agrasya-voice-sdk 1.0.0 → 1.0.2

package/dist/index.d.mts

@@ -30,6 +30,8 @@ interface AgrasyaEvents {
     onAudioLevel?: (level: number) => void;
     /** Emitted when the connection status changes */
     onStatusChange?: (status: "idle" | "initializing" | "connecting" | "connected" | "error") => void;
+    /** Emitted when the call duration updates */
+    onCallDurationUpdate?: (duration: number) => void;
 }
 
 /**
@@ -41,22 +43,22 @@ declare class AgrasyaVoiceSDK {
     private peerConnection;
     private dataChannel;
     private audioStream;
+    private socket;
     private sessionId;
-    private status;
+    private ephemeralKey;
+    private callDuration;
+    private durationInterval;
     constructor(config: AgrasyaSDKConfig, events?: AgrasyaEvents);
     private log;
     private setStatus;
-    /**
-     * Starts a voice session.
-     */
-    start(): Promise<void>;
-    /**
-     * Stops the current voice session.
-     */
-    stop(): Promise<void>;
+    private startTimer;
+    private stopTimer;
     private initializeBackendSession;
-    private connectWebRTC;
-    private notifySessionEnd;
+    private setupWebRTC;
+    private setupSocketListeners;
+    startVoiceSession(language: string): Promise<void>;
+    stopVoiceSession(): Promise<void>;
+    private cleanup;
 }
 
 export { AgrasyaVoiceSDK };

package/dist/index.d.ts

@@ -30,6 +30,8 @@ interface AgrasyaEvents {
     onAudioLevel?: (level: number) => void;
     /** Emitted when the connection status changes */
     onStatusChange?: (status: "idle" | "initializing" | "connecting" | "connected" | "error") => void;
+    /** Emitted when the call duration updates */
+    onCallDurationUpdate?: (duration: number) => void;
 }
 
 /**
@@ -41,22 +43,22 @@ declare class AgrasyaVoiceSDK {
     private peerConnection;
     private dataChannel;
     private audioStream;
+    private socket;
     private sessionId;
-    private status;
+    private ephemeralKey;
+    private callDuration;
+    private durationInterval;
     constructor(config: AgrasyaSDKConfig, events?: AgrasyaEvents);
     private log;
     private setStatus;
-    /**
-     * Starts a voice session.
-     */
-    start(): Promise<void>;
-    /**
-     * Stops the current voice session.
-     */
-    stop(): Promise<void>;
+    private startTimer;
+    private stopTimer;
     private initializeBackendSession;
-    private connectWebRTC;
-    private notifySessionEnd;
+    private setupWebRTC;
+    private setupSocketListeners;
+    startVoiceSession(language: string): Promise<void>;
+    stopVoiceSession(): Promise<void>;
+    private cleanup;
 }
 
 export { AgrasyaVoiceSDK };

package/dist/index.js

@@ -1,7 +1,9 @@
 "use strict";
+var __create = Object.create;
 var __defProp = Object.defineProperty;
 var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
 var __getOwnPropNames = Object.getOwnPropertyNames;
+var __getProtoOf = Object.getPrototypeOf;
 var __hasOwnProp = Object.prototype.hasOwnProperty;
 var __export = (target, all) => {
   for (var name in all)
@@ -15,6 +17,14 @@ var __copyProps = (to, from, except, desc) => {
   }
   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);
 
