yixiong/MorphDoc
4
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 7 Wiki Activity

feat: 完成 DOCX R4 元素级视觉门禁

Browse Source 
建立 Chromium、Word 与 WPS 的可编辑内容、段落几何、页眉页脚、封面、媒体和逐页视觉比较链路。

支持封面独立分节、正文页码重启、主题页边距与横向纸张，并将自然分页差异降为诊断项。

修正代码块内边距和折叠表格单元格边框映射，14 套主题生产矩阵与六组布局视觉矩阵通过。
This commit is contained in:
SkyJourney
2026-08-02 03:27:11 +08:00
parent f9f5fccfc9
commit 58f87cc19f
100 changed files with 10409 additions and 386 deletions
Download Patch File Download Diff File Expand all files Collapse all files
packages/document-visual-diff/tests/report.test.ts
+401
View File
@@ -9,6 +9,8 @@ import {
serializePdfVisualDiffJson,
type PdfDocumentSnapshot,
type PdfPageRaster,
type PdfTextLineSnapshot,
type VisualPageSemanticExpectation,
} from "../src/index.js";
function raster(color: string): PdfPageRaster {
@@ -50,6 +52,57 @@ function snapshot(label: string, pageRaster: PdfPageRaster): PdfDocumentSnapshot
};
}
function contentLine(text: string, y: number): PdfTextLineSnapshot {
return {
text,
normalizedText: text,
bounds: { x: 4, y, width: 8, height: 2 },
baselineY: y + 2,
role: "content",
items: [],
};
}
function flowSnapshot(
label: string,
pageLines: readonly (readonly PdfTextLineSnapshot[])[],
color: string,
): PdfDocumentSnapshot {
return {
schemaVersion: 1,
source: { kind: "custom", label },
sha256: label.padEnd(64, "0").slice(0, 64),
pageCount: pageLines.length,
contentText: pageLines.flat().map((line) => line.normalizedText).join(""),
pages: pageLines.map((lines, index) => ({
pageNumber: index + 1,
widthPt: 20,
heightPt: 25,
rotation: 0,
items: [],
lines: [...lines],
contentText: lines.map((line) => line.normalizedText).join(""),
raster: raster(color),
})),
};
}
function bodySemantics(pageCount: number): VisualPageSemanticExpectation[] {
return Array.from({ length: pageCount }, (_, index) => ({
physicalPageNumber: index + 1,
kind: index === 0 ? "body-first" : "body-rest",
logicalPageNumber: index + 1,
logicalPageCount: pageCount,
headerVisible: false,
footerVisible: false,
}));
}
const flowParagraphs = [
{ index: 0, text: "甲", role: "body" as const, section: "body" as const },
{ index: 1, text: "乙丙", role: "body" as const, section: "body" as const },
];
describe("PDF 视觉差异报告", () => {
it("汇总栅格门禁并生成内嵌图片的安全 HTML", async () => {
const report = await createPdfVisualDiffReport(
@@ -97,4 +150,352 @@ describe("PDF 视觉差异报告", () => {
expect(overflow?.unpairedPng?.byteLength).toBeGreaterThan(0);
expect(renderPdfVisualDiffHtml(report)).toContain("未配对页面");
});
it("纸张物理尺寸容差内不因一像素舍入差重复失败", async () => {
const baseline = snapshot("基线", raster("#111111"));
const candidate = snapshot("候选", {
...raster("#111111"),
widthPx: 41,
});
candidate.pages[0]!.widthPt = baseline.pages[0]!.widthPt + 0.36;
const report = await createPdfVisualDiffReport(baseline, candidate, {
thresholds: {
maxMeanAbsoluteError: 255,
maxChangedPixelRatio: 1,
minInkIou: 0,
minEdgeIou: 0,
},
});
expect(report.pages[0]?.metrics).toMatchObject({
dimensionsMatch: false,
geometryNormalized: true,
comparedWidthPx: 40,
});
expect(report.issues.map((issue) => issue.code)).not.toContain(
"RASTER_SIZE_MISMATCH",
);
});
it("展示并校验逐页逻辑页码和位置", async () => {
const baseline = snapshot("Chromium", raster("#111111"));
const candidate = snapshot("Word", raster("#111111"));
const pageNumberLine = (x: number) => ({
text: "— 1 —",
normalizedText: "— 1 —",
bounds: { x, y: 22, width: 2, height: 1 },
baselineY: 23,
role: "page-number" as const,
items: [],
});
baseline.pages[0]!.lines = [pageNumberLine(17)];
baseline.pages[0]!.pageNumberText = "— 1 —";
candidate.pages[0]!.lines = [pageNumberLine(9)];
candidate.pages[0]!.pageNumberText = "— 1 —";
const report = await createPdfVisualDiffReport(baseline, candidate, {
pageSemantics: [
{
physicalPageNumber: 1,
kind: "body-first",
logicalPageNumber: 1,
