761 lines
33 KiB
HTML
761 lines
33 KiB
HTML
<!doctype html>
|
|
<html lang="en">
|
|
<head>
|
|
<meta charset="UTF-8" />
|
|
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
|
<meta http-equiv="Cross-Origin-Opener-Policy" content="same-origin" />
|
|
<meta http-equiv="Cross-Origin-Embedder-Policy" content="credentialless" />
|
|
<title>Kreuzberg | Live Demo</title>
|
|
<!-- SEO Meta Tags -->
|
|
<link rel="icon" type="image/x-icon" href="./assets/favicon.ico" />
|
|
<meta
|
|
name="description"
|
|
content="Extract text, tables, and metadata from 90+ file formats directly in your browser using WebAssembly. No uploading to a server required."
|
|
/>
|
|
<meta property="og:title" content="Kreuzberg | Live Demo" />
|
|
<meta
|
|
property="og:description"
|
|
content="Secure, in-browser document intelligence and text extraction."
|
|
/>
|
|
<meta property="og:url" content="https://kreuzberg.dev" />
|
|
<meta property="og:type" content="website" />
|
|
|
|
<!-- Google Fonts -->
|
|
<link rel="preconnect" href="https://fonts.googleapis.com" />
|
|
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin />
|
|
<link
|
|
href="https://fonts.googleapis.com/css2?family=Sora:wght@100..800&family=Exo+2:ital,wght@0,100..900;1,100..900&family=JetBrains+Mono:wght@400;500&display=swap"
|
|
rel="stylesheet"
|
|
/>
|
|
|
|
<!-- Tailwind CDN -->
|
|
<script src="https://cdn.tailwindcss.com"></script>
|
|
<script>
|
|
tailwind.config = {
|
|
theme: {
|
|
extend: {
|
|
fontFamily: {
|
|
sora: ["Sora", "sans-serif"],
|
|
exo: ["Exo 2", "sans-serif"],
|
|
mono: ["JetBrains Mono", "monospace"],
|
|
},
|
|
colors: {
|
|
black: "#323040",
|
|
green: "#58FBDA",
|
|
"green-100": "#e6f6f3",
|
|
purple: "#da2ae0",
|
|
"dark-purple": "#26203a",
|
|
"grey-50": "#f6f5f9",
|
|
"app-red": "#ff456d",
|
|
},
|
|
},
|
|
},
|
|
};
|
|
</script>
|
|
|
|
<style>
|
|
body {
|
|
font-family: "Exo 2", sans-serif;
|
|
background-color: #323040;
|
|
color: white;
|
|
font-optical-sizing: auto;
|
|
}
|
|
|
|
/* Premium conic gradient border from Na'aman's React component */
|
|
.gradient-border-conic {
|
|
position: relative;
|
|
border-radius: 22px;
|
|
}
|
|
|
|
.gradient-border-conic::before {
|
|
content: "";
|
|
pointer-events: none;
|
|
position: absolute;
|
|
inset: 0;
|
|
z-index: 10;
|
|
border-radius: 22px;
|
|
padding: 2px;
|
|
background-image: conic-gradient(from 90deg, #58fbda, #da2ae0, #58fbda);
|
|
-webkit-mask:
|
|
linear-gradient(#fff 0 0) content-box,
|
|
linear-gradient(#fff 0 0);
|
|
mask:
|
|
linear-gradient(#fff 0 0) content-box,
|
|
linear-gradient(#fff 0 0);
|
|
-webkit-mask-composite: xor;
|
|
mask-composite: exclude;
|
|
}
|
|
|
|
@keyframes blob {
|
|
0% {
|
|
transform: translate(0px, 0px) scale(1);
|
|
}
|
|
33% {
|
|
transform: translate(30px, -50px) scale(1.1);
|
|
}
|
|
66% {
|
|
transform: translate(-20px, 20px) scale(0.9);
|
|
}
|
|
100% {
|
|
transform: translate(0px, 0px) scale(1);
|
|
}
|
|
}
|
|
.animate-blob {
|
|
animation: blob 7s infinite;
|
|
}
|
|
.animation-delay-2000 {
|
|
animation-delay: 2s;
|
|
}
|
|
</style>
|
|
</head>
|
|
|
|
<body class="min-h-screen flex flex-col bg-black relative overflow-x-hidden">
|
|
<!-- Animated Background Blobs -->
|
|
<div class="pointer-events-none fixed inset-0 z-[-1]">
|
|
<div
|
|
class="absolute top-0 right-0 -translate-x-1/4 -translate-y-1/4 md:left-auto md:translate-x-[20%] md:translate-y-[-20%]"
|
|
>
|
|
<div class="relative">
|
|
<div
|
|
