@livekit/agents 0.5.0 → 0.5.1

package/dist/job.cjs

@@ -214,9 +214,9 @@ class JobRequest {
     await this.#onReject();
   }
   /** Accepts the job, launching it on an idle child process. */
-  async accept(name = "", identity = "", metadata = "") {
+  async accept(name = "", identity = "", metadata = "", attributes) {
     if (identity === "") identity = "agent-" + this.id;
-    this.#onAccept({ name, identity, metadata });
+    this.#onAccept({ name, identity, metadata, attributes });
   }
 }
 // Annotate the CommonJS export names for ESM import in node:

package/dist/job.cjs.map

@@ -1 +1 @@
-{"version":3,"sources":["../src/job.ts"],"sourcesContent":["// SPDX-FileCopyrightText: 2024 LiveKit, Inc.\n//\n// SPDX-License-Identifier: Apache-2.0\nimport type * as proto from '@livekit/protocol';\nimport type {\n  E2EEOptions,\n  LocalParticipant,\n  RemoteParticipant,\n  Room,\n  RtcConfiguration,\n} from '@livekit/rtc-node';\nimport { ParticipantKind, RoomEvent, TrackKind } from '@livekit/rtc-node';\nimport type { Logger } from 'pino';\nimport { log } from './log.js';\n\n/** Which tracks, if any, should the agent automatically subscribe to? */\nexport enum AutoSubscribe {\n  SUBSCRIBE_ALL,\n  SUBSCRIBE_NONE,\n  VIDEO_ONLY,\n  AUDIO_ONLY,\n}\n\nexport type JobAcceptArguments = {\n  name: string;\n  identity: string;\n  metadata: string;\n};\n\nexport type RunningJobInfo = {\n  acceptArguments: JobAcceptArguments;\n  job: proto.Job;\n  url: string;\n  token: string;\n};\n\n/** Attempted to add a function callback, but the function already exists. */\nexport class FunctionExistsError extends Error {\n  constructor(msg?: string) {\n    super(msg);\n    Object.setPrototypeOf(this, new.target.prototype);\n  }\n}\n\n/** The job and environment context as seen by the agent, accessible by the entrypoint function. */\nexport class JobContext {\n  #proc: JobProcess;\n  #info: RunningJobInfo;\n  #room: Room;\n  #onConnect: () => void;\n  #onShutdown: (s: string) => void;\n  /** @internal */\n  shutdownCallbacks: (() => Promise<void>)[] = [];\n  #participantEntrypoints: ((job: JobContext, p: RemoteParticipant) => Promise<void>)[] = [];\n  #participantTasks: {\n    [id: string]: {\n      callback: (job: JobContext, p: RemoteParticipant) => Promise<void>;\n      result: Promise<void>;\n    };\n  } = {};\n  #logger: Logger;\n\n  constructor(\n    proc: JobProcess,\n    info: RunningJobInfo,\n    room: Room,\n    onConnect: () => void,\n    onShutdown: (s: string) => void,\n  ) {\n    this.#proc = proc;\n    this.#info = info;\n    this.#room = room;\n    this.#onConnect = onConnect;\n    this.#onShutdown = onShutdown;\n    this.onParticipantConnected = this.onParticipantConnected.bind(this);\n    this.#room.on(RoomEvent.ParticipantConnected, this.onParticipantConnected);\n    this.#logger = log().child({ info: this.#info });\n  }\n\n  get proc(): JobProcess {\n    return this.#proc;\n  }\n\n  get job(): proto.Job {\n    return this.#info.job;\n  }\n\n  /** @returns The room the agent was called into */\n  get room(): Room {\n    return this.#room;\n  }\n\n  /** @returns The agent's participant if connected to the room, otherwise `undefined` */\n  get agent(): LocalParticipant | undefined {\n    return this.#room.localParticipant;\n  }\n\n  /** Adds a promise to be awaited when {@link JobContext.shutdown | shutdown} is called. */\n  addShutdownCallback(callback: () => Promise<void>) {\n    this.shutdownCallbacks.push(callback);\n  }\n\n  async waitForParticipant(identity?: string): Promise<RemoteParticipant> {\n    if (!this.#room.isConnected) {\n      throw new Error('room is not connected');\n    }\n\n    for (const p of this.#room.remoteParticipants.values()) {\n      if ((!identity || p.identity === identity) && p.info.kind != ParticipantKind.AGENT) {\n        return p;\n      }\n    }\n\n    return new Promise((resolve, reject) => {\n      const onParticipantConnected = (participant: RemoteParticipant) => {\n        if (\n          (!identity || participant.identity === identity) &&\n          participant.info.kind != ParticipantKind.AGENT\n        ) {\n          clearHandlers();\n          resolve(participant);\n        }\n      };\n      const onDisconnected = () => {\n        clearHandlers();\n        reject(new Error('Room disconnected while waiting for participant'));\n      };\n\n      const clearHandlers = () => {\n        this.#room.off(RoomEvent.ParticipantConnected, onParticipantConnected);\n        this.#room.off(RoomEvent.Disconnected, onDisconnected);\n      };\n\n      this.#room.on(RoomEvent.ParticipantConnected, onParticipantConnected);\n      this.#room.on(RoomEvent.Disconnected, onDisconnected);\n    });\n  }\n\n  /**\n   * Connects the agent to the room.\n   *\n   * @remarks\n   * It is recommended to run this command as early in the function as possible, as executing it\n   * later may cause noticeable delay between user and agent joins.\n   *\n   * @see {@link https://github.com/livekit/node-sdks/tree/main/packages/livekit-rtc#readme |\n   * @livekit/rtc-node} for more information about the parameters.\n   */\n  async connect(\n    e2ee?: E2EEOptions,\n    autoSubscribe: AutoSubscribe = AutoSubscribe.SUBSCRIBE_ALL,\n    rtcConfig?: RtcConfiguration,\n  ) {\n    const opts = {\n      e2ee,\n      autoSubscribe: autoSubscribe == AutoSubscribe.SUBSCRIBE_ALL,\n      rtcConfig,\n      dynacast: false,\n    };\n\n    await this.#room.connect(this.#info.url, this.#info.token, opts);\n    this.#onConnect();\n\n    this.#room.remoteParticipants.forEach(this.onParticipantConnected);\n\n    if ([AutoSubscribe.AUDIO_ONLY, AutoSubscribe.VIDEO_ONLY].includes(autoSubscribe)) {\n      this.#room.remoteParticipants.forEach((p) => {\n        p.trackPublications.forEach((pub) => {\n          if (\n            (autoSubscribe === AutoSubscribe.AUDIO_ONLY && pub.kind === TrackKind.KIND_AUDIO) ||\n            (autoSubscribe === AutoSubscribe.VIDEO_ONLY && pub.kind === TrackKind.KIND_VIDEO)\n          ) {\n            pub.setSubscribed(true);\n          }\n        });\n      });\n    }\n  }\n\n  /**\n   * Gracefully shuts down the job, and runs all shutdown promises.