@cartesia/cartesia-js 0.0.3 → 1.0.0-alpha.1

package/dist/react/index.js

@@ -1,514 +1,136 @@
-"use strict";
-var __create = Object.create;
-var __defProp = Object.defineProperty;
-var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
-var __getOwnPropNames = Object.getOwnPropertyNames;
-var __getOwnPropSymbols = Object.getOwnPropertySymbols;
-var __getProtoOf = Object.getPrototypeOf;
-var __hasOwnProp = Object.prototype.hasOwnProperty;
-var __propIsEnum = Object.prototype.propertyIsEnumerable;
-var __knownSymbol = (name, symbol) => {
-  return (symbol = Symbol[name]) ? symbol : Symbol.for("Symbol." + name);
-};
-var __defNormalProp = (obj, key, value) => key in obj ? __defProp(obj, key, { enumerable: true, configurable: true, writable: true, value }) : obj[key] = value;
-var __spreadValues = (a, b) => {
-  for (var prop in b || (b = {}))
-    if (__hasOwnProp.call(b, prop))
-      __defNormalProp(a, prop, b[prop]);
-  if (__getOwnPropSymbols)
-    for (var prop of __getOwnPropSymbols(b)) {
-      if (__propIsEnum.call(b, prop))
-        __defNormalProp(a, prop, b[prop]);
-    }
-  return a;
-};
-var __export = (target, all) => {
-  for (var name in all)
-    __defProp(target, name, { get: all[name], enumerable: true });
-};
-var __copyProps = (to, from, except, desc) => {
-  if (from && typeof from === "object" || typeof from === "function") {
-    for (let key of __getOwnPropNames(from))
-      if (!__hasOwnProp.call(to, key) && key !== except)
-        __defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
-  }
-  return to;
-};
-var __toESM = (mod, isNodeMode, target) => (target = mod != null ? __create(__getProtoOf(mod)) : {}, __copyProps(
-  // If the importer is in node compatibility mode or this is not an ESM
-  // file that has been converted to a CommonJS file using a Babel-
-  // compatible transform (i.e. "__esModule" has not been set), then set
-  // "default" to the CommonJS "module.exports" for node compatibility.
-  isNodeMode || !mod || !mod.__esModule ? __defProp(target, "default", { value: mod, enumerable: true }) : target,
-  mod
-));
-var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
-var __async = (__this, __arguments, generator) => {
-  return new Promise((resolve, reject) => {
-    var fulfilled = (value) => {
-      try {
-        step(generator.next(value));
-      } catch (e) {
-        reject(e);
-      }
-    };
-    var rejected = (value) => {
-      try {
-        step(generator.throw(value));
-      } catch (e) {
-        reject(e);
-      }
-    };
-    var step = (x) => x.done ? resolve(x.value) : Promise.resolve(x.value).then(fulfilled, rejected);
-    step((generator = generator.apply(__this, __arguments)).next());
-  });
-};
-var __forAwait = (obj, it, method) => (it = obj[__knownSymbol("asyncIterator")]) ? it.call(obj) : (obj = obj[__knownSymbol("iterator")](), it = {}, method = (key, fn) => (fn = obj[key]) && (it[key] = (arg) => new Promise((yes, no, done) => (arg = fn.call(obj, arg), done = arg.done, Promise.resolve(arg.value).then((value) => yes({ value, done }), no)))), method("next"), method("return"), it);
-
-// src/react/index.ts
-var react_exports = {};
-__export(react_exports, {
-  useAudio: () => useAudio
-});
-module.exports = __toCommonJS(react_exports);
-var import_react = require("react");
-
-// src/audio/index.ts
-var import_emittery = __toESM(require("emittery"));
-var import_human_id = require("human-id");
-var import_partysocket = require("partysocket");
-
-// src/lib/constants.ts
-var BASE_URL = "https://api.cartesia.ai/v0";
-var SAMPLE_RATE = 44100;
-var constructWebsocketUrl = (baseUrl) => {
-  return new URL(`${baseUrl.replace(/^http/, "ws")}/audio/websocket`);
