morphconv 1.0.3__tar.gz → 1.0.4__tar.gz

{morphconv-1.0.3 → morphconv-1.0.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: morphconv
-Version: 1.0.3
+Version: 1.0.4
 Summary: morph — convert anything to anything, from the CLI
 License: MIT
 Requires-Python: >=3.10
@@ -30,6 +30,20 @@ Requires-Dist: pytesseract>=0.3.10
 Requires-Dist: trafilatura>=1.8.0
 Requires-Dist: vtracer>=0.6.0
 Requires-Dist: yt-dlp>=2024.03.10
+Requires-Dist: pymupdf>=1.24.0
+Requires-Dist: pdfplumber>=0.11.0
+Requires-Dist: pdf2docx>=0.5.8
+Requires-Dist: py7zr>=0.21.0
+Requires-Dist: rarfile>=4.1
+Requires-Dist: pysubs2>=1.6.0
+Requires-Dist: tomlkit>=0.12.0
+Requires-Dist: nbconvert>=7.0.0
+Requires-Dist: img2pdf>=0.5.0
+Requires-Dist: rawpy>=0.19.0
+Requires-Dist: qrcode[pil]>=7.4.0
+Requires-Dist: pyzbar>=0.1.9
+Requires-Dist: markdownify>=0.13.0
+Requires-Dist: weasyprint>=62.0
 Provides-Extra: dev
 Requires-Dist: pytest>=8.0.0; extra == "dev"
 Requires-Dist: pytest-mock>=3.12.0; extra == "dev"
@@ -62,6 +76,7 @@ morph report.docx report.pdf
 morph data.csv data.xlsx --table-style TableStyleMedium2 --header-bg 2E7D32
 morph logo.png logo.svg --mode spline
 morph batch '*.mp4' mp3 --workers 4
+morph pack '*.png' photos/ doc.pdf archive.zip
 ```
 
 </div>
@@ -85,6 +100,7 @@ Whether you are rendering a 3D `.blend` file to a `.png`, scraping a Wikipedia a
 - **Output Verification.** Pass the `--verify` flag to calculate and display the exact SHA-256 checksum of the resulting file right alongside the success message.
 - **Transparent Remote Downloads.** Pass any `http://` link and `morph` handles the rest. It uses `trafilatura` to scrape web articles into clean Markdown, or `yt-dlp` to natively download high-quality media (`morph <url> --audio`).
 - **Batch Processing Engine.** `morph batch '*.mp4' mp3` converts files concurrently with a live progress table, smart skip logic (`--newer-only`, `--skip-existing`), and output structure mirroring.
+- **Pack Command.** Run `morph pack *.png output.zip` to collect any number of files, globs, or directories into a single archive (zip/7z/tar.gz), with auto-stripped common paths.
 - **Interactive TUI.** Type `morph` with no arguments to launch a stunning, keyboard-driven Textual UI. Navigate your filesystem, preview conversion options, and watch live progress bars without touching your mouse.
 - **Automated Dependency Management.** Missing `ffmpeg` or `pandoc`? `morph` detects your OS and package manager (brew/apt/winget/etc.), shows you the exact install command, and asks before running it. Massive dependencies (like 3D rendering engines) are cleanly isolated as optional extras!