\n   *\n   * @param reason - Optional reason for shutdown\n   */\n  shutdown(reason = '') {\n    this.#onShutdown(reason);\n  }\n\n  /** @internal */\n  onParticipantConnected(p: RemoteParticipant) {\n    for (const callback of this.#participantEntrypoints) {\n      if (this.#participantTasks[p.identity]?.callback == callback) {\n        this.#logger.warn(\n          'a participant has joined before a prior prticipant task matching the same identity has finished:',\n          p.identity,\n        );\n      }\n      const result = callback(this, p);\n      result.finally(() => delete this.#participantTasks[p.identity]);\n      this.#participantTasks[p.identity] = { callback, result };\n    }\n  }\n\n  /**\n   * Adds a promise to be awaited whenever a new participant joins the room.\n   *\n   * @throws {@link FunctionExistsError} if an entrypoint already exists\n   */\n  addParticipantEntrypoint(callback: (job: JobContext, p: RemoteParticipant) => Promise<void>) {\n    if (this.#participantEntrypoints.includes(callback)) {\n      throw new FunctionExistsError('entrypoints cannot be added more than once');\n    }\n\n    this.#participantEntrypoints.push(callback);\n  }\n}\n\nexport class JobProcess {\n  #pid = process.pid;\n  userData: { [id: string]: unknown } = {};\n\n  get pid(): number {\n    return this.#pid;\n  }\n}\n\n/**\n * A request sent by the server to spawn a new agent job.\n *\n * @remarks\n * For most applications, this is best left to the default, which simply accepts the job and\n * handles the logic inside the entrypoint function. This class is useful for vetting which\n * requests should fill idle processes and which should be outright rejected.\n */\nexport class JobRequest {\n  #job: proto.Job;\n  #onReject: () => Promise<void>;\n  #onAccept: (args: JobAcceptArguments) => Promise<void>;\n\n  /** @internal */\n  constructor(\n    job: proto.Job,\n    onReject: () => Promise<void>,\n    onAccept: (args: JobAcceptArguments) => Promise<void>,\n  ) {\n    this.#job = job;\n    this.#onReject = onReject;\n    this.#onAccept = onAccept;\n  }\n\n  /** @returns The ID of the job, set by the LiveKit server */\n  get id(): string {\n    return this.#job.id;\n  }\n\n  /** @see {@link https://www.npmjs.com/package/@livekit/protocol | @livekit/protocol} */\n  get job(): proto.Job {\n    return this.#job;\n  }\n\n  /** @see {@link https://www.npmjs.com/package/@livekit/protocol | @livekit/protocol} */\n  get room(): proto.Room | undefined {\n    return this.#job.room;\n  }\n\n  /** @see {@link https://www.npmjs.com/package/@livekit/protocol | @livekit/protocol} */\n  get publisher(): proto.ParticipantInfo | undefined {\n    return this.#job.participant;\n  }\n\n  /** @returns The agent's name, as set in {@link WorkerOptions} */\n  get agentName(): string {\n    return this.#job.agentName;\n  }\n\n  /** Rejects the job. */\n  async reject() {\n    await this.#onReject();\n  }\n\n  /** Accepts the job, launching it on an idle child process. */\n  async accept(name = '', identity = '', metadata = '') {\n    if (identity === '') identity = 'agent-' + this.id;\n\n    this.#onAccept({ name, identity, metadata });\n  }\n}\n"],"mappings":";;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAWA,sBAAsD;AAEtD,iBAAoB;AAGb,IAAK,gBAAL,kBAAKA,mBAAL;AACL,EAAAA,8BAAA;AACA,EAAAA,8BAAA;AACA,EAAAA,8BAAA;AACA,EAAAA,8BAAA;AAJU,SAAAA;AAAA,GAAA;AAqBL,MAAM,4BAA4B,MAAM;AAAA,EAC7C,YAAY,KAAc;AACxB,UAAM,GAAG;AACT,WAAO,eAAe,MAAM,WAAW,SAAS;AAAA,EAClD;AACF;AAGO,MAAM,WAAW;AAAA,EACtB;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA;AAAA,EAEA,oBAA6C,CAAC;AAAA,EAC9C,0BAAwF,CAAC;AAAA,EACzF,oBAKI,CAAC;AAAA,EACL;AAAA,EAEA,YACE,MACA,MACA,MACA,WACA,YACA;AACA,SAAK,QAAQ;AACb,SAAK,QAAQ;AACb,SAAK,QAAQ;AACb,SAAK,aAAa;AAClB,SAAK,cAAc;AACnB,SAAK,yBAAyB,KAAK,uBAAuB,KAAK,IAAI;AACnE,SAAK,MAAM,GAAG,0BAAU,sBAAsB,KAAK,sBAAsB;AACzE,SAAK,cAAU,gBAAI,EAAE,MAAM,EAAE,MAAM,KAAK,MAAM,CAAC;AAAA,EACjD;AAAA,EAEA,IAAI,OAAmB;AACrB,WAAO,KAAK;AAAA,EACd;AAAA,EAEA,IAAI,MAAiB;AACnB,WAAO,KAAK,MAAM;AAAA,EACpB;AAAA;AAAA,EAGA,IAAI,OAAa;AACf,WAAO,KAAK;AAAA,EACd;AAAA;AAAA,EAGA,IAAI,QAAsC;AACxC,WAAO,KAAK,MAAM;AAAA,EACpB;AAAA;AAAA,EAGA,oBAAoB,UAA+B;AACjD,SAAK,kBAAkB,KAAK,QAAQ;AAAA,EACtC;AAAA,EAEA,MAAM,mBAAmB,UAA+C;AACtE,QAAI,CAAC,KAAK,MAAM,aAAa;AAC3B,YAAM,IAAI,MAAM,uBAAuB;AAAA,IACzC;AAEA,eAAW,KAAK,KAAK,MAAM,mBAAmB,OAAO,GAAG;AACtD,WAAK,CAAC,YAAY,EAAE,aAAa,aAAa,EAAE,KAAK,QAAQ,gCAAgB,OAAO;AAClF,eAAO;AAAA,MACT;AAAA,IACF;AAEA,WAAO,IAAI,QAAQ,CAAC,SAAS,WAAW;AACtC,YAAM,yBAAyB,CAAC,gBAAmC;AACjE,aACG,CAAC,YAAY,YAAY,aAAa,aACvC,YAAY,KAAK,QAAQ,gCAAgB,OACzC;AACA,wBAAc;AACd,kBAAQ,WAAW;AAAA,QACrB;AAAA,MACF;AACA,YAAM,iBAAiB,MAAM;AAC3B,sBAAc;AACd,eAAO,IAAI,MAAM,iDAAiD,CAAC;AAAA,MACrE;AAEA,YAAM,gBAAgB,MAAM;AAC1B,aAAK,MAAM,IAAI,0BAAU,sBAAsB,sBAAsB;AACrE,aAAK,MAAM,IAAI,0BAAU,cAAc,cAAc;AAAA,MACvD;AAEA,WAAK,MAAM,GAAG,0BAAU,sBAAsB,sBAAsB;AACpE,WAAK,MAAM,GAAG,0BAAU,cAAc,cAAc;AAAA,IACtD,CAAC;AAAA,EACH;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,MAAM,QACJ,MACA,gBAA+B,uBAC/B,WACA;AACA,UAAM,OAAO;AAAA,MACX;AAAA,MACA,eAAe,iBAAiB;AAAA,MAChC;AAAA,MACA,UAAU;AAAA,IACZ;AAEA,UAAM,KAAK,MAAM,QAAQ,KAAK,MAAM,KAAK,KAAK,MAAM,OAAO,IAAI;AAC/D,SAAK,WAAW;AAEhB,SAAK,MAAM,mBAAmB,QAAQ,KAAK,sBAAsB;AAEjE,QAAI,CAAC,oBAA0B,kBAAwB,EAAE,SAAS,aAAa,GAAG;AAChF,WAAK,MAAM,mBAAmB,QAAQ,CAAC,MAAM;AAC3C,UAAE,kBAAkB,QAAQ,CAAC,QAAQ;AACnC,cACG,kBAAkB,sBAA4B,IAAI,SAAS,0BAAU,cACrE,kBAAkB,sBAA4B,IAAI,SAAS,0BAAU,YACtE;AACA,gBAAI,cAAc,IAAI;AAAA,UACxB;AAAA,QACF,CAAC;AAAA,MACH,CAAC;AAAA,IACH;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAS,SAAS,IAAI;AACpB,SAAK,YAAY,MAAM;AAAA,EACzB;AAAA;AAAA,EAGA,uBAAuB,GAAsB;AA7L/C;AA8LI,eAAW,YAAY,KAAK,yBAAyB;AACnD,YAAI,UAAK,kBAAkB,EAAE,QAAQ,MAAjC,mBAAoC,aAAY,UAAU;AAC5D,aAAK,QAAQ;AAAA,UACX;AAAA,UACA,EAAE;AAAA,QACJ;AAAA,MACF;AACA,YAAM,SAAS,SAAS,MAAM,CAAC;AAC/B,aAAO,QAAQ,MAAM,OAAO,KAAK,kBAAkB,EAAE,QAAQ,CAAC;AAC9D,WAAK,kBAAkB,EAAE,QAAQ,IAAI,EAAE,UAAU,OAAO;AAAA,IAC1D;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,yBAAyB,UAAoE;AAC3F,QAAI,KAAK,wBAAwB,SAAS,QAAQ,GAAG;AACnD,YAAM,IAAI,oBAAoB,4CAA4C;AAAA,IAC5E;AAEA,SAAK,wBAAwB,KAAK,QAAQ;AAAA,EAC5C;AACF;AAEO,MAAM,WAAW;AAAA,EACtB,OAAO,QAAQ;AAAA,EACf,WAAsC,CAAC;AAAA,EAEvC,IAAI,MAAc;AAChB,WAAO,KAAK;AAAA,EACd;AACF;AAUO,MAAM,WAAW;AAAA,EACtB;AAAA,EACA;AAAA,EACA;AAAA;AAAA,EAGA,YACE,KACA,UACA,UACA;AACA,SAAK,OAAO;AACZ,SAAK,YAAY;AACjB,SAAK,YAAY;AAAA,EACnB;AAAA;AAAA,EAGA,IAAI,KAAa;AACf,WAAO,KAAK,KAAK;AAAA,EACnB;AAAA;AAAA,EAGA,IAAI,MAAiB;AACnB,WAAO,KAAK;AAAA,EACd;AAAA;AAAA,EAGA,IAAI,OAA+B;AACjC,WAAO,KAAK,KAAK;AAAA,EACnB;AAAA;AAAA,EAGA,IAAI,YAA+C;AACjD,WAAO,KAAK,KAAK;AAAA,EACnB;AAAA;AAAA,EAGA,IAAI,YAAoB;AACtB,WAAO,KAAK,KAAK;AAAA,EACnB;AAAA;AAAA,EAGA,MAAM,SAAS;AACb,UAAM,KAAK,UAAU;AAAA,EACvB;AAAA;AAAA,EAGA,MAAM,OAAO,OAAO,IAAI,WAAW,IAAI,WAAW,IAAI;AACpD,QAAI,aAAa,GAAI,YAAW,WAAW,KAAK;AAEhD,SAAK,UAAU,EAAE,MAAM,UAAU,SAAS,CAAC;AAAA,EAC7C;AACF;","names":["AutoSubscribe"]}
