main.py

"""
ClearWave AI — API Space (FastAPI)
===================================

Pipeline:
1. Download audio from URL
2. Denoise → Cleanvoice SDK (tested & working)
3. Transcribe → Groq Whisper large-v3 / faster-whisper fallback
4. Translate → Helsinki-NLP / NLLB / Google fallback
5. Summarize → Extractive (position-scored)
6. Upload to Cloudinary

Environment vars: CLEANVOICE_API_KEY, CLOUD_NAME, API_KEY, API_SECRET, GROQ_API_KEY
"""

import os
import json
import time
import tempfile
import logging
import shutil
import requests
import cloudinary
import cloudinary.uploader
from fastapi import FastAPI, Request
from fastapi.responses import StreamingResponse, JSONResponse
from fastapi.middleware.cors import CORSMiddleware

from denoiser import Denoiser        # ✅ Your working denoiser
from transcriber import Transcriber  # Your transcriber
from translator import Translator    # Your translator

# ── Cloudinary config ─────────────────────────────────────────────
cloudinary.config(
    cloud_name=os.environ.get("CLOUD_NAME"),
    api_key=os.environ.get("API_KEY"),
    api_secret=os.environ.get("API_SECRET"),
)

logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

# ── Singletons (loaded once at startup) ──────────────────────────
denoiser = Denoiser()
transcriber = Transcriber()
translator = Translator()

app = FastAPI(title="ClearWave AI API")
app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],
    allow_methods=["*"],
    allow_headers=["*"],
)


def _safe_cleanup(out_dir):
    """Remove temp directory safely — never raises."""
    try:
        if hasattr(denoiser, 'cleanup_temp_files'):
            denoiser.cleanup_temp_files(out_dir)
        elif out_dir and os.path.exists(out_dir):
            shutil.rmtree(out_dir, ignore_errors=True)
    except Exception as e:
        logger.warning(f"Cleanup warning (non-fatal): {e}")


def run_pipeline(audio_path, src_lang="auto", tgt_lang="te",
                 opt_fillers=True, opt_stutters=True, opt_silences=True):
    """
    Generator — yields SSE-style dicts at each step.
    """
    out_dir = tempfile.mkdtemp()
    word_segs = []
    stats = {}

    try:
        # ── Step 1: Cleanvoice API ────────────────────────────────
        yield {"status": "processing", "step": 1,
               "message": "Step 1/4 — Enhancing audio with Cleanvoice..."}
        
        result = denoiser.process(
            audio_path, out_dir,
            fillers=opt_fillers,
            stutters=opt_stutters,
            long_silences=opt_silences,
        )
        clean_audio = result["audio_path"]
        stats = {
            "noise_method": "Cleanvoice API",
            "fillers": opt_fillers,
            "stutters": opt_stutters,
            "silences": opt_silences,
        }
        logger.info("Cleanvoice enhancement complete")

        # ── Step 2: Transcribe ─────────────────────────────────────
        yield {"status": "processing", "step": 2,
               "message": "Step 2/4 — Transcribing..."}
        transcript, detected_lang, t_method = transcriber.transcribe(clean_audio, src_lang)
        word_segs = transcriber._last_segments

        # Clean transcript fillers too
        if opt_fillers:
            transcript = denoiser.clean_transcript_fillers(transcript)

        logger.info(f"Transcription: {len(transcript.split())} words, lang={detected_lang}")

        # ── Step 3: Translate ──────────────────────────────────────
        translation = transcript
        tl_method = "same language"
        if tgt_lang != "auto" and detected_lang != tgt_lang:
            yield {"status": "processing", "step": 3,
                   "message": "Step 3/4 — Translating..."}
            translation, tl_method = translator.translate(transcript, detected_lang, tgt_lang)
            logger.info(f"Translation done via {tl_method}")
        else:
            yield {"status": "processing", "step": 3,
                   "message": "Step 3/4 — Skipping translation (same language)"}

        # ── Step 4: Summarize & Cloudinary ────────────────────────
        yield {"status": "processing", "step": 4,
               "message": "Step 4/4 — Summarizing & uploading..."}
        summary = translator.summarize(transcript)

        enhanced_url = None
        try:
            upload_result = cloudinary.uploader.upload(
                clean_audio,
                resource_type="video",
                folder="clearwave_enhanced",
            )
            enhanced_url = upload_result["secure_url"]
            logger.info(f"Cloudinary upload: {enhanced_url}")
        except Exception as e:
            logger.error(f"Cloudinary failed: {e}")

        # ✅ yield done INSIDE try so cleanup never interrupts it
        yield {
            "status": "done",
            "step": 4,
            "message": "✅ Complete!",
            "transcript": transcript,
            "translation": translation,
            "summary": summary,
            "enhancedAudio": enhanced_url,
            "stats": {
                "language": detected_lang.upper(),
                "denoiser": stats,
                "transcription_method": t_method,
                "translation_method": tl_method,
                "word_segments": len(word_segs),
                "transcript_words": len(transcript.split()),
            },
        }

    except Exception as e:
        logger.error(f"Pipeline error: {e}", exc_info=True)
        yield {"status": "error", "message": f"Pipeline failed: {str(e)}"}

    finally:
        # ✅ Cleanup is now in finally — runs after done is yielded, never crashes pipeline
        _safe_cleanup(out_dir)


@app.get("/api/health")
async def health():
    return JSONResponse({"status": "ok", "service": "ClearWave AI API"})


@app.post("/api/process-url")
async def process_url(request: Request):
    data = await request.json()
    audio_url = data.get("audioUrl")
    audio_id = data.get("audioId", "")
    src_lang = data.get("srcLang", "auto")
    tgt_lang = data.get("tgtLang", "te")
    opt_fillers = data.get("optFillers", True)
    opt_stutters = data.get("optStutters", True)
    opt_silences = data.get("optSilences", True)

    if not audio_url:
        return JSONResponse({"error": "audioUrl required"}, status_code=400)

    async def generate():
        def sse(obj):
            return "data: " + json.dumps(obj) + "\n\n"

        yield sse({"status": "processing", "step": 0, "message": "Downloading audio..."})

        # Download audio from URL
        try:
            resp = requests.get(audio_url, timeout=60, stream=True)
            resp.raise_for_status()

            lower_url = audio_url.lower().split("?")[0]
            suffix = ".mp3"
            if ".opus" in lower_url:   suffix = ".opus"
            elif ".ogg" in lower_url:  suffix = ".ogg"
            elif ".aac" in lower_url:  suffix = ".aac"
            elif ".m4a" in lower_url:  suffix = ".m4a"
            elif ".wav" in lower_url:  suffix = ".wav"

            tmp = tempfile.NamedTemporaryFile(delete=False, suffix=suffix)
            for chunk in resp.iter_content(chunk_size=65536):
                if chunk:
                    tmp.write(chunk)
            tmp.close()

        except Exception as e:
            yield sse({"status": "error", "message": f"Download failed: {e}"})
            return

        # Run full pipeline
        for result in run_pipeline(
            tmp.name, src_lang, tgt_lang,
            opt_fillers, opt_stutters, opt_silences
        ):
            result["audioId"] = audio_id
            yield sse(result)

        try:
            os.unlink(tmp.name)
        except Exception:
            pass

    return StreamingResponse(
        generate(),
        media_type="text/event-stream",
        headers={"Cache-Control": "no-cache", "X-Accel-Buffering": "no"},
    )