|
| 1 | +"""Container Type Specification.""" |
| 2 | + |
| 3 | +from __future__ import annotations |
| 4 | + |
| 5 | +import io |
| 6 | +from typing import IO, Any, Dict, List, Tuple, Type, cast |
| 7 | + |
| 8 | +from pydantic import BaseModel, ConfigDict |
| 9 | +from typing_extensions import Self |
| 10 | + |
| 11 | +from lean_spec.types.constants import OFFSET_BYTE_LENGTH |
| 12 | + |
| 13 | +from .ssz_base import SSZType |
| 14 | +from .uint import Uint32 |
| 15 | + |
| 16 | + |
| 17 | +class Container(BaseModel, SSZType): |
| 18 | + """ |
| 19 | + A strict SSZ Container type: an ordered, heterogeneous collection of fields. |
| 20 | +
|
| 21 | + Inherit from this class to define a new container structure. Field types |
| 22 | + must be valid SSZ types. |
| 23 | +
|
| 24 | + Example: |
| 25 | + class BeaconBlockHeader(Container): |
| 26 | + slot: Uint64 |
| 27 | + proposer_index: Uint64 |
| 28 | + parent_root: Bytes32 |
| 29 | + """ |
| 30 | + |
| 31 | + # Configure all container subclasses to be strict and immutable by default. |
| 32 | + model_config = ConfigDict(strict=True, frozen=True) |
| 33 | + |
| 34 | + # --- SSZType Implementation --- |
| 35 | + |
| 36 | + @classmethod |
| 37 | + def is_fixed_size(cls) -> bool: |
| 38 | + """ |
| 39 | + Determine if the container is a fixed-size type. |
| 40 | +
|
| 41 | + A container is fixed-size if and only if all of its fields are fixed-size. |
| 42 | +
|
| 43 | + Returns: |
| 44 | + bool: True if all fields are fixed-size, False otherwise. |
| 45 | + """ |
| 46 | + # Iterate through the types of all fields defined in the model. |
| 47 | + for field_type in cls.model_fields.values(): |
| 48 | + # The `annotation` attribute holds the actual type hint (e.g., `Uint64`). |
| 49 | + # We assume all field types are valid SSZ types and have `is_fixed_size`. |
| 50 | + if not cast(Type[SSZType], field_type.annotation).is_fixed_size(): |
| 51 | + # If any field is variable-size, the container is variable-size. |
| 52 | + return False |
| 53 | + # If the loop completes, all fields are fixed-size. |
| 54 | + return True |
| 55 | + |
| 56 | + @classmethod |
| 57 | + def get_byte_length(cls) -> int: |
| 58 | + """ |
| 59 | + Get the byte length of the container if it is fixed-size. |
| 60 | +
|
| 61 | + Raises: |
| 62 | + TypeError: If the container is not fixed-size. |
| 63 | +
|
| 64 | + Returns: |
| 65 | + int: The total byte length of all fields. |
| 66 | + """ |
| 67 | + # A byte length can only be determined for fixed-size containers. |
| 68 | + if not cls.is_fixed_size(): |
| 69 | + raise TypeError(f"{cls.__name__} is not a fixed-size type.") |
| 70 | + |
| 71 | + # The total length is the sum of the byte lengths of all its fields. |
| 72 | + return sum( |
| 73 | + cast(Type[SSZType], field.annotation).get_byte_length() |
| 74 | + for field in cls.model_fields.values() |
| 75 | + ) |
| 76 | + |
| 77 | + def serialize(self, stream: IO[bytes]) -> int: |
| 78 | + """ |
| 79 | + Serialize the container to a binary stream according to SSZ rules. |
| 80 | +
|
| 81 | + This method correctly handles the mixed serialization of fixed-size fields |
| 82 | + and offsets for variable-size fields. |
| 83 | +
|
| 84 | + Args: |
| 85 | + stream (IO[bytes]): The stream to write the serialized data to. |
| 86 | +
|
| 87 | + Returns: |
| 88 | + int: The total number of bytes written. |
| 89 | + """ |
| 90 | + # Separate fields into fixed and variable parts for serialization. |
| 91 | + fixed_parts: List[bytes] = [] |
| 92 | + variable_parts: List[bytes] = [] |
| 93 | + |
| 94 | + # Iterate through all defined fields to process them in order. |
| 95 | + for field_name, field_info in type(self).model_fields.items(): |
| 96 | + # Get the actual value of the field from the instance. |
| 97 | + value = getattr(self, field_name) |
| 98 | + # The field's type is its annotation (e.g., `Uint64`). |
| 99 | + field_type = cast(Type[SSZType], field_info.annotation) |
| 100 | + |
| 101 | + # Check if the field type is fixed or variable size. |
| 102 | + if field_type.is_fixed_size(): |
| 103 | + # For fixed-size fields, serialize the value directly. |
| 104 | + fixed_parts.append(value.encode_bytes()) |
| 105 | + else: |
| 106 | + # For variable-size fields, add a placeholder for the offset |
| 107 | + # in the fixed part and serialize the value's data into the variable part. |
| 108 | + fixed_parts.append(b"") # Placeholder, will be replaced with offset. |
| 109 | + variable_parts.append(value.encode_bytes()) |
| 110 | + |
| 111 | + # Calculate the starting offset for the variable data. It begins after all fixed parts. |
| 112 | + current_offset = sum( |
| 113 | + part_len if part_len > 0 else OFFSET_BYTE_LENGTH for part_len in map(len, fixed_parts) |
| 114 | + ) |
| 115 | + |
| 116 | + # Write the fixed parts to the stream, replacing placeholders with calculated offsets. |