 // src/index.ts
@@ -23,18 +33,24 @@ __export(index_exports, {
   AgrasyaVoiceSDK: () => AgrasyaVoiceSDK
 });
 module.exports = __toCommonJS(index_exports);
+var import_socket = __toESM(require("socket.io-client"));
+var import_dotenv = __toESM(require("dotenv"));
+import_dotenv.default.config();
 var AgrasyaVoiceSDK = class {
   config;
   events;
   peerConnection = null;
   dataChannel = null;
   audioStream = null;
+  socket = null;
   sessionId = null;
-  status = "idle";
+  ephemeralKey = null;
+  callDuration = 0;
+  durationInterval = null;
   constructor(config, events = {}) {
     this.config = {
-      baseUrl: "http://localhost:4001",
-      language: "en",
+      baseUrl: process.env.API_BASE_URL,
+      language: "",
       debug: false,
       ...config
     };
@@ -46,56 +62,20 @@ var AgrasyaVoiceSDK = class {
     }
   }
   setStatus(status) {
-    this.status = status;
     this.events.onStatusChange?.(status);
   }
-  /**
-   * Starts a voice session.
-   */
-  async start() {
-    if (this.status !== "idle") {
-      throw new Error("Session is already in progress or initializing");
-    }
-    try {
-      this.setStatus("initializing");
-      this.log("Initializing session...");
-      const sessionData = await this.initializeBackendSession();
-      this.sessionId = sessionData.sessionId;
-      await this.connectWebRTC(sessionData.ephemeralKey);
-      this.setStatus("connected");
-      this.events.onStart?.();
-    } catch (error) {
-      this.setStatus("error");
-      const err = error instanceof Error ? error : new Error(String(error));
-      this.events.onError?.(err);
-      throw err;
-    }
+  startTimer() {
+    this.callDuration = 0;
+    this.durationInterval = setInterval(() => {
+      this.callDuration += 1;
+      this.events.onCallDurationUpdate?.(this.callDuration);
+    }, 1e3);
   }
-  /**
-   * Stops the current voice session.
-   */
-  async stop() {
-    this.log("Stopping session...");
-    if (this.peerConnection) {
-      this.peerConnection.close();
-      this.peerConnection = null;
-    }
-    if (this.dataChannel) {
-      this.dataChannel.close();
-      this.dataChannel = null;
+  stopTimer() {
+    if (this.durationInterval) {
+      clearInterval(this.durationInterval);
+      this.durationInterval = null;
     }
-    if (this.audioStream) {
-      this.audioStream.getTracks().forEach((track) => track.stop());
-      this.audioStream = null;
-    }
-    if (this.sessionId) {
-      this.notifySessionEnd(this.sessionId).catch(
-        (err) => this.log("Failed to notify session end", err)
-      );
-      this.sessionId = null;
-    }
-    this.setStatus("idle");
-    this.events.onStop?.();
   }
   async initializeBackendSession() {
     const url = `${this.config.baseUrl}/v1/agrasya/client-token?language=${this.config.language}`;
@@ -118,38 +98,37 @@ var AgrasyaVoiceSDK = class {
     }
     return result.data;
   }
-  async connectWebRTC(ephemeralKey) {
-    this.setStatus("connecting");
-    this.log("Connecting to AI Provider via WebRTC...");
+  async setupWebRTC(ephemeralKey) {
     this.peerConnection = new RTCPeerConnection();
+    this.dataChannel = this.peerConnection.createDataChannel(
+      "oai-events"
+    );
+    if (this.dataChannel) {
+      this.dataChannel.onmessage = (event) => {
+        const data = JSON.parse(event.data);
+        this.events.onTranscript?.(data);
+      };
+    }
     this.peerConnection.ontrack = (event) => {
-      this.log("Received remote audio track");
-      const audio = new Audio();
-      audio.srcObject = event.streams[0];
-      audio.play().catch((err) => this.log("Failed to play audio", err));
+      const audioElement = document.createElement("audio");