 - **Pluggable Architecture.** Drop a Python script in `morph/converters/`, add a `@register` decorator, and it's instantly live in the routing graph.
@@ -249,20 +265,24 @@ During batch conversions, `morph` displays a live status table:
 
 | Domain | Formats | Backend |
 |---|---|---|
-| **Data** | `csv`, `xlsx`, `json`, `yaml`, `parquet`, `feather`, `ods`, `xml`, `html`, `sqlite` | Native (`pandas` / `openpyxl`) |
-| **Documents** | `md`, `html`, `docx`, `odt`, `rtf`, `epub`, `latex`, `rst`, `txt`, `ipynb`, `pptx`, `adoc`, `org`, `opml`, `man`, `pdf`* | `pandoc` |
-| **Images (Raster)** | `png`, `jpg/jpeg`, `webp`, `bmp`, `gif`, `tiff`, `ico`, `heic`, `avif`, `icns`, `pdf` | `Pillow` / `cairosvg` |
+| **Data** | `csv`, `xlsx`, `json`, `yaml`, `toml`, `parquet`, `feather`, `ods`, `xml`, `html`, `sqlite` | Native (`pandas` / `tomlkit`) |
+| **Documents** | `md`, `html`, `docx`, `odt`, `rtf`, `epub`, `latex`, `rst`, `txt`, `pptx`, `adoc`, `org`, `opml`, `man` | `pandoc` |
+| **Legacy Office** | `doc`, `xls`, `ppt` | `LibreOffice (soffice)` |
+| **Notebooks** | `ipynb` | `nbconvert` (executes & renders outputs) |
+| **PDF** | `pdf` (real text, layout, tables, generation) | `pymupdf` / `pdfplumber` / `pdf2docx` |
+| **Images (Raster)** | `png`, `jpg/jpeg`, `webp`, `bmp`, `gif`, `tiff`, `ico`, `heic`, `avif`, `icns` | `Pillow` / `cairosvg` |
 | **Vectorization** | `png`/`jpg`/etc. → `svg` | `vtracer` |
 | **3D Models** | `obj`, `stl`, `fbx`, `gltf`, `glb`, `blend`, `any → png` (Render) | `bpy` (Blender) |
 | **Web Extraction** | `url` → `md`, `txt`, `xml`, `html` | `trafilatura` / `crawl4ai` |
 | **Audio** | `mp3`, `wav`, `flac`, `ogg`, `aac`, `m4a`, `opus`, `wma` | `ffmpeg` |
 | **Video** | `mp4`, `mkv`, `mov`, `webm`, `avi`, `flv`, `wmv`, `mpeg` | `ffmpeg` |
+| **Subtitles** | `srt`, `vtt`, `ass`, `ssa`, `sub`, `sami` | `pysubs2` |
 | **Ebooks** | `epub`, `mobi`, `azw3` | `ebook-convert` |
 | **Fonts** | `ttf`, `otf`, `woff`, `woff2` | `fontTools` |
 | **OCR** | `png`, `jpg`, `webp`, `pdf`, etc. → `txt` | `tesseract` |
-| **Archives** | `zip`, `tar`, `tar.gz`, `tar.bz2`, `tar.xz` | Python stdlib |
+| **Archives** | `zip`, `7z`, `tar`, `tar.gz`, `tar.bz2`, `tar.xz`, `rar`* | Python stdlib / `py7zr` / `rarfile` |
 
-*\* Note: PDF is an output-only format for documents. Extracting structured text back out of a PDF is handled via the OCR pipeline.*
+*\* Note: RAR is an extract-only format (RAR creation requires proprietary tools).*
 
 ---
 

{morphconv-1.0.3 → morphconv-1.0.4}/README.md

@@ -18,6 +18,7 @@ morph report.docx report.pdf
 morph data.csv data.xlsx --table-style TableStyleMedium2 --header-bg 2E7D32
 morph logo.png logo.svg --mode spline
 morph batch '*.mp4' mp3 --workers 4
+morph pack '*.png' photos/ doc.pdf archive.zip
 ```
 
