denoiser.py

import os
import requests
import logging
import subprocess
import re

from cleanvoice import Cleanvoice

logger = logging.getLogger(__name__)


class Denoiser:
    def __init__(self):
        self.api_key = os.environ.get("CLEANVOICE_API_KEY", "")
        if self.api_key:
            print("[Denoiser] ☁️ Cleanvoice API Ready (WhatsApp Support Active)")
        else:
            print("[Denoiser] ⚠️ No CLEANVOICE_API_KEY found — will use preprocessed WAV only")

    def process(self, audio_path: str, out_dir: str, **kwargs) -> dict:
        # Step A: Pre-convert to standard WAV (Handles .opus, .aac, .m4a)
        standard_input = os.path.join(out_dir, "input_converted.wav")
        result = subprocess.run([
            "ffmpeg", "-y", "-i", audio_path, 
            "-ar", "16000", "-ac", "1", standard_input
        ], capture_output=True)

        # Fallback to original if FFmpeg fails
        if result.returncode != 0 or not os.path.exists(standard_input):
            logger.warning("[Denoiser] FFmpeg conversion failed — using original file")
            standard_input = audio_path

        try:
            from cleanvoice import Cleanvoice
            cv = Cleanvoice({'api_key': self.api_key})
            
            # Step B: Process via Cleanvoice API
            supported_kwargs = {k: v for k, v in kwargs.items() if k in ["fillers", "stutters", "long_silences"]}
            result = cv.process(
                standard_input, 
                remove_noise=True, 
                studio_sound=True, 
                **supported_kwargs
            )
            
            # Step C: Download the processed WAV file
            audio_data = requests.get(result.audio.url, timeout=60).content
            wav_output = os.path.join(out_dir, "clean_output.wav")
            with open(wav_output, "wb") as f:
                f.write(audio_data)
            
            # ✅ NEW: Convert WAV → MP3 (high quality)
            mp3_output = os.path.join(out_dir, "clean_output.mp3")
            subprocess.run([
                "ffmpeg", "-y", "-i", wav_output,
                "-codec:a", "libmp3lame", "-b:a", "192k",
                "-ar", "44100", "-ac", "2",
                mp3_output
            ], capture_output=True)
            
            return {'audio_path': mp3_output}  # Returns MP3!
            
        except Exception as e:
            logger.error(f"Denoiser error: {e}")
            
            # Fallback: Convert original to MP3
            fallback_mp3 = os.path.join(out_dir, "fallback.mp3")
            subprocess.run([
                "ffmpeg", "-y", "-i", audio_path,
                "-codec:a", "libmp3lame", "-b:a", "192k",
                fallback_mp3
            ], capture_output=True)
            return {'audio_path': fallback_mp3}

    def clean_transcript_fillers(self, transcript: str) -> str:
        fillers = {"um", "umm", "uh", "basically", "like", "ante", "ane"}
        words = transcript.split()
        return " ".join([w for w in words if re.sub(r'[^a-z]', '', w.lower()) not in fillers])

    # Important: Dummy methods to prevent "AttributeError" in app.py
    def _remove_fillers(self, audio, sr, segments): return audio, 0
    def _remove_stutters(self, audio, sr, segments): return audio, 0