-};
-
-// src/lib/client.ts
-var Client = class {
-  constructor(options = {}) {
-    if (!(options.apiKey || process.env.CARTESIA_API_KEY)) {
-      throw new Error("Missing Cartesia API key.");
-    }
-    this.apiKey = options.apiKey || process.env.CARTESIA_API_KEY;
-    this.baseUrl = options.baseUrl || BASE_URL;
-  }
-};
-
-// src/audio/utils.ts
-var import_base64_js = __toESM(require("base64-js"));
-function getBufferDuration(b64) {
-  const floats = base64ToArray(b64);
-  return floats.length / SAMPLE_RATE;
-}
-function base64ToArray(b64) {
-  return filterSentinel(b64).reduce((acc, b) => {
-    const floats = new Float32Array(import_base64_js.default.toByteArray(b).buffer);
-    const newAcc = new Float32Array(acc.length + floats.length);
-    newAcc.set(acc, 0);
-    newAcc.set(floats, acc.length);
-    return newAcc;
-  }, new Float32Array(0));
-}
-function playAudioBuffer(b64, context, maybeStartAt = null, onEnded = null) {
-  const startAt = maybeStartAt != null ? maybeStartAt : context.currentTime;
-  const floats = base64ToArray(b64);
-  const source = context.createBufferSource();
-  const buffer = context.createBuffer(1, floats.length, SAMPLE_RATE);
-  buffer.getChannelData(0).set(floats);
-  source.buffer = buffer;
-  source.connect(context.destination);
-  source.start(startAt);
-  source.onended = onEnded;
-  return buffer.duration;
-}
-function createMessageHandlerForContextId(contextId, handler) {
-  return (event) => {
-    const message = JSON.parse(event.data);
-    if (message.context_id !== contextId) {
-      return;
-    }
-    let chunk;
-    if (message.done) {
-      chunk = getSentinel();
-    } else {
-      chunk = message.data;
-    }
-    handler({ chunk, message });
-  };
-}
-function getSentinel() {
-  return null;
-}
-function isSentinel(x) {
-  return x === getSentinel();
-}
-function filterSentinel(collection) {
-  return collection.filter(
-    (x) => !isSentinel(x)
-  );
-}
-function isComplete(chunks) {
-  return isSentinel(chunks[chunks.length - 1]);
-}
-function getEmitteryCallbacks(emitter) {
-  return {
-    on: emitter.on.bind(emitter),
-    off: emitter.off.bind(emitter),
-    once: emitter.once.bind(emitter),
-    events: emitter.events.bind(emitter)
-  };
-}
-function bufferToWav(sampleRate, channelBuffers) {
-  const totalSamples = channelBuffers[0].length * channelBuffers.length;
-  const buffer = new ArrayBuffer(44 + totalSamples * 2);
-  const view = new DataView(buffer);
-  const writeString = (view2, offset2, string) => {
-    for (let i = 0; i < string.length; i++) {
-      view2.setUint8(offset2 + i, string.charCodeAt(i));
-    }
-  };
-  writeString(view, 0, "RIFF");
-  view.setUint32(4, 36 + totalSamples * 2, true);
-  writeString(view, 8, "WAVE");
-  writeString(view, 12, "fmt ");
-  view.setUint32(16, 16, true);
-  view.setUint16(20, 1, true);
-  view.setUint16(22, channelBuffers.length, true);
-  view.setUint32(24, sampleRate, true);
-  view.setUint32(28, sampleRate * 4, true);
-  view.setUint16(32, channelBuffers.length * 2, true);
-  view.setUint16(34, 16, true);
-  writeString(view, 36, "data");
-  view.setUint32(40, totalSamples * 2, true);
-  let offset = 44;
-  for (let i = 0; i < channelBuffers[0].length; i++) {
-    for (let channel = 0; channel < channelBuffers.length; channel++) {
-      const s = Math.max(-1, Math.min(1, channelBuffers[channel][i]));
-      view.setInt16(offset, s < 0 ? s * 32768 : s * 32767, true);
-      offset += 2;
-    }
-  }
-  return buffer;
-}
-
-// src/audio/index.ts
-var audio_default = class extends Client {
-  constructor() {
-    super(...arguments);
-    this.isConnected = false;
-  }
-  /**
-   * Stream audio from a model.