class="animation-delay-2000 absolute top-0 right-0 h-[300px] w-[300px] animate-blob rounded-full bg-green opacity-50 mix-blend-multiply blur-[80px] filter md:top-30 md:h-[600px] md:w-[600px]"
|
|
></div>
|
|
<div
|
|
class="absolute top-20 right-0 h-[300px] w-[300px] animate-blob rounded-full bg-purple opacity-50 mix-blend-multiply blur-[80px] filter md:top-90 md:h-[600px] md:w-[600px]"
|
|
></div>
|
|
</div>
|
|
</div>
|
|
</div>
|
|
|
|
<!-- Header -->
|
|
<header class="w-full px-4 py-4 lg:px-6 lg:py-6">
|
|
<div
|
|
class="grid grid-cols-[1fr_1fr] items-center gap-4 w-full max-w-9xl mx-auto px-4 md:px-0"
|
|
>
|
|
<a href="/" aria-label="Go to home" class="justify-self-start">
|
|
<img
|
|
src="./assets/icons/mobile-logo.svg"
|
|
alt="Kreuzberg Logo"
|
|
class="h-10 w-10 lg:h-[65px] lg:w-[65px]"
|
|
/>
|
|
</a>
|
|
|
|
<div class="flex items-center gap-3 justify-self-end">
|
|
<a
|
|
href="https://discord.gg/xt9WY3GnKR"
|
|
rel="noopener noreferrer"
|
|
target="_blank"
|
|
aria-label="Join our Discord"
|
|
>
|
|
<button
|
|
class="flex size-9 cursor-pointer items-center justify-center rounded-[8px] border border-green bg-dark-purple p-0 transition-colors hover:bg-black lg:size-10 xl:size-11"
|
|
>
|
|
<img src="./assets/icons/discord-white-bg.svg" alt="Discord" class="w-6 h-6" />
|
|
</button>
|
|
</a>
|
|
|
|
<a
|
|
href="https://github.com/kreuzberg-dev/kreuzberg/"
|
|
rel="noopener noreferrer"
|
|
target="_blank"
|
|
aria-label="View on GitHub"
|
|
>
|
|
<button
|
|
class="group flex cursor-pointer items-center gap-2 border border-green transition-colors hover:bg-black h-9 rounded-[8px] bg-dark-purple px-3 py-2 lg:h-12 lg:rounded-[12px] lg:px-4 lg:py-3"
|
|
>
|
|
<img src="./assets/icons/github-white-bg.svg" alt="GitHub" class="w-6 h-6" />
|
|
<span class="font-normal text-base text-white hidden font-sora lg:block"
|
|
>View on GitHub</span
|
|
>
|
|
</button>
|
|
</a>
|
|
</div>
|
|
</div>
|
|
</header>
|
|
|
|
<!-- Main Content -->
|
|
<main class="flex flex-1 w-full">
|
|
<div class="w-full h-full px-4 md:px-[60px] pb-[40px]">
|
|
<div class="flex flex-col items-center gap-10 md:gap-[60px]">
|
|
<div class="flex flex-col gap-6 text-center">
|
|
<h1
|
|
class="text-4xl md:text-[100px] font-bold text-grey-50 leading-tight md:leading-[120px]"
|
|
>
|
|
Live Demo
|
|
</h1>
|
|
<p class="text-base md:text-2xl font-light text-grey-50 px-4 md:px-0">
|
|
Drop a file. Kreuzberg extracts text, tables, and metadata - no server, no upload, no
|
|
API <br class="hidden md:block" />
|
|
key. Everything runs in your browser via WASM.
|
|
</p>
|
|
</div>
|
|
<div
|
|
class="relative w-full max-w-[1063px] bg-black rounded-[22px] gradient-border-conic p-6 md:p-10 flex flex-col items-center gap-8"
|
|
>
|
|
<div
|
|
id="btn-back"
|
|
style="display: none"
|
|
class="flex gap-1 px-3 py-2 items-center self-start cursor-pointer hover:opacity-80 transition-opacity"
|
|
>
|
|
<img src="./assets/icons/chrevon-left.svg" alt="Back" class="size-[16px]" />
|
|
<p class="font-sora text-base text-green leading-[22px]">Back to Upload</p>
|
|
</div>
|
|
<div class="w-full h-full flex flex-col items-center gap-2">
|
|
<img src="./assets/icons/flask-fill.svg" alt="Flask" class="size-[37px]" />
|
|
<p class="text-[28px] font-bold text-white leading-[32px]">Try it yourself!</p>
|
|
<p class="text-[14px] font-light text-white leading-[18px]">
|
|
Processed internally in your browser via WebAssembly. Your file never leaves your
|
|
device.
