ScaleSerializable mixin (#125)

* ScaleSerializable mixin

* Drop 3.7 support

Author: arjanz

.github/workflows/unittests.yml

@@ -13,7 +13,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: ['3.7', '3.8', '3.9', '3.10']
+        python-version: ['3.8', '3.9', '3.10', '3.11', '3.12']
 
     steps:
     - uses: actions/checkout@v2

pyproject.toml

@@ -11,7 +11,7 @@ authors = [
     { name = "Polkascan Foundation", email = "[email protected]" }
 ]
 
-requires-python = ">=3.6, <4"
+requires-python = ">=3.8, <4"
 classifiers = [
     "Development Status :: 5 - Production/Stable",
     "Intended Audience :: Developers",

scalecodec/mixins.py

@@ -0,0 +1,223 @@
+import dataclasses
+from dataclasses import is_dataclass
+import enum
+from typing import Type, TypeVar, Union
+import typing
+import json
+
+from scalecodec.base import ScaleTypeDef, ScaleType, ScaleBytes
+from scalecodec.types import Struct, Option, Vec, Enum
+
+T = TypeVar('T')
+
+
+class ScaleSerializable:
+    @classmethod
+    def scale_type_def(cls) -> ScaleTypeDef:
+        if is_dataclass(cls):
+
+            arguments = {}
+            for field in dataclasses.fields(cls):
+                arguments[field.name] = cls.dataclass_field_to_scale_typ_def(field)
+
+            return Struct(**arguments)
+        elif issubclass(cls, enum.Enum):
+            variants = {status.name: None for status in cls}
+            return Enum(**variants)
+
+        raise NotImplementedError
+
+    def serialize(self) -> Union[str, int, float, bool, dict, list]:
+        scale_type = self.to_scale_type()
+        return scale_type.serialize()
+
+    @classmethod
+    def deserialize(cls: Type[T], data: Union[str, int, float, bool, dict, list]) -> T:
+        scale_type = cls.scale_type_def().new()
+        scale_type.deserialize(data)
+        return cls.from_scale_type(scale_type)
+
+    def to_scale_type(self) -> ScaleType:
+
+        if not is_dataclass(self) and not issubclass(self.__class__, enum.Enum):
+            raise NotImplementedError("Type not supported.")
+
+        scale_type = self.scale_type_def().new()
+
+        if issubclass(self.__class__, enum.Enum):
+            scale_type.deserialize(self.name)
+        elif is_dataclass(self):
+            value = {}
+            for field in dataclasses.fields(self):
+
+                actual_type = field.type
+                field_name = field.name[:-1] if field.name.endswith('_') else field.name
+
+                if typing.get_origin(actual_type) is typing.Union:
+                    # Extract the arguments of the Union type
+                    args = typing.get_args(actual_type)
+                    if type(None) in args:
+                        # If NoneType is in the args, it's an Optional
+                        actual_type = [arg for arg in args if arg is not type(None)][0]
+
+                if getattr(self, field.name) is None:
+                    value[field_name] = None
+                else:
+
+                    if typing.get_origin(actual_type) is list:
+                        actual_type = typing.get_args(actual_type)[0]
+
+                        if issubclass(actual_type, ScaleSerializable):
+                            value[field_name] = [i.serialize() for i in getattr(self, field.name)]
+                        else:
+                            value[field_name] = getattr(self, field.name)
+
+                    # TODO too simplified now
+                    elif issubclass(actual_type, ScaleSerializable):
+
+                        value[field_name] = getattr(self, field.name).serialize()
+                    else:
+                        value[field_name] = getattr(self, field.name)
+
+            scale_type.deserialize(value)
+
+        return scale_type
+
+    @classmethod
+    def from_scale_type(cls: Type[T], scale_type: ScaleType) -> T:
+        if is_dataclass(cls):
+
+            fields = {}
+
+            for field in dataclasses.fields(cls):
+
+                scale_field_name = field.name[:-1] if field.name.endswith('_') else field.name
+
+                actual_type = field.type
+
+                if typing.get_origin(field.type) is typing.Union:
+                    # Extract the arguments of the Union type
+                    args = typing.get_args(field.type)
+                    if type(None) in args:
+                        # If NoneType is in the args, it's an Optional
+                        if field.name in scale_type.value:
+                            if scale_type.value[field.name] is None:
+                                fields[field.name] = None
+                                continue
+                            else:
+                                actual_type = [arg for arg in args if arg is not type(None)][0]
+                        else:
+                            # print(field.name)
+                            continue