 </div>
@@ -41,6 +42,7 @@ Whether you are rendering a 3D `.blend` file to a `.png`, scraping a Wikipedia a
 - **Output Verification.** Pass the `--verify` flag to calculate and display the exact SHA-256 checksum of the resulting file right alongside the success message.
 - **Transparent Remote Downloads.** Pass any `http://` link and `morph` handles the rest. It uses `trafilatura` to scrape web articles into clean Markdown, or `yt-dlp` to natively download high-quality media (`morph <url> --audio`).
 - **Batch Processing Engine.** `morph batch '*.mp4' mp3` converts files concurrently with a live progress table, smart skip logic (`--newer-only`, `--skip-existing`), and output structure mirroring.
+- **Pack Command.** Run `morph pack *.png output.zip` to collect any number of files, globs, or directories into a single archive (zip/7z/tar.gz), with auto-stripped common paths.
 - **Interactive TUI.** Type `morph` with no arguments to launch a stunning, keyboard-driven Textual UI. Navigate your filesystem, preview conversion options, and watch live progress bars without touching your mouse.
 - **Automated Dependency Management.** Missing `ffmpeg` or `pandoc`? `morph` detects your OS and package manager (brew/apt/winget/etc.), shows you the exact install command, and asks before running it. Massive dependencies (like 3D rendering engines) are cleanly isolated as optional extras!
 - **Pluggable Architecture.** Drop a Python script in `morph/converters/`, add a `@register` decorator, and it's instantly live in the routing graph.
@@ -205,20 +207,24 @@ During batch conversions, `morph` displays a live status table:
 