|
|
</p>
|
|
</div>
|
|
|
|
<div
|
|
id="file-info"
|
|
style="display: none"
|
|
class="p-3 border border-green rounded-[12px] w-full flex items-center justify-between"
|
|
>
|
|
<div class="flex items-center gap-3">
|
|
<div class="flex items-center gap-1">
|
|
<img src="./assets/icons/file-fill.svg" alt="File" class="size-[24px]" />
|
|
<p id="file-name-display" class="font-semibold text-base text-white">
|
|
my-file.pdf
|
|
</p>
|
|
</div>
|
|
<p id="file-size-display" class="font-light text-sm text-white">4.00MB</p>
|
|
</div>
|
|
<img
|
|
id="btn-remove-file"
|
|
src="./assets/icons/close.svg"
|
|
alt="Close"
|
|
class="size-[24px] cursor-pointer hover:opacity-80 transition-opacity"
|
|
/>
|
|
</div>
|
|
|
|
<div
|
|
id="error-message"
|
|
style="display: none"
|
|
class="w-full p-6 flex flex-col bg-app-red/10 border border-app-red rounded-[12px]"
|
|
>
|
|
<div class="space-y-1">
|
|
<p id="error-title" class="text-base font-semibold text-white">File too large</p>
|
|
<p id="error-desc" class="text-sm font-light text-white">
|
|
File exceeds the 1MB sandbox limit. Reduce file size or use the API for larger
|
|
documents.
|
|
</p>
|
|
</div>
|
|
</div>
|
|
|
|
<div
|
|
id="output-screen"
|
|
style="display: none"
|
|
class="min-h-[500px] h-[865px] max-h-[90vh] md:max-h-none w-full flex flex-col gap-8 bg-dark-purple p-6 md:p-10 rounded-[12px]"
|
|
>
|
|
<div
|
|
class="flex flex-col sm:flex-row items-start sm:items-center justify-between gap-4"
|
|
>
|
|
<p class="font-bold text-2xl text-green-100">Output</p>
|
|
<div
|
|
class="relative inline-flex w-full sm:w-[238px] h-[38px] items-center rounded-[16px] bg-dark-purple border border-green p-1 text-slate-400"
|
|
>
|
|
<div
|
|
id="tab-indicator"
|
|
class="absolute left-1 top-1 bottom-1 w-[calc(50%-4px)] sm:w-[114px] rounded-[12px] bg-green transition-transform duration-300 ease-in-out"
|
|
></div>
|
|
<button
|
|
id="tab-markdown"
|
|
onclick="switchTab('markdown')"
|
|
class="relative z-10 flex w-1/2 h-full items-center justify-center whitespace-nowrap text-base font-normal font-sora text-black hover:opacity-80 transition-colors focus:outline-none"
|
|
>
|
|
Markdown
|
|
</button>
|
|
<button
|
|
id="tab-json"
|
|
onclick="switchTab('json')"
|
|
class="relative z-10 flex w-1/2 h-full items-center justify-center whitespace-nowrap text-base font-normal font-sora text-green hover:opacity-80 transition-colors focus:outline-none"
|
|
>
|
|
JSON
|
|
</button>
|
|
</div>
|
|
</div>
|
|
<div
|
|
id="tab-content"
|
|
class="w-full h-full bg-black rounded-[12px] px-6 py-8 overflow-y-auto text-sm"
|
|
>
|
|
<div class="w-full h-6 flex items-center justify-end gap-2">
|
|
<span
|
|
id="copy-feedback"
|
|
class="text-green text-xs font-sora font-semibold opacity-0 transition-opacity duration-300"
|
|
>
|
|
Copied!
