#!/usr/bin/env python3
"""
Qwen-Based Audiobook Converter
Converts TXT, PDF and EPUB files into audiobooks using a local Qwen3-TTS server.

Edit converter/config.py to change voice and processing settings.
"""

import argparse
import sys
import traceback

# Fix Windows console encoding for unicode output
if sys.platform == "win32":
    try:
        sys.stdout.reconfigure(encoding="utf-8")
        sys.stderr.reconfigure(encoding="utf-8")
    except AttributeError:
        pass

from converter import config
from converter.converter import (
    AUDIO_FORMATS,
    AudiobookConverter,
    setup_directories,
    setup_logging,
)
from converter.tts import (
    BACKEND_AUDIOCPP,
    BACKEND_FASTER,
    BACKEND_GRADIO,
    VOICE_MODE_CLONE,
    VOICE_MODE_CUSTOM,
    normalize_language,
)


def main() -> None:
    """Entry point with argparse."""
    parser = argparse.ArgumentParser(
        description="Convert books to audiobooks using the Qwen3-TTS voice model",
        formatter_class=argparse.RawDescriptionHelpFormatter,
        epilog="""
Examples:
  # Use the Qwen Gradio demo with a custom voice (default - Vivian speaker)
  python audiobook.py

  # Use the Qwen Gradio demo with voice cloning from reference audio
  python audiobook.py --clone path/to/reference.wav

  # Use the faster-qwen3-tts server (voice cloning, configured server-side)
  python audiobook.py --backend faster [--voice NAME]

  # Use an audio.cpp audiocpp_server (speaker mode, or a server-side voice)
  python audiobook.py --backend audiocpp [--voice NAME]
        """
    )

    parser.add_argument(
        "--clone",
        type=str,
        metavar="PATH",
        help=("Path to reference audio file for voice cloning (WAV format). "
              "Passing this flag switches the converter to voice clone mode.")
    )

    parser.add_argument(
        "--transcription",
        type=str,
        default=None,
        help=("Transcript of the reference audio for in-context cloning (recommended for "
              "highest quality). If omitted, a local Whisper backend is used if installed; "
              "otherwise the converter falls back to x-vector-only mode.")
    )

    parser.add_argument(
        "--no-transcription",
        action="store_true",
        help=("Skip automatic transcription of the reference audio (use x-vector-only "
              "cloning). Ignored when --transcription is provided.")
    )

    parser.add_argument(
        "--language",
        type=str,
        default=None,
        metavar="LANG",
        help=("Output language for the synthesized speech, e.g. English, Japanese, "
              "or Auto (language names and short codes like en/ja are accepted). "
              "With --backend audiocpp the language is adapted to the model "
              "family: sent as a code (e.g. 'en') for families that take one, or "
              "omitted when the model detects the language itself. Defaults to "
              "the LANGUAGE setting in converter/config.py (English).")
    )

    parser.add_argument(
        "--speed",
        type=float,
        default=1.0,
        help="Playback speed factor for the final audiobook (1.0 = normal). Pitch-preserving."
    )

    parser.add_argument(
        "--format",
        choices=list(AUDIO_FORMATS),
        default=config.AUDIO_FORMAT,
        help=f"Output container format (default: {config.AUDIO_FORMAT}). m4b uses AAC audio."
    )

    parser.add_argument(
        "--single-file",
        action="store_true",
        help=("Combine all chapters into a single audio file. By default books with "
              "chapters (e.g. EPUB) are converted to one file per chapter. "
              "Ignored for m4b, which is always a single file.")
    )

    parser.add_argument(
        "--backend",
        choices=[BACKEND_GRADIO, BACKEND_FASTER, BACKEND_AUDIOCPP],
        default=config.BACKEND,
        help=("TTS server to talk to: the Qwen3-TTS Gradio demos (gradio), the "
              "faster-qwen3-tts OpenAI-compatible server (faster), or an "
              "audio.cpp audiocpp_server (audiocpp) hosting any of its TTS "
              "model families — Qwen3-TTS, Higgs Audio, VoxCPM2, IndexTTS2, "
              "and more. Defaults to the BACKEND setting in "
              "converter/config.py (gradio).")
    )

    parser.add_argument(
        "--voice",
        type=str,
        default=None,
        metavar="NAME",
        help=("Voice to request from a server-side voice configuration. faster: "
              "a key in the server's voices.json ('default' when it was started "
              "with --ref-audio). audiocpp: a voice_preset or voice_dir entry "
              "(cloning); required for audio.cpp families without built-in "
              "speakers (everything except Qwen3-TTS CustomVoice). Not used by "
              "the gradio backend (use converter/config.py SPEAKER or --clone "
              "there).")
    )

    parser.add_argument(
        "--debug",
        action="store_true",
        help=("Troubleshooting mode: dump each chunk's raw audio and the exact text "
              "sent for it under the debug/ folder (organized per book and chapter), "
              "and log every TTS request and response to the console and log file.")
    )

    parser.add_argument(
        "--chunk",
        action="store_true",
        help=("Force client-side chunking into CHUNK_SIZE-word requests (see "
              "converter/config.py). Only matters for --backend audiocpp, which "
              "otherwise sends each chapter as one request and lets the server "
              "chunk long text itself; the gradio and faster backends always "
              "chunk.")
    )

