src/api/routes/transcribe.py

"""POST /api/v1/transcribe — convert uploaded audio to text."""
from __future__ import annotations

import logging
import os
import tempfile
from typing import Annotated

from fastapi import APIRouter, Depends, File, Form, HTTPException, UploadFile

from src.api.dependencies import get_transcriber
from src.api.schemas import TranscribeResponse
from src.engine.transcriber import Transcriber

logger = logging.getLogger(__name__)
router = APIRouter()

SUPPORTED_LANGUAGES = {"bam", "ful"}
SUPPORTED_EXTENSIONS = {".wav", ".mp3", ".ogg", ".m4a", ".flac", ".webm"}
MAX_AUDIO_BYTES = 10 * 1024 * 1024  # 10 MB


@router.post("/transcribe", response_model=TranscribeResponse)
async def transcribe_audio(
    audio_file: Annotated[UploadFile, File(description="Audio file (wav/mp3/ogg/m4a/flac/webm)")],
    language: Annotated[str, Form(description="Language code: 'bam' (Bambara) or 'ful' (Fula)")] = "bam",
    transcriber: Transcriber = Depends(get_transcriber),
) -> TranscribeResponse:
    # Validate language
    if language not in SUPPORTED_LANGUAGES:
        raise HTTPException(
            status_code=422,
            detail=f"Unsupported language '{language}'. Supported: {sorted(SUPPORTED_LANGUAGES)}",
        )

    # Validate file extension
    filename = audio_file.filename or "audio.wav"
    ext = os.path.splitext(filename)[1].lower()
    if ext not in SUPPORTED_EXTENSIONS:
        raise HTTPException(
            status_code=422,
            detail=f"Unsupported file type '{ext}'. Supported: {sorted(SUPPORTED_EXTENSIONS)}",
        )

    # Read and size-check
    audio_bytes = await audio_file.read()
    if len(audio_bytes) > MAX_AUDIO_BYTES:
        raise HTTPException(
            status_code=413,
            detail=f"File too large ({len(audio_bytes) / 1e6:.1f} MB). Max 10 MB.",
        )

    # Windows-safe temp file: delete=False + manual unlink in finally
    tmp_path = None
    try:
        with tempfile.NamedTemporaryFile(delete=False, suffix=ext) as tmp:
            tmp.write(audio_bytes)
            tmp_path = tmp.name

        result = transcriber.transcribe_file(tmp_path, language)
    except Exception as e:
        logger.error("Transcription failed: %s", e, exc_info=True)
        raise HTTPException(status_code=500, detail=str(e))
    finally:
        if tmp_path and os.path.exists(tmp_path):
            os.unlink(tmp_path)

    return TranscribeResponse(
        text=result.text,
        language=result.language,
        duration_s=result.duration_s,
        processing_time_ms=result.processing_time_ms,
        confidence=result.confidence,
    )