+
+                if typing.get_origin(actual_type) is list:
+                    items = []
+                    actual_type = typing.get_args(actual_type)[0]
+
+                    if issubclass(type(scale_type.type_def), (Struct, Option)):
+                        list_items = scale_type.value_object[scale_field_name].value_object
+                    elif issubclass(type(scale_type.type_def), (Vec, Enum)):
+                        list_items = scale_type.value_object[1].value_object
+                    else:
+                        raise ValueError(f'Unsupported type: {type(scale_type.type_def)}')
+
+                    for item in list_items:
+                        if actual_type in [str, int, float, bool]:
+                            items.append(item.value)
+                        elif actual_type is bytes:
+                            items.append(item.to_bytes())
+                        elif is_dataclass(actual_type):
+                            items.append(actual_type.from_scale_type(item))
+
+                    fields[field.name] = items
+
+                elif actual_type in [str, int, float, bool]:
+                    fields[field.name] = scale_type.value[scale_field_name]
+                elif actual_type is bytes:
+                    fields[field.name] = scale_type.value_object[scale_field_name].to_bytes()
+                elif is_dataclass(actual_type):
+                    try:
+
+                        # TODO unwrap Option
+                        if issubclass(type(scale_type.type_def), (Struct, Option)):
+
+                            field_scale_type = scale_type.value_object[scale_field_name]
+                        elif issubclass(type(scale_type.type_def), Enum):
+                            field_scale_type = scale_type.value_object[1]
+                        else:
+                            raise ValueError(f"Unexpected type {type(scale_type.type_def)}")
+
+                        fields[field.name] = actual_type.from_scale_type(field_scale_type)
+                    except (KeyError, TypeError) as e:
+                        print('oeps', str(e))
+                elif issubclass(actual_type, enum.Enum):
+                    fields[field.name] = actual_type[scale_type.value_object[1].value]
+            return cls(**fields)
+        raise NotImplementedError
+
+    def to_scale_bytes(self) -> ScaleBytes:
+        scale_obj = self.to_scale_type()
+        return scale_obj.encode()
+
+    @classmethod
+    def from_scale_bytes(cls: Type[T], scale_bytes: ScaleBytes) -> T:
+        scale_obj = cls.scale_type_def().new()
+        scale_obj.decode(scale_bytes)
+        return cls.from_scale_type(scale_obj)
+
+    def to_json(self) -> str:
+        return json.dumps(self.serialize(), indent=4)
+
+    @classmethod
+    def from_json(cls: Type[T], json_data: str) -> T:
+        # data = json.loads(json_data)
+        return cls.deserialize(json_data)
+
+    @classmethod
+    def dataclass_field_to_scale_typ_def(cls, field) -> ScaleTypeDef:
+
+        if 'scale' in field.metadata:
+            return field.metadata['scale']
+
+        # Check if the field type is an instance of Optional
+        actual_type = field.type
+        wrap_option = False
+        wrap_vec = False
+
+        if typing.get_origin(field.type) is typing.Union:
+            # Extract the arguments of the Union type
+            args = typing.get_args(field.type)
+            if type(None) in args:
+                # If NoneType is in the args, it's an Optional
+                wrap_option = True
+                actual_type = [arg for arg in args if arg is not type(None)][0]
+                # print(f"The field '{field.name}' is Optional with inner type: {actual_type}")
+
+        if typing.get_origin(actual_type) is list:
+            wrap_vec = True
+            actual_type = typing.get_args(actual_type)[0]
+
+        if is_dataclass(actual_type):
+            if issubclass(actual_type, ScaleSerializable):
+                scale_def = actual_type.scale_type_def()
+            else:
+                raise ValueError(f"Cannot serialize dataclass {field.type.__class__}")
+
+        elif actual_type is bytes:
+            raise ValueError("bytes is ambiguous; specify SCALE type def in metadata e.g. {'scale': H256}")
+        elif actual_type is int:
+            raise ValueError("int is ambiguous; specify SCALE type def in metadata e.g. {'scale': U32}")
+
+        elif issubclass(actual_type, enum.Enum):
+            variants = {status.name: None for status in actual_type}
+            scale_def = Enum(**variants)
+
+        else:
+            raise ValueError(f"Cannot convert {actual_type} to ScaleTypeDef")
+
+        if wrap_vec:
+            scale_def = Vec(scale_def)
+        if wrap_option:
+            scale_def = Option(scale_def)
+
+        return scale_def

