[GitHub] [beam] TheNeuralBit commented on a diff in pull request #17253: [BEAM-14213] Add API and construction time validation for Batched DoFns

[GitBox <gi...@apache.org>]

TheNeuralBit commented on code in PR #17253:
URL: https://github.com/apache/beam/pull/17253#discussion_r842236823


##########
sdks/python/apache_beam/typehints/batch.py:
##########
@@ -0,0 +1,301 @@
+#
+# Licensed to the Apache Software Foundation (ASF) under one or more
+# contributor license agreements.  See the NOTICE file distributed with
+# this work for additional information regarding copyright ownership.
+# The ASF licenses this file to You under the Apache License, Version 2.0
+# (the "License"); you may not use this file except in compliance with
+# the License.  You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+"""Utilities for type-hinting batched types for use in the Beam SDK.
+
+A batched type is a type B that is logically equivalent to Sequence[E], where E
+is some other type. Typically B has a different physical representation than
+Sequence[E] for performance reasons.
+
+A trivial example is B=np.array(dtype=np.int64), E=int.
+
+Batched type hints are used to enable more efficient processing of
+a PCollection[E], by allowing users to write DoFns that operate on
+multi-element partitions of the PCollection represented with type B."""
+
+from typing import Generic
+from typing import Iterator
+from typing import Optional
+from typing import Sequence
+from typing import TypeVar
+
+import numpy as np
+
+import apache_beam as beam
+from apache_beam.typehints import row_type
+from apache_beam.typehints.typehints import TypeConstraint
+
+try:
+  import pandas as pd
+except ImportError:
+  pd = None
+
+B = TypeVar('B')
+E = TypeVar('E')
+
+BATCH_CONVERTER_REGISTRY = []
+
+
+class BatchConverter(Generic[B, E]):
+  def __init__(self, batch_type, element_type):
+    self._batch_type = batch_type
+    self._element_type = element_type
+
+  def produce_batch(self, elements: Sequence[E]) -> B:
+    """Convert an instance of List[E] to a single instance of B."""
+    raise NotImplementedError
+
+  def explode_batch(self, batch: B) -> Iterator[E]:
+    """Convert an instance of B to Iterator[E]."""
+    raise NotImplementedError
+
+  def combine_batches(self, batches: Sequence[B]) -> B:
+    raise NotImplementedError
+
+  def get_length(self, batch: B) -> int:
+    raise NotImplementedError
+
+  @staticmethod
+  def register(batching_util_fn):
+    BATCH_CONVERTER_REGISTRY.append(batching_util_fn)
+    return batching_util_fn
+
+  # TODO: Consider required kwargs here. Easy to mix up element/batch_type
+  @staticmethod
+  def from_typehints(*, element_type, batch_type) -> 'BatchConverter':
+    for constructor in BATCH_CONVERTER_REGISTRY:
+      result = constructor(element_type, batch_type)
+      if result is not None:
+        return result
+
+    # TODO: Include explanations for realistic candidate BatchConverter
+    raise TypeError(
+        f"Unable to find BatchConverter for element_type {element_type!r} and "
+        f"batch_type {batch_type!r}")
+
+  @property
+  def batch_type(self):
+    return self._batch_type
+
+  @property
+  def element_type(self):
+    return self._element_type
+
+
+N = "ARBITRARY LENGTH DIMENSION"
+
+
+class BatchTypeConstraint(TypeConstraint):

Review Comment:
   TODO: Remove BatchTypeConstraint concept. This is from a previous iteration of the Batched DoFn concept.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: github-unsubscribe@beam.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org