Spaces:

HelloSun
/

dosllm

Running

dosllm / bridge.js

Add 小貓中文: Chinese questions and answers on the DOS screen (Big5, Zhuyin); v86 patch of XiaoMao; updated dev package

1cd0943 verified 12 days ago

8.57 kB

/*
 * HTTP -> HTTPS bridge for a v86 DOS guest, restricted to Gradio Spaces.
 *
 * DOS (mTCP) can only speak plain TCP: no TLS. v86's "wisp" network backend terminates the guest's TCP
 * streams; this file plays the WISP *server* inside the page. For every TCP stream the guest opens to
 * port 80 it collects one HTTP request, performs the same request with fetch() over HTTPS, and writes the
 * HTTP response back to the guest. Hugging Face Spaces answer CORS requests, so this works from a static page.
 *
 * Chinese: XiaoMao Chinese shows Big5 on the DOS screen, Gradio speaks UTF-8/JSON. When the DOS request body
 * contains Big5 bytes the bridge decodes them (TextDecoder('big5')) before the HTTPS request, and the JSON reply
 * (\uXXXX escapes) is converted to Big5 bytes (see big5.js) before it is handed to DOS.
 *
 * Safety: only hosts matching *.hf.space and only paths starting with /gradio_api/ are relayed (GET/POST,
 * request body <= 8 KB). It is not an open proxy. No credentials are ever added.
 *
 * Install with GradioBridge.install(); then give v86 `relay_url: GradioBridge.relayUrl('dos')`.
 */
(function () {
  'use strict';
  const RealWebSocket = window.WebSocket;
  const FAKE_HOST = 'dosllm-bridge.invalid';   // fake WISP endpoint: never resolved, only recognised by this file
  const HOST_RE = /^[a-z0-9][a-z0-9-]*\.hf\.space$/i;
  const MAX_BODY = 8192;
  const TIMEOUT_MS = 150000;
  const WINDOW = 128;

  const bridge = { onEvent: null, requests: 0 };
  const emit = (machine, text) => { if (typeof bridge.onEvent === 'function') bridge.onEvent(machine, text); };
  const u32 = (n) => [n & 255, (n >> 8) & 255, (n >> 16) & 255, (n >>> 24) & 255];
  const frame = (type, id, payload) => {
    const body = payload || [];
    const out = new Uint8Array(5 + body.length);
    out[0] = type; out.set(u32(id), 1); out.set(body, 5);
    return out;
  };
  const concat = (parts) => {
    const out = new Uint8Array(parts.reduce((n, p) => n + p.length, 0));
    let o = 0; for (const p of parts) { out.set(p, o); o += p.length; }
    return out;
  };
  const indexOfCrlf2 = (b) => { for (let i = 0; i + 3 < b.length; i++) if (b[i] === 13 && b[i + 1] === 10 && b[i + 2] === 13 && b[i + 3] === 10) return i; return -1; };

