zzy
4e39a4f2ac
- 新增 CITE_PATTERN 正则表达式匹配 [N] 引用格式 - 添加 base_dir 参数支持相对图片路径解析 - 实现书签创建和超链接替换功能 - 添加 link_body_citations 函数处理正文引用链接 - 在参考文献段落中添加书签标识 - 支持将 [N] 引用替换为指向参考文献的超链接
277 lines
8.7 KiB
Python
277 lines
8.7 KiB
Python
"""
|
||
Markdown 正文 → Word 段落转换。
|
||
|
||
将正文 Markdown 按标题层级拆分为带样式的段落序列，
|
||
再注入到渲染后 docx 文档的占位符位置。
|
||
"""
|
||
|
||
import re
|
||
from copy import deepcopy
|
||
from pathlib import Path
|
||
|
||
from docx import Document
|
||
from docx.oxml.ns import qn
|
||
|
||
from .images import make_image_paragraph, is_figure_caption, insert_image_paragraphs
|
||
|
||
_PAT_HEADING = re.compile(r"^(#{1,6})\s+(.+)$", re.MULTILINE)
|
||
|
||
# 匹配正文中的引用标记 [1] / [1,2,3]
|
||
_CITE_PATTERN = re.compile(r"\[(\d+(?:[,，\s]*\d+)*)\]")
|
||
|
||
|
||
def body_to_paragraphs(
|
||
md_text: str,
|
||
*,
|
||
level_offset: int = 0,
|
||
body_style: str = "Body Text Indent",
|
||
base_dir: str | Path | None = None,
|
||
) -> list[dict]:
|
||
"""将 Markdown 正文按标题和段落拆分为结构化列表。
|
||
|
||
Parameters
|
||
----------
|
||
md_text : str
|
||
正文 Markdown。
|
||
level_offset : int
|
||
标题级别偏移量（正文从 ``##`` 开始时传 ``-1``，使其输出为 ``Heading 1``）。
|
||
body_style : str
|
||
正文段落的 Word 样式名。
|
||
base_dir : str | Path | None
|
||
Markdown 文件所在目录，用于解析图片相对路径。
|
||
"""
|
||
paragraphs: list[dict] = []
|
||
last_end = 0
|
||
|
||
def _add_block(block: str) -> None:
|
||
block = block.strip()
|
||
if not block:
|
||
return
|
||
# 图片段落
|
||
img = make_image_paragraph(block, base_dir)
|
||
if img:
|
||
paragraphs.append(img)
|
||
return
|
||
# 跳过紧跟在图片后的重复图标题
|
||
if paragraphs and paragraphs[-1].get("type") == "image" and is_figure_caption(block):
|
||
return
|
||
# 普通正文段落
|
||
paragraphs.append({"text": block, "level": 0, "style": body_style})
|
||
|
||
for m in _PAT_HEADING.finditer(md_text):
|
||
# 标题前的普通文本
|
||
if m.start() > last_end:
|
||
pre = md_text[last_end : m.start()].strip()
|
||
if pre:
|
||
for block in re.split(r"\n\s*\n", pre):
|
||
_add_block(block)
|
||
|
||
level = len(m.group(1)) + level_offset
|
||
heading_text = m.group(2).strip()
|
||
paragraphs.append(
|
||
{"text": heading_text, "level": level, "style": f"Heading {level}"}
|
||
)
|
||
last_end = m.end()
|
||
|
||
# 最后一段 / 尾部文本
|
||
tail = md_text[last_end:].strip()
|
||
if tail:
|
||
for block in re.split(r"\n\s*\n", tail):
|
||
_add_block(block)
|
||
|
||
return paragraphs
|
||
|
||
|
||
def replace_placeholder(
|
||
doc: Document,
|
||
placeholder: str,
|
||
paragraphs: list[dict],
|
||
*,
|
||
default_body_style: str | None = None,
|
||
):
|
||
"""在 *doc* 中找到包含 *placeholder* 的段落，替换为 *paragraphs* 列表。
|
||
|
||
正文段落的样式优先级：
