fix: protect editor content and harden document exports

Protect pending edits on window close, keep save-as paths and drafts consistent,
copy relative image resources, reject oversized table imports without truncation,
and preserve independent preview scrolling during refresh.

Archive rendered MHTML resources with portable references and embedded local
fonts. Preserve existing outputs on export cancellation or failure. Convert
editor match positions and filter directory results in linear passes.

Include the previously completed editor image/table/history improvements,
undoable replacement, export snapshots, shared page settings, and cancellable
PDF/Word export workflow. Update documentation and version to 0.3.1.

Validation: 142 pytest cases passed; GUI smoke, improvements integration, and
offline MHTML/scroll reliability smoke checks passed.
This commit is contained in:
2026-10-29 18:22:42 +08:00
parent fac5cbb011
commit 8dae91c2ca
26 changed files with 2722 additions and 385 deletions
+152
View File
@@ -0,0 +1,152 @@
from __future__ import annotations
from dataclasses import asdict, dataclass, fields
PAPER_MM = {"A4": (210.0, 297.0), "A3": (297.0, 420.0), "Letter": (215.9, 279.4)}
@dataclass(frozen=True)
class PageSettings:
paper: str = "A4"
landscape: bool = False
top: float = 15.0
bottom: float = 15.0
left: float = 15.0
right: float = 15.0
include_toc: bool = True
toc_depth: int = 3
header: str = ""
footer: str = ""
page_numbers: bool = True
@classmethod
def from_dict(cls, data: dict) -> "PageSettings":
known = {field.name for field in fields(cls)}
clean = {key: value for key, value in data.items() if key in known}
defaults = cls()
for name in ("landscape", "include_toc", "page_numbers"):
clean[name] = clean.get(name, getattr(defaults, name)) in (True, 1, "true")
for name in ("top", "bottom", "left", "right"):
try:
clean[name] = max(0.0, min(50.0, float(clean.get(name, 15))))
except (ValueError, TypeError):
clean[name] = 15.0
clean["paper"] = clean.get("paper") if clean.get("paper") in PAPER_MM else "A4"
try:
clean["toc_depth"] = max(1, min(6, int(clean.get("toc_depth", 3))))
except (ValueError, TypeError):
clean["toc_depth"] = 3
for name in ("header", "footer"):
clean[name] = str(clean.get(name, "")).replace("\n", " ")[:100]
return cls(**clean)
def normalized(self) -> "PageSettings":
result = asdict(self)
if self.header:
result["top"] = max(12.0, self.top)
if self.footer or self.page_numbers:
result["bottom"] = max(12.0, self.bottom)
return self.from_dict(result)
def size_mm(self) -> tuple[float, float]:
width, height = PAPER_MM.get(self.paper, PAPER_MM["A4"])
return (height, width) if self.landscape else (width, height)
def apply_docx_page_settings(path: object, settings: PageSettings) -> None:
from docx import Document
from docx.enum.text import WD_BREAK, WD_ALIGN_PARAGRAPH
from docx.enum.section import WD_ORIENT
from docx.oxml import OxmlElement
from docx.oxml.ns import qn
from docx.shared import Mm, Pt
settings = settings.normalized()
document = Document(str(path))
width, height = settings.size_mm()
for section in document.sections:
section.orientation = WD_ORIENT.LANDSCAPE if settings.landscape else WD_ORIENT.PORTRAIT
section.page_width, section.page_height = Mm(width), Mm(height)
for name in ("top", "bottom", "left", "right"):
setattr(section, name + "_margin", Mm(getattr(settings, name)))
section.header_distance = Mm(5)
section.footer_distance = Mm(5)
for container, text in ((section.header, settings.header), (section.footer, settings.footer)):
paragraph = container.paragraphs[0]
paragraph.clear()
paragraph.alignment = WD_ALIGN_PARAGRAPH.CENTER