  class FakeWispServer {
    constructor(url) {
      this.machine = (/[?&]m=([^&]+)/.exec(url) || [])[1] || 'dos';
      this.binaryType = 'arraybuffer';
      this.readyState = 0;
      this.streams = new Map();
      this.onopen = this.onmessage = this.onclose = this.onerror = null;
      setTimeout(() => {
        this.readyState = 1;
        if (this.onopen) this.onopen({ type: 'open' });
        this.reply(frame(3, 0, u32(WINDOW)));
      }, 0);
    }
    reply(bytes) {
      if (this.readyState !== 1 || !this.onmessage) return;
      this.onmessage({ data: bytes.buffer.slice(bytes.byteOffset, bytes.byteOffset + bytes.byteLength) });
    }
    send(data) {
      const f = new Uint8Array(data.buffer ? data.buffer.slice(data.byteOffset, data.byteOffset + data.byteLength) : data);
      const type = f[0], id = new DataView(f.buffer).getUint32(1, true);
      if (type === 1) this.connect(id, f);
      else if (type === 2) this.data(id, f.subarray(5));
      else if (type === 4) this.streams.delete(id);
    }
    connect(id, f) {
      const port = f[6] | (f[7] << 8);
      if (port !== 80) {
        emit(this.machine, `TCP port ${port} refused (this bridge only relays plain HTTP on port 80)`);
        this.reply(frame(4, id, [0x44]));
        return;
      }
      this.streams.set(id, { chunks: [], handled: false });
    }
    data(id, bytes) {
      const s = this.streams.get(id);
      if (!s || s.handled) return;
      s.chunks.push(bytes.slice());
      this.reply(frame(3, id, u32(WINDOW)));
      const all = concat(s.chunks);
      const end = indexOfCrlf2(all);
      if (end < 0) return;
      const head = new TextDecoder().decode(all.subarray(0, end));
      const len = Math.min(parseInt((/content-length:\s*(\d+)/i.exec(head) || [])[1] || '0', 10), MAX_BODY + 1);
      if (all.length < end + 4 + len) return;                 // body not complete yet
      s.handled = true;
      this.relay(id, head, all.subarray(end + 4, end + 4 + len));
    }
    async respond(id, status, text, type, bodyBytes) {
      const body = bodyBytes || new TextEncoder().encode(text);
      const head = new TextEncoder().encode(`HTTP/1.0 ${status}\r\nContent-Type: ${type || 'text/plain'}\r\nContent-Length: ${body.length}\r\nConnection: close\r\n\r\n`);
      const all = concat([head, body]);
      for (let o = 0; o < all.length; o += 1024) this.reply(frame(2, id, all.subarray(o, o + 1024)));
      this.reply(frame(4, id, [0x02]));                        // close: guest sees FIN after the last data
      this.streams.delete(id);
    }
    async relay(id, head, body) {
      const m = /^(GET|POST) (\S+) HTTP\/1\.[01]/.exec(head);
      const host = ((/^host:\s*(\S+)/im.exec(head) || [])[1] || '').toLowerCase();
      if (!m || !HOST_RE.test(host) || !m[2].startsWith('/gradio_api/') || body.length > MAX_BODY) {
        emit(this.machine, `refused ${m ? m[1] + ' ' + host + m[2] : 'malformed request'} (only *.hf.space /gradio_api/ is relayed)`);
        return this.respond(id, '403 Forbidden', 'blocked by bridge');
      }
      const url = `https://${host}${m[2]}`;
      bridge.requests += 1;
      const t0 = performance.now();
      emit(this.machine, `DOS -> ${m[1]} ${url}${m[1] === 'POST' ? ` (${body.length} bytes)` : ''}`);
      const ctl = new AbortController();
      const timer = setTimeout(() => ctl.abort(), TIMEOUT_MS);
      try {
        const init = { method: m[1], signal: ctl.signal, headers: {} };
        if (m[1] === 'POST') {
          const big5 = window.Big5 && window.Big5.available && body.some((b) => b >= 0x80);
          init.body = big5 ? window.Big5.decode(body) : new TextDecoder().decode(body);
          if (big5) emit(this.machine, `request body has Big5 bytes: decoded to UTF-8 ("${init.body.replace(/^.*?\[\"/, '').slice(0, 60)}...")`);
          init.headers['Content-Type'] = 'application/json';
        }
        const resp = await fetch(url, init);
        const bytes = new Uint8Array(await resp.arrayBuffer());
        const secs = ((performance.now() - t0) / 1000).toFixed(1);
        emit(this.machine, `HTTPS ${resp.status} ${resp.headers.get('content-type') || ''} ${bytes.length} bytes in ${secs}s`);
        if (/text\/event-stream/.test(resp.headers.get('content-type') || '')) {
          const text = new TextDecoder().decode(bytes);
          const frames = (text.match(/^event: generating/gm) || []).length;
          const done = /event: complete\ndata: (.*)/.exec(text);
          let answer = '';
          try { answer = JSON.parse(done[1])[0][0]; } catch (e) { /* not a chat reply */ }
          emit(this.machine, `Gradio stream: ${frames} generating frames, ${done ? 'complete' : 'NO complete event'}`);
          if (answer) emit(this.machine, `Gradio answer: ${answer}`);
          bridge.lastAnswer = answer;
        }
        let out = bytes;
        const ct = resp.headers.get('content-type') || '';
        if (window.Big5 && window.Big5.available && /json|event-stream/.test(ct)) {
          out = window.Big5.jsonToBig5Bytes(new TextDecoder().decode(bytes));   // UTF-8/JSON -> Big5 bytes for the DOS screen
          if (out.some((b) => b >= 0x80)) emit(this.machine, `reply contains Chinese: converted to Big5 for DOS (${bytes.length} -> ${out.length} bytes)`);
        }
        await this.respond(id, `${resp.status} ${resp.statusText || 'OK'}`, '', ct || 'text/plain', out);
      } catch (e) {
        emit(this.machine, `HTTPS request failed: ${e && e.message ? e.message : e}`);
        await this.respond(id, '502 Bad Gateway', 'bridge: upstream request failed');
      } finally { clearTimeout(timer); }
    }
    close() { this.readyState = 3; this.streams.clear(); if (this.onclose) this.onclose({ type: 'close' }); }
  }
  bridge.install = function (opts) {
    Object.assign(bridge, opts || {});
    if (window.WebSocket.__gradioBridge) return;
    const patched = new Proxy(RealWebSocket, {
      construct(target, args) {
        if (String(args[0]).indexOf(FAKE_HOST) >= 0) return new FakeWispServer(String(args[0]));
        return Reflect.construct(target, args);
      },
    });
    patched.__gradioBridge = true;
    window.WebSocket = patched;
  };
  bridge.relayUrl = (machine) => `wisp://${FAKE_HOST}/?m=${encodeURIComponent(machine || 'dos')}`;
  window.GradioBridge = bridge;
})();

