Overview

// The two voice endpoints, as fetch handlers. Each framework's snippet mounts
// them; see the framework folders.

import { createTask } from "@pipedog/asyncio";
import { TransportParams } from "@pipedog/core/transports/base_transport";
import { ProtobufFrameSerializer } from "@pipedog/serializer-protobuf/protobuf";
import { smallWebRTCOfferHandler } from "@pipedog/transport-smallwebrtc/http";
import { SmallWebRTCRequestHandler } from "@pipedog/transport-smallwebrtc/request_handler";
import { SmallWebRTCTransport } from "@pipedog/transport-smallwebrtc/transport";
import {
  FastAPIWebsocketParams,
  FastAPIWebsocketTransport,
} from "@pipedog/transport-websocket/fastapi";
import { websocketFetchHandler } from "@pipedog/transport-websocket/fetch";
import type { WebsocketConnection } from "@pipedog/transport-websocket/socket";
import { runAgent } from "./agent.ts";

/**
 * WebRTC signalling: POST an SDP offer, get the answer. The media runs over UDP,
 * so this needs a long-lived server, not serverless functions.
 */
export const offer = smallWebRTCOfferHandler(new SmallWebRTCRequestHandler(), (connection) => {
  createTask(
    () =>
      runAgent(
        new SmallWebRTCTransport(
          connection,
          new TransportParams({ audioInEnabled: true, audioOutEnabled: true }),
        ),
      ),
    "voice-agent",
  );
});

/** Runs one agent on a WebSocket connection, with protobuf audio frames. */
export function runSocketAgent(connection: WebsocketConnection): Promise<void> {
  return runAgent(
    new FastAPIWebsocketTransport(
      connection,
      new FastAPIWebsocketParams({
        audioInEnabled: true,
        audioOutEnabled: true,
        serializer: new ProtobufFrameSerializer(),
      }),
    ),
  );
}

/** A WebSocket endpoint as a fetch handler, where the runtime upgrades from fetch. */
export const socket = websocketFetchHandler(runSocketAgent);

/** Rejects requests without a valid session token before the endpoint runs. */
function requireSession(endpoint: (request: Request) => Promise<Response>) {
  return async (request: Request): Promise<Response> => {
    const url = new URL(request.url);
    // Browsers can't set headers on a WebSocket, so a socket brings its token in the URL.
    const token =
      request.headers.get("authorization")?.replace(/^Bearer /, "") ??
      url.searchParams.get("token");
    if (!token || !(await verifySessionToken(token))) {
      return new Response("unauthorized", { status: 401 });
    }
    return endpoint(request);
  };
}

export const authedOffer = requireSession(offer);
export const authedSocket = requireSession(socket);