+      audioElement.srcObject = event.streams[0];
+      audioElement.autoplay = true;
+      document.body.appendChild(audioElement);
     };
     this.audioStream = await navigator.mediaDevices.getUserMedia({
       audio: true
     });
-    this.audioStream.getTracks().forEach((track) => {
-      if (this.peerConnection && this.audioStream) {
-        this.peerConnection.addTrack(track, this.audioStream);
-      }
-    });
-    this.dataChannel = this.peerConnection.createDataChannel("oai-events");
-    this.dataChannel.onmessage = (event) => {
-      const data = JSON.parse(event.data);
-      this.log("Received event:", data.type);
-      if (data.type === "response.audio_transcript.delta") {
-        this.events.onTranscript?.({ text: data.delta, isFinal: false });
-      } else if (data.type === "response.audio_transcript.done") {
-        this.events.onTranscript?.({ text: data.transcript, isFinal: true });
-      }
-    };
+    if (this.audioStream) {
+      const track = this.audioStream.getTracks()[0];
+      this.peerConnection.addTrack(
+        track,
+        this.audioStream
+      );
+    }
     const offer = await this.peerConnection.createOffer();
     await this.peerConnection.setLocalDescription(offer);
-    const baseUrl = "https://api.openai.com/v1/realtime";
-    const model = "gpt-4o-realtime-preview-2024-12-17";
+    const baseUrl = process.env.API_URL;
+    const model = process.env.MODEL_NAME;
     const sdpResponse = await fetch(`${baseUrl}?model=${model}`, {
       method: "POST",
       body: offer.sdp,
@@ -165,19 +144,60 @@ var AgrasyaVoiceSDK = class {
       type: "answer",
       sdp: await sdpResponse.text()
     };
-    await this.peerConnection.setRemoteDescription(answer);
-    this.log("WebRTC connection established");
+    const remoteDescription = {
+      type: answer.type,
+      sdp: answer.sdp
+    };
+    await this.peerConnection.setRemoteDescription(remoteDescription);
   }
-  async notifySessionEnd(sessionId) {
-    const url = `${this.config.baseUrl}/v1/agrasya/session-end/${sessionId}`;
-    await fetch(url, {
-      method: "PUT",
-      headers: {
-        "x-agrasya-client-api-key": this.config.apiKey,
-        "Content-Type": "application/json"
-      }
+  setupSocketListeners(socket) {
+    socket.on("disconnect", () => {
+      this.cleanup();
+    });
+    socket.on("session_ended", () => {
+      this.cleanup();
     });
   }
+  async startVoiceSession(language) {
+    try {
+      this.setStatus("initializing");
+      const sessionData = await this.initializeBackendSession();
+      this.sessionId = sessionData.sessionId;
+      this.ephemeralKey = sessionData.ephemeralKey;
+      await this.setupWebRTC(this.ephemeralKey);
+      this.socket = (0, import_socket.default)(this.config.baseUrl, { transports: ["websocket"] });
+      this.socket.emit("join_session", { sessionId: this.sessionId });
+      this.setupSocketListeners(this.socket);
+      this.setStatus("connected");
+      this.startTimer();
+    } catch (error) {
+      this.setStatus("error");
+      throw error;
+    }
+  }
+  async stopVoiceSession() {
+    this.cleanup();
+  }
+  cleanup() {
+    if (this.peerConnection) {
+      this.peerConnection.close();
+      this.peerConnection = null;
+    }
+    if (this.dataChannel) {
+      this.dataChannel.close();
+      this.dataChannel = null;
+    }
+    if (this.audioStream) {
+      this.audioStream.getTracks().forEach((track) => track.stop());
+      this.audioStream = null;
+    }
+    if (this.socket) {
+      this.socket.disconnect();
+      this.socket = null;
+    }
+    this.stopTimer();
+    this.setStatus("idle");
+  }
 };
 // Annotate the CommonJS export names for ESM import in node:
 0 && (module.exports = {