-   *
-   * @param inputs - Stream options. Includes a `model` key and some `parameters`, which
-   * are model-specific and can be found in the model's documentation.
-   * @param options - Options for the stream.
-   * @param options.timeout - The maximum time to wait for a chunk before cancelling the stream.
-   * If `0`, the stream will not time out.
-   * @returns An object with a method `play` of type `(bufferDuration: number) => Promise<void>`
-   * that plays the audio as it arrives, with `bufferDuration` seconds of audio buffered before
-   * starting playback.
-   */
-  stream(inputs, { timeout = 0 } = {}) {
-    var _a, _b, _c, _d;
-    if (!this.isConnected) {
-      throw new Error("Not connected to WebSocket. Call .connect() first.");
-    }
-    const contextId = this.generateId();
-    (_a = this.socket) == null ? void 0 : _a.send(
-      JSON.stringify({
-        data: inputs,
-        context_id: contextId
-      })
-    );
-    const streamCompleteController = new AbortController();
-    let timeoutId = null;
-    if (timeout > 0) {
-      timeoutId = setTimeout(streamCompleteController.abort, timeout);
-    }
-    const chunks = [];
-    const emitter = new import_emittery.default();
-    const handleMessage = createMessageHandlerForContextId(
-      contextId,
-      (_0) => __async(this, [_0], function* ({ chunk, message }) {
-        chunks.push(chunk);
-        yield emitter.emit("chunk", {
-          chunk,
-          chunks
-        });
-        yield emitter.emit("message", message);
-        if (isSentinel(chunk)) {
-          yield emitter.emit("streamed", {
-            chunks
-          });
-          streamCompleteController.abort();
-        } else if (timeoutId) {
-          clearTimeout(timeoutId);
-          timeoutId = setTimeout(streamCompleteController.abort, timeout);
-        }
-      })
-    );
-    (_b = this.socket) == null ? void 0 : _b.addEventListener("message", handleMessage, {
-      signal: streamCompleteController.signal
-    });
-    (_c = this.socket) == null ? void 0 : _c.addEventListener(
-      "close",
-      () => {
-        streamCompleteController.abort();
-      },
-      {
-        once: true
-      }
-    );
-    (_d = this.socket) == null ? void 0 : _d.addEventListener(
-      "error",
-      () => {
-        streamCompleteController.abort();
-      },
-      {
-        once: true
-      }
-    );
-    streamCompleteController.signal.addEventListener("abort", () => {
-      if (timeoutId) {
-        clearTimeout(timeoutId);
-      }
-      emitter.clearListeners();
-    });
-    const play = (_0) => __async(this, [_0], function* ({ bufferDuration }) {
-      const context = new AudioContext({
-        sampleRate: SAMPLE_RATE
-      });
-      let startNextPlaybackAt = 0;
-      const playLatestChunk = (chunk) => {
-        if (isSentinel(chunk)) {
-          return true;
-        }
-        startNextPlaybackAt = playAudioBuffer([chunk], context, startNextPlaybackAt) + Math.max(context.currentTime, startNextPlaybackAt);
-        return false;
-      };
-      const playChunks = (chunks2) => {
-        startNextPlaybackAt += playAudioBuffer(
-          chunks2,
-          context,
-          startNextPlaybackAt
-        );
-        if (isComplete(chunks2)) {
-          return;
-        }
-      };
-      const tryStart = (chunks2) => __async(this, null, function* () {
-        startNextPlaybackAt = context.currentTime;
-        if (isComplete(chunks2) || streamCompleteController.signal.aborted) {
-          playChunks(chunks2);
-          return true;
-        }
-        if (getBufferDuration(chunks2) > bufferDuration) {
-          playChunks(chunks2);
-          try {
-            for (var iter2 = __forAwait(emitter.events("chunk")), more2, temp2, error2; more2 = !(temp2 = yield iter2.next()).done; more2 = false) {
-              const { chunk } = temp2.value;
-              if (playLatestChunk(chunk)) {
-                break;
-              }
-            }
-          } catch (temp2) {
-            error2 = [temp2];
-          } finally {
-            try {
-              more2 && (temp2 = iter2.return) && (yield temp2.call(iter2));
-            } finally {
-              if (error2)
-                throw error2[0];
-            }
-          }
-          return true;
-        }
-        return false;
-      });
-      if (!(yield tryStart(chunks))) {
-        try {
-          for (var iter = __forAwait(emitter.events("chunk")), more, temp, error; more = !(temp = yield iter.next()).done; more = false) {
-            const { chunks: chunks2 } = temp.value;
-            if (yield tryStart(chunks2)) {
-              break;
-            }
-          }
-        } catch (temp) {
-          error = [temp];
-        } finally {
-          try {
-            more && (temp = iter.return) && (yield temp.call(iter));
-          } finally {
-            if (error)
-              throw error[0];
-          }
-        }
-      }
-    });
-    return __spreadValues({
-      play
-    }, getEmitteryCallbacks(emitter));
-  }
-  /**
-   * Generate a unique ID suitable for a streaming context.
