epub2audio_piper.py

#!/usr/bin/env python3
"""
$ pip install piper-tts loguru typed-argparse bs4 ebooklib mutagen
$ apt install ffmpeg

Assuming both piper and ffmpeg are in your path, this script will convert an epub file to an audiobook.
"""
import math
import os
import queue
import re
import subprocess
import sys
import threading
from dataclasses import dataclass, field
from multiprocessing import cpu_count
from typing import List, Tuple

import ebooklib
import typed_argparse as tap
from bs4 import BeautifulSoup
from ebooklib import epub
from loguru import logger
from mutagen.easyid3 import EasyID3

logger.remove(0)
logger.add(
    sys.stderr,
    format="<green>{time:HH:mm:ss}</green>\t{thread.name}\t{message}",
    level="INFO",
)


@dataclass
class RecordingJob:
    text: str
    output_file_prefix: str
    title: str
    author: str
    book_title: str
    chapter_number: int
    voice: str
    speed: str
    pause: str
    wav_filename: str = field(init=False)
    mp3_filename: str = field(init=False)

    def __post_init__(self):
        self.wav_filename = self.output_file_prefix + ".wav"
        self.mp3_filename = self.output_file_prefix + ".mp3"


def sanitize_title(title: str) -> str:
    sanitized_title = re.sub(r"[^\w\s]", "", title, flags=re.UNICODE)
    sanitized_title = re.sub(r"\s+", "_", sanitized_title.strip())
    return sanitized_title


def extract_chapters(epub_book: epub.EpubBook) -> List[Tuple[str, str]]:
    chapters = []
    for item in epub_book.get_items():
        if item.get_type() == ebooklib.ITEM_DOCUMENT:
            content = item.get_content()
            soup = BeautifulSoup(content, features="lxml")
            title = soup.title.string if soup.title else ""
            raw = soup.get_text(strip=False)
            logger.debug(f"Raw text: <{raw[:]}>")

            # Replace excessive whitespaces and newline characters based on the mode
            cleaned_text = re.sub(r"\s+", " ", raw.strip())
            logger.info(f"Cleaned text step 1: <{cleaned_text[:100]}>")

            # fill in the title if it's missing
            if not title:
                title = cleaned_text[:60]
            logger.debug(f"Raw title: <{title}>")
            title = sanitize_title(title)
            logger.info(f"Sanitized title: <{title}>")

            chapters.append((title, cleaned_text))
            soup.decompose()
    return chapters


def epub_to_audiobook(
    input_file: str,
    output_folder: str,
    voice: str,
    speed: str,
    pause: str,
    chapter_start: int,
    chapter_end: int,
) -> None:
    book = epub.read_epub(input_file)
    chapters = extract_chapters(book)

    os.makedirs(output_folder, exist_ok=True)

    # Get the book title and author from metadata or use fallback values
    book_title = "Untitled"
    author = "Unknown"
    if book.get_metadata("DC", "title"):
        book_title = book.get_metadata("DC", "title")[0][0]
    if book.get_metadata("DC", "creator"):
        author = book.get_metadata("DC", "creator")[0][0]

    # Filter out empty or very short chapters
    chapters = [(title, text) for title, text in chapters if text.strip()]

    logger.info(f"Chapters count: {len(chapters)}.")

    # Check chapter start and end args
    if chapter_start < 1 or chapter_start > len(chapters):
        raise ValueError(
            f"Chapter start index {chapter_start} is out of range. Check your input."
        )
    if chapter_end < -1 or chapter_end > len(chapters):
        raise ValueError(
            f"Chapter end index {chapter_end} is out of range. Check your input."
        )
    if chapter_end == -1:
        chapter_end = len(chapters)
    if chapter_start > chapter_end:
        raise ValueError(
            f"Chapter start index {chapter_start} is larger than chapter end index {chapter_end}. Check your input."
        )

    logger.info(f"Converting chapters {chapter_start} to {chapter_end}.")

