python · picnixz · Jul 12, 2025 · Apr 18, 2024 · Apr 18, 2024 · Apr 18, 2024
diff --git a/Doc/library/shelve.rst b/Doc/library/shelve.rst
@@ -17,7 +17,8 @@ This includes most class instances, recursive data types, and objects containing
 lots of shared  sub-objects.  The keys are ordinary strings.
 
 
-.. function:: open(filename, flag='c', protocol=None, writeback=False)
+.. function:: open(filename, flag='c', protocol=None, writeback=False, *, \
+                   serializer=None, deserializer=None)
 
    Open a persistent dictionary.  The filename specified is the base filename for
    the underlying database.  As a side-effect, an extension may be added to the
@@ -41,13 +42,24 @@ lots of shared  sub-objects.  The keys are ordinary strings.
    determine which accessed entries are mutable, nor which ones were actually
    mutated).
 
+   By default, :mod:`shelve` uses :func:`pickle.dumps` and :func:`pickle.loads`
+   for serializing and deserializing. This can be changed by supplying
+   *serializer* and *deserializer*, respectively. The *serializer* argument
+   should be a function that takes an object and returns its representation
+   as a :term:`bytes-like object`; *deserializer* should be a function that
+   takes :class:`bytes` and returns the corresponding object.
+   If one of these is given, the other must be given as well.
+
    .. versionchanged:: 3.10
       :const:`pickle.DEFAULT_PROTOCOL` is now used as the default pickle
       protocol.
 
    .. versionchanged:: 3.11
       Accepts :term:`path-like object` for filename.
 
+   .. versionchanged:: 3.13
+      Accepts *serializer* and *deserializer* as parameters.
+
    .. note::
 
       Do not rely on the shelf being closed automatically; always call
@@ -117,7 +129,8 @@ Restrictions
   which can cause hard crashes when trying to read from the database.
 
 
-.. class:: Shelf(dict, protocol=None, writeback=False, keyencoding='utf-8')
+.. class:: Shelf(dict, protocol=None, writeback=False, keyencoding='utf-8', \
+                 serializer=None, deserializer=None)
 
    A subclass of :class:`collections.abc.MutableMapping` which stores pickled
    values in the *dict* object.
@@ -135,6 +148,11 @@ Restrictions
    The *keyencoding* parameter is the encoding used to encode keys before they
    are used with the underlying dict.
 
+   The *deserializer* parameter can be the function that takes the
+   :term:`bytes-like object` and the *protocol* parameter and returns the
+   object. *serializer* parameter can be the function that takes the object
+   and returns :class:`bytes`.
+
    A :class:`Shelf` object can also be used as a context manager, in which
    case it will be automatically closed when the :keyword:`with` block ends.
 
@@ -149,8 +167,11 @@ Restrictions
       :const:`pickle.DEFAULT_PROTOCOL` is now used as the default pickle
       protocol.
 
+   .. versionchanged:: 3.13
+      Accepts *serializer* and *deserializer* as parameters.
 
-.. class:: BsdDbShelf(dict, protocol=None, writeback=False, keyencoding='utf-8')
+.. class:: BsdDbShelf(dict, protocol=None, writeback=False, \
+                      keyencoding='utf-8',  serializer=None, deserializer=None)
 
    A subclass of :class:`Shelf` which exposes :meth:`!first`, :meth:`!next`,
    :meth:`!previous`, :meth:`!last` and :meth:`!set_location` methods.
@@ -160,18 +181,20 @@ Restrictions
    modules.  The *dict* object passed to the constructor must support those
    methods.  This is generally accomplished by calling one of
    :func:`!bsddb.hashopen`, :func:`!bsddb.btopen` or :func:`!bsddb.rnopen`.  The
-   optional *protocol*, *writeback*, and *keyencoding* parameters have the same
-   interpretation as for the :class:`Shelf` class.
+   optional *protocol*, *writeback*, *keyencoding*, *serializer* and *deserializer*
+   parameters have the same interpretation as for the :class:`Shelf` class.
 
