Record Decoder

The record decoder reads an individual dataset record for the decode service and extracts its tensors: the node features, the targets, the graph connectivity, and any auxiliary columns.

Usage

Selected under services.decode.records.

services:
  decode:
    records:
      name: standard
      kwargs: {}

Variants

• Standard: extracts tensors from records written in the standard schema.

Registering a new record decoder

A record decoder is a subclass of RecordDecoder that declares a name and version and implements extraction of a tensor from a record by key. Register it with RecordDecoderFactory.

The RecordDecoder additionally provides optional hooks for label-specific extraction overrides:

_extract_features(self, record, key) -> Tensor | None

_extract_targets(self, record, key) -> Tensor | None

_extract_auxiliary(self, record, key) -> Tensor | None

_extract_edge_index(self, record, key) -> Tensor | None

_extract_edge_attr(self, record, key) -> Tensor | None

_extract_simweights(self, record, key) -> Tensor | None

from typing import Any, ClassVar

from torch import Tensor

from icegraph.common.record import Record
from icegraph.engine.services.decode.records import RecordDecoder, RecordDecoderFactory

from .config import MyConfig

class MyRecordDecoder(RecordDecoder[MyConfig]):
    name: ClassVar[str] = "my-records"
    version: ClassVar[int] = 1

    @classmethod
    def validate_config(cls, config: dict[str, Any]) -> MyConfig:
        return MyConfig(**config)

    def build(self) -> None:
        ...

    def extract(self, record: Record, key: str) -> Tensor | None:
        ...

RecordDecoderFactory.register(MyRecordDecoder)