v0.5.6

md2gost/office/uno_cli.py

@@ -0,0 +1,125 @@
+"""Headless soffice CLI helpers (convert/import) for UNO post-process."""
+
+from __future__ import annotations
+
+import os
+import subprocess
+import sys
+from pathlib import Path
+
+from .uno_helpers import file_url_from_path
+from .uno_profile import ensure_profile_seeded, profile_dir
+
+
+def soffice_cli_path(soffice_path: str) -> str:
+    """On Windows prefer ``soffice.com`` — ``soffice.exe`` opens the Start Center GUI."""
+    path = Path(soffice_path)
+    if sys.platform == "win32":
+        com = path.with_name("soffice.com")
+        if com.is_file():
+            return str(com)
+    return str(path)
+
+
+def _base_args(soffice_path: str, profile_root: Path) -> list[str]:
+    ensure_profile_seeded(profile_root)
+    return [
+        soffice_cli_path(soffice_path),
+        f"-env:UserInstallation={file_url_from_path(str(profile_root))}",
+        "--headless",
+        "--nologo",
+        "--norestore",
+        "--nodefault",
+        "--nofirststartwizard",
+        "--nolockcheck",
+    ]
+
+
+def convert_file(
+    soffice_path: str,
+    *,
+    engine: str,
+    src: str,
+    out_dir: str,
+    convert_to: str,
+    timeout_s: float = 300.0,
+) -> None:
+    """Run ``soffice --convert-to …`` and wait until it finishes."""
+    profile_root = profile_dir(engine)
+    cmd = [
+        *_base_args(soffice_path, profile_root),
+        "--convert-to",
+        convert_to,
+        "--outdir",
+        os.path.abspath(out_dir),
+        os.path.abspath(src),
+    ]
+    completed = subprocess.run(
+        cmd,
+        capture_output=True,
+        text=True,
+        timeout=timeout_s,
+        cwd=str(Path(soffice_path).parent),
+        creationflags=getattr(subprocess, "CREATE_NO_WINDOW", 0) if sys.platform == "win32" else 0,
+    )
+    if completed.returncode != 0:
+        tail = (completed.stderr or completed.stdout or "").strip()
+        if len(tail) > 500:
+            tail = tail[-500:]
+        raise RuntimeError(
+            f"convert-to {convert_to} failed (code {completed.returncode})"
+            + (f": {tail}" if tail else "")
+        )
+    stem = Path(src).stem
+    expected = Path(out_dir) / f"{stem}.docx"
+    if convert_to.startswith("docx") and not expected.is_file():
+        found = sorted(Path(out_dir).glob("*.docx"))
+        if len(found) == 1:
+            return
+        raise RuntimeError(
+            f"convert-to {convert_to}: output missing ({expected.name})"
+            + (f", got {[p.name for p in found]}" if found else "")
+        )
+
+
+def import_docx_to_odt(soffice_path: str, *, engine: str, docx_path: str, out_dir: str) -> Path:
+    """OpenOffice imports DOCX more reliably via native ODT round-trip."""
+    for spec in ("odt", 'odt:"writer8"'):
+        try:
+            convert_file(
+                soffice_path,
+                engine=engine,
+                src=docx_path,
+                out_dir=out_dir,
+                convert_to=spec,
+            )
+            odt = Path(out_dir) / f"{Path(docx_path).stem}.odt"
+            if odt.is_file():
+                return odt
+        except RuntimeError:
+            continue
+    raise RuntimeError(f"OpenOffice: не удалось импортировать DOCX → ODT: {docx_path}")
+
+
+def export_odt_to_docx(soffice_path: str, *, engine: str, odt_path: str, out_dir: str) -> Path:
+    # OpenOffice on Windows treats spaces in ``docx:MS Word 2007 XML`` as extra
+    # file arguments (Word / 2007 / XML). Use plain ``docx`` there.
+    if engine == "openoffice":
+        specs = ("docx",)
+    else:
+        specs = ("docx", "MS Word 2007 XML")
+    for spec in specs:
+        try:
+            convert_file(
+                soffice_path,
+                engine=engine,
+                src=odt_path,
+                out_dir=out_dir,
+                convert_to=spec,
+            )
+            docx = Path(out_dir) / f"{Path(odt_path).stem}.docx"
+            if docx.is_file():
+                return docx
+        except RuntimeError:
+            continue
+    raise RuntimeError(f"OpenOffice: не удалось экспортировать ODT → DOCX: {odt_path}")