|
|
</span>
|
|
<img
|
|
id="btn-copy"
|
|
src="./assets/icons/file-copy-fill-2.svg"
|
|
alt="Copy"
|
|
class="size-[24px] cursor-pointer hover:opacity-80 transition-opacity"
|
|
/>
|
|
</div>
|
|
<!-- Markdown Content -->
|
|
<div id="content-markdown" class="block text-white whitespace-pre-wrap">
|
|
# Markdown output will appear here...
|
|
</div>
|
|
<!-- JSON Content -->
|
|
<div id="content-json" class="hidden text-green whitespace-pre-wrap">
|
|
{ "status": "waiting", "message": "JSON output will appear here..." }
|
|
</div>
|
|
</div>
|
|
</div>
|
|
|
|
<div id="upload-screen" class="w-full h-full">
|
|
<label
|
|
id="dropZone"
|
|
for="fileInput"
|
|
class="flex w-full h-full cursor-pointer flex-col items-center justify-center gap-2 rounded-[12px] border border-dashed border-gray-500 bg-[#302D3F] px-[12px] py-[24px] transition-all hover:border-green hover:bg-[#302D3F]/80"
|
|
>
|
|
<div class="relative size-6">
|
|
<img
|
|
src="./assets/icons/upload.svg"
|
|
alt="Upload"
|
|
class="w-full h-full object-contain"
|
|
/>
|
|
</div>
|
|
<div class="space-y-1 text-center">
|
|
<h3 class="font-bold text-[18px] text-white">Drop a file here</h3>
|
|
<p class="font-light text-sm text-white">Or click to browse</p>
|
|
</div>
|
|
|
|
<input type="file" id="fileInput" class="hidden" />
|
|
</label>
|
|
</div>
|
|
</div>
|
|
</div>
|
|
</div>
|
|
</main>
|
|
|
|
<footer class="px-6 py-2.5 h-[63px] bg-black/20 flex items-center">
|
|
<a
|
|
href="https://kreuzberg.dev"
|
|
class="font-normal text-sm text-green hover:underline transition-all"
|
|
>
|
|
Join Kreuzberg Cloud Waitlist
|
|
</a>
|
|
</footer>
|
|
|
|
<script type="module">
|
|
const dropZone = document.getElementById("dropZone");
|
|
const fileInput = document.getElementById("fileInput");
|
|
const uploadScreen = document.getElementById("upload-screen");
|
|
const fileInfo = document.getElementById("file-info");
|
|
const outputScreen = document.getElementById("output-screen");
|
|
const btnBack = document.getElementById("btn-back");
|
|
|
|
// ── Structured logger ──────────────────────────────────────────────────
|
|
// All demo output is prefixed with [kreuzberg/wasm] so you can filter
|
|
// the DevTools console to that string and see only demo events.
|
|
// Use console.time labels to measure phases; timing shows in DevTools.
|
|
const log = (() => {
|
|
const NS = "[kreuzberg/wasm]";
|
|
return {
|
|
info: (msg, data) => console.info( `${NS} ${msg}`, ...(data !== undefined ? [data] : [])),
|
|
warn: (msg, data) => console.warn( `${NS} ${msg}`, ...(data !== undefined ? [data] : [])),
|
|
error: (msg, err) => console.error(`${NS} ${msg}`, err),
|
|
// Returns an object whose .end() stops the timer and returns elapsed ms.
|
|
phase: (label) => {
|
|
const key = `${NS} ${label}`;
|
|
const t0 = performance.now();
|
|
console.time(key);
|
|
return { end: () => { console.timeEnd(key); return performance.now() - t0; } };
|
|
},
|
|
};
|
|
})();
|
|
|
|
// ── Resolve CDN base URL ───────────────────────────────────────────────
|
|
// WASM_CDN_ORIGIN is the only URL that needs to change between prod and dev.
|
|
// patch-demo-dev.mjs rewrites it to http://localhost:9000 for local dev —
|
|
// that substitution covers the version fetch AND the URL passed into the
|
|
// Worker, so the Worker always loads from the same origin as the main thread.
|
|
// In production the origin stays @latest; after version is confirmed we pin
|
|
// to the exact version so the Worker and the version log agree.
|
|
const WASM_CDN_ORIGIN = "https://cdn.jsdelivr.net/npm/@kreuzberg/wasm@latest";
|
|
let resolvedVersion = null;
|
|
let wasmCdnBase = null;
|
|
try {
|
|
const t = log.phase("version resolve");
|
|
const pkg = await fetch(WASM_CDN_ORIGIN + "/package.json")
|
|
.then((r) => {
|
|
if (!r.ok) throw new Error(`CDN version fetch failed: ${r.status}`);
|
|
return r.json();
|
|
});
|
|
t.end();
|
|
resolvedVersion = pkg.version;
|
|
if (!resolvedVersion || typeof resolvedVersion !== "string")
|
|
throw new Error("Invalid version in @kreuzberg/wasm package.json");
|
|
// Dev: keep localhost base as-is. Prod: pin to resolved version so the
|
|
// Worker loads the exact same release that was verified above.