+{"version":3,"sources":["../src/job.ts"],"sourcesContent":["// SPDX-FileCopyrightText: 2024 LiveKit, Inc.\n//\n// SPDX-License-Identifier: Apache-2.0\nimport type * as proto from '@livekit/protocol';\nimport type {\n  E2EEOptions,\n  LocalParticipant,\n  RemoteParticipant,\n  Room,\n  RtcConfiguration,\n} from '@livekit/rtc-node';\nimport { ParticipantKind, RoomEvent, TrackKind } from '@livekit/rtc-node';\nimport type { Logger } from 'pino';\nimport { log } from './log.js';\n\n/** Which tracks, if any, should the agent automatically subscribe to? */\nexport enum AutoSubscribe {\n  SUBSCRIBE_ALL,\n  SUBSCRIBE_NONE,\n  VIDEO_ONLY,\n  AUDIO_ONLY,\n}\n\nexport type JobAcceptArguments = {\n  name: string;\n  identity: string;\n  metadata: string;\n  attributes?: { [key: string]: string };\n};\n\nexport type RunningJobInfo = {\n  acceptArguments: JobAcceptArguments;\n  job: proto.Job;\n  url: string;\n  token: string;\n};\n\n/** Attempted to add a function callback, but the function already exists. */\nexport class FunctionExistsError extends Error {\n  constructor(msg?: string) {\n    super(msg);\n    Object.setPrototypeOf(this, new.target.prototype);\n  }\n}\n\n/** The job and environment context as seen by the agent, accessible by the entrypoint function. */\nexport class JobContext {\n  #proc: JobProcess;\n  #info: RunningJobInfo;\n  #room: Room;\n  #onConnect: () => void;\n  #onShutdown: (s: string) => void;\n  /** @internal */\n  shutdownCallbacks: (() => Promise<void>)[] = [];\n  #participantEntrypoints: ((job: JobContext, p: RemoteParticipant) => Promise<void>)[] = [];\n  #participantTasks: {\n    [id: string]: {\n      callback: (job: JobContext, p: RemoteParticipant) => Promise<void>;\n      result: Promise<void>;\n    };\n  } = {};\n  #logger: Logger;\n\n  constructor(\n    proc: JobProcess,\n    info: RunningJobInfo,\n    room: Room,\n    onConnect: () => void,\n    onShutdown: (s: string) => void,\n  ) {\n    this.#proc = proc;\n    this.#info = info;\n    this.#room = room;\n    this.#onConnect = onConnect;\n    this.#onShutdown = onShutdown;\n    this.onParticipantConnected = this.onParticipantConnected.bind(this);\n    this.#room.on(RoomEvent.ParticipantConnected, this.onParticipantConnected);\n    this.#logger = log().child({ info: this.#info });\n  }\n\n  get proc(): JobProcess {\n    return this.#proc;\n  }\n\n  get job(): proto.Job {\n    return this.#info.job;\n  }\n\n  /** @returns The room the agent was called into */\n  get room(): Room {\n    return this.#room;\n  }\n\n  /** @returns The agent's participant if connected to the room, otherwise `undefined` */\n  get agent(): LocalParticipant | undefined {\n    return this.#room.localParticipant;\n  }\n\n  /** Adds a promise to be awaited when {@link JobContext.shutdown | shutdown} is called. */\n  addShutdownCallback(callback: () => Promise<void>) {\n    this.shutdownCallbacks.push(callback);\n  }\n\n  async waitForParticipant(identity?: string): Promise<RemoteParticipant> {\n    if (!this.#room.isConnected) {\n      throw new Error('room is not connected');\n    }\n\n    for (const p of this.#room.remoteParticipants.values()) {\n      if ((!identity || p.identity === identity) && p.info.kind != ParticipantKind.AGENT) {\n        return p;\n      }\n    }\n\n    return new Promise((resolve, reject) => {\n      const onParticipantConnected = (participant: RemoteParticipant) => {\n        if (\n          (!identity || participant.identity === identity) &&\n          participant.info.kind != ParticipantKind.AGENT\n        ) {\n          clearHandlers();\n          resolve(participant);\n        }\n      };\n      const onDisconnected = () => {\n        clearHandlers();\n        reject(new Error('Room disconnected while waiting for participant'));\n      };\n\n      const clearHandlers = () => {\n        this.#room.off(RoomEvent.ParticipantConnected, onParticipantConnected);\n        this.#room.off(RoomEvent.Disconnected, onDisconnected);\n      };\n\n      this.#room.on(RoomEvent.ParticipantConnected, onParticipantConnected);\n      this.#room.on(RoomEvent.Disconnected, onDisconnected);\n    });\n  }\n\n  /**\n   * Connects the agent to the room.\n   *\n   * @remarks\n   * It is recommended to run this command as early in the function as possible, as executing it\n   * later may cause noticeable delay between user and agent joins.\n   *\n   * @see {@link https://github.com/livekit/node-sdks/tree/main/packages/livekit-rtc#readme |\n   * @livekit/rtc-node} for more information about the parameters.\n   */\n  async connect(\n    e2ee?: E2EEOptions,\n    autoSubscribe: AutoSubscribe = AutoSubscribe.SUBSCRIBE_ALL,\n    rtcConfig?: RtcConfiguration,\n  ) {\n    const opts = {\n      e2ee,\n      autoSubscribe: autoSubscribe == AutoSubscribe.SUBSCRIBE_ALL,\n      rtcConfig,\n      dynacast: false,\n    };\n\n    await this.#room.connect(this.#info.url, this.#info.token, opts);\n    this.#onConnect();\n\n    this.#room.remoteParticipants.forEach(this.onParticipantConnected);\n\n    if ([AutoSubscribe.AUDIO_ONLY, AutoSubscribe.VIDEO_ONLY].includes(autoSubscribe)) {\n      this.