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
164 lines
5.0 KiB
Python
164 lines
5.0 KiB
Python
"""
|
||
图片处理模块。
|
||
|
||
将 Markdown 中的 ``<img>`` 标签解析为独立段落，
|
||
并在 Word 文档中插入图片及居中图标题。
|
||
"""
|
||
|
||
import struct
|
||
import re
|
||
from pathlib import Path
|
||
|
||
from docx import Document
|
||
from docx.enum.text import WD_ALIGN_PARAGRAPH
|
||
|
||
# 匹配 <img src="..." alt="...">
|
||
_IMG_TAG = re.compile(
|
||
r'<img\s+([^>]*?)src=["\']([^"\']+)["\']([^>]*?)>', re.IGNORECASE
|
||
)
|
||
|
||
# 从标签属性块中提取名值对
|
||
_ATTR = re.compile(r'(\w+)\s*=\s*["\']([^"\']*)["\']')
|
||
|
||
# 匹配 **图X ...** 重复图标题（注入时跳过）
|
||
_FIG_CAPTION = re.compile(
|
||
r'^\*\*(?:图|Table|Figure|Fig\.?)\s*\d+.*\*\*$', re.IGNORECASE
|
||
)
|
||
|
||
|
||
def make_image_paragraph(block: str, base_dir: str | Path | None = None) -> dict | None:
|
||
"""若 *block* 包含 ``<img>`` 标签，返回图片段落字典；否则返回 ``None``。
|
||
|
||
Parameters
|
||
----------
|
||
block : str
|
||
文本块。
|
||
base_dir : str | Path | None
|
||
Markdown 文件所在目录，用于解析图片相对路径。
|
||
"""
|
||
m = _IMG_TAG.search(block)
|
||
if not m:
|
||
return None
|
||
attrs = dict(_ATTR.findall(block))
|
||
src = attrs.get("src", m.group(2))
|
||
if base_dir and not Path(src).is_absolute():
|
||
src = str(Path(base_dir) / src)
|
||
return {
|
||
"type": "image",
|
||
"src": src,
|
||
"alt": attrs.get("alt", ""),
|
||
}
|
||
|
||
|
||
def is_figure_caption(block: str) -> bool:
|
||
"""检查 *block* 是否为 ``**图X ...**`` 格式的重复图标题。"""
|
||
return bool(_FIG_CAPTION.match(block.strip()))
|
||
|
||
|
||
def _get_image_dimensions(image_path: str) -> tuple[int, int] | None:
|
||
"""读取图片文件头返回 ``(width_px, height_px)``，不支持格式返回 ``None``。
|
||
|
||
仅读取文件头，不依赖第三方库。支持 PNG / JPEG / GIF / BMP。
|
||
"""
|
||
try:
|
||
with open(image_path, "rb") as f:
|
||
header = f.read(32)
|
||
except Exception:
|
||
return None
|
||
|
||
# PNG: 8-byte signature, then IHDR chunk
|
||
if header[:8] == b"\x89PNG\r\n\x1a\n":
|
||
w, h = struct.unpack_from(">II", header, 16)
|
||
return w, h
|
||
|
||
# JPEG: starts with FF D8, scan for SOF marker
|
||
if header[:2] == b"\xff\xd8":
|
||
pos = 2
|
||
while pos < len(header):
|
||
if header[pos] != 0xFF:
|
||
return None
|
||
marker = header[pos + 1]
|
||
if marker in (0xC0, 0xC1, 0xC2):
|
||
h, w = struct.unpack_from(">HH", header, pos + 5)
|
||
return w, h
|
||
seg_len = struct.unpack_from(">H", header, pos + 2)[0]
|
||
pos += 2 + seg_len
|
||
return None
|
||
|
||
# GIF: "GIF87a" or "GIF89a"
|
||
if header[:6] in (b"GIF87a", b"GIF89a"):
|
||
w, h = struct.unpack_from("<HH", header, 6)
|
||
return w, h
|
||
|
||
# BMP: "BM" signature
|
||
if header[:2] == b"BM":
|
||
w, h = struct.unpack_from("<ii", header, 18)
|
||
return w, abs(h)
|
||
|
||
return None
|
||
|
||
|
||
def _get_native_emu(image_path: str) -> int | None:
|
||
"""读取图片的原生宽度（EMU），失败返回 ``None``。
|
||
|
||
Word 默认以 72 DPI 渲染图片，1 px = 914400 / 72 = 12700 EMU。
|
||
"""
|
||
dims = _get_image_dimensions(image_path)
|
||
if dims is None:
|
||
return None
|
||
w_px, _ = dims
|
||
return w_px * 12700
|
||
|
||
|
||
def _constrain_width(image_path: str, page_text_width: int) -> int | None:
|
||
"""返回图片宽度（EMU），超出页宽时缩至页宽。
|
||
|
||
Parameters
|
||
----------
|
||
image_path : str
|
||
图片路径。
|
||
page_text_width : int
|
||
页面正文区宽度（EMU），来自 ``section.page_width - margins``。
|
||
"""
|
||
native = _get_native_emu(image_path)
|
||
if native is None:
|
||
return None
|
||
return min(native, page_text_width)
|
||
|
||
|
||
def insert_image_paragraphs(
|
||
doc: Document,
|
||
paragraphs: list[dict],
|
||
*,
|
||
idx: int,
|
||
parent,
|
||
):
|
||
"""在 *doc* 的指定位置插入图片段落序列。
|
||
|
||
每条图片段落生成两个 Word 段落：
|
||
1. 居中图片（超出页宽时自动缩放至页宽，否则保持原尺寸）
|
||
2. 居中图标题（从 ``alt`` 提取）
|
||
|
||
插入顺序保持 ``paragraphs`` 的原有顺序。
|
||
"""
|
||
section = doc.sections[0]
|
||
page_text_width = section.page_width - section.left_margin - section.right_margin
|
||
|
||
for pd_data in reversed(paragraphs):
|
||
# 图标题（在 reversed 中先插入，最终位于图片下方）
|
||
cap_p = doc.add_paragraph(pd_data.get("alt", ""))
|
||
cap_p.alignment = WD_ALIGN_PARAGRAPH.CENTER
|
||
parent.insert(idx, cap_p._element)
|
||
|
||
# 图片段落（max-width 行为：超出页宽时压缩，否则原尺寸）
|
||
img_p = doc.add_paragraph()
|
||
img_p.alignment = WD_ALIGN_PARAGRAPH.CENTER
|
||
img_run = img_p.add_run()
|
||
try:
|
||
img_width = _constrain_width(pd_data["src"], page_text_width)
|
||
img_run.add_picture(pd_data["src"], width=img_width)
|
||
except Exception as exc:
|
||
print(f"警告：图片加载失败 {pd_data['src']} — {exc}")
|
||
img_run.add_text(f"[图片加载失败: {pd_data['src']}]")
|
||
parent.insert(idx, img_p._element)
|