Merge pull request 'Merge/create epub into main' (#2) from merge/create-epub-into-main into main

Reviewed-on: #2

.gitignore

@@ -136,6 +136,7 @@ venv/
 ENV/
 env.bak/
 venv.bak/
+.tmp-samples/
 
 # Spyder project settings
 .spyderproject

.gitmodules

@@ -10,3 +10,9 @@
 [submodule "src/vendor/epub2tts-kokoro"]
 	path = src/vendor/epub2tts-kokoro
 	url = https://github.com/aedocw/epub2tts-kokoro
+[submodule "src/vendor/epub3-samples"]
+	path = src/vendor/epub3-samples
+	url = https://github.com/IDPF/epub3-samples
+[submodule "src/vendor/sample-epub-minimal"]
+	path = src/vendor/sample-epub-minimal
+	url = https://github.com/thansen0/sample-epub-minimal

README.md

@@ -18,6 +18,18 @@ Recent Python (>=3.11), pip (>=22.3), some sort of virtula environment recommend
 3. Test install running `epub-tts --version`
     - Make sure the venv is activated or that pip installed the package in your PATH
 
+# Sample EPUB download
+
+After cloning with submodules, the repository now includes the EPUB 3 sample collection under `src/vendor/epub3-samples`.
+
+Use the new console entrypoints to export the Moby Dick samples as EPUB files:
+
+- `download-moby-dick` creates `.tmp-samples/moby-dick.epub`
+- `download-moby-dick-media-overlays` creates `.tmp-samples/moby-dick-mo.epub`
+- `download-epub-minimal` creates `.tmp-samples/minimal.epub`
+
+The staging directory `.tmp-samples/` is included in `.gitignore` so generated artifacts stay out of version control.
+
 # 
 
 # Citation

pyproject.toml

@@ -41,6 +41,9 @@ Issues = "https://git.silveirarosa.com/renatoxsr/epub-tts/issues"
 
 [project.scripts]
 epub-tts = "epub_tts.__main__:main"
+download-moby-dick = "epub_tts.sample_download:main"
+download-moby-dick-media-overlays = "epub_tts.sample_download:main_with_media_overlay"
+download-epub-minimal = "epub_tts.sample_download:main_minimal_epub"
 
