Normalize captured ChatGPT rendering

This commit is contained in:
Codex
2026-07-08 04:29:09 +00:00
parent a84ceb8065
commit a86050f005
2 changed files with 75 additions and 16 deletions
+46 -6
View File
@@ -96,6 +96,45 @@ async function localizeAssets(conversation: NormalizedConversation, slug: string
return { failures };
}
function normalizeCopiedHtml(html: string): string {
const $ = cheerio.load(html, {}, false);
$("script, style, iframe, object, embed, form, button, textarea, input, nav").remove();
$("a").each((_, anchor) => {
const link = $(anchor);
const decorativeImages = link.find("img").filter((_, image) => {
const alt = ($(image).attr("alt") || "").trim();
return alt.length === 0;
});
if (decorativeImages.length > 0) {
decorativeImages.remove();
link.attr("data-citation", "true");
}
});
$("span").each((_, span) => {
const element = $(span);
if (Object.keys(span.attribs || {}).length === 0) element.replaceWith(element.contents());
});
$("p, li, blockquote").each((_, node) => {
const element = $(node);
element.html(
(element.html() || "").replace(/\*\*([^*<>\n][^*<>]*?)\*\*/g, (_match, text: string) => `<strong>${text}</strong>`)
);
});
$("a").each((_, anchor) => {
const link = $(anchor);
const text = link.text().replace(/\s+/g, " ").trim();
if (link.attr("data-citation") && text) link.text(text);
});
$(":empty").not("br,img,hr").remove();
return $.html().trim();
}
function findChromiumExecutable(): string {
const preferred = chromium.executablePath();
if (nodeFs.existsSync(preferred)) return preferred;
@@ -256,12 +295,10 @@ async function extractConversation(sourceUrl: string): Promise<ExtractedConversa
: index % 2 === 0
? "user"
: "assistant";
const contentRoot =
node.querySelector(".markdown") ||
node.querySelector("[data-message-content]") ||
node;
const markdownRoot = node.querySelector(".markdown") || node.querySelector("[class*=markdown]");
const contentRoot = markdownRoot || node.querySelector("[data-message-content]") || node;
const text = (contentRoot.textContent || "").replace(/\s+\n/g, "\n").trim();
const html = cleanHtml(contentRoot) || textToHtml(text);
const html = markdownRoot ? cleanHtml(markdownRoot) || textToHtml(text) : textToHtml(text);
return { id: `m-${index + 1}`, role, text, html };
})
.filter((message) => message.text.length > 0);
@@ -348,7 +385,10 @@ export async function runCapture(row: CaptureRow): Promise<void> {
title: extracted.title || "ChatGPT Share",
sourceUrl: row.source_url,
capturedAt: new Date().toISOString(),
messages: extracted.messages
messages: extracted.messages.map((message) => ({
...message,
html: normalizeCopiedHtml(message.html)
}))
};
store.updateStatus(row.id, "rendering");
+29 -10
View File
@@ -6,8 +6,8 @@ import { config } from "./config.js";
const chatGptAvatar = `
<svg viewBox="0 0 40 40" role="img" aria-label="ChatGPT" xmlns="http://www.w3.org/2000/svg">
<circle cx="20" cy="20" r="20" fill="#10a37f"/>
<path fill="#fff" d="M29.6 17.9a7 7 0 0 0-9.4-8.2 7 7 0 0 0-10.4 6.1 7 7 0 0 0 1.4 12.7 7 7 0 0 0 9.4 1.8 7 7 0 0 0 10.3-6.1 7 7 0 0 0-1.3-6.3Zm-9.4-6.2a5 5 0 0 1 6.9 5.4l-5.6-3.2a1 1 0 0 0-1 0l-7.6 4.4v-2.7l6.8-3.9h.5Zm-8.4 4.1a5 5 0 0 1 6.2-4.9l-5.5 3.2a1 1 0 0 0-.5.9v8.8l-2.4-1.4v-7.9c.6-.7 1.3-1.3 2.2-1.7Zm-1.2 10.5a5 5 0 0 1-2.8-8l.1 6.4c0 .4.2.7.5.9l7.6 4.4-2.4 1.4-6.8-3.9c-.2-.4-.2-.8-.2-1.2Zm9.2 3.9a5 5 0 0 1-6.9-5.4l5.6 3.2c.3.2.7.2 1 0l7.6-4.4v2.8l-6.8 3.9-.5-.1Zm8.4-4.1a5 5 0 0 1-6.2 4.9l5.5-3.2c.3-.2.5-.5.5-.9v-8.8l2.4 1.4v7.9c-.6.7-1.3 1.3-2.2 1.7Zm-8.2-.2-3-1.7v-3.5l3-1.7 3 1.7v3.5l-3 1.7Zm4-7-3.5-2a1 1 0 0 0-1 0l-3.5 2v-2.8l4-2.3 6.8 3.9c.2.4.2.8.2 1.2l-3-1.7Z"/>