logicalPageCount: 1,
headerVisible: false,
footerVisible: true,
footerAlignment: "right",
pageNumberText: "— 1 —",
},
],
});
const html = renderPdfVisualDiffHtml(report);
expect(report.status).toBe("failed");
expect(report.pages[0]?.semantics?.candidate).toMatchObject({
pageNumberAlignment: "center",
});
expect(report.issues.map((issue) => issue.code)).toContain(
"PAGE_NUMBER_ALIGNMENT_MISMATCH",
);
expect(html).toContain("页面语义");
expect(html).toContain("正文首页");
expect(html).toContain("逻辑页 1 / 1");
});
it("拒绝本应隐藏却仍然出现的页眉", async () => {
const baseline = snapshot("Chromium", raster("#111111"));
const candidate = snapshot("Word", raster("#111111"));
const headerLine = {
text: "左页眉",
normalizedText: "左页眉",
bounds: { x: 2, y: 1, width: 4, height: 1 },
baselineY: 2,
role: "content" as const,
items: [],
};
baseline.pages[0]!.lines = [headerLine];
candidate.pages[0]!.lines = [headerLine];
const report = await createPdfVisualDiffReport(baseline, candidate, {
pageSemantics: [
{
physicalPageNumber: 1,
kind: "cover",
logicalPageCount: 1,
headerVisible: false,
headerSlots: [{ alignment: "left", text: "左页眉" }],
footerVisible: false,
},
],
});
expect(report.status).toBe("failed");
expect(report.issues.map((issue) => issue.code)).toContain(
"HEADER_VISIBILITY_MISMATCH",
);
});
it("拒绝页眉和页码纵向基线偏差超过 2pt", async () => {
const baseline = snapshot("Chromium", raster("#111111"));
const candidate = snapshot("Word", raster("#111111"));
baseline.pages[0]!.heightPt = 100;
candidate.pages[0]!.heightPt = 100;
const lines = (headerY: number, footerY: number) => [
{
text: "左页眉",
normalizedText: "左页眉",
bounds: { x: 2, y: headerY - 1, width: 4, height: 1 },
baselineY: headerY,
role: "content" as const,
items: [],
},
{
text: "1 / 1",
normalizedText: "1 / 1",
bounds: { x: 9, y: footerY - 1, width: 2, height: 1 },
baselineY: footerY,
role: "page-number" as const,
items: [],
},
];
baseline.pages[0]!.lines = lines(2, 95);
candidate.pages[0]!.lines = lines(5, 92);
const report = await createPdfVisualDiffReport(baseline, candidate, {
pageSemantics: [
{
physicalPageNumber: 1,
kind: "body-first",
logicalPageNumber: 1,
logicalPageCount: 1,
headerVisible: true,
headerSlots: [{ alignment: "left", text: "左页眉" }],
footerVisible: true,
footerAlignment: "center",
pageNumberText: "1 / 1",
},
],
});
expect(report.issues.map((issue) => issue.code)).toEqual(
expect.arrayContaining([
"HEADER_VERTICAL_POSITION_MISMATCH",
"PAGE_NUMBER_VERTICAL_POSITION_MISMATCH",
]),
);
});
it("按文档锚点与页内相对偏移校验页码并容纳 Chromium 自身逐页抖动", async () => {
const createTwoPageSnapshot = (
label: string,
baselines: readonly [number, number],
): PdfDocumentSnapshot => {
const base = snapshot(label, raster("#111111"));
base.pageCount = 2;
base.pages = baselines.map((baselineY, index) => ({
...base.pages[0]!,
pageNumber: index + 1,
heightPt: 100,
lines: [
{
text: `${index + 1} / 2`,
normalizedText: `${index + 1} / 2`,
bounds: { x: 9, y: baselineY - 1, width: 2, height: 1 },
baselineY,
role: "page-number" as const,
items: [],
},
],
}));
return base;
};
const baseline = createTwoPageSnapshot("Chromium", [95, 93.5]);
const candidate = createTwoPageSnapshot("Word", [95.8, 95.8]);
const semantics: VisualPageSemanticExpectation[] = [1, 2].map(
(pageNumber) => ({
physicalPageNumber: pageNumber,
kind: pageNumber === 1 ? "body-first" : "body-rest",
logicalPageNumber: pageNumber,
logicalPageCount: 2,
headerVisible: false,
footerVisible: true,
footerAlignment: "center",
pageNumberText: `${pageNumber} / 2`,
}),
);
const report = await createPdfVisualDiffReport(baseline, candidate, {
pageSemantics: semantics,
});
expect(report.issues.map((issue) => issue.code)).not.toContain(
"PAGE_NUMBER_VERTICAL_POSITION_MISMATCH",
);
});
it("将格式一致的整段跨页栅格差异标记为正文流动警告", async () => {
const baseline = flowSnapshot(
"baseline",