 | Domain | Formats | Backend |
 |---|---|---|
-| **Data** | `csv`, `xlsx`, `json`, `yaml`, `parquet`, `feather`, `ods`, `xml`, `html`, `sqlite` | Native (`pandas` / `openpyxl`) |
-| **Documents** | `md`, `html`, `docx`, `odt`, `rtf`, `epub`, `latex`, `rst`, `txt`, `ipynb`, `pptx`, `adoc`, `org`, `opml`, `man`, `pdf`* | `pandoc` |
-| **Images (Raster)** | `png`, `jpg/jpeg`, `webp`, `bmp`, `gif`, `tiff`, `ico`, `heic`, `avif`, `icns`, `pdf` | `Pillow` / `cairosvg` |
+| **Data** | `csv`, `xlsx`, `json`, `yaml`, `toml`, `parquet`, `feather`, `ods`, `xml`, `html`, `sqlite` | Native (`pandas` / `tomlkit`) |
+| **Documents** | `md`, `html`, `docx`, `odt`, `rtf`, `epub`, `latex`, `rst`, `txt`, `pptx`, `adoc`, `org`, `opml`, `man` | `pandoc` |
+| **Legacy Office** | `doc`, `xls`, `ppt` | `LibreOffice (soffice)` |
+| **Notebooks** | `ipynb` | `nbconvert` (executes & renders outputs) |
+| **PDF** | `pdf` (real text, layout, tables, generation) | `pymupdf` / `pdfplumber` / `pdf2docx` |
+| **Images (Raster)** | `png`, `jpg/jpeg`, `webp`, `bmp`, `gif`, `tiff`, `ico`, `heic`, `avif`, `icns` | `Pillow` / `cairosvg` |
 | **Vectorization** | `png`/`jpg`/etc. → `svg` | `vtracer` |
 | **3D Models** | `obj`, `stl`, `fbx`, `gltf`, `glb`, `blend`, `any → png` (Render) | `bpy` (Blender) |
 | **Web Extraction** | `url` → `md`, `txt`, `xml`, `html` | `trafilatura` / `crawl4ai` |
 | **Audio** | `mp3`, `wav`, `flac`, `ogg`, `aac`, `m4a`, `opus`, `wma` | `ffmpeg` |
 | **Video** | `mp4`, `mkv`, `mov`, `webm`, `avi`, `flv`, `wmv`, `mpeg` | `ffmpeg` |
+| **Subtitles** | `srt`, `vtt`, `ass`, `ssa`, `sub`, `sami` | `pysubs2` |
 | **Ebooks** | `epub`, `mobi`, `azw3` | `ebook-convert` |
 | **Fonts** | `ttf`, `otf`, `woff`, `woff2` | `fontTools` |
 | **OCR** | `png`, `jpg`, `webp`, `pdf`, etc. → `txt` | `tesseract` |
-| **Archives** | `zip`, `tar`, `tar.gz`, `tar.bz2`, `tar.xz` | Python stdlib |
+| **Archives** | `zip`, `7z`, `tar`, `tar.gz`, `tar.bz2`, `tar.xz`, `rar`* | Python stdlib / `py7zr` / `rarfile` |
 
-*\* Note: PDF is an output-only format for documents. Extracting structured text back out of a PDF is handled via the OCR pipeline.*
+*\* Note: RAR is an extract-only format (RAR creation requires proprietary tools).*
 
 ---
 

{morphconv-1.0.3 → morphconv-1.0.4}/morph/cli.py

@@ -6,13 +6,15 @@ morph — convert anything to anything, from the CLI.
   morph data.csv data.xlsx --table-style TableStyleMedium2 --header-bg 2E7D32
   morph data.csv data.xlsx --help      (shows flags for THIS pair only)
   morph batch '*.mp4' mp3 --workers 4
+  morph pack '*.png' photos/ report.docx archive.zip
+  morph pack project/ backup.tar.gz --recursive
   morph formats docx
   morph history
   morph deps
   morph                                 (no args -> launches the interactive TUI)
 
 There is no "convert" subcommand — morph already means convert. Anything
-that isn't a recognized subcommand (formats, deps, batch, history) is
+that isn't a recognized subcommand (formats, deps, batch, history, pack) is
 treated as a conversion job and routed through the engine.
 """
 