 
-.. class:: DbfilenameShelf(filename, flag='c', protocol=None, writeback=False)
+.. class:: DbfilenameShelf(filename, flag='c', protocol=None, writeback=False, \
+                           serializer=None, deserializer=None)
 
    A subclass of :class:`Shelf` which accepts a *filename* instead of a dict-like
    object.  The underlying file will be opened using :func:`dbm.open`.  By
    default, the file will be created and opened for both read and write.  The
-   optional *flag* parameter has the same interpretation as for the :func:`.open`
-   function.  The optional *protocol* and *writeback* parameters have the same
-   interpretation as for the :class:`Shelf` class.
+   optional *flag* parameter has the same interpretation as for the
+   :func:`.open` function.  The optional *protocol*, *writeback*, *serializer*
+   and *deserializer* parameters have the same interpretation as for the
-   and *deserializer* parameters have the same interpretation as for the
+   and *deserializer* parameters have the same interpretation as in
-   and *deserializer* parameters have the same interpretation as for the
+   and *deserializer* parameters have the same interpretation as in
+   :class:`Shelf` class.
 
 
 .. _shelve-example:

diff --git a/Lib/shelve.py b/Lib/shelve.py
@@ -56,13 +56,18 @@
 the persistent dictionary on disk, if feasible).
 """
 
-from pickle import DEFAULT_PROTOCOL, Pickler, Unpickler
+from pickle import DEFAULT_PROTOCOL, Unpickler, dumps, loads
 from io import BytesIO
 
 import collections.abc
 
 __all__ = ["Shelf", "BsdDbShelf", "DbfilenameShelf", "open"]
 
+
+class ShelveError(Exception):
+    pass
+
+
 class _ClosedDict(collections.abc.MutableMapping):
     'Marker for a closed dict.  Access attempts raise a ValueError.'
 
@@ -82,7 +87,7 @@ class Shelf(collections.abc.MutableMapping):
     """
 
     def __init__(self, dict, protocol=None, writeback=False,
-                 keyencoding="utf-8"):
+                 keyencoding="utf-8", *, serializer=None, deserializer=None):
         self.dict = dict
         if protocol is None:
             protocol = DEFAULT_PROTOCOL
@@ -91,6 +96,16 @@ def __init__(self, dict, protocol=None, writeback=False,
         self.cache = {}
         self.keyencoding = keyencoding
 
+        if serializer is None and deserializer is None:
+            self.serializer = dumps
+            self.deserializer = loads
+        elif (serializer is None) ^ (deserializer is None):
+            raise ShelveError("Serializer and deserializer must be"
+                              "defined together.")
+        else:
+            self.serializer = serializer
+            self.deserializer = deserializer
+
     def __iter__(self):
         for k in self.dict.keys():
             yield k.decode(self.keyencoding)
@@ -110,19 +125,17 @@ def __getitem__(self, key):
         try:
             value = self.cache[key]
         except KeyError:
-            f = BytesIO(self.dict[key.encode(self.keyencoding)])
-            value = Unpickler(f).load()
+            f = self.dict[key.encode(self.keyencoding)]
+            value = self.deserializer(f)
             if self.writeback:
                 self.cache[key] = value
         return value
 
     def __setitem__(self, key, value):
         if self.writeback:
             self.cache[key] = value
-        f = BytesIO()
-        p = Pickler(f, self._protocol)
-        p.dump(value)
-        self.dict[key.encode(self.keyencoding)] = f.getvalue()
+        serialized_value = self.serializer(value, self._protocol)
+        self.dict[key.encode(self.keyencoding)] = serialized_value
 
     def __delitem__(self, key):
         del self.dict[key.encode(self.keyencoding)]
@@ -186,33 +199,29 @@ class BsdDbShelf(Shelf):
     """
 
     def __init__(self, dict, protocol=None, writeback=False,
-                 keyencoding="utf-8"):
-        Shelf.__init__(self, dict, protocol, writeback, keyencoding)
+                 keyencoding="utf-8", *, serializer=None, deserializer=None):
+        Shelf.__init__(self, dict, protocol, writeback, keyencoding,
+                       serializer=serializer, deserializer=deserializer)
 