    # Calculate the number of digits needed for zero padding the file name
    max_digits = int(math.log10(len(chapters))) + 1

    tts_queue = queue.Queue()
    for idx, (title, text) in enumerate(chapters, start=1):
        if idx < chapter_start:
            continue
        if idx > chapter_end:
            break

        padded_chap_number = str(idx).zfill(max_digits)
        file_name = f"{padded_chap_number}_{title}"

        tts_queue.put(
            RecordingJob(
                title=title,
                text=text,
                author=author,
                output_file_prefix=os.path.join(output_folder, file_name),
                book_title=book_title,
                chapter_number=idx,
                voice=voice,
                speed=speed,
                pause=pause,
            )
        )

    pool = [
        threading.Thread(
            target=worker, args=(tts_queue,), daemon=True, name=f"worker-{i}"
        )
        for i in range(0, cpu_count())
    ]
    for thread in pool:
        thread.start()

    tts_queue.join()


def worker(tts_queue: queue.Queue) -> None:
    while True:
        try:
            job: RecordingJob = tts_queue.get(block=False)
        except queue.Empty:
            return
        try:
            convert_chapter(job)
        except Exception as e:
            logger.error(
                f"Failed to convert chapter {job.chapter_number} to speech. Error: {e}"
            )
        else:
            logger.info(
                f"Converted chapter {job.chapter_number} to file {job.output_file_prefix}.mp3."
            )
        finally:
            tts_queue.task_done()


def convert_chapter(job: RecordingJob) -> None:
    logger.info(f"Converting chapter {job.chapter_number} to wav.")

    subprocess.run(
        [
            "piper",
            "--output_file",
            job.wav_filename,
            "--model",
            job.voice,
            "--length-scale",
            job.speed,
            "--sentence-silence",
            job.pause,
        ],
        input=job.text.encode("utf-8"),
        stdout=subprocess.DEVNULL,
        stderr=subprocess.DEVNULL,
    )
    logger.info(f"Converting chapter {job.chapter_number} to mp3.")
    subprocess.run(
        [
            "ffmpeg",
            "-i",
            job.wav_filename,
            "-codec:a",
            "libmp3lame",
            "-b:a",
            "64k",
            job.mp3_filename,
        ],
        stdout=subprocess.DEVNULL,
        stderr=subprocess.DEVNULL,
    )
    os.remove(job.wav_filename)

    tag = EasyID3(job.mp3_filename)
    tag["artist"] = job.author
    tag["title"] = job.title
    tag["album"] = job.book_title
    tag["tracknumber"] = str(job.chapter_number)
    tag.save(v2_version=3)


class Args(tap.TypedArgs):
    epub: str = tap.arg(
        positional=True,
        help="Epub file",
    )
    audiobook_folder: str = tap.arg(
        positional=True,
        help="Destination folder for the mp3 files",
    )
    start: int = tap.arg(
        default=1,
        help="chapter to start from",
    )
    end: int = tap.arg(
        default=-1,
        help="chapter to finish at",
    )
    speed: str = tap.arg(
        default="0.75",
        help="speed of the generated audio",
    )
    voice: str = tap.arg(
        default="en_US-joe-medium",
        help="voice to use for the generated audio. To see valid options, see the docs for piper",
    )
    pause: str = tap.arg(
        default="0.5",
        help="length of pauses between sentences",
    )


def main(args: Args):
    epub_to_audiobook(
        input_file=args.epub,
        output_folder=args.audiobook_folder,
        chapter_start=args.start,
        chapter_end=args.end,
        voice=args.voice,
        speed=args.speed,
        pause=args.pause,
    )

    logger.info("Done! 👍")


if __name__ == "__main__":
    try:
        tap.Parser(Args).bind(main).run()
    except KeyboardInterrupt:
        sys.exit(1)

Is there a way to use another voice like en_US-libritts_r-medium? I have the onnx files locally. It's supported on their site.

Used chatgpt, to add another command line argument -t,--threads, so you can choose the threads used. Extra time doesn't bother me, as I have it run while I'm asleep. Computer went from feeling like it was going to lock up. To acceptably usable and not likely to freeze and need a reboot. Here's the fork for anyone interested: https://gist.github.com/bonelifer/498bcf3b47a1a9c8881074602104d98c

If you check the version in the git repo, it defaults to only two threads. This version is much buggier.

I can see your issue in that repo. I'll look at it when I have time.

This version worked for me, without the problems the repo version had. Even with it maxing out my cores, the mp3's actually completed and worked. Thanks for replying. Look forward to new commits :)