#room.remoteParticipants.forEach((p) => {\n        p.trackPublications.forEach((pub) => {\n          if (\n            (autoSubscribe === AutoSubscribe.AUDIO_ONLY && pub.kind === TrackKind.KIND_AUDIO) ||\n            (autoSubscribe === AutoSubscribe.VIDEO_ONLY && pub.kind === TrackKind.KIND_VIDEO)\n          ) {\n            pub.setSubscribed(true);\n          }\n        });\n      });\n    }\n  }\n\n  /**\n   * Gracefully shuts down the job, and runs all shutdown promises.\n   *\n   * @param reason - Optional reason for shutdown\n   */\n  shutdown(reason = '') {\n    this.#onShutdown(reason);\n  }\n\n  /** @internal */\n  onParticipantConnected(p: RemoteParticipant) {\n    for (const callback of this.#participantEntrypoints) {\n      if (this.#participantTasks[p.identity]?.callback == callback) {\n        this.#logger.warn(\n          'a participant has joined before a prior prticipant task matching the same identity has finished:',\n          p.identity,\n        );\n      }\n      const result = callback(this, p);\n      result.finally(() => delete this.#participantTasks[p.identity]);\n      this.#participantTasks[p.identity] = { callback, result };\n    }\n  }\n\n  /**\n   * Adds a promise to be awaited whenever a new participant joins the room.\n   *\n   * @throws {@link FunctionExistsError} if an entrypoint already exists\n   */\n  addParticipantEntrypoint(callback: (job: JobContext, p: RemoteParticipant) => Promise<void>) {\n    if (this.#participantEntrypoints.includes(callback)) {\n      throw new FunctionExistsError('entrypoints cannot be added more than once');\n    }\n\n    this.#participantEntrypoints.push(callback);\n  }\n}\n\nexport class JobProcess {\n  #pid = process.pid;\n  userData: { [id: string]: unknown } = {};\n\n  get pid(): number {\n    return this.#pid;\n  }\n}\n\n/**\n * A request sent by the server to spawn a new agent job.\n *\n * @remarks\n * For most applications, this is best left to the default, which simply accepts the job and\n * handles the logic inside the entrypoint function. This class is useful for vetting which\n * requests should fill idle processes and which should be outright rejected.\n */\nexport class JobRequest {\n  #job: proto.Job;\n  #onReject: () => Promise<void>;\n  #onAccept: (args: JobAcceptArguments) => Promise<void>;\n\n  /** @internal */\n  constructor(\n    job: proto.Job,\n    onReject: () => Promise<void>,\n    onAccept: (args: JobAcceptArguments) => Promise<void>,\n  ) {\n    this.#job = job;\n    this.#onReject = onReject;\n    this.#onAccept = onAccept;\n  }\n\n  /** @returns The ID of the job, set by the LiveKit server */\n  get id(): string {\n    return this.#job.id;\n  }\n\n  /** @see {@link https://www.npmjs.com/package/@livekit/protocol | @livekit/protocol} */\n  get job(): proto.Job {\n    return this.#job;\n  }\n\n  /** @see {@link https://www.npmjs.com/package/@livekit/protocol | @livekit/protocol} */\n  get room(): proto.Room | undefined {\n    return this.#job.room;\n  }\n\n  /** @see {@link https://www.npmjs.com/package/@livekit/protocol | @livekit/protocol} */\n  get publisher(): proto.ParticipantInfo | undefined {\n    return this.#job.participant;\n  }\n\n  /** @returns The agent's name, as set in {@link WorkerOptions} */\n  get agentName(): string {\n    return this.#job.agentName;\n  }\n\n  /** Rejects the job. */\n  async reject() {\n    await this.#onReject();\n  }\n\n  /** Accepts the job, launching it on an idle child process. */\n  async accept(name = '', identity = '', metadata = '', attributes?: { [key: string]: string }) {\n    if (identity === '') identity = 'agent-' + this.id;\n\n    this.#onAccept({ name, identity, metadata, attributes });\n  }\n}\n"],"mappings":";;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAWA,sBAAsD;AAEtD,iBAAoB;AAGb,IAAK,gBAAL,kBAAKA,mBAAL;AACL,EAAAA,8BAAA;AACA,EAAAA,8BAAA;AACA,EAAAA,8BAAA;AACA,EAAAA,8BAAA;AAJU,SAAAA;AAAA,GAAA;AAsBL,MAAM,4BAA4B,MAAM;AAAA,EAC7C,YAAY,KAAc;AACxB,UAAM,GAAG;AACT,WAAO,eAAe,MAAM,WAAW,SAAS;AAAA,EAClD;AACF;AAGO,MAAM,WAAW;AAAA,EACtB;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA;AAAA,EAEA,oBAA6C,CAAC;AAAA,EAC9C,0BAAwF,CAAC;AAAA,EACzF,oBAKI,CAAC;AAAA,EACL;AAAA,EAEA,YACE,MACA,MACA,MACA,WACA,YACA;AACA,SAAK,QAAQ;AACb,SAAK,QAAQ;AACb,SAAK,QAAQ;AACb,SAAK,aAAa;AAClB,SAAK,cAAc;AACnB,SAAK,yBAAyB,KAAK,uBAAuB,KAAK,IAAI;AACnE,SAAK,MAAM,GAAG,0BAAU,sBAAsB,KAAK,sBAAsB;AACzE,SAAK,cAAU,gBAAI,EAAE,MAAM,EAAE,MAAM,KAAK,MAAM,CAAC;AAAA,EACjD;AAAA,EAEA,IAAI,OAAmB;AACrB,WAAO,KAAK;AAAA,EACd;AAAA,EAEA,IAAI,MAAiB;AACnB,WAAO,KAAK,MAAM;AAAA,EACpB;AAAA;AAAA,EAGA,IAAI,OAAa;AACf,WAAO,KAAK;AAAA,EACd;AAAA;AAAA,EAGA,IAAI,QAAsC;AACxC,WAAO,KAAK,MAAM;AAAA,EACpB;AAAA;AAAA,EAGA,oBAAoB,UAA+B;AACjD,SAAK,kBAAkB,KAAK,QAAQ;AAAA,EACtC;AAAA,EAEA,MAAM,mBAAmB,UAA+C;AACtE,QAAI,CAAC,KAAK,MAAM,aAAa;AAC3B,YAAM,IAAI,MAAM,uBAAuB;AAAA,IACzC;AAEA,eAAW,KAAK,KAAK,MAAM,mBAAmB,OAAO,GAAG;AACtD,WAAK,CAAC,YAAY,EAAE,aAAa,aAAa,EAAE,KAAK,QAAQ,gCAAgB,OAAO;AAClF,eAAO;AAAA,MACT;AAAA,IACF;AAEA,WAAO,IAAI,QAAQ,CAAC,SAAS,WAAW;AACtC,YAAM,yBAAyB,CAAC,gBAAmC;AACjE,aACG,CAAC,YAAY,YAAY,aAAa,aACvC,YAAY,KAAK,QAAQ,gCAAgB,OACzC;AACA,wBAAc;AACd,kBAAQ,WAAW;AAAA,QACrB;AAAA,MACF;AACA,YAAM,iBAAiB,MAAM;AAC3B,sBAAc;AACd,eAAO,IAAI,MAAM,iDAAiD,CAAC;AAAA,MACrE;AAEA,YAAM,gBAAgB,MAAM;AAC1B,aAAK,MAAM,IAAI,0BAAU,sBAAsB,sBAAsB;AACrE,aAAK,MAAM,IAAI,0BAAU,cAAc,cAAc;AAAA,MACvD;AAEA,WAAK,MAAM,GAAG,0BAAU,sBAAsB,sBAAsB;AACpE,WAAK,MAAM,GAAG,0BAAU,cAAc,cAAc;AAAA,IACtD,CAAC;AAAA,EACH;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,MAAM,QACJ,MACA,gBAA+B,uBAC/B,WACA;AACA,UAAM,OAAO;AAAA,MACX;AAAA,MACA,eAAe,iBAAiB;AAAA,MAChC;AAAA,MACA,UAAU;AAAA,IACZ;AAEA,UAAM,KAAK,MAAM,QAAQ,KAAK,MAAM,KAAK,KAAK,MAAM,OAAO,IAAI;AAC/D,SAAK,WAAW;AAEhB,SAAK,MAAM,mBAAmB,QAAQ,KAAK,sBAAsB;AAEjE,QAAI,CAAC,oBAA0B,kBAAwB,EAAE,SAAS,aAAa,GAAG;AAChF,WAAK,MAAM,mBAAmB,QAAQ,CAAC,MAAM;AAC3C,UAAE,kBAAkB,QAAQ,CAAC,QAAQ;AACnC,cACG,kBAAkB,sBAA4B,IAAI,SAAS,0BAAU,cACrE,kBAAkB,sBAA4B,IAAI,SAAS,0BAAU,YACtE;AACA,gBAAI,cAAc,IAAI;AAAA,UACxB;AAAA,QACF,CAAC;AAAA,MACH,CAAC;AAAA,IACH;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAS,SAAS,IAAI;AACpB,SAAK,YAAY,MAAM;AAAA,