|
|
// NOTE: string concatenation (not a template literal) is intentional —
|
|
// the patch regex stops at '"' so the prod URL is never misreplaced.
|
|
wasmCdnBase = WASM_CDN_ORIGIN.startsWith("http://localhost")
|
|
? WASM_CDN_ORIGIN
|
|
: "https://cdn.jsdelivr.net/npm/@kreuzberg/wasm@" + resolvedVersion;
|
|
log.info("CDN version resolved", { version: resolvedVersion, base: wasmCdnBase });
|
|
} catch (e) {
|
|
log.error("version resolution failed", e);
|
|
}
|
|
|
|
// ── Drag & Drop ────────────────────────────────────────────────────────
|
|
["dragenter", "dragover", "dragleave", "drop"].forEach((eventName) => {
|
|
dropZone.addEventListener(eventName, preventDefaults, false);
|
|
});
|
|
function preventDefaults(e) {
|
|
e.preventDefault();
|
|
e.stopPropagation();
|
|
}
|
|
["dragenter", "dragover"].forEach((eventName) => {
|
|
dropZone.addEventListener(
|
|
eventName,
|
|
() => {
|
|
dropZone.classList.add("border-green");
|
|
dropZone.classList.remove("border-gray-500");
|
|
},
|
|
false,
|
|
);
|
|
});
|
|
["dragleave", "drop"].forEach((eventName) => {
|
|
dropZone.addEventListener(
|
|
eventName,
|
|
() => {
|
|
dropZone.classList.remove("border-green");
|
|
dropZone.classList.add("border-gray-500");
|
|
},
|
|
false,
|
|
);
|
|
});
|
|
fileInput.addEventListener(
|
|
"change",
|
|
(e) => {
|
|
if (e.target.files && e.target.files.length > 0) processFile(e.target.files[0]);
|
|
},
|
|
false,
|
|
);
|
|
dropZone.addEventListener(
|
|
"drop",
|
|
(e) => {
|
|
const dt = e.dataTransfer;
|
|
if (dt?.files && dt.files.length > 0) processFile(dt.files[0]);
|
|
},
|
|
false,
|
|
);
|
|
|
|
// ── Markdown Renderer ──────────────────────────────────────────────────
|
|
function renderMarkdown(md) {
|
|
let html = md
|
|
.replace(/&/g, "&")
|
|
.replace(/</g, "<")
|
|
.replace(/>/g, ">")
|
|
.replace(
|
|
/```([\w]*)?\n([\s\S]*?)```/gm,
|
|
(_, lang, code) =>
|
|
`<pre style="background:#1a172a;padding:12px;border-radius:8px;overflow-x:auto"><code${lang ? ` class="language-${lang}"` : ""}>${code.trimEnd()}</code></pre>`,
|
|
)
|
|
.replace(
|
|
/^###### (.+)$/gm,
|
|
'<h6 style="font-size:0.85em;font-weight:700;margin:8px 0">$1</h6>',
|
|
)
|
|
.replace(
|
|
/^##### (.+)$/gm,
|
|
'<h5 style="font-size:0.9em;font-weight:700;margin:8px 0">$1</h5>',
|
|
)
|
|
.replace(
|
|
/^#### (.+)$/gm,
|
|
'<h4 style="font-size:1em;font-weight:700;margin:10px 0">$1</h4>',
|
|
)
|
|
.replace(
|
|
/^### (.+)$/gm,
|
|
'<h3 style="font-size:1.1em;font-weight:700;margin:12px 0">$1</h3>',
|
|
)
|
|
.replace(
|
|
/^## (.+)$/gm,
|
|
'<h2 style="font-size:1.2em;font-weight:700;margin:14px 0;color:#58FBDA">$1</h2>',
|
|
)
|
|
.replace(
|
|
/^# (.+)$/gm,
|
|
'<h1 style="font-size:1.4em;font-weight:700;margin:16px 0;color:#58FBDA">$1</h1>',
|
|
)
|
|
.replace(/^---+$/gm, '<hr style="border-color:#58FBDA33;margin:16px 0">')
|
|
.replace(/\*\*\*(.+?)\*\*\*/g, "<strong><em>$1</em></strong>")
|
|
.replace(/\*\*(.+?)\*\*/g, "<strong>$1</strong>")
|
|
.replace(/\*(.+?)\*/g, "<em>$1</em>")
|
|
.replace(
|
|
/`([^`]+)`/g,
|
|
'<code style="background:#1a172a;padding:2px 6px;border-radius:4px;font-family:JetBrains Mono,monospace">$1</code>',
|
|
)
|
|
.replace(
|
|