@@ -85,7 +87,7 @@ class MorphGroup(TyperGroup):
     def resolve_command(self, ctx, args):
         # We explicitly allow "config" alongside whatever is in self.commands
         # just in case Typer hasn't fully populated the commands dict yet.
-        known_commands = list(self.commands.keys()) + ["config", "init"]
+        known_commands = list(self.commands.keys()) + ["config", "init", "pack"]
         if args and args[0] not in known_commands and not args[0].startswith("-"):
             args = ["run", *args]
         return super().resolve_command(ctx, args)
@@ -552,6 +554,163 @@ def batch_cmd(
     raise typer.Exit(0 if result.failed == 0 else 1)
 
 
+
+# ── pack ──────────────────────────────────────────────────────────────────────
+
+@app.command(
+    "pack",
+    help="Pack any files or globs into a single archive (zip, 7z, tar.gz, tar.bz2, tar.xz).",
+)
+def pack_cmd(
+    inputs: List[str] = typer.Argument(..., help="Files, globs, or directories to pack. Last argument is the output archive path."),
+    recursive: bool = typer.Option(False, "-r", "--recursive", help="Recurse into directories."),
+    strip_top_level: bool = typer.Option(False, "--strip-top-level", help="Drop a single shared top-level directory from paths inside the archive."),
+    quiet: bool = typer.Option(False, "-q", "--quiet", help="Suppress output."),
+) -> None:
+    import glob as _glob
+    import zipfile
+    import tarfile
+    import os
+
+    if not quiet:
+        console.print("\n[bold cyan]⚡ MORPH[/bold cyan] [dim]— Pack anything.[/dim]\n")
+
+    if len(inputs) < 2:
+        err_console.print("[error]✗ Provide at least one input and an output archive path.[/error]")
+        err_console.print("  [muted]Example: morph pack *.png photos.zip[/muted]")
+        err_console.print("  [muted]Example: morph pack docs/ report.tar.gz --recursive[/muted]")
+        raise typer.Exit(1)
+
+    *patterns, output_str = inputs
+    output_path = Path(output_str)
+
+    # Determine archive format from extension
+    name_lower = output_path.name.lower()
+    if name_lower.endswith(".tar.gz") or name_lower.endswith(".tgz"):
+        fmt = "tar.gz"
+    elif name_lower.endswith(".tar.bz2") or name_lower.endswith(".tbz2"):
+        fmt = "tar.bz2"
+    elif name_lower.endswith(".tar.xz") or name_lower.endswith(".txz"):
+        fmt = "tar.xz"
+    elif name_lower.endswith(".tar"):
+        fmt = "tar"
+    elif name_lower.endswith(".7z"):
+        fmt = "7z"
+    elif name_lower.endswith(".zip"):
+        fmt = "zip"
+    else:
+        err_console.print(
+            f"[error]✗ Cannot determine archive format from:[/error] [bold]{output_path.name}[/bold]\n"
+            "  [muted]Supported: .zip  .7z  .tar  .tar.gz  .tar.bz2  .tar.xz[/muted]"
+        )
+        raise typer.Exit(1)
+
+    # Collect all matching files, preserving relative paths
+    collected: list[Path] = []
+    seen: set[Path] = set()
+
+    for pattern in patterns:
+        p = Path(pattern)
+        if p.is_dir():
+            glob_pat = "**/*" if recursive else "*"
+            matches = [f for f in p.glob(glob_pat) if f.is_file()]
+        else:
+            # Expand shell-style globs
+            expanded = [Path(m) for m in _glob.glob(pattern, recursive=recursive)]
+            if not expanded:
+                err_console.print(f"[warning]⚠ No files matched:[/warning] {pattern}")
+                continue
+            matches = []
+            for m in expanded:
+                if m.is_dir():
+                    if recursive:
+                        matches.extend(f for f in m.rglob("*") if f.is_file())
+                    else:
+                        err_console.print(f"[warning]⚠ Skipping directory (use --recursive):[/warning] {m}")
+                else:
+                    matches.append(m)
+
+        for f in matches:
+            resolved = f.resolve()
+            if resolved not in seen:
+                seen.add(resolved)
+                collected.append(f)
+
+    if not collected:
+        err_console.print("[error]✗ No files matched the given patterns.[/error]")
+        raise typer.Exit(1)
+
+    # Compute archive-internal paths: strip common root so paths are relative
+    abs_collected = [f.resolve() for f in collected]
+    if len(abs_collected) == 1:
+        common = abs_collected[0].parent
+    else:
+        # Find longest common ancestor directory
+        from pathlib import PurePath
+        parts_list = [p.parts for p in abs_collected]
+        common_parts = []
+        for parts in zip(*parts_list):
+            if len(set(parts)) == 1:
+                common_parts.append(parts[0])
+            else:
+                break
+        common = Path(*common_parts) if common_parts else Path("/")
+
+    def arcname(f: Path) -> str:
+        try:
+            rel = f.resolve().relative_to(common)
+        except ValueError:
+            rel = Path(f.name)
+        return str(rel)
+
+    output_path.parent.mkdir(parents=True, exist_ok=True)
+
+    if not quiet:
+        console.print(Panel.fit(
+            f"Packing [bold]{len(collected)}[/bold] file(s) into [bold]{output_path.name}[/bold]  [muted]({fmt})[/muted]",
+            title="morph pack", border_style="cyan",
+        ))
+
+    t_start = time.perf_counter()
+
+    try:
+        if fmt == "zip":
+            with zipfile.ZipFile(output_path, "w", zipfile.ZIP_DEFLATED) as zf:
+                for f in abs_collected:
+                    zf.write(f, arcname(f))
+
+        elif fmt in ("tar", "tar.gz", "tar.bz2", "tar.xz"):
+            _modes = {"tar": "w:", "tar.gz": "w:gz", "tar.bz2": "w:bz2", "tar.xz": "w:xz"}
+            with tarfile.open(output_path, _modes[fmt]) as tf:
+                for f in abs_collected:
+                    tf.add(f, arcname=arcname(f))
+
+        elif fmt == "7z":
+            try:
+                import py7zr
+            except ImportError:
+                err_console.print("[error]✗ py7zr is required for .7z packing.[/error]")
+                err_console.print("  [muted]Install: pip install py7zr[/muted]")
+                raise typer.Exit(1)
+            with py7zr.SevenZipFile(output_path, mode="w") as sz:
+                for f in abs_collected:
+                    sz.write(f, arcname(f))
+
+    except Exception as exc:
+        err_console.print(f"[error]✗ Pack failed:[/error] {exc}")
+        raise typer.Exit(1)
+
+    elapsed = time.perf_counter() - t_start
+    size_kb = output_path.stat().st_size / 1024
+
+    if not quiet:
+        console.print(
+            f"[success]✓ Done![/success]  → [accent]{output_path}[/accent]"
+            f"  ({len(collected)} files, {size_kb:.1f} KB, {elapsed:.2f}s)"
+        )
+        console.print("\n[dim]Built with 💖 by [link=https://hariharen.site]Hariharen[/link][/dim]\n")
+
+
 # ── formats ───────────────────────────────────────────────────────────────────
 