EACzB;AAAA;AAAA,EAGA,uBAAuB,GAAsB;AA9L/C;AA+LI,eAAW,YAAY,KAAK,yBAAyB;AACnD,YAAI,UAAK,kBAAkB,EAAE,QAAQ,MAAjC,mBAAoC,aAAY,UAAU;AAC5D,aAAK,QAAQ;AAAA,UACX;AAAA,UACA,EAAE;AAAA,QACJ;AAAA,MACF;AACA,YAAM,SAAS,SAAS,MAAM,CAAC;AAC/B,aAAO,QAAQ,MAAM,OAAO,KAAK,kBAAkB,EAAE,QAAQ,CAAC;AAC9D,WAAK,kBAAkB,EAAE,QAAQ,IAAI,EAAE,UAAU,OAAO;AAAA,IAC1D;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,yBAAyB,UAAoE;AAC3F,QAAI,KAAK,wBAAwB,SAAS,QAAQ,GAAG;AACnD,YAAM,IAAI,oBAAoB,4CAA4C;AAAA,IAC5E;AAEA,SAAK,wBAAwB,KAAK,QAAQ;AAAA,EAC5C;AACF;AAEO,MAAM,WAAW;AAAA,EACtB,OAAO,QAAQ;AAAA,EACf,WAAsC,CAAC;AAAA,EAEvC,IAAI,MAAc;AAChB,WAAO,KAAK;AAAA,EACd;AACF;AAUO,MAAM,WAAW;AAAA,EACtB;AAAA,EACA;AAAA,EACA;AAAA;AAAA,EAGA,YACE,KACA,UACA,UACA;AACA,SAAK,OAAO;AACZ,SAAK,YAAY;AACjB,SAAK,YAAY;AAAA,EACnB;AAAA;AAAA,EAGA,IAAI,KAAa;AACf,WAAO,KAAK,KAAK;AAAA,EACnB;AAAA;AAAA,EAGA,IAAI,MAAiB;AACnB,WAAO,KAAK;AAAA,EACd;AAAA;AAAA,EAGA,IAAI,OAA+B;AACjC,WAAO,KAAK,KAAK;AAAA,EACnB;AAAA;AAAA,EAGA,IAAI,YAA+C;AACjD,WAAO,KAAK,KAAK;AAAA,EACnB;AAAA;AAAA,EAGA,IAAI,YAAoB;AACtB,WAAO,KAAK,KAAK;AAAA,EACnB;AAAA;AAAA,EAGA,MAAM,SAAS;AACb,UAAM,KAAK,UAAU;AAAA,EACvB;AAAA;AAAA,EAGA,MAAM,OAAO,OAAO,IAAI,WAAW,IAAI,WAAW,IAAI,YAAwC;AAC5F,QAAI,aAAa,GAAI,YAAW,WAAW,KAAK;AAEhD,SAAK,UAAU,EAAE,MAAM,UAAU,UAAU,WAAW,CAAC;AAAA,EACzD;AACF;","names":["AutoSubscribe"]}

package/dist/job.d.ts

@@ -11,6 +11,9 @@ export type JobAcceptArguments = {
     name: string;
     identity: string;
     metadata: string;
+    attributes?: {
+        [key: string]: string;
+    };
 };
 export type RunningJobInfo = {
     acceptArguments: JobAcceptArguments;
@@ -95,6 +98,8 @@ export declare class JobRequest {
     /** Rejects the job. */
     reject(): Promise<void>;
     /** Accepts the job, launching it on an idle child process. */
-    accept(name?: string, identity?: string, metadata?: string): Promise<void>;
+    accept(name?: string, identity?: string, metadata?: string, attributes?: {
+        [key: string]: string;
+    }): Promise<void>;
 }
 //# sourceMappingURL=job.d.ts.map

package/dist/job.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"job.d.ts","sourceRoot":"","sources":["../src/job.ts"],"names":[],"mappings":"AAGA,OAAO,KAAK,KAAK,KAAK,MAAM,mBAAmB,CAAC;AAChD,OAAO,KAAK,EACV,WAAW,EACX,gBAAgB,EAChB,iBAAiB,EACjB,IAAI,EACJ,gBAAgB,EACjB,MAAM,mBAAmB,CAAC;AAK3B,yEAAyE;AACzE,oBAAY,aAAa;IACvB,aAAa,IAAA;IACb,cAAc,IAAA;IACd,UAAU,IAAA;IACV,UAAU,IAAA;CACX;AAED,MAAM,MAAM,kBAAkB,GAAG;IAC/B,IAAI,EAAE,MAAM,CAAC;IACb,QAAQ,EAAE,MAAM,CAAC;IACjB,QAAQ,EAAE,MAAM,CAAC;CAClB,CAAC;AAEF,MAAM,MAAM,cAAc,GAAG;IAC3B,eAAe,EAAE,kBAAkB,CAAC;IACpC,GAAG,EAAE,KAAK,CAAC,GAAG,CAAC;IACf,GAAG,EAAE,MAAM,CAAC;IACZ,KAAK,EAAE,MAAM,CAAC;CACf,CAAC;AAEF,6EAA6E;AAC7E,qBAAa,mBAAoB,SAAQ,KAAK;gBAChC,GAAG,CAAC,EAAE,MAAM;CAIzB;AAED,mGAAmG;AACnG,qBAAa,UAAU;;IAMrB,gBAAgB;IAChB,iBAAiB,EAAE,CAAC,MAAM,OAAO,CAAC,IAAI,CAAC,CAAC,EAAE,CAAM;gBAW9C,IAAI,EAAE,UAAU,EAChB,IAAI,EAAE,cAAc,EACpB,IAAI,EAAE,IAAI,EACV,SAAS,EAAE,MAAM,IAAI,EACrB,UAAU,EAAE,CAAC,CAAC,EAAE,MAAM,KAAK,IAAI;IAYjC,IAAI,IAAI,IAAI,UAAU,CAErB;IAED,IAAI,GAAG,IAAI,KAAK,CAAC,GAAG,CAEnB;IAED,kDAAkD;IAClD,IAAI,IAAI,IAAI,IAAI,CAEf;IAED,uFAAuF;IACvF,IAAI,KAAK,IAAI,gBAAgB,GAAG,SAAS,CAExC;IAED,0FAA0F;IAC1F,mBAAmB,CAAC,QAAQ,EAAE,MAAM,OAAO,CAAC,IAAI,CAAC;IAI3C,kBAAkB,CAAC,QAAQ,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,iBAAiB,CAAC;IAoCvE;;;;;;;;;OASG;IACG,OAAO,CACX,IAAI,CAAC,EAAE,WAAW,EAClB,aAAa,GAAE,aAA2C,EAC1D,SAAS,CAAC,EAAE,gBAAgB;IA4B9B;;;;OAIG;IACH,QAAQ,CAAC,MAAM,SAAK;IAIpB,gBAAgB;IAChB,sBAAsB,CAAC,CAAC,EAAE,iBAAiB;IAc3C;;;;OAIG;IACH,wBAAwB,CAAC,QAAQ,EAAE,CAAC,GAAG,EAAE,UAAU,EAAE,CAAC,EAAE,iBAAiB,KAAK,OAAO,CAAC,IAAI,CAAC;CAO5F;AAED,qBAAa,UAAU;;IAErB,QAAQ,EAAE;QAAE,CAAC,EAAE,EAAE,MAAM,GAAG,OAAO,CAAA;KAAE,CAAM;IAEzC,IAAI,GAAG,IAAI,MAAM,CAEhB;CACF;AAED;;;;;;;GAOG;AACH,qBAAa,UAAU;;IAKrB,gBAAgB;gBAEd,GAAG,EAAE,KAAK,CAAC,GAAG,EACd,QAAQ,EAAE,MAAM,OAAO,CAAC,IAAI,CAAC,EAC7B,QAAQ,EAAE,CAAC,IAAI,EAAE,kBAAkB,KAAK,OAAO,CAAC,IAAI,CAAC;IAOvD,4DAA4D;IAC5D,IAAI,EAAE,IAAI,MAAM,CAEf;IAED,uFAAuF;IACvF,IAAI,GAAG,IAAI,KAAK,CAAC,GAAG,CAEnB;IAED,uFAAuF;IACvF,IAAI,IAAI,IAAI,KAAK,CAAC,IAAI,GAAG,SAAS,CAEjC;IAED,uFAAuF;IACvF,IAAI,SAAS,IAAI,KAAK,CAAC,eAAe,GAAG,SAAS,CAEjD;IAED,iEAAiE;IACjE,IAAI,SAAS,IAAI,MAAM,CAEtB;IAED,uBAAuB;IACjB,MAAM;IAIZ,8DAA8D;IACxD,MAAM,CAAC,IAAI,SAAK,EAAE,QAAQ,SAAK,EAAE,QAAQ,SAAK;CAKrD"}
+{"version":3,"file":"job.d.ts","sourceRoot":"","sources":["../src/job.ts"],"names":[],"mappings":"AAGA,OAAO,KAAK,KAAK,KAAK,MAAM,mBAAmB,CAAC;AAChD,OAAO,KAAK,EACV,WAAW,EACX,gBAAgB,EAChB,iBAAiB,EACjB,IAAI,EACJ,gBAAgB,EACjB,MAAM,mBAAmB,CAAC;AAK3B,yEAAyE;AACzE,oBAAY,aAAa;IACvB,aAAa,IAAA;IACb,cAAc,IAAA;IACd,UAAU,IAAA;IACV,UAAU,IAAA;CACX;AAED,MAAM,MAAM,kBAAkB,GAAG;IAC/B,IAAI,EAAE,MAAM,CAAC;IACb,QAAQ,EAAE,MAAM,CAAC;IACjB,QAAQ,EAAE,MAAM,CAAC;IACjB,UAAU,CAAC,EAAE;QAAE,CAAC,GAAG,EAAE,MAAM,GAAG,MAAM,CAAA;KAAE,CAAC;CACxC,CAAC;AAEF,MAAM,MAAM,cAAc,GAAG;IAC3B,eAAe,EAAE,kBAAkB,CAAC;IACpC,GAAG,EAAE,KAAK,CAAC,GAAG,CAAC;IACf,GAAG,EAAE,MAAM,CAAC;IACZ,KAAK,EAAE,MAAM,CAAC;CACf,CAAC;AAEF,6EAA6E;AAC7E,qBAAa,mBAAoB,SAAQ,KAAK;gBAChC,GAAG,CAAC,EAAE,MAAM;CAIzB;AAED,mGAAmG;AACnG,qBAAa,UAAU;;IAMrB,gBAAgB;IAChB,iBAAiB,EAAE,CAAC,MAAM,OAAO,CAAC,IAAI,CAAC,CAAC,EAAE,CAAM;gBAW9C,IAAI,EAAE,UAAU,EAChB,IAAI,EAAE,cAAc,EACpB,IAAI,EAAE,IAAI,EACV,SAAS,EAAE,MAAM,IAAI,EACrB,UAAU,EAAE,CAAC,CAAC,EAAE,MAAM,KAAK,IAAI;IAYjC,IAAI,IAAI,IAAI,UAAU,CAErB;IAED,IAAI,GAAG,IAAI,KAAK,CAAC,GAAG,CAEnB;IAED,kDAAkD;IAClD,IAAI,IAAI,IAAI,IAAI,CAEf;IAED,uFAAuF;IACvF,IAAI,KAAK,IAAI,gBAAgB,GAAG,SAAS,CAExC;IAED,0FAA0F;IAC1F,mBAAmB,CAAC,QAAQ,EAAE,MAAM,OAAO,CAAC,IAAI,CAAC;IAI3C,kBAAkB,CAAC,QAAQ,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,iBAAiB,CAAC;IAoCvE;;;;;;;;;OASG;IACG,OAAO,CACX,IAAI,CAAC,EAAE,WAAW,EAClB,aAAa,GAAE,aAA2C,EAC1D,SAAS,CAAC,EAAE,gBAAgB;IA4B9B;;;;OAIG;IACH,QAAQ,CAAC,MAAM,SAAK;IAIpB,gBAAgB;IAChB,sBAAsB,CAAC,CAAC,EAAE,iBAAiB;IAc3C;;;;OAIG;IACH,wBAAwB,CAAC,QAAQ,EAAE,CAAC,GAAG,EAAE,UAAU,EAAE,CAAC,EAAE,iBAAiB,KAAK,OAAO,CAAC,IAAI,CAAC;CAO5F;AAED,qBAAa,UAAU;;IAErB,QAAQ,EAAE;QAAE,CAAC,EAAE,EAAE,MAAM,GAAG,OAAO,CAAA;KAAE,CAAM;IAEzC,IAAI,GAAG,IAAI,MAAM,CAEhB;CACF;AAED;;;;;;;GAOG;AACH,qBAAa,UAAU;;IAKrB,gBAAgB;gBAEd,GAAG,EAAE,KAAK,CAAC,GAAG,EACd,QAAQ,EAAE,MAAM,OAAO,CAAC,IAAI,CAAC,EAC7B,QAAQ,EAAE,CAAC,IAAI,EAAE,kBAAkB,KAAK,OAAO,CAAC,IAAI,CAAC;IAOvD,4DAA4D;IAC5D,IAAI,EAAE,IAAI,MAAM,CAEf;IAED,uFAAuF;IACvF,IAAI,GAAG,IAAI,KAAK,CAAC,GAAG,CAEnB;IAED,uFAAuF;IACvF,IAAI,IAAI,IAAI,KAAK,CAAC,IAAI,GAAG,SAAS,CAEjC;IAED,uFAAuF;IACvF,IAAI,SAAS,IAAI,KAAK,CAAC,eAAe,GAAG,SAAS,CAEjD;IAED,iEAAiE;IACjE,IAAI,SAAS,IAAI,MAAM,CAEtB;IAED,uBAAuB;IACjB,MAAM;IAIZ,8DAA8D;IACxD,MAAM,CAAC,IAAI,SAAK,EAAE,QAAQ,SAAK,EAAE,QAAQ,SAAK,EAAE,UAAU,CAAC,EAAE;QAAE,CAAC,GAAG,EAAE,MAAM,GAAG,MAAM,CAAA;KAAE;CAK7F"}