	/*
	* HTTP -> HTTPS bridge for a v86 DOS guest, restricted to Gradio Spaces.
	*
	* DOS (mTCP) can only speak plain TCP: no TLS. v86's "wisp" network backend terminates the guest's TCP
	* streams; this file plays the WISP server inside the page. For every TCP stream the guest opens to
	* port 80 it collects one HTTP request, performs the same request with fetch() over HTTPS, and writes the
	* HTTP response back to the guest. Hugging Face Spaces answer CORS requests, so this works from a static page.
	*
	* Chinese: XiaoMao Chinese shows Big5 on the DOS screen, Gradio speaks UTF-8/JSON. When the DOS request body
	* contains Big5 bytes the bridge decodes them (TextDecoder('big5')) before the HTTPS request, and the JSON reply
	* (\uXXXX escapes) is converted to Big5 bytes (see big5.js) before it is handed to DOS.
	*
	* Safety: only hosts matching *.hf.space and only paths starting with /gradio_api/ are relayed (GET/POST,
	* request body <= 8 KB). It is not an open proxy. No credentials are ever added.
	*
	* Install with GradioBridge.install(); then give v86 `relay_url: GradioBridge.relayUrl('dos')`.
	*/
	(function () {
	'use strict';
	const RealWebSocket = window.WebSocket;
	const FAKE_HOST = 'dosllm-bridge.invalid'; // fake WISP endpoint: never resolved, only recognised by this file
	const HOST_RE = /^[a-z0-9][a-z0-9-]*\.hf\.space$/i;
	const MAX_BODY = 8192;
	const TIMEOUT_MS = 150000;
	const WINDOW = 128;