     def set_location(self, key):
         (key, value) = self.dict.set_location(key)
-        f = BytesIO(value)
-        return (key.decode(self.keyencoding), Unpickler(f).load())
+        return (key.decode(self.keyencoding), self.deserializer(value))
 
     def next(self):
         (key, value) = next(self.dict)
-        f = BytesIO(value)
-        return (key.decode(self.keyencoding), Unpickler(f).load())
+        return (key.decode(self.keyencoding), self.deserializer(value))
 
     def previous(self):
         (key, value) = self.dict.previous()
-        f = BytesIO(value)
-        return (key.decode(self.keyencoding), Unpickler(f).load())
+        return (key.decode(self.keyencoding), self.deserializer(value))
 
     def first(self):
         (key, value) = self.dict.first()
-        f = BytesIO(value)
-        return (key.decode(self.keyencoding), Unpickler(f).load())
+        return (key.decode(self.keyencoding), self.deserializer(value))
 
     def last(self):
         (key, value) = self.dict.last()
-        f = BytesIO(value)
-        return (key.decode(self.keyencoding), Unpickler(f).load())
+        return (key.decode(self.keyencoding), self.deserializer(value))
 
 
 class DbfilenameShelf(Shelf):
@@ -222,9 +231,11 @@ class DbfilenameShelf(Shelf):
     See the module's __doc__ string for an overview of the interface.
     """
 
-    def __init__(self, filename, flag='c', protocol=None, writeback=False):
+    def __init__(self, filename, flag='c', protocol=None, writeback=False,
+                 serializer=None, deserializer=None):
         import dbm
-        Shelf.__init__(self, dbm.open(filename, flag), protocol, writeback)
+        Shelf.__init__(self, dbm.open(filename, flag), protocol, writeback,
+                       serializer=serializer, deserializer=deserializer)
 
     def clear(self):
         """Remove all items from the shelf."""
@@ -233,8 +244,8 @@ def clear(self):
         self.cache.clear()
         self.dict.clear()
 
-
-def open(filename, flag='c', protocol=None, writeback=False):
+def open(filename, flag='c', protocol=None, writeback=False, *,
+         serializer=None, deserializer=None):
     """Open a persistent dictionary for reading and writing.
 
     The filename parameter is the base filename for the underlying
@@ -247,4 +258,5 @@ def open(filename, flag='c', protocol=None, writeback=False):
     See the module's __doc__ string for an overview of the interface.
     """
 
-    return DbfilenameShelf(filename, flag, protocol, writeback)
+    return DbfilenameShelf(filename, flag, protocol, writeback,
+                           serializer, deserializer)
diff --git a/Lib/test/test_shelve.py b/Lib/test/test_shelve.py
@@ -3,8 +3,10 @@
 import shelve
 import pickle
 import os
+from io import BytesIO
+from pydoc import locate
 
-from test.support import os_helper
+from test.support import os_helper, import_helper
 from collections.abc import MutableMapping
 from test.test_dbm import dbm_iterator
 
@@ -165,6 +167,145 @@ def test_default_protocol(self):
         with shelve.Shelf({}) as s:
             self.assertEqual(s._protocol, pickle.DEFAULT_PROTOCOL)
 