    parser.add_argument(
        "--model",
        type=str,
        default=None,
        metavar="ID",
        help=("audio.cpp server model entry id to use for this run "
              "(--backend audiocpp only). Overrides AUDIOCPP_MODEL_ID in "
              "converter/config.py, which is useful for a server hosting "
              "several lazily-loaded models: generate one server.json with "
              "tools/make_audiocpp_server_json.py, then pick the model per "
              "run with --model. Leave unset to use the config id, or to "
              "auto-select when the server hosts exactly one entry.")
    )

    args = parser.parse_args()

    if args.speed <= 0:
        parser.error(f"--speed must be a positive number (got {args.speed:g})")

    if args.chunk:
        if args.backend == BACKEND_AUDIOCPP:
            print("[WARNING] --chunk: the audio.cpp server already splits long text "
                  "internally (its text_chunk_size); forcing client-side chunking "
                  "may cause needless double-chunking")
        else:
            print(f"[INFO] --chunk has no effect with --backend {args.backend}: "
                  "that backend always chunks")

    if args.backend == BACKEND_FASTER:
        if args.clone:
            print("[WARNING] --clone is ignored with --backend faster: that backend "
                  "always uses voice cloning, and the reference voice is configured "
                  "on the server (see README)")
            args.clone = None
        if args.transcription or args.no_transcription:
            print("[WARNING] --transcription/--no-transcription are ignored with "
                  "--backend faster: the reference transcript is configured on the "
                  "server (--ref-text or voices.json, see README)")
            args.transcription = None
            args.no_transcription = False
        if args.language is not None:
            print("[WARNING] --language is ignored with --backend faster: language "
                  "is configured on the server (see README)")
            args.language = None
    elif args.backend == BACKEND_AUDIOCPP:
        if args.clone:
            print("[WARNING] --clone is ignored with --backend audiocpp: cloning "
                  "uses a voice configured on the server (voice_presets or "
                  "voice_dir in its config); select it with --voice (see README)")
            args.clone = None
        if args.transcription or args.no_transcription:
            print("[WARNING] --transcription/--no-transcription are ignored with "
                  "--backend audiocpp: the reference transcript is configured on "
                  "the server (see README)")
            args.transcription = None
            args.no_transcription = False
        if args.language is not None:
            try:
                args.language = normalize_language(args.language)
            except ValueError as exc:
                parser.error(str(exc))
    else:
        if args.voice is not None:
            parser.error("--voice requires --backend faster or audiocpp; the "
                         "gradio backend uses built-in speakers "
                         "(converter/config.py SPEAKER) or --clone")
        if args.language is not None:
            try:
                args.language = normalize_language(args.language)
            except ValueError as exc:
                parser.error(str(exc))

        if not args.clone and (args.transcription or args.no_transcription):
            print("[WARNING] --transcription/--no-transcription "
                  "are ignored without --clone")

    if args.model is not None and args.backend != BACKEND_AUDIOCPP:
        parser.error("--model requires --backend audiocpp; it selects an "
                     "audio.cpp server model entry id")

    setup_logging(debug=args.debug)
    setup_directories()

    if args.backend == BACKEND_FASTER:
        voice_mode = VOICE_MODE_CLONE
    elif args.backend == BACKEND_AUDIOCPP:
        voice_mode = VOICE_MODE_CLONE if args.voice else VOICE_MODE_CUSTOM
    else:
        voice_mode = VOICE_MODE_CLONE if args.clone else VOICE_MODE_CUSTOM

    # Ask every overwrite question up front, before spending time connecting
    # to a TTS server: a user who declines (or has nothing to convert) never
    # waits on a slow server handshake. Nothing in this step needs the server.
    book_files, planned = AudiobookConverter.preflight_overwrites(
        backend=args.backend,
        voice=args.voice,
        voice_mode=voice_mode,
        voice_clone_ref_audio=args.clone,
        output_format=args.format,
    )

    if not book_files:
        print("[INFO] Nothing to convert. Add a .txt, .pdf, or .epub file "
              "to the input folder and run again.")
        sys.exit(0)

    if not planned:
        print("[INFO] Nothing to convert (all books skipped)")
        sys.exit(0)

    try:
        converter = AudiobookConverter(
            voice_mode=voice_mode,
            voice_clone_ref_audio=args.clone,
            voice_clone_ref_text=args.transcription,
            skip_transcription=args.no_transcription,
            speed=args.speed,
            single_file=args.single_file,
            output_format=args.format,
            language=args.language,
            backend=args.backend,
            voice=args.voice,
            debug=args.debug,
            chunk=args.chunk,
            model_id=args.model,
        )
        converter._book_files = book_files
        converter._planned = planned
        ok = converter.run()
    except KeyboardInterrupt:
        print("\n[WARNING] Shutdown requested by user")
        sys.exit(130)
    except Exception as exc:
        print(f"[FATAL] Fatal error: {exc}")
        traceback.print_exc()
        sys.exit(1)

    sys.exit(0 if ok else 1)


if __name__ == "__main__":
    main()