	const bridge = { onEvent: null, requests: 0 };
	const emit = (machine, text) => { if (typeof bridge.onEvent === 'function') bridge.onEvent(machine, text); };
	const u32 = (n) => [n & 255, (n >> 8) & 255, (n >> 16) & 255, (n >>> 24) & 255];
	const frame = (type, id, payload) => {
	const body = payload \|\| [];
	const out = new Uint8Array(5 + body.length);
	out[0] = type; out.set(u32(id), 1); out.set(body, 5);
	return out;
	};
	const concat = (parts) => {
	const out = new Uint8Array(parts.reduce((n, p) => n + p.length, 0));
	let o = 0; for (const p of parts) { out.set(p, o); o += p.length; }
	return out;
	};
	const indexOfCrlf2 = (b) => { for (let i = 0; i + 3 < b.length; i++) if (b[i] === 13 && b[i + 1] === 10 && b[i + 2] === 13 && b[i + 3] === 10) return i; return -1; };

	class FakeWispServer {
	constructor(url) {
	this.machine = (/[?&]m=([^&]+)/.exec(url) \|\| [])[1] \|\| 'dos';
	this.binaryType = 'arraybuffer';
	this.readyState = 0;
	this.streams = new Map();
	this.onopen = this.onmessage = this.onclose = this.onerror = null;
	setTimeout(() => {
	this.readyState = 1;
	if (this.onopen) this.onopen({ type: 'open' });
	this.reply(frame(3, 0, u32(WINDOW)));
	}, 0);
	}
	reply(bytes) {
	if (this.readyState !== 1 \|\| !this.onmessage) return;
	this.onmessage({ data: bytes.buffer.slice(bytes.byteOffset, bytes.byteOffset + bytes.byteLength) });
	}
	send(data) {
	const f = new Uint8Array(data.buffer ? data.buffer.slice(data.byteOffset, data.byteOffset + data.byteLength) : data);
	const type = f[0], id = new DataView(f.buffer).getUint32(1, true);
	if (type === 1) this.connect(id, f);
	else if (type === 2) this.data(id, f.subarray(5));
	else if (type === 4) this.streams.delete(id);
	}
	connect(id, f) {
	const port = f[6] \| (f[7] << 8);
	if (port !== 80) {
	emit(this.machine, `TCP port ${port} refused (this bridge only relays plain HTTP on port 80)`);
	this.reply(frame(4, id, [0x44]));
	return;
	}
	this.streams.set(id, { chunks: [], handled: false });
	}
	data(id, bytes) {
	const s = this.streams.get(id);
	if (!s \|\| s.handled) return;
	s.chunks.push(bytes.slice());
	this.reply(frame(3, id, u32(WINDOW)));
	const all = concat(s.chunks);
	const end = indexOfCrlf2(all);
	if (end < 0) return;
	const head = new TextDecoder().decode(all.subarray(0, end));
	const len = Math.min(parseInt((/content-length:\s*(\d+)/i.exec(head) \|\| [])[1] \|\| '0', 10), MAX_BODY + 1);
	if (all.length < end + 4 + len) return; // body not complete yet
	s.handled = true;
	this.relay(id, head, all.subarray(end + 4, end + 4 + len));
	}
	async respond(id, status, text, type, bodyBytes) {
	const body = bodyBytes \|\| new TextEncoder().encode(text);
	const head = new TextEncoder().encode(`HTTP/1.0 ${status}\r\nContent-Type: ${type \|\| 'text/plain'}\r\nContent-Length: ${body.length}\r\nConnection: close\r\n\r\n`);
	const all = concat([head, body]);
	for (let o = 0; o < all.length; o += 1024) this.reply(frame(2, id, all.subarray(o, o + 1024)));
	this.reply(frame(4, id, [0x02])); // close: guest sees FIN after the last data
	this.streams.delete(id);
	}
	async relay(id, head, body) {
	const m = /^(GET\|POST) (\S+) HTTP\/1\.[01]/.exec(head);
	const host = ((/^host:\s*(\S+)/im.exec(head) \|\| [])[1] \|\| '').toLowerCase();
