diff --git a/src/lib/clean-pdf/__tests__/template.test.ts b/src/lib/clean-pdf/__tests__/template.test.ts new file mode 100644 index 0000000..2a74221 --- /dev/null +++ b/src/lib/clean-pdf/__tests__/template.test.ts @@ -0,0 +1,59 @@ +import { describe, expect, it } from "vitest"; +import { buildCleanHtml, safePdfFilename } from "@/lib/clean-pdf/template"; + +const base = { + title: "Заголовок статьи", + url: "https://example.com/article", + theme: "light" as const, +}; + +describe("buildCleanHtml", () => { + it("экранирует HTML в метаполях", () => { + const html = buildCleanHtml({ ...base, title: ``, contentHtml: "

ок

" }); + expect(html).not.toContain(""); + expect(html).toContain("<script>"); + }); + + it("вставляет контент и шапку", () => { + const html = buildCleanHtml({ ...base, author: "Автор", site: "Example", contentHtml: "

Текст статьи

" }); + expect(html).toContain("

Текст статьи

"); + expect(html).toContain("Заголовок статьи"); + expect(html).toContain("Автор"); + expect(html).toContain("https://example.com/article"); + }); + + it("строит оглавление при 3+ заголовках и проставляет якоря", () => { + const content = "

Один

a

Два

b

Три

c

"; + const html = buildCleanHtml({ ...base, contentHtml: content }); + expect(html).toContain("Содержание"); + expect(html).toMatch(/

Один<\/h2>/); + expect((html.match(/class="toc-item/g) ?? []).length).toBe(3); + }); + + it("не строит оглавление при <3 заголовках", () => { + const html = buildCleanHtml({ ...base, contentHtml: "

Один

a

" }); + expect(html).not.toContain("Содержание"); + }); + + it("уникализирует одинаковые якоря", () => { + const content = "

Раздел

Раздел

Раздел

"; + const html = buildCleanHtml({ ...base, contentHtml: content }); + const ids = [...html.matchAll(/

m[1]); + expect(new Set(ids).size).toBe(3); + }); + + it("переключает тёмную тему", () => { + const light = buildCleanHtml({ ...base, contentHtml: "

x

" }); + const dark = buildCleanHtml({ ...base, theme: "dark", contentHtml: "

x

" }); + expect(light).toContain('data-theme="light"'); + expect(dark).toContain('data-theme="dark"'); + }); +}); + +describe("safePdfFilename", () => { + it("убирает опасные символы и ограничивает длину", () => { + expect(safePdfFilename('Статья: как/не\\надо "делать"?')).toBe("Статья_ как_не_надо _делать_"); + expect(safePdfFilename("")).toBe("document"); + expect(safePdfFilename("a".repeat(300)).length).toBeLessThanOrEqual(140); + }); +}); diff --git a/src/lib/clean-pdf/template.ts b/src/lib/clean-pdf/template.ts new file mode 100644 index 0000000..da29cb4 --- /dev/null +++ b/src/lib/clean-pdf/template.ts @@ -0,0 +1,107 @@ +import { JSDOM } from "jsdom"; + +export interface CleanArticle { + title: string; + author?: string; + site?: string; + url: string; + contentHtml: string; + theme: "light" | "dark"; +} + +function escapeHtml(s: string): string { + return s + .replace(/&/g, "&") + .replace(//g, ">") + .replace(/"/g, """); +} + +export function safePdfFilename(title: string): string { + const cleaned = title.replace(/[\\/:"*?<>|\n\r]+/g, "_").trim().slice(0, 140); + return cleaned || "document"; +} + +interface TocEntry { id: string; text: string; level: 2 | 3 } + +/** Проставляет id заголовкам h2/h3 и возвращает контент + записи оглавления. */ +function prepareContent(contentHtml: string): { html: string; toc: TocEntry[] } { + const dom = new JSDOM(`${contentHtml}`); + const doc = dom.window.document; + const used = new Set(); + const toc: TocEntry[] = []; + + doc.querySelectorAll("h2, h3").forEach((h) => { + const text = (h.textContent ?? "").trim(); + if (!text) return; + let id = text.toLowerCase().replace(/[^\p{L}\p{N}]+/gu, "-").replace(/^-+|-+$/g, "") || "section"; + let i = 2; + while (used.has(id)) id = `${id}-${i++}`; + used.add(id); + h.id = id; + toc.push({ id, text, level: h.tagName === "H2" ? 2 : 3 }); + }); + + return { html: doc.body.innerHTML, toc }; +} + +const CSS = ` + :root[data-theme="light"] { --bg: #faf7f0; --fg: #1f1d1a; --muted: #6b6459; --line: #d8d2c4; --accent: #7a2e2e; } + :root[data-theme="dark"] { --bg: #201e1b; --fg: #e8e4dc; --muted: #a39b8d; --line: #3d3a34; --accent: #d4a0a0; } + * { box-sizing: border-box; } + body { background: var(--bg); color: var(--fg); font-family: Georgia, "Times New Roman", serif; + font-size: 12.5pt; line-height: 1.65; margin: 0; } + main { max-width: 100%; } + h1 { font-size: 22pt; line-height: 1.25; margin: 0 0 6pt; } + h2 { font-size: 16pt; margin: 20pt 0 8pt; } + h3 { font-size: 13.5pt; margin: 16pt 0 6pt; } + p { margin: 0 0 9pt; } + a { color: var(--accent); text-decoration: none; } + img, video, iframe { max-width: 100%; height: auto; } + pre { background: rgba(127,127,127,.08); border: 1px solid var(--line); padding: 8pt; + overflow-x: hidden; white-space: pre-wrap; word-wrap: break-word; + font-family: "SF Mono", Menlo, Consolas, monospace; font-size: 9.5pt; } + code { font-family: "SF Mono", Menlo, Consolas, monospace; font-size: 0.9em; } + blockquote { border-left: 3px solid var(--line); margin: 0 0 9pt; padding: 2pt 0 2pt 12pt; color: var(--muted); } + table { border-collapse: collapse; width: 100%; font-size: 10.5pt; } + th, td { border: 1px solid var(--line); padding: 4pt 6pt; text-align: left; } + figure { margin: 0 0 9pt; } figcaption { color: var(--muted); font-size: 9.5pt; } + .meta { color: var(--muted); font-size: 10pt; margin-bottom: 4pt; } + .meta a { color: var(--muted); } + .head-rule { border: 0; border-top: 1px solid var(--line); margin: 12pt 0 16pt; } + .toc { border: 1px solid var(--line); padding: 10pt 14pt; margin: 0 0 16pt; } + .toc-title { font-weight: bold; margin-bottom: 6pt; } + .toc-item { display: block; margin: 2pt 0; } + .toc-item.lvl3 { padding-left: 14pt; font-size: 0.92em; } + h2, h3 { break-after: avoid; } pre, blockquote, figure, table { break-inside: avoid; } +`; + +export function buildCleanHtml(article: CleanArticle): string { + const { html, toc } = prepareContent(article.contentHtml); + const metaParts = [article.site, article.author].filter(Boolean).map((s) => escapeHtml(s!)); + + const tocHtml = toc.length >= 3 + ? `` + : ""; + + return ` + + + +${escapeHtml(article.title)} + + + +
+

${escapeHtml(article.title)}

+ ${metaParts.length ? `
${metaParts.join(" · ")}
` : ""} +
${escapeHtml(article.url)}
+
+ ${tocHtml} + ${html} +
+ +`; +}