 @app.command("formats", help="List formats morph can convert, optionally filtered by source format.")

morphconv-1.0.4/morph/converters/archives.py

@@ -0,0 +1,188 @@
+"""
+converters/archives.py — archive family (zip <-> tar <-> 7z, rar extract-only).
+
+Conversion means: extract every member from the source archive, then repack
+them into the destination format, preserving relative paths.
+
+Supported formats and their backing libraries:
+  zip       — stdlib zipfile        (read + write)
+  tar.*     — stdlib tarfile        (read + write)
+  7z        — py7zr (optional)      (read + write)
+  rar       — rarfile (optional)    (read only — RAR creation is proprietary)
+
+py7zr and rarfile are optional; a clear error with install hint is raised if
+the user tries a conversion that needs one and it isn't installed.
+"""
+
+from __future__ import annotations
+
+import tarfile
+import tempfile
+import zipfile
+from pathlib import Path
+
+from ..registry import ConversionResult, OptionSpec, register
+
+# ── format tables ──────────────────────────────────────────────────────────
+
+# tarfile compression mode suffix (empty string = uncompressed tar)
+_TAR_MODES = {"tar": "", "tar.gz": "gz", "tar.bz2": "bz2", "tar.xz": "xz"}
+
+# Formats that can be both read and written (appear as src AND dst)
+RW_FORMATS = ["zip", *_TAR_MODES.keys(), "7z"]
+
+# Formats that can only be read (appear as src only — never a pack target)
+RO_FORMATS = ["rar"]
+
+ALL_SRC_FORMATS = RW_FORMATS + RO_FORMATS
+
+OPTIONS = [
+    OptionSpec("strip_top_level", ("--strip-top-level",),
+               "Drop a single shared top-level folder when repacking, if every member has one.",
+               default=False, action="store_true"),
+]
+
+
+# ── optional library helpers ───────────────────────────────────────────────
+
+def _require_py7zr():
+    try:
+        import py7zr
+        return py7zr
+    except ImportError:
+        raise RuntimeError(
+            "py7zr is required for 7z archive support.\n"
+            "Install it with: pip install py7zr"
+        )
+
+
+def _require_rarfile():
+    try:
+        import rarfile
+        return rarfile
+    except ImportError:
+        raise RuntimeError(
+            "rarfile is required for RAR archive extraction.\n"
+            "Install it with: pip install rarfile\n"
+            "Note: unrar or bsdtar must also be on your PATH."
+        )
+
+
+# ── extract ───────────────────────────────────────────────────────────────
+
+def _extract(input_path: Path, dst_dir: Path, fmt: str) -> None:
+    if fmt == "zip":
+        with zipfile.ZipFile(input_path) as zf:
+            zf.extractall(dst_dir)
+    elif fmt in _TAR_MODES:
+        mode = f"r:{_TAR_MODES[fmt]}" if _TAR_MODES[fmt] else "r:"
+        with tarfile.open(input_path, mode) as tf:
+            tf.extractall(dst_dir, filter="data")
+    elif fmt == "7z":
+        py7zr = _require_py7zr()
+        with py7zr.SevenZipFile(input_path, mode="r") as sz:
+            sz.extractall(path=dst_dir)
+    elif fmt == "rar":
+        rarfile = _require_rarfile()
+        with rarfile.RarFile(input_path) as rf:
+            rf.extractall(path=dst_dir)
+    else:
+        raise ValueError(f"Unsupported source archive format: {fmt!r}")
+
+
+# ── strip top-level ───────────────────────────────────────────────────────
+
+def _maybe_strip_top_level(root: Path) -> Path:
+    entries = list(root.iterdir())
+    if len(entries) == 1 and entries[0].is_dir():
+        return entries[0]
+    return root
+
+
+# ── pack ──────────────────────────────────────────────────────────────────
+
+def _pack(src_dir: Path, output_path: Path, fmt: str) -> int:
+    count = 0
+    if fmt == "zip":
+        with zipfile.ZipFile(output_path, "w", zipfile.ZIP_DEFLATED) as zf:
+            for f in src_dir.rglob("*"):
+                if f.is_file():
+                    zf.write(f, f.relative_to(src_dir))
+                    count += 1
+    elif fmt in _TAR_MODES:
+        mode = f"w:{_TAR_MODES[fmt]}" if _TAR_MODES[fmt] else "w"
+        with tarfile.open(output_path, mode) as tf:
+            for f in src_dir.rglob("*"):
+                if f.is_file():
+                    tf.add(f, f.relative_to(src_dir))
+                    count += 1
+    elif fmt == "7z":
+        py7zr = _require_py7zr()
+        with py7zr.SevenZipFile(output_path, mode="w") as sz:
+            for f in src_dir.rglob("*"):
+                if f.is_file():
+                    sz.write(f, f.relative_to(src_dir))
+                    count += 1
+    else:
+        raise ValueError(f"Unsupported destination archive format: {fmt!r}")
+    return count
+
+
+# ── converter factory ─────────────────────────────────────────────────────
+
+def _make_converter(src_fmt: str, dst_fmt: str):
+    def _convert(
+        input_path: Path,
+        output_path: Path,
+        *,
+        strip_top_level: bool = False,
+        **_options,
+    ) -> ConversionResult:
+        with tempfile.TemporaryDirectory() as tmp:
+            tmp_path = Path(tmp)
+            _extract(input_path, tmp_path, src_fmt)
+            pack_root = _maybe_strip_top_level(tmp_path) if strip_top_level else tmp_path
+            output_path.parent.mkdir(parents=True, exist_ok=True)
+            n_files = _pack(pack_root, output_path, dst_fmt)
+        return ConversionResult(output=output_path, extra={"files": n_files})
+    _convert.__name__ = f"convert_{src_fmt}_to_{dst_fmt}"
+    return _convert
+
+
+# ── registration ──────────────────────────────────────────────────────────
+
+# Backend label per source format
+_SRC_BACKEND = {
+    "zip":     "stdlib",
+    "tar":     "stdlib",
+    "tar.gz":  "stdlib",
+    "tar.bz2": "stdlib",
+    "tar.xz":  "stdlib",
+    "7z":      "py7zr",
+    "rar":     "rarfile",
+}
+
+# Backend label per destination format
+_DST_BACKEND = {
+    "zip":     "stdlib",
+    "tar":     "stdlib",
+    "tar.gz":  "stdlib",
+    "tar.bz2": "stdlib",
+    "tar.xz":  "stdlib",
+    "7z":      "py7zr",
+}
+
+for _src in ALL_SRC_FORMATS:
+    for _dst in RW_FORMATS:          # RAR is extract-only; never a destination
+        if _dst == _src:
+            continue
+        _sb = _SRC_BACKEND[_src]
+        _db = _DST_BACKEND[_dst]
+        _backend = f"{_sb}+{_db}" if _sb != _db else _sb
+        register(
+            _src, _dst,
+            backend=_backend,
+            family="archive",
+            description=f"{_src} -> {_dst}",
+            options=OPTIONS,
+        )(_make_converter(_src, _dst))