	if (!m \|\| !HOST_RE.test(host) \|\| !m[2].startsWith('/gradio_api/') \|\| body.length > MAX_BODY) {
	emit(this.machine, `refused ${m ? m[1] + ' ' + host + m[2] : 'malformed request'} (only *.hf.space /gradio_api/ is relayed)`);
	return this.respond(id, '403 Forbidden', 'blocked by bridge');
	}
	const url = `https://${host}${m[2]}`;
	bridge.requests += 1;
	const t0 = performance.now();
	emit(this.machine, `DOS -> ${m[1]} ${url}${m[1] === 'POST' ? ` (${body.length} bytes)` : ''}`);
	const ctl = new AbortController();
	const timer = setTimeout(() => ctl.abort(), TIMEOUT_MS);
	try {
	const init = { method: m[1], signal: ctl.signal, headers: {} };
	if (m[1] === 'POST') {
	const big5 = window.Big5 && window.Big5.available && body.some((b) => b >= 0x80);
	init.body = big5 ? window.Big5.decode(body) : new TextDecoder().decode(body);
	if (big5) emit(this.machine, `request body has Big5 bytes: decoded to UTF-8 ("${init.body.replace(/^.*?\[\"/, '').slice(0, 60)}...")`);
	init.headers['Content-Type'] = 'application/json';
	}
	const resp = await fetch(url, init);
	const bytes = new Uint8Array(await resp.arrayBuffer());
	const secs = ((performance.now() - t0) / 1000).toFixed(1);
	emit(this.machine, `HTTPS ${resp.status} ${resp.headers.get('content-type') \|\| ''} ${bytes.length} bytes in ${secs}s`);
	if (/text\/event-stream/.test(resp.headers.get('content-type') \|\| '')) {
	const text = new TextDecoder().decode(bytes);
	const frames = (text.match(/^event: generating/gm) \|\| []).length;
	const done = /event: complete\ndata: (.*)/.exec(text);
	let answer = '';
	try { answer = JSON.parse(done[1])[0][0]; } catch (e) { /* not a chat reply */ }
	emit(this.machine, `Gradio stream: ${frames} generating frames, ${done ? 'complete' : 'NO complete event'}`);
	if (answer) emit(this.machine, `Gradio answer: ${answer}`);
	bridge.lastAnswer = answer;
	}
	let out = bytes;
	const ct = resp.headers.get('content-type') \|\| '';
	if (window.Big5 && window.Big5.available && /json\|event-stream/.test(ct)) {
	out = window.Big5.jsonToBig5Bytes(new TextDecoder().decode(bytes)); // UTF-8/JSON -> Big5 bytes for the DOS screen
	if (out.some((b) => b >= 0x80)) emit(this.machine, `reply contains Chinese: converted to Big5 for DOS (${bytes.length} -> ${out.length} bytes)`);
	}
	await this.respond(id, `${resp.status} ${resp.statusText \|\| 'OK'}`, '', ct \|\| 'text/plain', out);
	} catch (e) {
	emit(this.machine, `HTTPS request failed: ${e && e.message ? e.message : e}`);
	await this.respond(id, '502 Bad Gateway', 'bridge: upstream request failed');
	} finally { clearTimeout(timer); }
	}
	close() { this.readyState = 3; this.streams.clear(); if (this.onclose) this.onclose({ type: 'close' }); }
	}

	bridge.install = function (opts) {
	Object.assign(bridge, opts \|\| {});
	if (window.WebSocket.__gradioBridge) return;
	const patched = new Proxy(RealWebSocket, {
	construct(target, args) {
	if (String(args[0]).indexOf(FAKE_HOST) >= 0) return new FakeWispServer(String(args[0]));
	return Reflect.construct(target, args);
	},
	});
	patched.__gradioBridge = true;
	window.WebSocket = patched;
	};
	bridge.relayUrl = (machine) => `wisp://${FAKE_HOST}/?m=${encodeURIComponent(machine \|\| 'dos')}`;
	window.GradioBridge = bridge;
	})();