test/test_mixins.py

@@ -0,0 +1,135 @@
+import enum
+import json
+import os
+import unittest
+from dataclasses import dataclass, field
+from os import path
+from typing import Optional, Type, Union, List
+
+from scalecodec.base import ScaleBytes, ScaleType
+from scalecodec.mixins import ScaleSerializable, T
+from scalecodec.types import H256, U8, Array, Enum
+
+
+# Test definitions
+
+
+@dataclass
+class ValidatorData(ScaleSerializable):
+    bandersnatch: bytes = field(metadata={'scale': H256})
+    ed25519: bytes = field(metadata={'scale': H256})
+    bls: bytes = field(metadata={'scale': Array(U8, 144)})
+    metadata: bytes = field(metadata={'scale': Array(U8, 128)})
+
+
+@dataclass
+class EpochMark(ScaleSerializable):
+    entropy: bytes = field(metadata={'scale': H256})
+    validators: List[bytes] = field(metadata={'scale': Array(H256, 6)})
+
+@dataclass
+class OutputMarks(ScaleSerializable):
+    epoch_mark: Optional[EpochMark] = None
+
+
+class CustomErrorCode(ScaleSerializable, enum.Enum):
+    bad_slot = 0  # Timeslot value must be strictly monotonic.
+    unexpected_ticket = 1  # Received a ticket while in epoch's tail.
+    bad_ticket_order = 2  # Tickets must be sorted.
+    bad_ticket_proof = 3  # Invalid ticket ring proof.
+    bad_ticket_attempt = 4  # Invalid ticket attempt value.
+    reserved = 5  # Reserved
+    duplicate_ticket = 6  # Found a ticket duplicate.
+    too_many_tickets = 7  # Found amount of tickets > K
+
+
+@dataclass
+class Output(ScaleSerializable):
+    ok: Optional[OutputMarks] = None  # Markers
+    err: Optional[CustomErrorCode] = None
+
+    @classmethod
+    def scale_type_def(cls):
+
+        return Enum(
+            ok=OutputMarks.scale_type_def(),
+            err=CustomErrorCode.scale_type_def()
+        )
+
+    def to_scale_type(self) -> ScaleType:
+        scale_type = self.scale_type_def().new()
+        scale_type.deserialize(self.serialize())
+        return scale_type
+
+    @classmethod
+    def deserialize(cls: Type[T], data: Union[str, int, float, bool, dict, list]) -> T:
+
+        return super().deserialize(data)
+
+    def serialize(self) -> Union[str, int, float, bool, dict, list]:
+        if self.err is not None:
+            return {'err': self.err.serialize()}
+        else:
+            return {'ok': self.ok.serialize()}
+
+
+class TestSerializableMixin(unittest.TestCase):
+
+    def setUp(self):
+        data = {
+            'bandersnatch': '0x5e465beb01dbafe160ce8216047f2155dd0569f058afd52dcea601025a8d161d',
+            'ed25519': '0x3b6a27bcceb6a42d62a3a8d02a6f0d73653215771de243a63ac048a18b59da29',
+            'bls': '0x000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000',
+            'metadata': '0x0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000'
+        }
+
+        self.test_obj = ValidatorData.deserialize(data)
+
+    def test_dataclass_serialization(self):
+        output = Output(ok=OutputMarks(epoch_mark=None))
+        value = output.serialize()
+        self.assertEqual({'ok': {'epoch_mark': None}}, value)
+
+        output = Output(err=CustomErrorCode.duplicate_ticket)
+        value = output.serialize()
+
+        self.assertEqual({'err': 'duplicate_ticket'}, value)
+
+    def test_dataclass_to_scale_type(self):
+        output = Output(
+            ok=OutputMarks(
+                epoch_mark=EpochMark(
+                    entropy=bytes(32),
+                    validators=[bytes(32), bytes(32), bytes(32), bytes(32), bytes(32), bytes(32)]
+                )
+            )
+        )
+        scale_type = output.to_scale_type()
+        output2 = Output.from_scale_type(scale_type)
+        self.assertEqual(output, output2)
+
+    def test_deserialize(self):
+
+        data = {
+            'bandersnatch': '0x5e465beb01dbafe160ce8216047f2155dd0569f058afd52dcea601025a8d161d',
+            'ed25519': '0x3b6a27bcceb6a42d62a3a8d02a6f0d73653215771de243a63ac048a18b59da29',
+            'bls': '0x000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000',
+            'metadata': '0x0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000'
+        }
+
+        validator_obj = ValidatorData.deserialize(data)
+
+        self.assertEqual(self.test_obj, validator_obj)
+        self.assertEqual(data, validator_obj.serialize())
+
+    def test_from_to_scale_bytes(self):
+
+        scale_data = self.test_obj.to_scale_bytes()
+
+        validator_obj = ValidatorData.from_scale_bytes(scale_data)
+
+        self.assertEqual(self.test_obj, validator_obj)
+
+
+if __name__ == '__main__':
+    unittest.main()