initial MVP

This commit is contained in:
Mikan
2025-12-08 16:57:25 +03:00
commit 848eb462b5
20 changed files with 398 additions and 0 deletions

0
src/core/__init__.py Normal file
View File

30
src/core/converter.py Normal file
View File

@@ -0,0 +1,30 @@
from pathlib import Path
from docx import Document
from .style_registry import StyleRegistry
from .renderer import render_markdown_to_docx
def process_document(input_dir: Path, output_dir: Path, style_config_path: Path):
style_reg = StyleRegistry(style_config_path)
for doc_dir in input_dir.iterdir():
if not doc_dir.is_dir():
continue
doc_name = doc_dir.name
md_path = doc_dir / "document.md"
images_dir = doc_dir / "images"
if not md_path.exists():
print(f"⚠️ Пропущено: {doc_name} — нет document.md")
continue
with open(md_path, encoding="utf-8") as f:
md_content = f.read()
doc = Document()
style_reg.ensure_styles_in_doc(doc) # добавляем стили
render_markdown_to_docx(doc, md_content, images_dir, style_reg)
output_path = output_dir / f"{doc_name}.docx"
output_path.parent.mkdir(parents=True, exist_ok=True)
doc.save(output_path)
print(f"{doc_name}{output_path}")

17
src/core/renderer.py Normal file
View File

@@ -0,0 +1,17 @@
import mistune
from .style_registry import StyleRegistry
from ..handlers import get_handler
def render_markdown_to_docx(doc, md_content: str, images_dir, style_reg: StyleRegistry):
markdown = mistune.create_markdown(renderer=None)
ast = markdown(md_content)
def walk(node):
handler = get_handler(node["type"])
if handler:
handler(node, doc, images_dir, style_reg)
elif node.get("children"):
for child in node["children"]:
walk(child)
walk(ast)

View File

@@ -0,0 +1,56 @@
import yaml
from docx import Document
from docx.enum.style import WD_STYLE_TYPE
from docx.shared import Pt, RGBColor
from docx.enum.text import WD_LINE_SPACING, WD_ALIGN_PARAGRAPH
from docx.styles.style import ParagraphStyle
class StyleRegistry:
def __init__(self, config_path):
with open(config_path, encoding="utf-8") as f:
self.config = yaml.safe_load(f)
def ensure_styles_in_doc(self, doc: Document):
"""Создаёт недостающие стили в документе, если их нет."""
for name, cfg in self.config["styles"].items():
try:
doc.styles[name]
except KeyError:
# Создаём стиль
style = doc.styles.add_style(name, WD_STYLE_TYPE.PARAGRAPH)
self.apply_style(style, name)
def apply_style(self, style, style_name: str):
cfg = self.config["styles"].get(style_name)
if not cfg:
return
font_cfg = cfg.get("font", {})
para_cfg = cfg.get("paragraph", {})
if font_cfg:
font = style.font
font.name = font_cfg.get("name", font.name)
if "size" in font_cfg:
font.size = Pt(font_cfg["size"])
font.bold = font_cfg.get("bold", font.bold)
font.italic = font_cfg.get("italic", font.italic)
if para_cfg:
pf = style.paragraph_format
if "space_before" in para_cfg:
pf.space_before = Pt(para_cfg["space_before"])
if "space_after" in para_cfg:
pf.space_after = Pt(para_cfg["space_after"])
if "line_spacing" in para_cfg:
pf.line_spacing = para_cfg["line_spacing"]
pf.line_spacing_rule = WD_LINE_SPACING.MULTIPLE
if "alignment" in para_cfg:
alignment_map = {"center": WD_ALIGN_PARAGRAPH.CENTER}
align = alignment_map.get(para_cfg["alignment"])
if align:
pf.alignment = align
def get_color(self, key: str):
hex_color = self.config.get("colors", {}).get(key, "000000")
return RGBColor.from_string(hex_color)