-   *
-   * Not suitable for security purposes or as a primary key, since
-   * it lacks the amount of entropy required for those use cases.
-   *
-   * @returns A unique ID.
-   */
-  generateId() {
-    return (0, import_human_id.humanId)({
-      separator: "-",
-      capitalize: false
-    });
-  }
-  /**
-   * Authenticate and connect to a Cartesia streaming WebSocket.
-   *
-   * @returns A promise that resolves when the WebSocket is connected.
-   * @throws {Error} If the WebSocket fails to connect.
-   */
-  connect() {
-    const url = constructWebsocketUrl(this.baseUrl);
-    url.searchParams.set("api_key", this.apiKey);
-    const emitter = new import_emittery.default();
-    this.socket = new import_partysocket.WebSocket(url.toString());
-    this.socket.onopen = () => {
-      this.isConnected = true;
-      emitter.emit("open");
-    };
-    this.socket.onclose = () => {
-      this.isConnected = false;
-      emitter.emit("close");
-    };
-    return new Promise(
-      (resolve, reject) => {
-        var _a, _b, _c;
-        (_a = this.socket) == null ? void 0 : _a.addEventListener(
-          "open",
-          () => {
-            resolve(getEmitteryCallbacks(emitter));
-          },
-          {
-            once: true
-          }
-        );
-        const aborter = new AbortController();
-        (_b = this.socket) == null ? void 0 : _b.addEventListener(
-          "error",
-          () => {
-            aborter.abort();
-            reject(new Error("WebSocket failed to connect."));
-          },
-          {
-            signal: aborter.signal
-          }
-        );
-        (_c = this.socket) == null ? void 0 : _c.addEventListener(
-          "close",
-          () => {
-            aborter.abort();
-            reject(new Error("WebSocket closed before it could connect."));
-          },
-          {
-            signal: aborter.signal
-          }
-        );
-      }
-    );
-  }
-  /**
-   * Disconnect from the Cartesia streaming WebSocket.