/^> (.+)$/gm,
|
|
'<blockquote style="border-left:3px solid #58FBDA;padding-left:12px;color:#aaa;margin:8px 0">$1</blockquote>',
|
|
)
|
|
.replace(/^[ \t]*[-*+] (.+)$/gm, "<li>$1</li>")
|
|
.replace(/^[ \t]*\d+\. (.+)$/gm, "<oli>$1</oli>");
|
|
|
|
html = html.replace(
|
|
/(<li>[\s\S]*?<\/li>\n?)+/g,
|
|
(m) => `<ul style="padding-left:20px;margin:8px 0">${m}</ul>`,
|
|
);
|
|
html = html.replace(
|
|
/(<oli>[\s\S]*?<\/oli>\n?)+/g,
|
|
(m) =>
|
|
`<ol style="padding-left:20px;margin:8px 0">${m.replace(/<oli>/g, "<li>").replace(/<\/oli>/g, "</li>")}</ol>`,
|
|
);
|
|
|
|
const blockTags = /^<(h[1-6]|ul|ol|li|pre|blockquote|hr)/;
|
|
html = html
|
|
.split("\n")
|
|
.map((line) => {
|
|
if (!line.trim()) return "";
|
|
if (blockTags.test(line.trim())) return line;
|
|
return `<p style="margin:4px 0;line-height:1.6">${line}</p>`;
|
|
})
|
|
.join("\n");
|
|
return html;
|
|
}
|
|
|
|
// ── JSON Syntax Highlighter ────────────────────────────────────────────
|
|
function highlightJson(obj) {
|
|
const raw = JSON.stringify(obj, null, 2);
|
|
return raw.replace(
|
|
/("[^"]*"\s*:)|("[^"]*")|(-?\d+\.?\d*(?:[eE][+-]?\d+)?)|(true|false)|(null)/g,
|
|
(_, key, str, num, bool, nil) => {
|
|
if (key) return `<span style="color:#58FBDA">${key}</span>`;
|
|
if (str) return `<span style="color:#e6f6f3">${str}</span>`;
|
|
if (num) return `<span style="color:#da2ae0">${num}</span>`;
|
|
if (bool) return `<span style="color:#ff456d">${bool}</span>`;
|
|
if (nil) return `<span style="color:#888">${nil}</span>`;
|
|
return _;
|
|
},
|
|
);
|
|
}
|
|
|
|
// ── Extraction via Web Worker ──────────────────────────────────────────
|
|
// WASM extraction is CPU-bound and blocks whichever thread runs it.
|
|
// Spawning a fresh inline Worker per extraction keeps the main thread free:
|
|
// - setTimeout, repaint, and input events all work during extraction
|
|
// - worker.terminate() is a real kill switch (unlike Promise.race +
|
|
// setTimeout, which can never fire on a blocked main-thread event loop)
|
|
// - ArrayBuffer is transferred (not copied) to avoid peak memory doubling
|
|
// - Worker emits progress phases so the UI can show what is happening
|
|
function extractInWorker(fileBytes, mimeType, cdnBase, onProgress) {
|
|
return new Promise((resolve, reject) => {
|
|
const workerCode = `
|
|
self.onmessage = async (e) => {
|
|
const { fileBytes, mimeType, cdnBase } = e.data;
|
|
try {
|
|
self.postMessage({ type: 'progress', msg: 'Loading WASM module…' });
|
|
const mod = await import(cdnBase + '/dist/index.js');
|
|
self.postMessage({ type: 'progress', msg: 'Initializing…' });
|
|
await mod.initWasm(cdnBase + '/dist/pkg/kreuzberg_wasm_bg.wasm');
|
|
await mod.enableOcr();
|
|
self.postMessage({ type: 'progress', msg: 'Extracting content…' });
|
|
const blob = new Blob([new Uint8Array(fileBytes)]);
|
|
const result = await mod.extractFromFile(blob, mimeType || null);
|
|
// WASM bindgen instances expose data via prototype getters only;
|
|
// structured clone (postMessage) and JSON.stringify both miss them.
|
|
// Build a plain object from known scalar/array getters, then
|
|
// JSON-round-trip it so postMessage never sees a non-cloneable value.