|
||
1. ``style`` 字段指定的样式名（来自 ``body_to_paragraphs`` 的 ``body_style``）
|
||
2. 占位符段落自身的样式（模板中已设好的样式）
|
||
3. ``Normal``
|
||
"""
|
||
placeholder_found = False
|
||
for para in doc.paragraphs:
|
||
if placeholder in para.text:
|
||
placeholder_found = True
|
||
placeholder_style = para.style.name if para.style else None
|
||
parent = para._element.getparent()
|
||
idx = list(parent).index(para._element)
|
||
|
||
# 保存原段落的编号属性（numPr），用于继承自动编号
|
||
orig_pPr = para._element.find(qn("w:pPr"))
|
||
numPr = orig_pPr.find(qn("w:numPr")) if orig_pPr is not None else None
|
||
|
||
parent.remove(para._element)
|
||
|
||
# 为参考文献段落准备书签 ID
|
||
bm_id = _max_bookmark_id(doc) + 1
|
||
|
||
for pd_data in reversed(paragraphs):
|
||
if pd_data.get("type") == "image":
|
||
insert_image_paragraphs(
|
||
doc, [pd_data], idx=idx, parent=parent
|
||
)
|
||
else:
|
||
new_p = doc.add_paragraph(pd_data["text"])
|
||
style_name = pd_data["style"]
|
||
|
||
# 尝试应用样式，逐步降级
|
||
applied = _apply_style(new_p, doc, style_name)
|
||
if not applied and style_name.startswith("Heading"):
|
||
new_p.style = doc.styles["Normal"]
|
||
elif not applied:
|
||
if placeholder_style:
|
||
_apply_style(new_p, doc, placeholder_style)
|
||
if new_p.style.name == "Normal" and placeholder_style:
|
||
new_p.style = doc.styles[placeholder_style]
|
||
|
||
# 继承原段落的编号属性（自动编号）
|
||
if numPr is not None:
|
||
new_pPr = new_p._element.find(qn("w:pPr"))
|
||
if new_pPr is None:
|
||
new_pPr = new_p._element.makeelement(qn("w:pPr"), {})
|
||
new_p._element.insert(0, new_pPr)
|
||
existing = new_pPr.find(qn("w:numPr"))
|
||
if existing is not None:
|
||
new_pPr.remove(existing)
|
||
new_pPr.append(deepcopy(numPr))
|
||
|
||
parent.insert(idx, new_p._element)
|
||
|
||
# 为参考文献条目添加书签
|
||
ref_id = pd_data.get("ref_id")
|
||
if ref_id is not None:
|
||
_add_bookmark(new_p, f"ref-{ref_id}", bm_id)
|
||
bm_id += 1
|
||
break
|
||
|
||
if not placeholder_found:
|
||
print(f"警告：未找到占位符 '{placeholder}'，正文段落未注入。")
|
||
|
||
|
||
def link_body_citations(doc: Document):
|
||
"""将文档中正文段落的 ``[N]`` 引用替换为指向对应书签的超链接。"""
|
||
for para in doc.paragraphs:
|
||
# 跳过已有超链接的段落（如目录页）
|
||
if para._element.findall(qn("w:hyperlink")):
|
||
continue
|
||
_link_paragraph(para)
|
||
|
||
|
||
def _max_bookmark_id(doc: Document) -> int:
|
||
"""扫描文档返回最大书签 ID。"""
|
||
max_id = 0
|
||
for para in doc.paragraphs:
|
||
for bm in para._element.iter(qn("w:bookmarkStart")):
|
||
try:
|
||
max_id = max(max_id, int(bm.get(qn("w:id"))))
|
||
except (ValueError, TypeError):
|
||
pass
|
||
return max_id
|
||
|
||
|
||
def _add_bookmark(paragraph, name: str, bm_id: int):