run = paragraph.add_run(text)
run.font.size = Pt(9)
if settings.page_numbers:
paragraph = section.footer.paragraphs[0]
paragraph.add_run(" " if settings.footer else "")
paragraph.add_run("第 ")
for instruction, following in [("PAGE", " 页 / 共 "), ("NUMPAGES", " 页")]:
field = OxmlElement("w:fldSimple")
field.set(qn("w:instr"), instruction)
paragraph._p.append(field)
paragraph.add_run(following)
available_width = Mm(width - settings.left - settings.right)
available_height = Mm(height - settings.top - settings.bottom - 5)
for shape in document.inline_shapes:
if shape.width and shape.height:
scale = min(1.0, available_width / shape.width, available_height / shape.height)
if scale < 1:
shape.width = round(shape.width * scale)
shape.height = round(shape.height * scale)
if settings.include_toc:
# Word calculates page numbers using its own layout engine when fields update.
body = document.element.body
heading = OxmlElement("w:p")
r = OxmlElement("w:r"); t = OxmlElement("w:t"); t.text = "目录"; r.append(t); heading.append(r)
body.insert(0, heading)
paragraph = OxmlElement("w:p")
field = OxmlElement("w:fldSimple")
field.set(qn("w:instr"), f'TOC \\o "1-{settings.toc_depth}" \\h \\z \\u')
field.set(qn("w:dirty"), "true")
run = OxmlElement("w:r"); text = OxmlElement("w:t")
text.text = "请在 Word 中右键此处,选择“更新域”生成目录和页码。"
run.append(text); field.append(run); paragraph.append(field); body.insert(1, paragraph)
page_break = OxmlElement("w:p"); run = OxmlElement("w:r"); br = OxmlElement("w:br")
br.set(qn("w:type"), "page"); run.append(br); page_break.append(run); body.insert(2, page_break)
update = document.settings.element.find(qn("w:updateFields"))
if update is None:
update = OxmlElement("w:updateFields"); document.settings.element.append(update)
update.set(qn("w:val"), "true")
document.save(str(path))
def decorate_pdf(data: bytes, settings: PageSettings, cancel: object = None) -> bytes:
if not (settings.header or settings.footer or settings.page_numbers):
return data
from io import BytesIO
from pypdf import PdfReader, PdfWriter
from PySide6.QtCore import QBuffer, QIODevice, QMarginsF, QRectF, QSizeF, Qt
from PySide6.QtGui import QColor, QFont, QPageLayout, QPageSize, QPainter, QPdfWriter
from .export_jobs import check_cancelled
reader, writer = PdfReader(BytesIO(data)), PdfWriter()
writer.clone_document_from_reader(reader)
for index, page in enumerate(writer.pages):
check_cancelled(cancel)
width, height = float(page.mediabox.width), float(page.mediabox.height)
buffer = QBuffer(); buffer.open(QIODevice.OpenModeFlag.ReadWrite)
pdf = QPdfWriter(buffer); pdf.setResolution(72)
pdf.setPageLayout(QPageLayout(QPageSize(QSizeF(width, height), QPageSize.Unit.Point),
QPageLayout.Orientation.Portrait, QMarginsF(0, 0, 0, 0)))
painter = QPainter(pdf); painter.setPen(QColor("#333333")); painter.setFont(QFont("Microsoft YaHei", 9))
left = settings.left * 72 / 25.4
right = width - settings.right * 72 / 25.4
header_box = QRectF(left, 4 * 72 / 25.4, right - left, 16)
footer_box = QRectF(left, height - 10 * 72 / 25.4, right - left, 16)
painter.drawText(header_box, Qt.AlignmentFlag.AlignCenter, settings.header)
if settings.page_numbers:
painter.drawText(footer_box, Qt.AlignmentFlag.AlignRight, f"第 {index + 1} 页 / 共 {len(writer.pages)} 页")
footer_box.setWidth(max(1, footer_box.width() - 150))
painter.drawText(footer_box, Qt.AlignmentFlag.AlignLeft if settings.page_numbers else Qt.AlignmentFlag.AlignCenter, settings.footer)
painter.end()
del painter
del pdf
page.merge_page(PdfReader(BytesIO(bytes(buffer.data()))).pages[0])
check_cancelled(cancel)
output = BytesIO(); writer.write(output)
return output.getvalue()