[[contentLine("甲", 8)], [contentLine("乙丙", 8)]],
"#111111",
);
const candidate = flowSnapshot(
"candidate",
[[contentLine("甲", 8), contentLine("乙丙", 12)], []],
"#cc0000",
);
const report = await createPdfVisualDiffReport(baseline, candidate, {
expectedEditableText: "甲乙丙",
expectedEditableParagraphs: flowParagraphs,
baselinePageSemantics: bodySemantics(2),
candidatePageSemantics: bodySemantics(2),
});
expect(report.status).toBe("warning");
expect(report.basic.bodyFlow).toMatchObject({
detected: true,
legal: true,
movedParagraphIndexes: [1],
});
expect(report.pages.map((page) => page.rasterComparisonMode)).toEqual([
"body-flow",
"body-flow",
]);
expect(report.issues.map((issue) => issue.code)).toContain(
"BODY_FLOW_RASTER_DIFFERENCE",
);
expect(report.issues.map((issue) => issue.code)).not.toContain(
"INK_IOU_BELOW_THRESHOLD",
);
});
it("整段跨页同时发生换行变化时仍保持严格失败", async () => {
const baseline = flowSnapshot(
"baseline",
[[contentLine("甲", 8)], [contentLine("乙丙", 8)]],
"#111111",
);
const candidate = flowSnapshot(
"candidate",
[
[contentLine("甲", 8), contentLine("乙", 12), contentLine("丙", 16)],
[],
],
"#cc0000",
);
const report = await createPdfVisualDiffReport(baseline, candidate, {
expectedEditableText: "甲乙丙",
expectedEditableParagraphs: flowParagraphs,
baselinePageSemantics: bodySemantics(2),
candidatePageSemantics: bodySemantics(2),
});
expect(report.status).toBe("failed");
expect(report.basic.bodyFlow).toMatchObject({ detected: true, legal: false });
expect(report.pages.every(
(page) => page.rasterComparisonMode === "strict",
)).toBe(true);
expect(report.issues.map((issue) => issue.code)).toContain(
"TEXT_BLOCK_LINE_COUNT_MISMATCH",
);
});
it("允许格式一致的正文自然分页产生物理页数差异", async () => {
const baseline = flowSnapshot(
"baseline",
[[contentLine("甲", 8)], [contentLine("乙丙", 8)]],
"#111111",
);
const candidate = flowSnapshot(
"candidate",
[[contentLine("甲", 8), contentLine("乙丙", 12)]],
"#cc0000",
);
const report = await createPdfVisualDiffReport(baseline, candidate, {
expectedEditableText: "甲乙丙",
expectedEditableParagraphs: flowParagraphs,
baselinePageSemantics: bodySemantics(2),
candidatePageSemantics: bodySemantics(1),
});
expect(report.status).toBe("warning");
expect(report.basic.bodyFlow).toMatchObject({
legal: true,
baselineBodyPageCount: 2,
candidateBodyPageCount: 1,
});
expect(report.issues.map((issue) => issue.code)).toContain(
"BODY_FLOW_PAGE_COUNT_DIFFERENCE",
);
expect(report.pages[1]).toMatchObject({
rasterComparisonMode: "body-flow",
status: "warning",
});
});
it("新增正文页缺少预期页码时不得认定为合法分页流动", async () => {
const pageNumber = (text: string): PdfTextLineSnapshot => ({
text,
normalizedText: text,
bounds: { x: 9, y: 22, width: 2, height: 1 },
baselineY: 23,
role: "page-number",
items: [],
});
const baseline = flowSnapshot(
"baseline",
[[contentLine("甲", 8), contentLine("乙丙", 12), pageNumber("1 / 1")]],
"#111111",
);
const candidate = flowSnapshot(
"candidate",
[
[contentLine("甲", 8), pageNumber("1 / 2")],
[contentLine("乙丙", 8)],
],
"#cc0000",
);
const baselineSemantics = bodySemantics(1).map((item) => ({
...item,
footerVisible: true,
footerAlignment: "center" as const,
pageNumberText: "1 / 1",
}));
const candidateSemantics = bodySemantics(2).map((item, index) => ({
...item,
footerVisible: true,
footerAlignment: "center" as const,
pageNumberText: `${index + 1} / 2`,
}));
const report = await createPdfVisualDiffReport(baseline, candidate, {
expectedEditableText: "甲乙丙",
expectedEditableParagraphs: flowParagraphs,
baselinePageSemantics: baselineSemantics,
candidatePageSemantics: candidateSemantics,
});
expect(report.status).toBe("failed");
expect(report.basic.bodyFlow).toMatchObject({ detected: true, legal: false });
expect(report.issues.map((issue) => issue.code)).toContain(
"PAGE_NUMBER_VISIBILITY_MISMATCH",
);
});
});