-   */
-  disconnect() {
-    var _a;
-    (_a = this.socket) == null ? void 0 : _a.close();
-  }
-};
+import {
+  Cartesia
+} from "../chunk-X7SJMF2R.js";
+import "../chunk-WBK6LLXX.js";
+import {
+  pingServer
+} from "../chunk-3FL2SNIR.js";
+import "../chunk-4RMSIQLG.js";
+import "../chunk-NDNN326Q.js";
+import "../chunk-JOHSCOLW.js";
+import "../chunk-BCQ63627.js";
+import "../chunk-3GBZUGUD.js";
+import {
+  Player
+} from "../chunk-WE63M7PJ.js";
+import "../chunk-LYPTISWL.js";
+import {
+  __async
+} from "../chunk-WIFMLPT5.js";
 
 // src/react/index.ts
-function useAudio({ apiKey, baseUrl }) {
+import { useCallback, useEffect, useMemo, useRef, useState } from "react";
+var PING_INTERVAL = 5e3;
+var DEFAULT_BUFFER_DURATION = 0.01;
+function useTTS({
+  apiKey,
+  baseUrl,
+  sampleRate
+}) {
+  var _a, _b;
   if (typeof window === "undefined") {
     return {
-      stream: () => {
-      },
+      buffer: () => __async(this, null, function* () {
+      }),
       play: () => __async(this, null, function* () {
       }),
-      download: () => null,
+      pause: () => __async(this, null, function* () {
+      }),
+      resume: () => __async(this, null, function* () {
+      }),
+      toggle: () => __async(this, null, function* () {
+      }),
+      playbackStatus: "inactive",
+      bufferStatus: "inactive",
+      isWaiting: false,
+      source: null,
       isConnected: false,
-      isPlaying: false,
-      isStreamed: false,
-      chunks: [],
-      messages: []
+      metrics: {
+        modelLatency: null
+      }
     };
   }
-  const audio = (0, import_react.useMemo)(() => {
+  const websocket = useMemo(() => {
     if (!apiKey) {
       return null;
     }
-    const audio2 = new audio_default({ apiKey, baseUrl });
-    return audio2;
-  }, [apiKey, baseUrl]);
-  const streamReturn = (0, import_react.useRef)(null);
-  const [isStreamed, setIsStreamed] = (0, import_react.useState)(false);
-  const [isPlaying, setIsPlaying] = (0, import_react.useState)(false);
-  const [isConnected, setIsConnected] = (0, import_react.useState)(false);
-  const [chunks, setChunks] = (0, import_react.useState)([]);
-  const [messages, setMessages] = (0, import_react.useState)([]);
-  const stream = (0, import_react.useCallback)(
+    const cartesia = new Cartesia({ apiKey, baseUrl });
+    baseUrl = baseUrl != null ? baseUrl : cartesia.baseUrl;
+    return cartesia.tts.websocket({ sampleRate });
+  }, [apiKey, baseUrl, sampleRate]);
+  const websocketReturn = useRef(null);
+  const player = useRef(null);
+  const [playbackStatus, setPlaybackStatus] = useState("inactive");
+  const [bufferStatus, setBufferStatus] = useState("inactive");
+  const [isWaiting, setIsWaiting] = useState(false);
+  const [isConnected, setIsConnected] = useState(false);
+  const [bufferDuration, setBufferDuration] = useState(null);
+  const [messages, setMessages] = useState([]);
+  const buffer = useCallback(
     (options) => __async(this, null, function* () {
-      var _a;
-      streamReturn.current = (_a = audio == null ? void 0 : audio.stream(options)) != null ? _a : null;
-      if (!streamReturn.current) {
+      var _a2;
+      setMessages([]);
+      setBufferStatus("buffering");
+      websocketReturn.current = (_a2 = websocket == null ? void 0 : websocket.send(options)) != null ? _a2 : null;
+      if (!websocketReturn.current) {
         return;
       }
-      setMessages([]);
-      streamReturn.current.on(
-        "chunk",
-        ({ chunks: chunks3 }) => {
-          setChunks(chunks3);
-        }
-      );
-      streamReturn.current.on(
-        "message",
-        (message) => {
-          setMessages((messages2) => [...messages2, message]);
-        }
-      );
-      const { chunks: chunks2 } = yield streamReturn.current.once("streamed");
-      setChunks(chunks2);
-      setIsStreamed(true);
+      websocketReturn.current.on("message", (message) => {
+        setMessages((messages2) => [...messages2, JSON.parse(message)]);
+      });
+      yield websocketReturn.current.source.once("close");
+      setBufferStatus("buffered");
     }),
-    [audio]
+    [websocket]
   );
-  const download = (0, import_react.useCallback)(() => {
-    if (!isStreamed) {
-      return null;
+  const metrics = useMemo(() => {
+    var _a2;
+    if (messages.length === 0) {
+      return {
+        modelLatency: null
+      };
     }
-    const audio2 = bufferToWav(SAMPLE_RATE, [base64ToArray(chunks)]);
