update

README.md

@@ -0,0 +1,172 @@
+# Piper-TTS Audiobook Script
+
+This project is a **one-command solution** to convert large `.txt` documents into high-quality **MP3 audiobooks** entirely **offline** using [Piper](https://github.com/rhasspy/piper) — a fast, local, open-source text-to-speech engine.
+
+The script:
+- Installs all prerequisites (`pipx`, `ffmpeg`, `piper-tts`)
+- Downloads a chosen Piper voice model (default: **female UK English — en_GB-cori-high**)
+- Splits large text into manageable chunks for Piper
+- Synthesizes each chunk into audio
+- Stitches the chunks together into a single **MP3** file
+
+---
+
+## 📂 Project Structure
+
+```
+
+Piper-TTS-Script/
+├── make\_audiobook.sh   # Main script
+└── README.md           # This file
+
+````
+
+---
+
+## ⚙️ Requirements
+
+- **Linux Mint / Ubuntu** (or other Debian-based distro)
+- Internet connection for the first run (to install packages and download the voice model)
+- A `.txt` file containing your book or text to be converted
+
+---
+
+## 📥 Installation
+
+1. **Clone or download this repo**:
+   ```bash
+   cd ~/Documents/GitHub
+   git clone https://github.com/yourusername/Piper-TTS-Script.git
+````
+
+*(Or just manually create the folder and place `make_audiobook.sh` inside)*
+
+2. **Make the script executable**:
+
+   ```bash
+   chmod +x ~/Documents/GitHub/Piper-TTS-Script/make_audiobook.sh
+   ```
+
+---
+
+## 🗂 Preparing Your Text File
+
+Place your text file somewhere accessible, e.g.:
+
+```
+~/Documents/audiobooks/mybook.txt
+```
+
+If you have a PDF or EPUB, convert it to plain text first:
+
+```bash
+# PDF to TXT
+pdftotext mybook.pdf mybook.txt
+
+# EPUB to TXT (requires calibre)
+ebook-convert mybook.epub mybook.txt
+```
+
+---
+
+## 🚀 Usage
+
+Basic usage:
+
+```bash
+~/Documents/GitHub/Piper-TTS-Script/make_audiobook.sh ~/Documents/audiobooks/mybook.txt
+```
+
+* **First run**: Installs all dependencies and downloads the default voice model.
+* **Output**: An MP3 file in the same directory as your text:
+
+  ```
+  ~/Documents/audiobooks/mybook.mp3
+  ```
+
+---
+
+## 🔧 Configuration
+
+At the top of `make_audiobook.sh` there is a **CONFIG** section where you can adjust:
+
+| Setting                             | Description                                            |
+| ----------------------------------- | ------------------------------------------------------ |
+| `MODEL_NAME`                        | Name of the Piper voice model                          |
+| `MODEL_FILE`                        | Path to your `.onnx` model file                        |
+| `MODEL_ONNX_URL` / `MODEL_JSON_URL` | Download URLs for the model and metadata               |
+| `LENGTH_SCALE`                      | Speech speed (1.0 = normal, <1 faster, >1 slower)      |
+| `NOISE_SCALE`                       | Expressiveness (lower = flatter, higher = more varied) |
+| `MAX_CHARS`                         | Max characters per chunk sent to Piper                 |
+| `SILENCE_MS`                        | Silence gap between chunks in milliseconds             |
+| `MP3_Q`                             | MP3 quality (VBR: 0=best, 2=high, 5=medium)            |
+
+---
+
+## 🎤 Changing the Voice
+
+This script defaults to:
+
+* **Female UK English** — `en_GB-cori-high`
+
+To change:
+
+1. Visit the official Piper voice list:
+   [https://github.com/rhasspy/piper/blob/master/VOICES.md](https://github.com/rhasspy/piper/blob/master/VOICES.md)
+
+2. Copy the `.onnx` and `.onnx.json` URLs for your chosen voice.