package/dist/job.js

@@ -187,9 +187,9 @@ class JobRequest {
     await this.#onReject();
   }
   /** Accepts the job, launching it on an idle child process. */
-  async accept(name = "", identity = "", metadata = "") {
+  async accept(name = "", identity = "", metadata = "", attributes) {
     if (identity === "") identity = "agent-" + this.id;
-    this.#onAccept({ name, identity, metadata });
+    this.#onAccept({ name, identity, metadata, attributes });
   }
 }
 export {

package/dist/job.js.map

@@ -1 +1 @@
-{"version":3,"sources":["../src/job.ts"],"sourcesContent":["// SPDX-FileCopyrightText: 2024 LiveKit, Inc.\n//\n// SPDX-License-Identifier: Apache-2.0\nimport type * as proto from '@livekit/protocol';\nimport type {\n  E2EEOptions,\n  LocalParticipant,\n  RemoteParticipant,\n  Room,\n  RtcConfiguration,\n} from '@livekit/rtc-node';\nimport { ParticipantKind, RoomEvent, TrackKind } from '@livekit/rtc-node';\nimport type { Logger } from 'pino';\nimport { log } from './log.js';\n\n/** Which tracks, if any, should the agent automatically subscribe to? */\nexport enum AutoSubscribe {\n  SUBSCRIBE_ALL,\n  SUBSCRIBE_NONE,\n  VIDEO_ONLY,\n  AUDIO_ONLY,\n}\n\nexport type JobAcceptArguments = {\n  name: string;\n  identity: string;\n  metadata: string;\n};\n\nexport type RunningJobInfo = {\n  acceptArguments: JobAcceptArguments;\n  job: proto.Job;\n  url: string;\n  token: string;\n};\n\n/** Attempted to add a function callback, but the function already exists. */\nexport class FunctionExistsError extends Error {\n  constructor(msg?: string) {\n    super(msg);\n    Object.setPrototypeOf(this, new.target.prototype);\n  }\n}\n\n/** The job and environment context as seen by the agent, accessible by the entrypoint function. */\nexport class JobContext {\n  #proc: JobProcess;\n  #info: RunningJobInfo;\n  #room: Room;\n  #onConnect: () => void;\n  #onShutdown: (s: string) => void;\n  /** @internal */\n  shutdownCallbacks: (() => Promise<void>)[] = [];\n  #participantEntrypoints: ((job: JobContext, p: RemoteParticipant) => Promise<void>)[] = [];\n  #participantTasks: {\n    [id: string]: {\n      callback: (job: JobContext, p: RemoteParticipant) => Promise<void>;\n      result: Promise<void>;\n    };\n  } = {};\n  #logger: Logger;\n\n  constructor(\n    proc: JobProcess,\n    info: RunningJobInfo,\n    room: Room,\n    onConnect: () => void,\n    onShutdown: (s: string) => void,\n  ) {\n    this.#proc = proc;\n    this.#info = info;\n    this.#room = room;\n    this.#onConnect = onConnect;\n    this.#onShutdown = onShutdown;\n    this.onParticipantConnected = this.onParticipantConnected.bind(this);\n    this.#room.on(RoomEvent.ParticipantConnected, this.onParticipantConnected);\n    this.#logger = log().child({ info: this.#info });\n  }\n\n  get proc(): JobProcess {\n    return this.#proc;\n  }\n\n  get job(): proto.Job {\n    return this.#info.job;\n  }\n\n  /** @returns The room the agent was called into */\n  get room(): Room {\n    return this.#room;\n  }\n\n  /** @returns The agent's participant if connected to the room, otherwise `undefined` */\n  get agent(): LocalParticipant | undefined {\n    return this.#room.localParticipant;\n  }\n\n  /** Adds a promise to be awaited when {@link JobContext.shutdown | shutdown} is called. */\n  addShutdownCallback(callback: () => Promise<void>) {\n    this.shutdownCallbacks.push(callback);\n  }\n\n  async waitForParticipant(identity?: string): Promise<RemoteParticipant> {\n    if (!this.#room.isConnected) {\n      throw new Error('room is not connected');\n    }\n\n    for (const p of this.#room.remoteParticipants.values()) {\n      if ((!identity || p.identity === identity) && p.info.kind != ParticipantKind.AGENT) {\n        return p;\n      }\n    }\n\n    return new Promise((resolve, reject) => {\n      const onParticipantConnected = (participant: RemoteParticipant) => {\n        if (\n          (!identity || participant.identity === identity) &&\n          participant.info.kind != ParticipantKind.AGENT\n        ) {\n          clearHandlers();\n          resolve(participant);\n        }\n      };\n      const onDisconnected = () => {\n        clearHandlers();\n        reject(new Error('Room disconnected while waiting for participant'));\n      };\n\n      const clearHandlers = () => {\n        this.#room.off(RoomEvent.ParticipantConnected, onParticipantConnected);\n        this.#room.off(RoomEvent.Disconnected, onDisconnected);\n      };\n\n      this.#room.on(RoomEvent.ParticipantConnected, onParticipantConnected);\n      this.#room.on(RoomEvent.Disconnected, onDisconnected);\n    });\n  }\n\n  /**\n   * Connects the agent to the room.\n   *\n   * @remarks\n   * It is recommended to run this command as early in the function as possible, as executing it\n   * later may cause noticeable delay between user and agent joins.\n   *\n   * @see {@link https://github.com/livekit/node-sdks/tree/main/packages/livekit-rtc#readme |\n   * @livekit/rtc-node} for more information about the parameters.\n   */\n  async connect(\n    e2ee?: E2EEOptions,\n    autoSubscribe: AutoSubscribe = AutoSubscribe.SUBSCRIBE_ALL,\n    rtcConfig?: RtcConfiguration,\n  ) {\n    const opts = {\n      e2ee,\n      autoSubscribe: autoSubscribe == AutoSubscribe.SUBSCRIBE_ALL,\n      rtcConfig,\n      dynacast: false,\n    };\n\n    await this.#room.connect(this.#info.url, this.#info.token, opts);\n    this.#onConnect();\n\n    this.