<circle cx="20" cy="20" r="19" fill="var(--assistant-avatar-bg)" stroke="var(--assistant-avatar-border)"/>
<path fill="var(--assistant-avatar-fg)" d="M29.6 17.9a7 7 0 0 0-9.4-8.2 7 7 0 0 0-10.4 6.1 7 7 0 0 0 1.4 12.7 7 7 0 0 0 9.4 1.8 7 7 0 0 0 10.3-6.1 7 7 0 0 0-1.3-6.3Zm-9.4-6.2a5 5 0 0 1 6.9 5.4l-5.6-3.2a1 1 0 0 0-1 0l-7.6 4.4v-2.7l6.8-3.9h.5Zm-8.4 4.1a5 5 0 0 1 6.2-4.9l-5.5 3.2a1 1 0 0 0-.5.9v8.8l-2.4-1.4v-7.9c.6-.7 1.3-1.3 2.2-1.7Zm-1.2 10.5a5 5 0 0 1-2.8-8l.1 6.4c0 .4.2.7.5.9l7.6 4.4-2.4 1.4-6.8-3.9c-.2-.4-.2-.8-.2-1.2Zm9.2 3.9a5 5 0 0 1-6.9-5.4l5.6 3.2c.3.2.7.2 1 0l7.6-4.4v2.8l-6.8 3.9-.5-.1Zm8.4-4.1a5 5 0 0 1-6.2 4.9l5.5-3.2c.3-.2.5-.5.5-.9v-8.8l2.4 1.4v7.9c-.6.7-1.3 1.3-2.2 1.7Zm-8.2-.2-3-1.7v-3.5l3-1.7 3 1.7v3.5l-3 1.7Zm4-7-3.5-2a1 1 0 0 0-1 0l-3.5 2v-2.8l4-2.3 6.8 3.9c.2.4.2.8.2 1.2l-3-1.7Z"/>
</svg>`;
function escapeHtml(value: string): string {
@@ -31,6 +31,9 @@ function css(): string {
--code: #0f172a;
--code-bg: #f4f4f5;
--accent: #10a37f;
--assistant-avatar-bg: #111827;
--assistant-avatar-fg: #ffffff;
--assistant-avatar-border: #111827;
font-family: ui-sans-serif, system-ui, -apple-system, BlinkMacSystemFont, "Segoe UI", "Noto Sans", "Noto Sans CJK SC", "Noto Sans CJK JP", "Noto Sans CJK KR", "Noto Sans Thai", "Helvetica Neue", Arial, sans-serif;
}
@media (prefers-color-scheme: dark) {
@@ -43,13 +46,16 @@ function css(): string {
--soft: #2f2f2f;
--code: #e5e7eb;
--code-bg: #171717;
--assistant-avatar-bg: #ececec;
--assistant-avatar-fg: #111827;
--assistant-avatar-border: #ececec;
}
}
html[data-theme="light"] {
--bg: #ffffff; --panel: #ffffff; --text: #1f2328; --muted: #6b7280; --border: #e5e7eb; --soft: #f7f7f8; --code: #0f172a; --code-bg: #f4f4f5;
--bg: #ffffff; --panel: #ffffff; --text: #1f2328; --muted: #6b7280; --border: #e5e7eb; --soft: #f7f7f8; --code: #0f172a; --code-bg: #f4f4f5; --assistant-avatar-bg: #111827; --assistant-avatar-fg: #ffffff; --assistant-avatar-border: #111827;
}
html[data-theme="dark"] {
--bg: #212121; --panel: #212121; --text: #ececec; --muted: #a8a8a8; --border: #3f3f46; --soft: #2f2f2f; --code: #e5e7eb; --code-bg: #171717;
--bg: #212121; --panel: #212121; --text: #ececec; --muted: #a8a8a8; --border: #3f3f46; --soft: #2f2f2f; --code: #e5e7eb; --code-bg: #171717; --assistant-avatar-bg: #ececec; --assistant-avatar-fg: #111827; --assistant-avatar-border: #ececec;
}
* { box-sizing: border-box; }
body { margin: 0; background: var(--bg); color: var(--text); font-size: 16px; line-height: 1.65; }
@@ -61,14 +67,23 @@ a { color: inherit; }
.title { max-width: 880px; margin: 0 auto; padding: 34px 20px 18px; }
h1 { margin: 0; font-size: clamp(28px, 5vw, 44px); line-height: 1.12; letter-spacing: 0; }
.conversation { max-width: 880px; margin: 0 auto; padding: 0 20px 36px; }
.message { display: grid; grid-template-columns: 38px minmax(0, 1fr); gap: 16px; padding: 22px 0; border-top: 1px solid var(--border); }
.avatar { width: 32px; height: 32px; display: grid; place-items: center; margin-top: 2px; }
