15067549553

Committed 16 May 2025 11:36AM UTC coverage: 90.475% (+0.03%) from 90.445%

Build # 15067549553

Build Type

Pull #9345

github

Committed by

web-flow

Commit Message

Merge 0c4204fda into 707573d96

Pull Request Pull Request #9345: feat: add serialization to `State` / move `State` to utils

Run Details

11000 of 12158 relevant lines covered (90.48%)

0.9 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

95.45

haystack/utils/base_serialization.py

# SPDX-FileCopyrightText: 2022-present deepset GmbH <info@deepset.ai>
#
# SPDX-License-Identifier: Apache-2.0

from typing import Any, Dict

from haystack.core.errors import DeserializationError, SerializationError
from haystack.core.serialization import generate_qualified_class_name, import_class_by_name


def serialize_class_instance(obj: Any) -> Dict[str, Any]:
    """
    Serializes an object that has a `to_dict` method into a dictionary.

    :param obj:
        The object to be serialized.
    :returns:
        A dictionary representation of the object.
    :raises SerializationError:
        If the object does not have a `to_dict` method.
    """
    if not hasattr(obj, "to_dict"):
        raise SerializationError(f"Object of class '{type(obj).__name__}' does not have a 'to_dict' method")

    output = obj.to_dict()
    return {"type": generate_qualified_class_name(type(obj)), "data": output}


def deserialize_class_instance(data: Dict[str, Any]) -> Any:
    """
    Deserializes an object from a dictionary representation generated by `auto_serialize_class_instance`.

    :param data:
        The dictionary to deserialize from.
    :returns:
        The deserialized object.
    :raises DeserializationError:
        If the serialization data is malformed, the class type cannot be imported, or the
        class does not have a `from_dict` method.
    """
    if "type" not in data:
        raise DeserializationError("Missing 'type' in serialization data")
    if "data" not in data:
        raise DeserializationError("Missing 'data' in serialization data")

    try:
        obj_class = import_class_by_name(data["type"])
    except ImportError as e:
        raise DeserializationError(f"Class '{data['type']}' not correctly imported") from e

    if not hasattr(obj_class, "from_dict"):
        raise DeserializationError(f"Class '{data['type']}' does not have a 'from_dict' method")

    return obj_class.from_dict(data["data"])


def serialize_value(value: Any) -> Any:
    """
    Serializes a value into a format suitable for storage or transmission.

    Handles various types including:
    - Haystack dataclass objects (Answer, Document, etc.)
    - Primitive types (returned as is)
    - Lists of primitives (returned as is)
    - Lists of complex types (recursively serialized)
    - Dictionaries (recursively serialized)

    :param value: The value to serialize
    :returns: The serialized representation of the value
    """

    if hasattr(value, "to_dict") and callable(getattr(value, "to_dict")):
        serialized_value = value.to_dict()
        serialized_value["_type"] = generate_qualified_class_name(type(value))

        return serialized_value

    # this is a hack to serialize inputs that don't have a to_dict
    elif hasattr(value, "__dict__"):
        return {"_type": value.__class__, "attributes": value.__dict__}

    # recursively serialize all inputs in a dict
    elif isinstance(value, dict):
        return {k: serialize_value(v) for k, v in value.items()}

    # recursively serialize all inputs in lists or tuples
    elif isinstance(value, (list, tuple, set)):
        return type(value)(serialize_value(item) for item in value)

    return value


# pylint: disable=too-many-return-statements
def deserialize_value(value: Any) -> Any:
    """
    Deserializes a value from its serialized representation.

    Handles various types including:
    - Haystack dataclass objects (Answer, Document, etc.)
    - Primitive types (returned as is)
    - Lists of primitives (returned as is)
    - Lists of complex types (recursively deserialized)
    - Dictionaries (recursively deserialized)

    :param value: The serialized value to deserialize
    :returns: The deserialized value
    """

    # None or primitive types are returned as is
    if not value or isinstance(value, (str, int, float, bool)):
        return value

    # check if the dictionary has a "_type" key and the class type has a "from_dict" method
    if isinstance(value, dict):
        if "_type" in value:
            obj_class = import_class_by_name(value.pop("_type"))
            if hasattr(obj_class, "from_dict"):
                return obj_class.from_dict(value)

