Add custom serialization support for pyarrow

Closes #2103

Author: dhirschf

distributed/protocol/__init__.py

@@ -37,3 +37,8 @@ def _register_keras():
 @partial(register_serialization_lazy, "sparse")
 def _register_sparse():
     from . import sparse
+
+
+@partial(register_serialization_lazy, "arrow")
+def _register_arrow():
+    from . import arrow

distributed/protocol/arrow.py

@@ -0,0 +1,53 @@
+from __future__ import print_function, division, absolute_import
+
+from .serialize import register_serialization
+
+
+def serialize_batch(batch):
+    import pyarrow as pa
+    sink = pa.BufferOutputStream()
+    writer = pa.RecordBatchStreamWriter(sink, batch.schema)
+    writer.write_batch(batch)
+    writer.close()
+    buf = sink.get_result()
+    header = {}
+    frames = [buf.to_pybytes()]
+    return header, frames
+
+
+def deserialize_batch(header, frames):
+    import pyarrow as pa
+    blob = frames[0]
+    reader = pa.RecordBatchStreamReader(pa.BufferReader(blob))
+    return reader.read_next_batch()
+
+
+def serialize_table(tbl):
+    import pyarrow as pa
+    sink = pa.BufferOutputStream()
+    writer = pa.RecordBatchStreamWriter(sink, tbl.schema)
+    writer.write_table(tbl)
+    writer.close()
+    buf = sink.get_result()
+    header = {}
+    frames = [buf.to_pybytes()]
+    return header, frames
+
+
+def deserialize_table(header, frames):
+    import pyarrow as pa
+    blob = frames[0]
+    reader = pa.RecordBatchStreamReader(pa.BufferReader(blob))
+    return reader.read_all()
+
+
+register_serialization(
+    'pyarrow.lib.RecordBatch',
+    serialize_batch,
+    deserialize_batch
+)
+register_serialization(
+    'pyarrow.lib.Table',
+    serialize_table,
+    deserialize_table
+)