utils/hparam_helpers.py

"""Small helpers used from HyperPyYAML files."""

from __future__ import annotations

import json
from pathlib import Path


def read_output_neurons(metadata_path: str) -> int:
    """Read output_neurons from a metadata JSON file."""
    path = Path(metadata_path)
    with path.open("r", encoding="utf-8") as f:
        metadata = json.load(f)
    return int(metadata["output_neurons"])


def read_subphonetic_topology_output_neurons(
    label_encoder_path: str,
    context_phone_mode: str = "mono",
    topology: str = "three_state_skip_two",
    split_silence: bool = False,
) -> int:
    """Compute output size for an auxiliary subphonetic topology head."""
    from utils.subphonetic_topology import topology_output_neurons

    return topology_output_neurons(
        label_encoder_path=label_encoder_path,
        context_phone_mode=context_phone_mode,
        topology=topology,
        split_silence=split_silence,
    )


def sum_ints(*values) -> int:
    """Return the sum of integer-like HyperPyYAML values."""
    return sum(int(value) for value in values if value is not None)


def load_ctc_text_encoder_from_file(path: str):
    """Construct and load a SpeechBrain CTCTextEncoder from a text file."""
    import speechbrain as sb

    encoder = sb.dataio.encoder.CTCTextEncoder()
    encoder.load(path)
    return encoder