|
|
const meta = result.metadata;
|
|
const plain = JSON.parse(JSON.stringify({
|
|
content: result.content ?? null,
|
|
mime_type: result.mimeType ?? null,
|
|
extraction_method: result.extractionMethod ?? null,
|
|
pages: result.pages
|
|
? Array.from(result.pages, p => ({ content: p.content, page_number: p.pageNumber }))
|
|
: [],
|
|
metadata: meta ? {
|
|
format: meta.format ?? null,
|
|
title: meta.title ?? null,
|
|
language: meta.language ?? null,
|
|
ocr_used: meta.ocrUsed ?? null,
|
|
is_empty: meta.isEmpty ?? null,
|
|
} : null,
|
|
}, (_k, v) => v instanceof Map ? Object.fromEntries(v) : v));
|
|
self.postMessage({ type: 'result', ok: true, result: plain });
|
|
} catch (err) {
|
|
self.postMessage({ type: 'result', ok: false, error: err?.message || String(err) });
|
|
}
|
|
};
|
|
`;
|
|
const blobUrl = URL.createObjectURL(
|
|
new Blob([workerCode], { type: "application/javascript" }),
|
|
);
|
|
const worker = new Worker(blobUrl, { type: "module" });
|
|
|
|
const TIMEOUT_MS = 30_000;
|
|
// cleanup() is the single teardown path — used by timeout, onmessage,
|
|
// onerror, and onmessageerror so terminate/revokeObjectURL are never
|
|
// called twice from different paths.
|
|
let cleaned = false;
|
|
const cleanup = () => {
|
|
if (cleaned) return;
|
|
cleaned = true;
|
|
clearTimeout(timer);
|
|
worker.terminate();
|
|
URL.revokeObjectURL(blobUrl);
|
|
};
|
|
const timer = setTimeout(() => {
|
|
cleanup();
|
|
reject(
|
|
new Error(
|
|
"Extraction timed out after 30s. The file may be too complex for in-browser processing. Try a simpler file or use the API.",
|
|
),
|
|
);
|
|
}, TIMEOUT_MS);
|
|
|
|
worker.onmessage = (e) => {
|
|
if (e.data.type === "progress") { onProgress?.(e.data.msg); return; }
|
|
cleanup();
|
|
if (e.data.ok) resolve(e.data.result);
|
|
else reject(new Error(e.data.error));
|
|
};
|
|
|
|
worker.onerror = (e) => { cleanup(); reject(new Error(e.message || "Worker error")); };
|
|
worker.onmessageerror = (e) => { cleanup(); reject(new Error("Worker message deserialisation failed")); };
|
|
|
|
// Transfer the ArrayBuffer so the worker owns the memory — avoids
|
|
// a full copy and halves peak memory for large files.
|
|
worker.postMessage({ fileBytes, mimeType, cdnBase }, [fileBytes]);
|
|
});
|
|
}
|
|
|
|
// ── Process File ───────────────────────────────────────────────────────
|
|
async function processFile(file) {
|
|
if (!wasmCdnBase) {
|
|
const errorBox = document.getElementById("error-message");
|
|
if (errorBox) {
|
|
errorBox.style.display = "flex";
|
|
document.getElementById("error-title").textContent = "Not ready";
|
|
document.getElementById("error-desc").textContent =
|
|
"WASM failed to load. Please refresh and try again.";
|
|
}
|
|
return;
|
|
}
|
|
|
|
uploadScreen.style.display = "none";
|
|
document.getElementById("file-name-display").textContent = file.name;
|
|
document.getElementById("file-size-display").textContent =
|
|
`${(file.size / 1024 / 1024).toFixed(2)}MB`;
|
|
btnBack.style.display = "flex";
|
|
fileInfo.style.display = "flex";
|
|
|
|
if (file.size > 1024 * 1024) {
|
|
const errorBox = document.getElementById("error-message");
|
|
if (errorBox) {
|
|
errorBox.style.display = "flex";
|
|
fileInfo.style.borderColor = "#FF456D";
|
|
document.getElementById("error-title").textContent = "File too large";
|
|
document.getElementById("error-desc").textContent =
|
|
"File exceeds the 1MB sandbox limit. Reduce file size or use the API for larger documents.";
|
|
}
|
|
return;
|
|
}
|
|
|
|
// Clear stale output immediately so the user never sees previous results
|
|
// while a new extraction is in progress.
