def from_source(self,
source: Source,
watermark_strategy: WatermarkStrategy,
source_name: str,
type_info: TypeInformation = None) -> 'DataStream':
"""
Adds a data :class:`~pyflink.datastream.connectors.Source` to the environment to get a
:class:`~pyflink.datastream.DataStream`.
The result will be either a bounded data stream (that can be processed in a batch way) or
an unbounded data stream (that must be processed in a streaming way), based on the
boundedness property of the source.
This method takes an explicit type information for the produced data stream, so that
callers can define directly what type/serializer will be used for the produced stream. For
sources that describe their produced type, the parameter type_info should not be specified
to avoid specifying the produced type redundantly.
.. versionadded:: 1.13.0
"""
if type_info:
j_type_info = type_info.get_java_type_info()
else:
j_type_info = None
j_data_stream = self._j_stream_execution_environment.fromSource(
source.get_java_function(),
watermark_strategy._j_watermark_strategy,
source_name,
j_type_info)
return DataStream(j_data_stream=j_data_stream)
def add_source(self, source: Source) -> 'HybridSourceBuilder':
self._j_hybrid_source_builder.addSource(source.get_java_function())
return self
def builder(first_source: Source) -> 'HybridSourceBuilder':
JHybridSource = get_gateway(
).jvm.org.apache.flink.connector.base.source.hybrid.HybridSource
return HybridSourceBuilder(
JHybridSource.builder(first_source.get_java_function()))