processor/custom_node.py,sha256=1nqbJEhNiMP1rmN9ufpUuKO1IkuI2BEM5auW4JceGMA,933
 datapipelab/app/node/processor/shell_node.py,sha256=s3dKgfEqbpUIEiwORERgvp7FNDE5JkFHBo7EnJYBPnA,669
 datapipelab/app/node/processor/spark_node.py,sha256=jzqdffIHUCgOfMFcoqjXdl8wFag-3gafxfNCdssKnwc,483
@@ -25,7 +25,7 @@ datapipelab/app/node/source/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMp
 datapipelab/app/node/source/delta_node.py,sha256=gg7SfuKBAAfjk6OX2jNrot9XX61HoBe3us3D8O-dscE,529
 datapipelab/app/node/source/hive_node.py,sha256=h_AMCnnmon7uLRIGsaHAPWEReD3VaWZXnz9r0TpLGNM,478
 datapipelab/app/node/source/spark_node.py,sha256=S_x2atRFPDnXmhCUtcmaLc4BDFd2H4uQq6wnEJb7Uug,480
-datapipelab-0.1.6.dist-info/METADATA,sha256=yuXR8PtBINVXBFVsFncHbu9cnvJ52Jp8ZCDi-Bsmrnw,220
-datapipelab-0.1.6.dist-info/WHEEL,sha256=Nw36Djuh_5VDukK0H78QzOX-_FQEo6V37m3nkm96gtU,91
-datapipelab-0.1.6.dist-info/top_level.txt,sha256=HgeBjHvXorKzvNqU5BNPutoI771HtiqVit9_-0Zyrb4,12
-datapipelab-0.1.6.dist-info/RECORD,,
+datapipelab-0.1.8.dist-info/METADATA,sha256=JhNtfMCVgxF4ZTiCcdYQNnzfFV7lq5Xkw3GEDlR5lz4,220
+datapipelab-0.1.8.dist-info/WHEEL,sha256=zaaOINJESkSfm_4HQVc5ssNzHCPXhJm0kEUakpsEHaU,91
+datapipelab-0.1.8.dist-info/top_level.txt,sha256=HgeBjHvXorKzvNqU5BNPutoI771HtiqVit9_-0Zyrb4,12
+datapipelab-0.1.8.dist-info/RECORD,,

{datapipelab-0.1.6.dist-info → datapipelab-0.1.8.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (80.7.1)
+Generator: setuptools (80.8.0)
 Root-Is-Purelib: true
 Tag: py3-none-any