+    def test_custom_serializer_and_deserializer(self):
+        def serializer(obj, protocol=None):
+            return bytes(f"{type(obj).__name__}", 'utf-8')
+
+        def deserializer(data):
+            value = BytesIO(data).read()
+            return locate(value.decode("utf-8"))
+
+        os.mkdir(self.dirname)
+        self.addCleanup(os_helper.rmtree, self.dirname)
+
+        with shelve.open(self.fn,
+                         serializer=serializer,
+                         deserializer=deserializer) as s:
+            num = 1
+            s['number'] = num
+            self.assertEqual(s['number'], type(num))
+
+        with self.assertRaises(AssertionError):
+            def serializer(obj, protocol=None):
+                return bytes(f"{type(obj).__name__}", 'utf-8')
+
+            def deserializer(data):
+                pass
+
+            with shelve.open(self.fn,
+                             serializer=serializer,
+                             deserializer=deserializer) as s:
+                s['number'] = 100
+                self.assertEqual(s['number'], 100)
+
+        with self.assertRaises(dbm.sqlite3.error):
+            def serializer(obj, protocol=None):
+                pass
+
+            def deserializer(data):
+                return BytesIO(data).read().decode("utf-8")
+
+            with shelve.open(self.fn,
+                             serializer=serializer,
+                             deserializer=deserializer) as s:
+                s['number'] = 100
+                self.assertEqual(s['number'], 100)
+
+    def test_custom_serializer_and_deserializer_bsd_db_shelf(self):
+        berkeleydb = import_helper.import_module('berkeleydb')
+
+        def serializer(obj, protocol=None):
+            return bytes(f"{type(obj).__name__}", 'utf-8')
+
+        def deserializer(data):
+            value = BytesIO(data).read()
+            return locate(value.decode("utf-8"))
+
+        os.mkdir(self.dirname)
+        self.addCleanup(os_helper.rmtree, self.dirname)
+
+        with shelve.BsdDbShelf(berkeleydb.btopen(self.fn),
+                               serializer=serializer,
+                               deserializer=deserializer) as s:
+            num = 1
+            s['number'] = num
+            num2 = 2
+            s['number2'] = num2
+            self.assertEqual(s['number'], type(num))
+
+            key, value = s.previous()
+            self.assertEqual("number2", key)
+            self.assertEqual(value, type(num))
+
+            key, value = s.set_location(b'number')
+            self.assertEqual("number", key)
+            self.assertEqual(value, type(num))
+
+            key, value = s.next()
+            self.assertEqual("number2", key)
+            self.assertEqual(value, type(num))
+
+            key, value = s.first()
+            self.assertEqual("number", key)
+            self.assertEqual(s['number'], value)
+
+            key, value = s.last()
+            self.assertEqual("number2", key)
+            self.assertEqual(s['number2'], value)
+
+        with self.assertRaises(AssertionError):
+            def serializer(obj, protocol=None):
+                return bytes(f"{type(obj).__name__}", 'utf-8')
+
+            def deserializer(data):
+                pass
+
+            with shelve.BsdDbShelf(berkeleydb.btopen(self.fn),
+                                   serializer=serializer,
+                                   deserializer=deserializer) as s:
+                s['number'] = 100
+                self.assertEqual(s['number'], 100)
+
+        def serializer(obj, protocol=None):
+            pass
+
+        def deserializer(data):
+            return BytesIO(data).read().decode("utf-8")
+
+        with shelve.BsdDbShelf(berkeleydb.btopen(self.fn),
+                               serializer=serializer,
+                               deserializer=deserializer) as s:
+            s['number'] = 100
+            self.assertNotEqual(s['number'], 100)
+            self.assertEqual(s['number'], "")
+
+    def test_missing_custom_deserializer(self):
+        def serializer(obj, protocol=None):
+            pass
+
+        with self.assertRaises(shelve.ShelveError):
+            shelve.Shelf({},
+                         protocol=2, writeback=False, serializer=serializer)
+
+        with self.assertRaises(shelve.ShelveError):
+            shelve.BsdDbShelf({},
+                              protocol=2,
+                              writeback=False, serializer=serializer)
+
+    def test_missing_custom_serializer(self):
+        def deserializer(data):
+            pass
+
+        with self.assertRaises(shelve.ShelveError):
+            shelve.Shelf({},
+                         protocol=2,
+                         writeback=False, deserializer=deserializer)
+
+        with self.assertRaises(shelve.ShelveError):
+            shelve.BsdDbShelf({},
+                              protocol=2,
+                              writeback=False, deserializer=deserializer)
+
 
 class TestShelveBase:
     type2test = shelve.Shelf