package/dist/index.mjs

@@ -1,16 +1,22 @@
 // src/index.ts
+import io from "socket.io-client";
+import dotenv from "dotenv";
+dotenv.config();
 var AgrasyaVoiceSDK = class {
   config;
   events;
   peerConnection = null;
   dataChannel = null;
   audioStream = null;
+  socket = null;
   sessionId = null;
-  status = "idle";
+  ephemeralKey = null;
+  callDuration = 0;
+  durationInterval = null;
   constructor(config, events = {}) {
     this.config = {
-      baseUrl: "http://localhost:4001",
-      language: "en",
+      baseUrl: process.env.API_BASE_URL,
+      language: "",
       debug: false,
       ...config
     };
@@ -22,56 +28,20 @@ var AgrasyaVoiceSDK = class {
     }
   }
   setStatus(status) {
-    this.status = status;
     this.events.onStatusChange?.(status);
   }
-  /**
-   * Starts a voice session.
-   */
-  async start() {
-    if (this.status !== "idle") {
-      throw new Error("Session is already in progress or initializing");
-    }
-    try {
-      this.setStatus("initializing");
-      this.log("Initializing session...");
-      const sessionData = await this.initializeBackendSession();
-      this.sessionId = sessionData.sessionId;
-      await this.connectWebRTC(sessionData.ephemeralKey);
-      this.setStatus("connected");
-      this.events.onStart?.();
-    } catch (error) {
-      this.setStatus("error");
-      const err = error instanceof Error ? error : new Error(String(error));
-      this.events.onError?.(err);
-      throw err;
-    }
+  startTimer() {
+    this.callDuration = 0;
+    this.durationInterval = setInterval(() => {
+      this.callDuration += 1;
+      this.events.onCallDurationUpdate?.(this.callDuration);
+    }, 1e3);
   }
-  /**
-   * Stops the current voice session.
-   */
-  async stop() {
-    this.log("Stopping session...");
-    if (this.peerConnection) {
-      this.peerConnection.close();
-      this.peerConnection = null;
-    }
-    if (this.dataChannel) {
-      this.dataChannel.close();
-      this.dataChannel = null;
+  stopTimer() {
+    if (this.durationInterval) {
+      clearInterval(this.durationInterval);
+      this.durationInterval = null;
     }
-    if (this.audioStream) {
-      this.audioStream.getTracks().forEach((track) => track.stop());
-      this.audioStream = null;
-    }
-    if (this.sessionId) {
-      this.notifySessionEnd(this.sessionId).catch(
-        (err) => this.log("Failed to notify session end", err)
-      );
-      this.sessionId = null;
-    }
-    this.setStatus("idle");
-    this.events.onStop?.();
   }
   async initializeBackendSession() {
     const url = `${this.config.baseUrl}/v1/agrasya/client-token?language=${this.config.language}`;
@@ -94,38 +64,37 @@ var AgrasyaVoiceSDK = class {
     }
     return result.data;
   }
-  async connectWebRTC(ephemeralKey) {
-    this.setStatus("connecting");
-    this.log("Connecting to AI Provider via WebRTC...");
+  async setupWebRTC(ephemeralKey) {
     this.peerConnection = new RTCPeerConnection();
+    this.dataChannel = this.peerConnection.createDataChannel(
+      "oai-events"
+    );
+    if (this.dataChannel) {
+      this.dataChannel.onmessage = (event) => {
+        const data = JSON.parse(event.data);
+        this.events.onTranscript?.(data);
+      };
+    }
     this.peerConnection.ontrack = (event) => {
-      this.log("Received remote audio track");
-      const audio = new Audio();
-      audio.srcObject = event.streams[0];
-      audio.play().catch((err) => this.log("Failed to play audio", err));
+      const audioElement = document.createElement("audio");
+      audioElement.srcObject = event.streams[0];
+      audioElement.autoplay = true;
+      document.body.appendChild(audioElement);
     };
     this.audioStream = await navigator.mediaDevices.getUserMedia({
       audio: true
     });
-    this.audioStream.getTracks().forEach((track) => {
-      if (this.peerConnection && this.audioStream) {
-        this.peerConnection.addTrack(track, this.audioStream);
-      }
-    });
-    this.dataChannel = this.peerConnection.createDataChannel("oai-events");
