A temporary quick fix for dataclass serialization (#500)

dill/_dill.py

@@ -193,6 +193,10 @@ def get_file_type(*args, **kwargs):
     singletontypes = []
 
 import inspect
+import dataclasses
+
+from pickle import GLOBAL
+
 
 ### Shims for different versions of Python and dill
 class Sentinel(object):
@@ -2024,6 +2028,41 @@ def save_capsule(pickler, obj):
 else:
     _testcapsule = None
 
+
+#############################
+# A quick fix for issue #500
+# This should be removed when a better solution is found.
+
+if hasattr(dataclasses, "_HAS_DEFAULT_FACTORY_CLASS"):
+    @register(dataclasses._HAS_DEFAULT_FACTORY_CLASS)
+    def save_dataclasses_HAS_DEFAULT_FACTORY_CLASS(pickler, obj):
+        logger.trace(pickler, "DcHDF: %s", obj)
+        pickler.write(GLOBAL + b"dataclasses\n_HAS_DEFAULT_FACTORY\n")
+        logger.trace(pickler, "# DcHDF")
+
+if hasattr(dataclasses, "MISSING"):
+    @register(type(dataclasses.MISSING))
+    def save_dataclasses_MISSING_TYPE(pickler, obj):
+        logger.trace(pickler, "DcM: %s", obj)
+        pickler.write(GLOBAL + b"dataclasses\nMISSING\n")
+        logger.trace(pickler, "# DcM")
+
+if hasattr(dataclasses, "KW_ONLY"):
+    @register(type(dataclasses.KW_ONLY))
+    def save_dataclasses_KW_ONLY_TYPE(pickler, obj):
+        logger.trace(pickler, "DcKWO: %s", obj)
+        pickler.write(GLOBAL + b"dataclasses\nKW_ONLY\n")
+        logger.trace(pickler, "# DcKWO")
+
+if hasattr(dataclasses, "_FIELD_BASE"):
+    @register(dataclasses._FIELD_BASE)
+    def save_dataclasses_FIELD_BASE(pickler, obj):
+        logger.trace(pickler, "DcFB: %s", obj)
+        pickler.write(GLOBAL + b"dataclasses\n" + obj.name.encode() + b"\n")
+        logger.trace(pickler, "# DcFB")
+
+#############################
+
 # quick sanity checking
 def pickles(obj,exact=False,safe=False,**kwds):
     """

dill/_objects.py

@@ -23,6 +23,7 @@
 import collections
 import codecs
 import struct
+import dataclasses
 import datetime
 import calendar
 import weakref
@@ -493,6 +494,18 @@ class _Struct(ctypes.Structure):
     d['PyCapsuleType'] = _testcapsule
 del _testcapsule
 
+if hasattr(dataclasses, '_HAS_DEFAULT_FACTORY'):
+    x['DataclassesHasDefaultFactoryType'] = dataclasses._HAS_DEFAULT_FACTORY
+
+if hasattr(dataclasses, 'MISSING'):
+    x['DataclassesMissingType'] = dataclasses.MISSING
+
+if hasattr(dataclasses, 'KW_ONLY'):
+    x['DataclassesKWOnlyType'] = dataclasses.KW_ONLY
+
+if hasattr(dataclasses, '_FIELD_BASE'):
+    x['DataclassesFieldBaseType'] = dataclasses._FIELD
+
 # -- cleanup ----------------------------------------------------------------
 a.update(d) # registered also succeed
 if sys.platform[:3] == 'win':

dill/tests/test_dataclasses.py

@@ -0,0 +1,35 @@
+#!/usr/bin/env python
+#
+# Author: Mike McKerns (mmckerns @caltech and @uqfoundation)
+# Author: Anirudh Vegesana (avegesan@cs.stanford.edu)
+# Copyright (c) 2022 The Uncertainty Quantification Foundation.
+# License: 3-clause BSD.  The full license text is available at:
+#  - https://github.com/uqfoundation/dill/blob/master/LICENSE
+"""
+test pickling a dataclass
+"""
+
+import dill
+import dataclasses
+
+def test_dataclasses():
+    # Issue #500
+    @dataclasses.dataclass
+    class A:
+        x: int
+        y: str
+
+    @dataclasses.dataclass
+    class B:
+        a: A
+
+    a = A(1, "test")
+    before = B(a)
+    save = dill.dumps(before)
+    after = dill.loads(save)
+    assert before != after # classes don't match
+    assert before == B(A(**dataclasses.asdict(after.a)))
+    assert dataclasses.asdict(before) == dataclasses.asdict(after)
+
+if __name__ == '__main__':
+    test_dataclasses()