+
+3. Edit the **CONFIG** section in `make_audiobook.sh` with the new:
+
+   * `MODEL_NAME`
+   * `MODEL_ONNX_URL`
+   * `MODEL_JSON_URL`
+
+Example for **Female US English (Lessac)**:
+
+```bash
+MODEL_NAME="en_US-lessac-high"
+MODEL_ONNX_URL="https://huggingface.co/rhasspy/piper-voices/resolve/main/en/en_US/lessac/high/en_US-lessac-high.onnx"
+MODEL_JSON_URL="https://huggingface.co/rhasspy/piper-voices/resolve/main/en/en_US/lessac/high/en_US-lessac-high.onnx.json"
+```
+
+---
+
+## 🛠 Advanced Usage
+
+Specify a custom voice model:
+
+```bash
+~/Documents/GitHub/Piper-TTS-Script/make_audiobook.sh ~/Documents/audiobooks/mybook.txt \
+  --model ~/.local/share/piper/voices/en_US-lessac-high.onnx
+```
+
+Name the output file:
+
+```bash
+~/Documents/GitHub/Piper-TTS-Script/make_audiobook.sh ~/Documents/audiobooks/mybook.txt \
+  --name my_custom_audiobook.mp3
+```
+
+---
+
+## ⚠️ Limitations
+
+* For **extremely large books** (>200k characters), processing can still take a while even with chunking.
+* If the voice model URLs change, you’ll need to update them from [VOICES.md](https://github.com/rhasspy/piper/blob/master/VOICES.md).
+* Piper’s pronunciation is very good but not perfect; some technical or foreign words may be read oddly.
+
+---
+
+## 📄 License
+
+This script is released under the MIT License. Piper itself is licensed under the Mozilla Public License 2.0.
+
+---
+
+## 🙋‍♂️ Credits
+
+* [Piper TTS](https://github.com/rhasspy/piper) — Open-source text-to-speech engine
+* [ffmpeg](https://ffmpeg.org) — Audio conversion
+* Script by **Your Name**
+

make_audiobook.sh

@@ -0,0 +1,233 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# ============================================================
+# make_audiobook.sh
+# One-shot installer + large-text → MP3 audiobook using Piper
+# Works on Linux Mint / Ubuntu (needs sudo for apt installs)
+# ============================================================
+
+# -----------------------
+# CONFIG (edit as needed)
+# -----------------------
+# Piper model (female by default). Set MODEL_FILE to a local .onnx if you already downloaded one.
+# Otherwise, the script will try to download the URLs below (you can swap them with your preferred voice).
+MODEL_NAME="en_GB-cori-high"  # Example: UK English, female, high quality
+MODEL_DIR="${HOME}/.local/share/piper/voices"
+MODEL_FILE="${MODEL_DIR}/${MODEL_NAME}.onnx"
+MODEL_JSON="${MODEL_FILE}.json"
+
+# If MODEL_FILE doesn't exist, try to fetch from these URLs:
+MODEL_ONNX_URL="https://huggingface.co/rhasspy/piper-voices/resolve/main/en/en_GB/cori/high/en_GB-cori-high.onnx"
+MODEL_JSON_URL="https://huggingface.co/rhasspy/piper-voices/resolve/main/en/en_GB/cori/high/en_GB-cori-high.onnx.json"
+
+# Piper synthesis knobs (tune to taste)
+LENGTH_SCALE="1.0"   # >1.0 = slower, <1.0 = faster
+NOISE_SCALE="0.60"   # lower = flatter; higher = more expressive
+MAX_CHARS="3000"     # characters per chunk sent to Piper
+SILENCE_MS="300"     # silence between chunks (ms)
+
+# MP3 quality (VBR via -q:a: 0=best, 2~high, 5=medium)
+MP3_Q="2"
+
+# -----------------------
+# END CONFIG
+# -----------------------
+
+# --- helpers ---
+say() { printf "\n\033[1;32m%s\033[0m\n" "$*"; }
+warn() { printf "\n\033[1;33m%s\033[0m\n" "$*"; }
+die() { printf "\n\033[1;31mERROR:\033[0m %s\n" "$*" >&2; exit 1; }
+
+# --- args ---
+if [[ $# -lt 1 ]]; then
+  cat >&2 <<EOF
+Usage: $0 /path/to/mybook.txt [--model /path/to/voice.onnx] [--name OutputName.mp3]
+
+Options:
+  --model PATH    Use a specific Piper .onnx model file (overrides CONFIG)
+  --name  NAME    Output mp3 name (default: input basename with .mp3)
+EOF
+  exit 1
+fi
+
+INPUT_TXT=""
+CUSTOM_MODEL=""
+OUT_NAME=""
+