 # SETUPTOOLS
 [tool.setuptools.packages.find]

src/epub_tts/sample_download.py

@@ -0,0 +1,160 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+# Copyright (c) 2025-2026 Renato Xavier da Silveira Rosa
+# See [LICENSE](./LICENSE) or [BSD-3-Clause-Clear](https://spdx.org/licenses/BSD-3-Clause-Clear.html)
+
+"""Download EPUB sample files from a vendored samples repository and package them as EPUB archives."""
+
+from __future__ import annotations
+
+import argparse
+import shutil
+import zipfile
+from pathlib import Path
+from typing import Optional
+
+from epub_tts.vendor_utils import get_repo_path
+
+
+def get_project_root() -> Path:
+    return Path(__file__).resolve().parents[1].parent
+
+
+def get_temp_sample_root() -> Path:
+    return get_project_root() / ".tmp-samples"
+
+
+def get_sample_directory(with_media_overlay: bool) -> Path:
+    sample_name = "moby-dick-mo" if with_media_overlay else "moby-dick"
+    sample_dir = get_repo_path("epub3-samples") / "30" / sample_name
+    if not sample_dir.exists():
+        raise FileNotFoundError(
+            f"Sample directory not found: {sample_dir}. Ensure the epub3-samples submodule is initialized."
+        )
+    return sample_dir
+
+
+def get_minimal_sample_file() -> Path:
+    sample_file = get_repo_path("sample-epub-minimal") / "minimal.epub"
+    if not sample_file.exists():
+        raise FileNotFoundError(
+            f"Sample EPUB file not found: {sample_file}. Ensure the sample-epub-minimal submodule is initialized."
+        )
+    return sample_file
+
+
+def copy_sample_to_temp(sample_dir: Path) -> Path:
+    temp_root = get_temp_sample_root()
+    temp_root.mkdir(parents=True, exist_ok=True)
+    dest_dir = temp_root / sample_dir.name
+    if dest_dir.exists():
+        shutil.rmtree(dest_dir)
+    shutil.copytree(sample_dir, dest_dir)
+    return dest_dir
+
+
+def copy_file_to_temp(sample_file: Path) -> Path:
+    temp_root = get_temp_sample_root()
+    temp_root.mkdir(parents=True, exist_ok=True)
+    dest_file = temp_root / sample_file.name
+    shutil.copy2(sample_file, dest_file)
+    return dest_file
+
+
+def zip_sample_directory(sample_dir: Path, output_path: Path) -> Path:
+    output_path.parent.mkdir(parents=True, exist_ok=True)
+    if output_path.exists():
+        output_path.unlink()
+
+    mimetype_file = sample_dir / "mimetype"
+    with zipfile.ZipFile(output_path, mode="w", compression=zipfile.ZIP_DEFLATED) as archive:
+        if mimetype_file.exists():
+            archive.write(mimetype_file, "mimetype", compress_type=zipfile.ZIP_STORED)
+        for path in sorted(sample_dir.rglob("*")):
+            if path.is_file() and path != mimetype_file:
+                archive.write(path, path.relative_to(sample_dir).as_posix())
+
+    return output_path
+
+
+def download_sample_moby_dick(
+    output: Optional[str] = None,
+    with_media_overlay: bool = False,
+) -> Path:
+    sample_dir = get_sample_directory(with_media_overlay=with_media_overlay)
+    temp_sample_dir = copy_sample_to_temp(sample_dir)
+    default_output = get_temp_sample_root() / f"{sample_dir.name}.epub"
+    output_path = Path(output) if output else default_output
+    output_path = output_path.with_suffix(".epub")
+    return zip_sample_directory(temp_sample_dir, output_path)
+
+
+def download_minimal_epub_sample(output: Optional[str] = None) -> Path:
+    sample_file = get_minimal_sample_file()
+    temp_sample_file = copy_file_to_temp(sample_file)
+    default_output = get_temp_sample_root() / sample_file.name
+    output_path = Path(output) if output else default_output
+    output_path = output_path.with_suffix(".epub")
+    if temp_sample_file != output_path:
+        output_path.parent.mkdir(parents=True, exist_ok=True)
+        shutil.copy2(temp_sample_file, output_path)
+    return output_path
+
+
+def main(argv: Optional[list[str]] = None) -> int:
+    parser = argparse.ArgumentParser(
+        description="Download the Moby Dick EPUB sample and package it as an EPUB file."
+    )
+    parser.add_argument(
+        "-o",
+        "--output",
+        help="Output EPUB file path. Defaults to .tmp-samples/moby-dick.epub or .tmp-samples/moby-dick-mo.epub.",
+    )
+    args = parser.parse_args(argv)
+    output_path = download_sample_moby_dick(
+        output=args.output,
+        with_media_overlay=False,
+    )
+    print(f"Created EPUB sample archive: {output_path}")
+    return 0
+
+
+def main_with_media_overlay(argv: Optional[list[str]] = None) -> int:
+    parser = argparse.ArgumentParser(
+        description=(
+            "Download the Moby Dick EPUB sample with media overlay and package it as an EPUB file."
+        )
+    )
+    parser.add_argument(
+        "-o",
+        "--output",
+        help="Output EPUB file path. Defaults to .tmp-samples/moby-dick-mo.epub.",
+    )
+    args = parser.parse_args(argv)
+    output_path = download_sample_moby_dick(
+        output=args.output,
+        with_media_overlay=True,
+    )
+    print(f"Created EPUB sample archive: {output_path}")
+    return 0
+
+
+def main_minimal_epub(argv: Optional[list[str]] = None) -> int:
+    parser = argparse.ArgumentParser(
+        description=(
+            "Download the minimal EPUB sample and copy it to an EPUB file."
+        )
+    )
+    parser.add_argument(
+        "-o",
+        "--output",
+        help="Output EPUB file path. Defaults to .tmp-samples/minimal.epub.",
+    )
+    args = parser.parse_args(argv)
+    output_path = download_minimal_epub_sample(output=args.output)
+    print(f"Created minimal EPUB sample archive: {output_path}")
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())