{morphconv-1.0.3 → morphconv-1.0.4}/morph/converters/images.py

@@ -119,3 +119,72 @@ for _src in FORMATS:
             lossy=(_dst_pil in _QUALITY_FORMATS or _dst == "ico"),
             options=_OPTIONS_BY_DST.get(_dst, _DEFAULT_OPTIONS),
         )(_make_converter(_dst_pil))
+
+
+# ── img2pdf: lossless image → PDF (overwrites Pillow's re-encoding routes) ──
+#
+# Pillow's PDF writer transcodes — a JPEG saved to PDF becomes a re-encoded
+# JPEG, losing quality. img2pdf embeds the original image bytes directly,
+# producing smaller files with zero quality loss. Supported source types:
+#   JPEG/JPG  — embedded as-is (best case)
+#   PNG       — embedded as-is if no alpha; RGB-flattened then embedded
+#   TIFF      — embedded as-is
+#   WEBP/BMP/GIF/etc. — converted to PNG internally by img2pdf, then embedded
+#
+# img2pdf does NOT support alpha channels, so we pre-flatten RGBA sources.
+
+def _require_img2pdf():
+    try:
+        import img2pdf
+        return img2pdf
+    except ImportError:
+        raise RuntimeError(
+            "img2pdf is required for lossless image-to-PDF conversion.\n"
+            "Install it with: pip install img2pdf"
+        )
+
+
+def _img_to_pdf(input_path: Path, output_path: Path, **_options) -> ConversionResult:
+    """Embed image bytes directly into PDF without transcoding."""
+    img2pdf = _require_img2pdf()
+
+    output_path.parent.mkdir(parents=True, exist_ok=True)
+
+    # Check for alpha channel — img2pdf requires opaque images
+    with Image.open(input_path) as img:
+        has_alpha = img.mode in ("RGBA", "LA", "P") and "transparency" in img.info or img.mode in ("RGBA", "LA")
+
+    if has_alpha:
+        # Flatten alpha onto white background, save to temp PNG, then embed
+        import tempfile
+        with Image.open(input_path) as img:
+            background = Image.new("RGB", img.size, (255, 255, 255))
+            if img.mode in ("RGBA", "LA"):
+                background.paste(img, mask=img.split()[-1])
+            else:
+                background.paste(img.convert("RGBA"), mask=img.convert("RGBA").split()[-1])
+            with tempfile.NamedTemporaryFile(suffix=".png", delete=False) as tmp:
+                tmp_path = Path(tmp.name)
+                background.save(tmp_path, format="PNG")
+        try:
+            pdf_bytes = img2pdf.convert(str(tmp_path))
+        finally:
+            tmp_path.unlink(missing_ok=True)
+    else:
+        pdf_bytes = img2pdf.convert(str(input_path))
+
+    output_path.write_bytes(pdf_bytes)
+    return ConversionResult(output=output_path)
+
+
+# Formats img2pdf handles (all raster sources except pdf itself)
+_IMG2PDF_SRCS = [s for s in FORMATS if s != "pdf"]
+
+for _src in _IMG2PDF_SRCS:
+    register(
+        _src, "pdf",
+        backend="img2pdf",
+        family="image",
+        description=f"{_src} -> pdf (lossless, img2pdf)",
+        lossy=False,  # img2pdf embeds original bytes — truly lossless
+    )(_img_to_pdf)