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --model) CUSTOM_MODEL="${2:-}"; shift 2;;
+    --name)  OUT_NAME="${2:-}"; shift 2;;
+    *) INPUT_TXT="$1"; shift;;
+  esac
+done
+
+[[ -f "$INPUT_TXT" ]] || die "Input text not found: $INPUT_TXT"
+INPUT_TXT="$(realpath "$INPUT_TXT")"
+OUT_DIR="$(dirname "$INPUT_TXT")"
+BASE="$(basename "${INPUT_TXT%.*}")"
+OUTPUT_MP3="${OUT_NAME:-${BASE}.mp3}"
+OUTPUT_MP3="${OUT_DIR}/${OUTPUT_MP3}"
+
+# --- install prereqs ---
+say "Installing prerequisites (sudo required for apt)…"
+sudo apt update -y
+sudo apt install -y python3-venv pipx ffmpeg curl wget
+
+# Ensure pipx path is active in this session
+if ! command -v pipx >/dev/null 2>&1; then
+  die "pipx not found after install. Please re-open your terminal and re-run this script."
+fi
+pipx ensurepath >/dev/null 2>&1 || true
+# shellcheck disable=SC1090
+source "${HOME}/.profile" 2>/dev/null || true
+hash -r
+
+# Install Piper CLI via pipx if missing
+if ! command -v piper >/dev/null 2>&1; then
+  say "Installing Piper CLI with pipx…"
+  pipx install piper-tts
+fi
+
+command -v piper >/dev/null 2>&1 || die "Piper CLI not found on PATH after install."
+
+# --- model setup ---
+mkdir -p "${MODEL_DIR}"
+
+if [[ -n "${CUSTOM_MODEL}" ]]; then
+  [[ -f "$CUSTOM_MODEL" ]] || die "Custom model not found: $CUSTOM_MODEL"
+  MODEL_FILE="$(realpath "$CUSTOM_MODEL")"
+  MODEL_JSON="${MODEL_FILE}.json"
+  [[ -f "$MODEL_JSON" ]] || warn "Note: ${MODEL_JSON} not found. Piper usually works without it, but voice metadata will be missing."
+else
+  if [[ ! -f "$MODEL_FILE" ]]; then
+    say "Downloading Piper model: ${MODEL_NAME}"
+    wget -O "${MODEL_FILE}" "${MODEL_ONNX_URL}" || die "Failed to download model .onnx"
+  fi
+  if [[ ! -f "$MODEL_JSON" ]]; then
+    say "Downloading Piper model metadata (.json)…"
+    wget -O "${MODEL_JSON}" "${MODEL_JSON_URL}" || warn "Could not download ${MODEL_JSON}. Continuing without metadata."
+  fi
+fi
+
+# --- temp working dir ---
+WORKDIR="$(mktemp -d -t piperbook-XXXXXX)"
+trap 'rm -rf "$WORKDIR"' EXIT
+
+# --- python chunk + synth ---
+say "Starting synthesis with Piper (this can take a while)…"
+PY_SCRIPT="${WORKDIR}/run_piper.py"
+cat > "${PY_SCRIPT}" <<'PYCODE'
+import argparse, os, re, subprocess, sys, tempfile, shutil
+
+def split_sentences(text: str):
+    text = re.sub(r'\s+', ' ', text.strip())
+    # basic sentence split: keep punctuation; avoid splitting on common decimals/abbrevs heuristically
+    pieces = re.split(r'(?<=[.!?])\s+(?=[A-Z0-9(])', text)
+    return [p.strip() for p in pieces if p.strip()]
+
+def chunk(sentences, max_chars=3000):
+    buf, n = [], 0
+    for s in sentences:
+        if n + len(s) + 1 > max_chars and buf:
+            yield ' '.join(buf)
+            buf, n = [], 0
+        buf.append(s)
+        n += len(s) + 1
+    if buf:
+        yield ' '.join(buf)
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument("--input", required=True)
+    ap.add_argument("--model", required=True)
+    ap.add_argument("--outdir", required=True)
+    ap.add_argument("--length_scale", type=float, default=1.0)
+    ap.add_argument("--noise_scale", type=float, default=0.667)
+    ap.add_argument("--max_chars", type=int, default=3000)
+    ap.add_argument("--silence_ms", type=int, default=300)
+    args = ap.parse_args()
+
+    if not shutil.which("piper"):
+        sys.exit("piper not found in PATH")
+    if not shutil.which("ffmpeg"):
+        sys.exit("ffmpeg not found in PATH")
+
+    with open(args.input, "r", encoding="utf-8") as f:
+        raw = f.read()
+    sents = split_sentences(raw)
+    chunks = list(chunk(sents, max_chars=args.max_chars))
+    if not chunks:
+        sys.exit("Nothing to synthesize.")