-    return new Blob([audio2], { type: "audio/wav" });
-  }, [isStreamed, chunks]);
-  (0, import_react.useEffect)(() => {
+    const modelLatency = (_a2 = messages[0].step_time) != null ? _a2 : null;
+    return {
+      modelLatency: Math.trunc(modelLatency)
+    };
+  }, [messages]);
+  useEffect(() => {
     let cleanup = () => {
     };
     function setupConnection() {
       return __async(this, null, function* () {
         try {
-          const connection = yield audio == null ? void 0 : audio.connect();
+          const connection = yield websocket == null ? void 0 : websocket.connect();
           if (!connection) {
             return;
           }
           setIsConnected(true);
+          connection.on("open", () => {
+            setIsConnected(true);
+          });
           const unsubscribe = connection.on("close", () => {
             setIsConnected(false);
           });
+          const intervalId = setInterval(() => {
+            if (baseUrl) {
+              pingServer(new URL(baseUrl).origin).then((ping) => {
+                let bufferDuration2;
+                if (ping < 300) {
+                  bufferDuration2 = 0.01;
+                } else if (ping > 1500) {
+                  bufferDuration2 = 6;
+                } else {
+                  bufferDuration2 = ping / 1e3 * 4;
+                }
+                setBufferDuration(bufferDuration2);
+              });
+            }
+          }, PING_INTERVAL);
           return () => {
             unsubscribe();
-            audio == null ? void 0 : audio.disconnect();
+            clearInterval(intervalId);
+            websocket == null ? void 0 : websocket.disconnect();
           };
         } catch (e) {
           console.error(e);
@@ -519,31 +141,69 @@ function useAudio({ apiKey, baseUrl }) {
       cleanup = cleanupConnection;
     });
     return () => cleanup == null ? void 0 : cleanup();
-  }, [audio]);
-  const play = (0, import_react.useCallback)(
-    (bufferDuration = 0) => __async(this, null, function* () {
-      var _a;
-      if (isPlaying || !streamReturn.current) {
-        return;
+  }, [websocket, baseUrl]);
+  const play = useCallback(() => __async(this, null, function* () {
+    if (playbackStatus === "playing" || !websocketReturn.current) {
+      return;
+    }
+    setPlaybackStatus("playing");
+    const unsubscribes = [];
+    unsubscribes.push(
+      websocketReturn.current.source.on("wait", () => {
+        setIsWaiting(true);
+      })
+    );
+    unsubscribes.push(
+      websocketReturn.current.source.on("read", () => {
+        setIsWaiting(false);
+      })
+    );
+    player.current = new Player({
+      bufferDuration: bufferDuration != null ? bufferDuration : DEFAULT_BUFFER_DURATION
+    });
+    yield player.current.play(websocketReturn.current.source);
+    for (const unsubscribe of unsubscribes) {
+      unsubscribe();
+    }
+    setPlaybackStatus("finished");
+  }), [playbackStatus, bufferDuration]);
+  const pause = useCallback(() => __async(this, null, function* () {
+    var _a2;
+    yield (_a2 = player.current) == null ? void 0 : _a2.pause();
+    setPlaybackStatus("paused");
+  }), []);
+  const resume = useCallback(() => __async(this, null, function* () {
+    var _a2;
+    yield (_a2 = player.current) == null ? void 0 : _a2.resume();
+    setPlaybackStatus("playing");
+  }), []);
+  const toggle = useCallback(() => __async(this, null, function* () {
+    var _a2;
+    yield (_a2 = player.current) == null ? void 0 : _a2.toggle();
+    setPlaybackStatus((status) => {
+      if (status === "playing") {
+        return "paused";
       }
-      setIsPlaying(true);
-      yield (_a = streamReturn.current) == null ? void 0 : _a.play({ bufferDuration });
-      setIsPlaying(false);
-    }),
-    [isPlaying]
-  );
+      if (status === "paused") {
+        return "playing";
+      }
+      return status;
+    });
+  }), []);
   return {
-    stream,
+    buffer,
     play,
-    download,
-    isPlaying,
+    pause,
+    source: (_b = (_a = websocketReturn.current) == null ? void 0 : _a.source) != null ? _b : null,
+    resume,
+    toggle,
+    playbackStatus,
+    bufferStatus,
+    isWaiting,
     isConnected,
-    isStreamed,
-    chunks,
-    messages
+    metrics
   };
 }
-// Annotate the CommonJS export names for ESM import in node:
-0 && (module.exports = {
-  useAudio
-});
+export {
+  useTTS
+};