v0.5.2

Что то сделал

word2md/pipeline.py

@@ -0,0 +1,110 @@
+"""DOCX → md2gost Markdown pipeline."""
+
+from __future__ import annotations
+
+import traceback
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Callable
+
+from docx import Document
+
+from .emit import emit_markdown
+from .media import MediaExtractor
+from .refs import apply_reference_pass
+from .walker import DocxWalker, postprocess_blocks
+
+LogFn = Callable[[str], None]
+
+
+@dataclass
+class ImportRequest:
+    filename: str = ""
+    output: str | None = None
+    media_dir: str | None = None
+    keep_toc_pages: bool = False
+    emit_page_breaks: bool = False
+
+
+@dataclass
+class ImportResult:
+    ok: bool
+    exit_code: int = 0
+    output_path: str | None = None
+    media_dir: str | None = None
+    message: str = ""
+
+
+def default_output_path(docx_path: str | Path) -> Path:
+    p = Path(docx_path)
+    return p.with_suffix(".md")
+
+
+def default_media_dir(md_path: Path) -> Path:
+    return md_path.parent / f"{md_path.stem}_media"
+
+
+def convert_docx(req: ImportRequest, log: LogFn | None = None) -> ImportResult:
+    """Convert DOCX to md2gost dialect Markdown. Errors go into ImportResult."""
+    emit: LogFn = log if log is not None else print
+
+    def fail(code: int, message: str) -> ImportResult:
+        emit(message)
+        return ImportResult(False, code, message=message)
+
+    filename = (req.filename or "").strip()
+    if not filename:
+        return fail(2, "Укажите исходный .docx")
+    path = Path(filename)
+    if not path.is_file():
+        return fail(2, f"Файл не найден: {path}")
+    if path.suffix.lower() != ".docx":
+        return fail(1, "Исходный файл должен быть в формате .docx")
+
+    out = Path(req.output) if req.output else default_output_path(path)
+    if out.suffix.lower() != ".md":
+        out = out.with_suffix(".md")
+    out.parent.mkdir(parents=True, exist_ok=True)
+
+    media_path = Path(req.media_dir) if req.media_dir else default_media_dir(out)
+    rel_prefix = media_path.name
+    # If media dir is not sibling of md, use relative path from md parent
+    try:
+        rel_prefix = str(media_path.resolve().relative_to(out.parent.resolve())).replace("\\", "/")
+    except ValueError:
+        rel_prefix = str(media_path).replace("\\", "/")
+
+    try:
+        document = Document(str(path))
+    except Exception as exc:
+        return fail(1, f"Не удалось открыть DOCX: {exc}")
+
+    try:
+        with MediaExtractor(path, media_path, rel_prefix=rel_prefix) as media:
+            walker = DocxWalker(
+                document,
+                media,
+                keep_toc_pages=req.keep_toc_pages,
+                emit_page_breaks=req.emit_page_breaks,
+                log=emit,
+            )
+            blocks = walker.walk()
+            blocks = postprocess_blocks(blocks)
+            blocks = apply_reference_pass(blocks)
+            md_text = emit_markdown(blocks, emit_page_breaks=req.emit_page_breaks)
+        out.write_text(md_text, encoding="utf-8")
+    except Exception as exc:
+        emit(traceback.format_exc())
+        return fail(1, f"Ошибка импорта: {exc}")
+
+    msg = f"Markdown: {out}"
+    emit(msg)
+    if media_path.is_dir() and any(media_path.iterdir()):
+        emit(f"Медиа: {media_path}")
+    return ImportResult(
+        True,
+        0,
+        output_path=str(out.resolve()),
+        media_dir=str(media_path.resolve()) if media_path.exists() else None,
+        message=msg,
+    )