+
+    os.makedirs(args.outdir, exist_ok=True)
+    wavs = []
+    for i, text in enumerate(chunks):
+        wav_path = os.path.join(args.outdir, f"part_{i:05d}.wav")
+        p = subprocess.run(
+            ["piper",
+             "--model", args.model,
+             "--length_scale", str(args.length_scale),
+             "--noise_scale", str(args.noise_scale),
+             "--output_file", wav_path],
+            input=text.encode("utf-8"),
+            stdout=subprocess.PIPE, stderr=subprocess.PIPE)
+        if p.returncode != 0 or not os.path.exists(wav_path):
+            sys.stderr.write(p.stderr.decode("utf-8", errors="ignore"))
+            sys.exit(f"Piper failed on chunk {i}")
+        wavs.append(wav_path)
+
+    # Build concat list with optional silence
+    list_path = os.path.join(args.outdir, "concat.txt")
+    with open(list_path, "w", encoding="utf-8") as lf:
+        for i, w in enumerate(wavs):
+            lf.write(f"file '{w}'\n")
+            if i != len(wavs)-1 and args.silence_ms > 0:
+                sil = os.path.join(args.outdir, f"sil_{i:05d}.wav")
+                # 22.05k mono silence
+                subprocess.run([
+                    "ffmpeg","-f","lavfi","-i",
+                    f"anullsrc=r=22050:cl=mono",
+                    "-t", str(args.silence_ms/1000.0),
+                    "-y", sil
+                ], stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL, check=True)
+                lf.write(f"file '{sil}'\n")
+
+    # Concatenate WAVs losslessly
+    joined_wav = os.path.join(args.outdir, "joined.wav")
+    subprocess.run([
+        "ffmpeg","-f","concat","-safe","0","-i", list_path,
+        "-c","copy","-y", joined_wav
+    ], stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL, check=True)
+
+    print(joined_wav)
+
+if __name__ == "__main__":
+    main()
+PYCODE
+
+# Run the Python worker to synthesize chunks → joined.wav
+JOINED_WAV="$(
+  python3 "${PY_SCRIPT}" \
+    --input "${INPUT_TXT}" \
+    --model "${MODEL_FILE}" \
+    --outdir "${WORKDIR}/audio" \
+    --length_scale "${LENGTH_SCALE}" \
+    --noise_scale "${NOISE_SCALE}" \
+    --max_chars "${MAX_CHARS}" \
+    --silence_ms "${SILENCE_MS}"
+)"
+
+[[ -f "${JOINED_WAV}" ]] || die "Synthesis did not produce ${JOINED_WAV}"
+
+# Convert to MP3
+say "Encoding MP3 → ${OUTPUT_MP3}"
+ffmpeg -y -i "${JOINED_WAV}" -codec:a libmp3lame -q:a "${MP3_Q}" "${OUTPUT_MP3}" >/dev/null 2>&1
+
+say "Done! MP3 written to:"
+echo "  ${OUTPUT_MP3}"
+
+echo
+warn "Tip: To use a different voice, edit the CONFIG section at the top or pass --model /path/to/voice.onnx"
+warn "Latest voice list (copy links): https://github.com/rhasspy/piper/blob/master/VOICES.md"