.message { display: grid; grid-template-columns: 36px minmax(0, 1fr); gap: 16px; padding: 22px 0; border-top: 1px solid var(--border); }
.avatar { width: 30px; height: 30px; display: grid; place-items: center; margin-top: 1px; }
.avatar.user { border-radius: 50%; background: var(--soft); color: var(--text); font-weight: 650; font-size: 13px; }
.avatar svg { width: 32px; height: 32px; display: block; }
.avatar svg { width: 30px; height: 30px; display: block; flex: 0 0 auto; }
.role { color: var(--muted); font-size: 13px; margin-bottom: 4px; }
.content { min-width: 0; overflow-wrap: anywhere; }
.content p { margin: 0 0 1em; }
.content h2, .content h3, .content h4 { margin: 1.35em 0 0.55em; line-height: 1.35; font-weight: 700; }
.content h2 { font-size: 1.28em; }
.content h3 { font-size: 1.12em; }
.content h4 { font-size: 1em; }
.content p { margin: 0 0 0.9em; }
.content p:last-child { margin-bottom: 0; }
.content ul, .content ol { margin: 0.6em 0 1em; padding-left: 1.45em; }
.content li { margin: 0.3em 0; padding-left: 0.1em; }
.content li > p { margin: 0.2em 0; }
.content li > ul, .content li > ol { margin: 0.35em 0 0.45em; padding-left: 1.25em; }
.content strong { font-weight: 700; }
.content pre { margin: 1em 0; overflow: auto; padding: 14px 16px; border-radius: 8px; background: var(--code-bg); color: var(--code); font-size: 14px; line-height: 1.55; }
.content code { font-family: ui-monospace, SFMono-Regular, Menlo, Monaco, Consolas, "Liberation Mono", "Noto Sans Mono", monospace; }
.content :not(pre) > code { background: var(--code-bg); border-radius: 5px; padding: 0.1em 0.32em; }
@@ -76,6 +91,8 @@ h1 { margin: 0; font-size: clamp(28px, 5vw, 44px); line-height: 1.12; letter-spa
.content th, .content td { border: 1px solid var(--border); padding: 8px 10px; text-align: left; vertical-align: top; }
.content blockquote { border-left: 3px solid var(--border); margin: 1em 0; padding-left: 14px; color: var(--muted); }
.content img { max-width: 100%; height: auto; border-radius: 8px; }
.content a[data-citation] { color: var(--muted); font-size: 0.92em; white-space: nowrap; text-decoration-thickness: 1px; text-underline-offset: 2px; }
.content a[data-citation] img { display: none; }
.downloads { border-top: 1px solid var(--border); background: var(--soft); }
.downloads-inner { max-width: 880px; margin: 0 auto; padding: 22px 20px; display: flex; flex-wrap: wrap; gap: 10px; align-items: center; }
.button, select { min-height: 38px; border: 1px solid var(--border); border-radius: 8px; background: var(--panel); color: var(--text); padding: 8px 12px; font: inherit; font-size: 14px; text-decoration: none; }
@@ -89,8 +106,10 @@ h1 { margin: 0; font-size: clamp(28px, 5vw, 44px); line-height: 1.12; letter-spa
.topbar-inner { padding: 12px 16px; }
.title { padding: 26px 16px 12px; }
.conversation { padding: 0 16px 28px; }
.message { grid-template-columns: 32px minmax(0, 1fr); gap: 12px; padding: 18px 0; }
.avatar, .avatar svg { width: 28px; height: 28px; }
.message { grid-template-columns: 30px minmax(0, 1fr); gap: 10px; padding: 18px 0; }
.avatar, .avatar svg { width: 26px; height: 26px; }
.avatar.user { font-size: 11px; }
.content ul, .content ol { padding-left: 1.25em; }
.downloads-inner { padding: 18px 16px; align-items: stretch; }
.button, select { width: 100%; }
}`;