|
||
"""为段落添加书签。"""
|
||
bm_start = paragraph._element.makeelement(qn("w:bookmarkStart"), {})
|
||
bm_start.set(qn("w:id"), str(bm_id))
|
||
bm_start.set(qn("w:name"), name)
|
||
|
||
bm_end = paragraph._element.makeelement(qn("w:bookmarkEnd"), {})
|
||
bm_end.set(qn("w:id"), str(bm_id))
|
||
|
||
pPr = paragraph._element.find(qn("w:pPr"))
|
||
if pPr is not None:
|
||
paragraph._element.insert(1, bm_start)
|
||
else:
|
||
paragraph._element.insert(0, bm_start)
|
||
paragraph._element.append(bm_end)
|
||
|
||
|
||
def _link_paragraph(para):
|
||
"""将单个段落中的 ``[N]`` 替换为 HYPERLINK 域。"""
|
||
runs = list(para._element.findall(qn("w:r")))
|
||
if not runs:
|
||
return
|
||
|
||
full_text = ""
|
||
for r in runs:
|
||
t = r.find(qn("w:t"))
|
||
if t is not None and t.text:
|
||
full_text += t.text
|
||
|
||
matches = list(_CITE_PATTERN.finditer(full_text))
|
||
if not matches:
|
||
return
|
||
|
||
first_rPr = runs[0].find(qn("w:rPr"))
|
||
for r in runs:
|
||
para._element.remove(r)
|
||
|
||
pos = 0
|
||
for m in matches:
|
||
before = full_text[pos : m.start()]
|
||
if before:
|
||
_add_run(para._element, before, first_rPr)
|
||
|
||
nums = re.findall(r"\d+", m.group(1))
|
||
_add_hlink(para._element, f"ref-{nums[0]}", m.group())
|
||
|
||
pos = m.end()
|
||
|
||
after = full_text[pos:]
|
||
if after:
|
||
_add_run(para._element, after, first_rPr)
|
||
|
||
|
||
def _add_run(parent, text: str, rPr):
|
||
r = parent.makeelement(qn("w:r"), {})
|
||
if rPr is not None:
|
||
r.append(deepcopy(rPr))
|
||
t = r.makeelement(qn("w:t"), {})
|
||
t.text = text
|
||
r.append(t)
|
||
parent.append(r)
|
||
|
||
|
||
def _add_hlink(parent, anchor: str, text: str):
|
||
hl = parent.makeelement(qn("w:hyperlink"), {})
|
||
hl.set(qn("w:anchor"), anchor)
|
||
|
||
r = parent.makeelement(qn("w:r"), {})
|
||
rPr = r.makeelement(qn("w:rPr"), {})
|
||
rStyle = rPr.makeelement(qn("w:rStyle"), {})
|
||
rStyle.set(qn("w:val"), "Hyperlink")
|
||
rPr.append(rStyle)
|
||
vertAlign = rPr.makeelement(qn("w:vertAlign"), {})
|
||
vertAlign.set(qn("w:val"), "superscript")
|
||
rPr.append(vertAlign)
|
||
r.append(rPr)
|
||
|
||
t = r.makeelement(qn("w:t"), {})
|
||
t.text = text
|
||
r.append(t)
|
||
hl.append(r)
|
||
parent.append(hl)
|
||
|
||
|
||
def _apply_style(paragraph, doc, style_name: str) -> bool:
|
||
"""尝试给段落应用样式，成功返回 ``True``。"""
|
||
try:
|
||
paragraph.style = doc.styles[style_name]
|
||
return True
|
||
except KeyError:
|
||
pass
|
||
# 大小写不敏感匹配
|
||
for s in doc.styles:
|
||
if s.name.lower() == style_name.lower():
|
||
paragraph.style = s
|
||
return True
|
||
return False
|