|
|
const mdEl = document.getElementById("content-markdown");
|
|
if (mdEl)
|
|
mdEl.innerHTML =
|
|
'<p style="color:#888; text-align:center; margin-top:2rem;">Extracting content...</p>';
|
|
const jsonEl = document.getElementById("content-json");
|
|
if (jsonEl) jsonEl.innerHTML = '{ "status": "processing" }';
|
|
|
|
outputScreen.style.display = "flex";
|
|
const extractPhase = log.phase(`extract ${file.name}`);
|
|
try {
|
|
const fileBytes = await file.arrayBuffer();
|
|
log.info("worker dispatched", { file: file.name, bytes: fileBytes.byteLength, mime: file.type || "(auto)" });
|
|
const result = await extractInWorker(
|
|
fileBytes,
|
|
file.type || null,
|
|
wasmCdnBase,
|
|
(msg) => {
|
|
if (mdEl) mdEl.innerHTML = `<p style="color:#888;text-align:center;margin-top:2rem">${msg}</p>`;
|
|
},
|
|
);
|
|
const elapsedMs = extractPhase.end();
|
|
log.info("extraction complete", { file: file.name, chars: result.content?.length ?? 0, ms: Math.round(elapsedMs) });
|
|
|
|
if (mdEl)
|
|
mdEl.innerHTML = result.content
|
|
? renderMarkdown(result.content)
|
|
: '<p style="color:#888">No content extracted.</p>';
|
|
if (jsonEl) jsonEl.innerHTML = highlightJson(result);
|
|
} catch (e) {
|
|
extractPhase.end();
|
|
log.error("extraction failed", e);
|
|
outputScreen.style.display = "none";
|
|
const errorBox = document.getElementById("error-message");
|
|
if (errorBox) {
|
|
errorBox.style.display = "flex";
|
|
fileInfo.style.borderColor = "#FF456D";
|
|
document.getElementById("error-title").textContent = "Extraction failed";
|
|
document.getElementById("error-desc").textContent =
|
|
e instanceof Error ? e.message : String(e);
|
|
}
|
|
}
|
|
}
|
|
|
|
// ── Reset ──────────────────────────────────────────────────────────────
|
|
function resetUI() {
|
|
fileInput.value = "";
|
|
fileInfo.style.display = "none";
|
|
fileInfo.style.borderColor = "";
|
|
outputScreen.style.display = "none";
|
|
btnBack.style.display = "none";
|
|
const errorBox = document.getElementById("error-message");
|
|
if (errorBox) errorBox.style.display = "none";
|
|
uploadScreen.style.display = "flex";
|
|
}
|
|
btnBack.addEventListener("click", resetUI, false);
|
|
document.getElementById("btn-remove-file").addEventListener("click", resetUI, false);
|
|
|
|
// ── Copy ───────────────────────────────────────────────────────────────
|
|
const copyFeedback = document.getElementById("copy-feedback");
|
|
document.getElementById("btn-copy").addEventListener("click", async () => {
|
|
const mdEl = document.getElementById("content-markdown");
|
|
const jsonEl = document.getElementById("content-json");
|
|
const isMd = mdEl?.classList.contains("block");
|
|
const text = isMd ? mdEl.innerText : jsonEl.innerText;
|
|
if (!text) return;
|
|
try {
|
|
await navigator.clipboard.writeText(text.trim());
|
|
if (copyFeedback) {
|
|
copyFeedback.classList.remove("opacity-0");
|
|
copyFeedback.classList.add("opacity-100");
|
|
setTimeout(() => {
|
|
copyFeedback.classList.remove("opacity-100");
|
|
copyFeedback.classList.add("opacity-0");
|
|
}, 2000);
|
|
}
|
|
} catch (err) {
|
|
log.warn("clipboard write failed", err);
|
|
}
|
|
});
|
|
|
|
// ── Tab Switcher ───────────────────────────────────────────────────────
|
|
window.switchTab = (tab) => {
|
|
const ind = document.getElementById("tab-indicator");
|
|
const mdBtn = document.getElementById("tab-markdown");
|
|
const jsonBtn = document.getElementById("tab-json");
|
|
const mdEl = document.getElementById("content-markdown");
|
|
const jsonEl = document.getElementById("content-json");
|
|
const isMd = tab === "markdown";
|
|
ind.style.transform = isMd ? "translateX(0)" : "translateX(100%)";
|
|
mdBtn.classList.toggle("text-black", isMd);
|
|
mdBtn.classList.toggle("text-green", !isMd);
|
|
jsonBtn.classList.toggle("text-black", !isMd);
|
|
jsonBtn.classList.toggle("text-green", isMd);
|
|
mdEl.classList.toggle("hidden", !isMd);
|
|
mdEl.classList.toggle("block", isMd);
|
|
jsonEl.classList.toggle("hidden", isMd);
|
|
jsonEl.classList.toggle("block", !isMd);
|
|
};
|
|
</script>
|
|
</body>
|
|
</html>
|