-    this.dataChannel.onmessage = (event) => {
-      const data = JSON.parse(event.data);
-      this.log("Received event:", data.type);
-      if (data.type === "response.audio_transcript.delta") {
-        this.events.onTranscript?.({ text: data.delta, isFinal: false });
-      } else if (data.type === "response.audio_transcript.done") {
-        this.events.onTranscript?.({ text: data.transcript, isFinal: true });
-      }
-    };
+    if (this.audioStream) {
+      const track = this.audioStream.getTracks()[0];
+      this.peerConnection.addTrack(
+        track,
+        this.audioStream
+      );
+    }
     const offer = await this.peerConnection.createOffer();
     await this.peerConnection.setLocalDescription(offer);
-    const baseUrl = "https://api.openai.com/v1/realtime";
-    const model = "gpt-4o-realtime-preview-2024-12-17";
+    const baseUrl = process.env.API_URL;
+    const model = process.env.MODEL_NAME;
     const sdpResponse = await fetch(`${baseUrl}?model=${model}`, {
       method: "POST",
       body: offer.sdp,
@@ -141,19 +110,60 @@ var AgrasyaVoiceSDK = class {
       type: "answer",
       sdp: await sdpResponse.text()
     };
-    await this.peerConnection.setRemoteDescription(answer);
-    this.log("WebRTC connection established");
+    const remoteDescription = {
+      type: answer.type,
+      sdp: answer.sdp
+    };
+    await this.peerConnection.setRemoteDescription(remoteDescription);
   }
-  async notifySessionEnd(sessionId) {
-    const url = `${this.config.baseUrl}/v1/agrasya/session-end/${sessionId}`;
-    await fetch(url, {
-      method: "PUT",
-      headers: {
-        "x-agrasya-client-api-key": this.config.apiKey,
-        "Content-Type": "application/json"
-      }
+  setupSocketListeners(socket) {
+    socket.on("disconnect", () => {
+      this.cleanup();
+    });
+    socket.on("session_ended", () => {
+      this.cleanup();
     });
   }
+  async startVoiceSession(language) {
+    try {
+      this.setStatus("initializing");
+      const sessionData = await this.initializeBackendSession();
+      this.sessionId = sessionData.sessionId;
+      this.ephemeralKey = sessionData.ephemeralKey;
+      await this.setupWebRTC(this.ephemeralKey);
+      this.socket = io(this.config.baseUrl, { transports: ["websocket"] });
+      this.socket.emit("join_session", { sessionId: this.sessionId });
+      this.setupSocketListeners(this.socket);
+      this.setStatus("connected");
+      this.startTimer();
+    } catch (error) {
+      this.setStatus("error");
+      throw error;
+    }
+  }
+  async stopVoiceSession() {
+    this.cleanup();
+  }
+  cleanup() {
+    if (this.peerConnection) {
+      this.peerConnection.close();
+      this.peerConnection = null;
+    }
+    if (this.dataChannel) {
+      this.dataChannel.close();
+      this.dataChannel = null;
+    }
+    if (this.audioStream) {
+      this.audioStream.getTracks().forEach((track) => track.stop());
+      this.audioStream = null;
+    }
+    if (this.socket) {
+      this.socket.disconnect();
+      this.socket = null;
+    }
+    this.stopTimer();
+    this.setStatus("idle");
+  }
 };
 export {
   AgrasyaVoiceSDK

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "agrasya-voice-sdk",
-  "version": "1.0.0",
+  "version": "1.0.2",
   "main": "dist/index.js",
   "module": "dist/index.mjs",
   "types": "dist/index.d.ts",
@@ -12,5 +12,22 @@
     "dev": "tsup src/index.ts --format cjs,esm --watch --dts",
     "lint": "eslint src/**/*.ts",
     "test": "jest"
+  },
+  "devDependencies": {
+    "@types/jest": "^29.5.0",
+    "@types/webrtc": "^0.0.47",
+    "@typescript-eslint/eslint-plugin": "^6.0.0",
+    "@typescript-eslint/parser": "^6.0.0",
+    "eslint": "^8.50.0",
+    "jest": "^29.7.0",
+    "jest-environment-jsdom": "^30.3.0",
+    "ts-jest": "^29.1.1",
+    "tsup": "^8.0.0",
+    "typescript": "^5.0.0"
+  },
+  "dependencies": {
+    "dotenv": "^17.3.1",
+    "react-native-webrtc": "^124.0.7",
+    "socket.io-client": "^4.8.3"
   }
 }