        # If not a known type, recursively deserialize each item in the dictionary
        return {k: deserialize_value(v) for k, v in value.items()}

    if isinstance(value, (list, tuple, set)):
        return type(value)(deserialize_value(i) for i in value)

    return value

1	# SPDX-FileCopyrightText: 2022-present deepset GmbH <info@deepset.ai>
2	#
3	# SPDX-License-Identifier: Apache-2.0
4
5	from typing import Any, Dict	1✔
6
7	from haystack.core.errors import DeserializationError, SerializationError	1✔
8	from haystack.core.serialization import generate_qualified_class_name, import_class_by_name	1✔
9
10
11	def serialize_class_instance(obj: Any) -> Dict[str, Any]:	1✔
12	"""
13	Serializes an object that has a `to_dict` method into a dictionary.
14
15	:param obj:
16	The object to be serialized.
17	:returns:
18	A dictionary representation of the object.
19	:raises SerializationError:
20	If the object does not have a `to_dict` method.
21	"""
22	if not hasattr(obj, "to_dict"):	1✔
23	raise SerializationError(f"Object of class '{type(obj).__name__}' does not have a 'to_dict' method")	1✔
24
25	output = obj.to_dict()	1✔
26	return {"type": generate_qualified_class_name(type(obj)), "data": output}	1✔
27
28
29	def deserialize_class_instance(data: Dict[str, Any]) -> Any:	1✔
30	"""
31	Deserializes an object from a dictionary representation generated by `auto_serialize_class_instance`.
32
33	:param data:
34	The dictionary to deserialize from.
35	:returns:
36	The deserialized object.
37	:raises DeserializationError:
38	If the serialization data is malformed, the class type cannot be imported, or the
39	class does not have a `from_dict` method.
40	"""
41	if "type" not in data:	1✔
42	raise DeserializationError("Missing 'type' in serialization data")	1✔
43	if "data" not in data:	1✔
44	raise DeserializationError("Missing 'data' in serialization data")	1✔
45
46	try:	1✔
47	obj_class = import_class_by_name(data["type"])	1✔
48	except ImportError as e:	1✔
49	raise DeserializationError(f"Class '{data['type']}' not correctly imported") from e	1✔
50
51	if not hasattr(obj_class, "from_dict"):	1✔
52	raise DeserializationError(f"Class '{data['type']}' does not have a 'from_dict' method")	1✔
53
54	return obj_class.from_dict(data["data"])	1✔
55
56
57	def serialize_value(value: Any) -> Any:	1✔
58	"""
59	Serializes a value into a format suitable for storage or transmission.
60
61	Handles various types including:
62	- Haystack dataclass objects (Answer, Document, etc.)
63	- Primitive types (returned as is)
64	- Lists of primitives (returned as is)
65	- Lists of complex types (recursively serialized)
66	- Dictionaries (recursively serialized)
67
68	:param value: The value to serialize
69	:returns: The serialized representation of the value
70	"""
71
72	if hasattr(value, "to_dict") and callable(getattr(value, "to_dict")):	1✔
73	serialized_value = value.to_dict()	1✔
74	serialized_value["_type"] = generate_qualified_class_name(type(value))	1✔
75
76	return serialized_value	1✔
77
78	# this is a hack to serialize inputs that don't have a to_dict
79	elif hasattr(value, "__dict__"):	1✔
80	return {"_type": value.__class__, "attributes": value.__dict__}	×
81
82	# recursively serialize all inputs in a dict
83	elif isinstance(value, dict):	1✔
84	return {k: serialize_value(v) for k, v in value.items()}	1✔
85
86	# recursively serialize all inputs in lists or tuples
87	elif isinstance(value, (list, tuple, set)):	1✔
88	return type(value)(serialize_value(item) for item in value)	1✔
89
90	return value	1✔
91
92
93	# pylint: disable=too-many-return-statements
94	def deserialize_value(value: Any) -> Any:	1✔
95	"""
96	Deserializes a value from its serialized representation.
97
98	Handles various types including:
99	- Haystack dataclass objects (Answer, Document, etc.)
100	- Primitive types (returned as is)
101	- Lists of primitives (returned as is)
102	- Lists of complex types (recursively deserialized)
103	- Dictionaries (recursively deserialized)
104
105	:param value: The serialized value to deserialize
106	:returns: The deserialized value
107	"""
108
109	# None or primitive types are returned as is
110	if not value or isinstance(value, (str, int, float, bool)):	1✔
111	return value	1✔
112
113	# check if the dictionary has a "_type" key and the class type has a "from_dict" method
114	if isinstance(value, dict):	1✔
115	if "_type" in value:	1✔
116	obj_class = import_class_by_name(value.pop("_type"))	1✔
117	if hasattr(obj_class, "from_dict"):	1✔
118	return obj_class.from_dict(value)	1✔
119
120	# If not a known type, recursively deserialize each item in the dictionary
121	return {k: deserialize_value(v) for k, v in value.items()}	1✔
122
123	if isinstance(value, (list, tuple, set)):	1✔
124	return type(value)(deserialize_value(i) for i in value)	1✔
125
126	return value	×

deepset-ai / haystack / 15067549553

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous