backends/registry.py

"""
Model family ids and lazy mapping to backend builders (heavy imports on first eval only).
"""

from __future__ import annotations

from collections.abc import Callable

from family_registry import (
    DROPDOWN_CHOICES,
    FAMILY_IDS,
    FAMILY_META,
    default_family_id,
    normalize_family_id,
    resolve_label,
)

__all__ = [
    "DROPDOWN_CHOICES",
    "FAMILY_IDS",
    "FAMILY_META",
    "build_transcriber",
    "default_family_id",
    "normalize_family_id",
    "resolve_label",
]


def build_transcriber(
    family_id: str,
    model_id: str,
    device_str: str,
    device_int: int,
) -> tuple[Callable[..., str], Callable[[], None]]:
    from family_registry import _FAMILY_ALIASES

    from . import (
        auto,
        granite_speech,
        speechbrain_asr,
        transformers_ctc,
        transformers_pipeline,
        universal,
    )

    from .family_resolve import resolve_auto_family

    fid = resolve_auto_family(model_id, _FAMILY_ALIASES.get(family_id, family_id))
    if fid == "auto":
        return auto.build_transcriber(model_id, device_str, device_int)
    if fid == "transformers_pipeline":
        return transformers_pipeline.build_transcriber(model_id, device_int)
    if fid == "universal":
        return universal.build_transcriber(model_id, device_str)
    if fid == "granite_speech":
        return granite_speech.build_transcriber(model_id, device_str)
    if fid == "transformers_ctc":
        return transformers_ctc.build_transcriber(model_id, device_str)
    if fid == "speechbrain":
        return speechbrain_asr.build_transcriber(model_id, device_str)
    raise ValueError(f"Unknown model family: {family_id!r}. Choose one of {FAMILY_IDS}.")