| 117 | + variable_part_index = 0 |
| 118 | + for part in fixed_parts: |
| 119 | + # If the part is not a placeholder, write it directly. |
| 120 | + if part: |
| 121 | + stream.write(part) |
| 122 | + # If it is a placeholder, write the calculated offset instead. |
| 123 | + else: |
| 124 | + Uint32(current_offset).serialize(stream) |
| 125 | + # Update the offset for the next variable part. |
| 126 | + current_offset += len(variable_parts[variable_part_index]) |
| 127 | + variable_part_index += 1 |
| 128 | + |
| 129 | + # Write all the serialized variable data at the end of the stream. |
| 130 | + for part in variable_parts: |
| 131 | + stream.write(part) |
| 132 | + |
| 133 | + # The final offset value is the total number of bytes written. |
| 134 | + return current_offset |
| 135 | + |
| 136 | + @classmethod |
| 137 | + def deserialize(cls, stream: IO[bytes], scope: int) -> Self: |
| 138 | + """ |
| 139 | + Deserialize a container from a binary stream. |
| 140 | +
|
| 141 | + Args: |
| 142 | + stream (IO[bytes]): The stream to read from. |
| 143 | + scope (int): The number of bytes available to read for this object. |
| 144 | +
|
| 145 | + Returns: |
| 146 | + Self: A new instance of the container with the deserialized data. |
| 147 | + """ |
| 148 | + # --- Phase 1: Read fixed data and gather variable field offsets --- |
| 149 | + deserialized_fields: Dict[str, Any] = {} |
| 150 | + variable_field_info: List[Tuple[str, Type[SSZType], int]] = [] |
| 151 | + |
| 152 | + # Read the fixed-size portion of the data from the stream. |
| 153 | + fixed_data_end = 0 |
| 154 | + for field_name, field_info in cls.model_fields.items(): |
| 155 | + field_type = cast(Type[SSZType], field_info.annotation) |
| 156 | + |
| 157 | + if field_type.is_fixed_size(): |
| 158 | + # Directly deserialize fixed-size fields. |
| 159 | + field_length = field_type.get_byte_length() |
| 160 | + field_data = stream.read(field_length) |
| 161 | + if len(field_data) != field_length: |
| 162 | + raise IOError(f"Stream ended prematurely while decoding field '{field_name}'") |
| 163 | + deserialized_fields[field_name] = field_type.decode_bytes(field_data) |
| 164 | + fixed_data_end += field_length |
| 165 | + else: |
| 166 | + # For variable fields, read the offset and store it for later processing. |
| 167 | + offset_data = stream.read(OFFSET_BYTE_LENGTH) |
| 168 | + if len(offset_data) != OFFSET_BYTE_LENGTH: |
| 169 | + raise IOError( |
| 170 | + f"Stream ended prematurely while reading offset for '{field_name}'" |
| 171 | + ) |
| 172 | + offset = int(Uint32.decode_bytes(offset_data)) |
| 173 | + variable_field_info.append((field_name, field_type, offset)) |
| 174 | + fixed_data_end += OFFSET_BYTE_LENGTH |
| 175 | + |
| 176 | + # --- Phase 2: Read variable data using the collected offsets --- |
| 177 | + if variable_field_info: |
| 178 | + # Add the total scope as the final offset boundary. |
| 179 | + offsets = [info[2] for info in variable_field_info] + [scope] |
| 180 | + |
| 181 | + # Read the entire variable data block into memory. |
| 182 | + variable_data_length = scope - fixed_data_end |
| 183 | + variable_data = stream.read(variable_data_length) |
| 184 | + if len(variable_data) != variable_data_length: |
| 185 | + raise IOError("Stream ended prematurely while reading variable data block.") |
| 186 | + |
| 187 | + # Deserialize each variable field from its slice of the data block. |
| 188 | + for i in range(len(variable_field_info)): |
| 189 | + field_name, field_type, start_offset = variable_field_info[i] |
| 190 | + end_offset = offsets[i + 1] |
| 191 | + |
| 192 | + # The actual data slice is relative to the start of the variable block. |
| 193 | + slice_start = start_offset - fixed_data_end |
| 194 | + slice_end = end_offset - fixed_data_end |
| 195 | + |
| 196 | + if slice_start > slice_end or slice_start < 0: |
| 197 | + raise ValueError( |
| 198 | + f"Invalid offsets for field '{field_name}': start > end or start < 0" |
| 199 | + ) |
| 200 | + |
| 201 | + field_data_slice = variable_data[slice_start:slice_end] |
| 202 | + deserialized_fields[field_name] = field_type.decode_bytes(field_data_slice) |
| 203 | + |
| 204 | + # Construct the final object instance from the deserialized fields. |
| 205 | + return cls(**deserialized_fields) |
| 206 | + |
| 207 | + def encode_bytes(self) -> bytes: |
| 208 | + """Serializes the Container to a byte string.""" |
| 209 | + with io.BytesIO() as stream: |
| 210 | + self.serialize(stream) |
| 211 | + return stream.getvalue() |
| 212 | + |
| 213 | + @classmethod |
| 214 | + def decode_bytes(cls, data: bytes) -> Self: |
| 215 | + """Deserializes a byte string into a Container instance.""" |
| 216 | + with io.BytesIO(data) as stream: |
| 217 | + return cls.deserialize(stream, len(data)) |
0 commit comments