#room.remoteParticipants.forEach(this.onParticipantConnected);\n\n    if ([AutoSubscribe.AUDIO_ONLY, AutoSubscribe.VIDEO_ONLY].includes(autoSubscribe)) {\n      this.#room.remoteParticipants.forEach((p) => {\n        p.trackPublications.forEach((pub) => {\n          if (\n            (autoSubscribe === AutoSubscribe.AUDIO_ONLY && pub.kind === TrackKind.KIND_AUDIO) ||\n            (autoSubscribe === AutoSubscribe.VIDEO_ONLY && pub.kind === TrackKind.KIND_VIDEO)\n          ) {\n            pub.setSubscribed(true);\n          }\n        });\n      });\n    }\n  }\n\n  /**\n   * Gracefully shuts down the job, and runs all shutdown promises.\n   *\n   * @param reason - Optional reason for shutdown\n   */\n  shutdown(reason = '') {\n    this.#onShutdown(reason);\n  }\n\n  /** @internal */\n  onParticipantConnected(p: RemoteParticipant) {\n    for (const callback of this.#participantEntrypoints) {\n      if (this.#participantTasks[p.identity]?.callback == callback) {\n        this.#logger.warn(\n          'a participant has joined before a prior prticipant task matching the same identity has finished:',\n          p.identity,\n        );\n      }\n      const result = callback(this, p);\n      result.finally(() => delete this.#participantTasks[p.identity]);\n      this.#participantTasks[p.identity] = { callback, result };\n    }\n  }\n\n  /**\n   * Adds a promise to be awaited whenever a new participant joins the room.\n   *\n   * @throws {@link FunctionExistsError} if an entrypoint already exists\n   */\n  addParticipantEntrypoint(callback: (job: JobContext, p: RemoteParticipant) => Promise<void>) {\n    if (this.#participantEntrypoints.includes(callback)) {\n      throw new FunctionExistsError('entrypoints cannot be added more than once');\n    }\n\n    this.#participantEntrypoints.push(callback);\n  }\n}\n\nexport class JobProcess {\n  #pid = process.pid;\n  userData: { [id: string]: unknown } = {};\n\n  get pid(): number {\n    return this.#pid;\n  }\n}\n\n/**\n * A request sent by the server to spawn a new agent job.\n *\n * @remarks\n * For most applications, this is best left to the default, which simply accepts the job and\n * handles the logic inside the entrypoint function. This class is useful for vetting which\n * requests should fill idle processes and which should be outright rejected.\n */\nexport class JobRequest {\n  #job: proto.Job;\n  #onReject: () => Promise<void>;\n  #onAccept: (args: JobAcceptArguments) => Promise<void>;\n\n  /** @internal */\n  constructor(\n    job: proto.Job,\n    onReject: () => Promise<void>,\n    onAccept: (args: JobAcceptArguments) => Promise<void>,\n  ) {\n    this.#job = job;\n    this.#onReject = onReject;\n    this.#onAccept = onAccept;\n  }\n\n  /** @returns The ID of the job, set by the LiveKit server */\n  get id(): string {\n    return this.#job.id;\n  }\n\n  /** @see {@link https://www.npmjs.com/package/@livekit/protocol | @livekit/protocol} */\n  get job(): proto.Job {\n    return this.#job;\n  }\n\n  /** @see {@link https://www.npmjs.com/package/@livekit/protocol | @livekit/protocol} */\n  get room(): proto.Room | undefined {\n    return this.#job.room;\n  }\n\n  /** @see {@link https://www.npmjs.com/package/@livekit/protocol | @livekit/protocol} */\n  get publisher(): proto.ParticipantInfo | undefined {\n    return this.#job.participant;\n  }\n\n  /** @returns The agent's name, as set in {@link WorkerOptions} */\n  get agentName(): string {\n    return this.#job.agentName;\n  }\n\n  /** Rejects the job. */\n  async reject() {\n    await this.#onReject();\n  }\n\n  /** Accepts the job, launching it on an idle child process. */\n  async accept(name = '', identity = '', metadata = '') {\n    if (identity === '') identity = 'agent-' + this.id;\n\n    this.#onAccept({ name, identity, metadata });\n  }\n}\n"],"mappings":"AAWA,SAAS,iBAAiB,WAAW,iBAAiB;AAEtD,SAAS,WAAW;AAGb,IAAK,gBAAL,kBAAKA,mBAAL;AACL,EAAAA,8BAAA;AACA,EAAAA,8BAAA;AACA,EAAAA,8BAAA;AACA,EAAAA,8BAAA;AAJU,SAAAA;AAAA,GAAA;AAqBL,MAAM,4BAA4B,MAAM;AAAA,EAC7C,YAAY,KAAc;AACxB,UAAM,GAAG;AACT,WAAO,eAAe,MAAM,WAAW,SAAS;AAAA,EAClD;AACF;AAGO,MAAM,WAAW;AAAA,EACtB;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA;AAAA,EAEA,oBAA6C,CAAC;AAAA,EAC9C,0BAAwF,CAAC;AAAA,EACzF,oBAKI,CAAC;AAAA,EACL;AAAA,EAEA,YACE,MACA,MACA,MACA,WACA,YACA;AACA,SAAK,QAAQ;AACb,SAAK,QAAQ;AACb,SAAK,QAAQ;AACb,SAAK,aAAa;AAClB,SAAK,cAAc;AACnB,SAAK,yBAAyB,KAAK,uBAAuB,KAAK,IAAI;AACnE,SAAK,MAAM,GAAG,UAAU,sBAAsB,KAAK,sBAAsB;AACzE,SAAK,UAAU,IAAI,EAAE,MAAM,EAAE,MAAM,KAAK,MAAM,CAAC;AAAA,EACjD;AAAA,EAEA,IAAI,OAAmB;AACrB,WAAO,KAAK;AAAA,EACd;AAAA,EAEA,IAAI,MAAiB;AACnB,WAAO,KAAK,MAAM;AAAA,EACpB;AAAA;AAAA,EAGA,IAAI,OAAa;AACf,WAAO,KAAK;AAAA,EACd;AAAA;AAAA,EAGA,IAAI,QAAsC;AACxC,WAAO,KAAK,MAAM;AAAA,EACpB;AAAA;AAAA,EAGA,oBAAoB,UAA+B;AACjD,SAAK,kBAAkB,KAAK,QAAQ;AAAA,EACtC;AAAA,EAEA,MAAM,mBAAmB,UAA+C;AACtE,QAAI,CAAC,KAAK,MAAM,aAAa;AAC3B,YAAM,IAAI,MAAM,uBAAuB;AAAA,IACzC;AAEA,eAAW,KAAK,KAAK,MAAM,mBAAmB,OAAO,GAAG;AACtD,WAAK,CAAC,YAAY,EAAE,aAAa,aAAa,EAAE,KAAK,QAAQ,gBAAgB,OAAO;AAClF,eAAO;AAAA,MACT;AAAA,IACF;AAEA,WAAO,IAAI,QAAQ,CAAC,SAAS,WAAW;AACtC,YAAM,yBAAyB,CAAC,gBAAmC;AACjE,aACG,CAAC,YAAY,YAAY,aAAa,aACvC,YAAY,KAAK,QAAQ,gBAAgB,OACzC;AACA,wBAAc;AACd,kBAAQ,WAAW;AAAA,QACrB;AAAA,MACF;AACA,YAAM,iBAAiB,MAAM;AAC3B,sBAAc;AACd,eAAO,IAAI,MAAM,iDAAiD,CAAC;AAAA,MACrE;AAEA,YAAM,gBAAgB,MAAM;AAC1B,aAAK,MAAM,IAAI,UAAU,sBAAsB,sBAAsB;AACrE,aAAK,MAAM,IAAI,UAAU,cAAc,cAAc;AAAA,MACvD;AAEA,WAAK,MAAM,GAAG,UAAU,sBAAsB,sBAAsB;AACpE,WAAK,MAAM,GAAG,UAAU,cAAc,cAAc;AAAA,IACtD,CAAC;AAAA,EACH;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,MAAM,QACJ,MACA,gBAA+B,uBAC/B,WACA;AACA,UAAM,OAAO;AAAA,MACX;AAAA,MACA,eAAe,iBAAiB;AAAA,MAChC;AAAA,MACA,UAAU;AAAA,IACZ;AAEA,UAAM,KAAK,MAAM,QAAQ,KAAK,MAAM,KAAK,KAAK,MAAM,OAAO,IAAI;AAC/D,SAAK,WAAW;AAEhB,SAAK,MAAM,mBAAmB,QAAQ,KAAK,sBAAsB;AAEjE,QAAI,CAAC,oBAA0B,kBAAwB,EAAE,SAAS,aAAa,GAAG;AAChF,WAAK,MAAM,mBAAmB,QAAQ,CAAC,MAAM;AAC3C,UAAE,kBAAkB,QAAQ,CAAC,QAAQ;AACnC,cACG,kBAAkB,sBAA4B,IAAI,SAAS,UAAU,cACrE,kBAAkB,sBAA4B,IAAI,SAAS,UAAU,YACtE;AACA,gBAAI,cAAc,IAAI;AAAA,UACxB;AAAA,QACF,CAAC;AAAA,MACH,CAAC;AAAA,IACH;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAS,SAAS,IAAI;AACpB,SAAK,YAAY,MAAM;AAAA,EACzB;AAAA;AAAA,EAGA,uBAAuB,GAAsB;AA7L/C;AA8LI,eAAW,YAAY,KAAK,yBAAyB;AACnD,YAAI,UAAK,kBAAkB,EAAE,QAAQ,MAAjC,mBAAoC,aAAY,UAAU;AAC5D,aAAK,QAAQ;AAAA,UACX;AAAA,UACA,EAAE;AAAA,QACJ;AAAA,MACF;AACA,YAAM,SAAS,SAAS,MAAM,CAAC;AAC/B,aAAO,QAAQ,MAAM,OAAO,KAAK,kBAAkB,EAAE,QAAQ,CAAC;AAC9D,WAAK,kBAAkB,EAAE,QAAQ,IAAI,EAAE,UAAU,OAAO;AAAA,IAC1D;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,yBAAyB,UAAoE;AAC3F,QAAI,KAAK,wBAAwB,SAAS,QAAQ,GAAG;AACnD,YAAM,IAAI,oBAAoB,4CAA4C;AAAA,IAC5E;AAEA,SAAK,wBAAwB,KAAK,QAAQ;AAAA,EAC5C;AACF;AAEO,MAAM,WAAW;AAAA,EACtB,OAAO,QAAQ;AAAA,EACf,WAAsC,CAAC;AAAA,EAEvC,IAAI,MAAc;AAChB,WAAO,KAAK;AAAA,EACd;AACF;AAUO,MAAM,WAAW;AAAA,EACtB;AAAA,EACA;AAAA,EACA;AAAA;AAAA,EAGA,YACE,KACA,UACA,UACA;AACA,SAAK,OAAO;AACZ,SAAK,YAAY;AACjB,SAAK,YAAY;AAAA,EACnB;AAAA;AAAA,EAGA,IAAI,KAAa;AACf,WAAO,KAAK,KAAK;AAAA,EACnB;AAAA;AAAA,EAGA,IAAI,MAAiB;AACnB,WAAO,KAAK;AAAA,EACd;AAAA;AAAA,EAGA,IAAI,OAA+B;AACjC,WAAO,KAAK,KAAK;AAAA,EACnB;AAAA;AAAA,EAGA,IAAI,YAA+C;AACjD,WAAO,KAAK,KAAK;AAAA,EACnB;AAAA;AAAA,EAGA,IAAI,YAAoB;AACtB,WAAO,KAAK,KAAK;AAAA,EACnB;AAAA;AAAA,EAGA,MAAM,SAAS;AACb,UAAM,KAAK,UAAU;AAAA,EACvB;AAAA;AAAA,EAGA,MAAM,OAAO,OAAO,IAAI,WAAW,IAAI,WAAW,IAAI;AACpD,QAAI,aAAa,GAAI,YAAW,WAAW,KAAK;AAEhD,SAAK,UAAU,EAAE,MAAM,UAAU,SAAS,CAAC;AAAA,EAC7C;AACF;","names":["AutoSubscribe"]}
+{"version":3,"sources":["../src/job.ts"],"sourcesContent":["// SPDX-FileCopyrightText: 2024 LiveKit, Inc.\n//\n// SPDX-License-Identifier: Apache-2.0\nimport type * as proto from '@livekit/protocol';\nimport type {\n  E2EEOptions,\n  LocalParticipant,\n  RemoteParticipant,\n  Room,\n  RtcConfiguration,\n} from '@livekit/rtc-node';\nimport { ParticipantKind, RoomEvent, TrackKind } from '@livekit/rtc-node';\nimport type { Logger } from 'pino';\nimport { log } from './log.js';\n\n/** Which tracks, if any, should the agent automatically subscribe to? */\nexport enum AutoSubscribe {\n  SUBSCRIBE_ALL,\n  SUBSCRIBE_NONE,\n  VIDEO_ONLY,\n  AUDIO_ONLY,\n}\n\nexport type JobAcceptArguments = {\n  name: string;\n  identity: string;\n  metadata: string;\n  attributes?: { [key: string]: string };\n};\n\nexport type RunningJobInfo = {\n  acceptArguments: JobAcceptArguments;\n  job: proto.Job;\n  url: string;\n  token: string;\n};\n\n/** Attempted to add a function callback, but the function already exists. */\nexport class FunctionExistsError extends Error {\n  constructor(msg?: string) {\n    super(msg);\n    Object.setPrototypeOf(this, new.target.prototype);\n  }\n}\n\n/** The job and environment context as seen by the agent, accessible by the entrypoint function. */\nexport class JobContext {\n  #proc: JobProcess;\n  #info: RunningJobInfo;\n  #room: Room;\n  #onConnect: () => void;\n  #onShutdown: (s: string) => void;\n  /** @internal */\n  shutdownCallbacks: (() => Promise<void>)[] = [];\n  #participantEntrypoints: ((job: JobContext, p: RemoteParticipant) => Promise<void>)[] = [];\n  #participantTasks: {\n    [id: string]: {\n      callback: (job: JobContext, p: RemoteParticipant) => Promise<void>;\n      result: Promise<void>;\n    };\n  } = {};\n  #logger: Logger;\n\n  constructor(\n    proc: JobProcess,\n    info: RunningJobInfo,\n    room: Room,\n    onConnect: () => void,\n    onShutdown: (s: string) => void,\n  ) {\n    this.#proc = proc;\n    this.#info = info;\n    this.#room = room;\n    this.#onConnect = onConnect;\n    this.#onShutdown = onShutdown;\n    this.onParticipantConnected = this.onParticipantConnected.bind(this);\n    this.#room.on(RoomEvent.ParticipantConnected, this.onParticipantConnected);\n    this.#logger = log().child({ info: this.#info });\n  }\n\n  get proc(): JobProcess {\n    return this.#proc;\n  }\n\n  get job(): proto.Job {\n    return this.#info.job;\n  }\n\n  /** @returns The room the agent was called into */\n  get room(): Room {\n    return this.#room;\n  }\n\n  /** @returns The agent's participant if connected to the room, otherwise `undefined` */\n  get agent(): LocalParticipant | undefined {\n    return this.#room.localParticipant;\n  }\n\n  /** Adds a promise to be awaited when {@link JobContext.shutdown | shutdown} is called. */\n  addShutdownCallback(callback: () => Promise<void>) {\n    this.shutdownCallbacks.push(callback);\n  }\n\n  async waitForParticipant(identity?: string): Promise<RemoteParticipant> {\n    if (!this.#room.isConnected) {\n      throw new Error('room is not connected');\n    }\n\n    for (const p of this.#room.remoteParticipants.values()) {\n      if ((!identity || p.identity === identity) && p.info.kind != ParticipantKind.AGENT) {\n        return p;\n      }\n    }\n\n    return new Promise((resolve, reject) => {\n      const onParticipantConnected = (participant: RemoteParticipant) => {\n        if (\n          (!identity || participant.identity === identity) &&\n          participant.info.kind != ParticipantKind.AGENT\n        ) {\n          clearHandlers();\n          resolve(participant);\n        }\n      };\n      const onDisconnected = () => {\n        clearHandlers();\n        reject(new Error('Room disconnected while waiting for participant'));\n      };\n\n      const clearHandlers = () => {\n        this.#room.off(RoomEvent.ParticipantConnected, onParticipantConnected);\n        this.#room.off(RoomEvent.Disconnected, onDisconnected);\n      };\n\n      this.#room.on(RoomEvent.ParticipantConnected, onParticipantConnected);\n      this.#room.on(RoomEvent.Disconnected, onDisconnected);\n    });\n  }\n\n  /**\n   * Connects the agent to the room.\n   *\n   * @remarks\n   * It is recommended to run this command as early in the function as possible, as executing it\n   * later may cause noticeable delay between user and agent joins.\n   *\n   * @see {@link https://github.com/livekit/node-sdks/tree/main/packages/livekit-rtc#readme |\n   * @livekit/rtc-node} for more information about the parameters.\n   */\n  async connect(\n    e2ee?: E2EEOptions,\n    autoSubscribe: AutoSubscribe = AutoSubscribe.SUBSCRIBE_ALL,\n    rtcConfig?: RtcConfiguration,\n  ) {\n    const opts = {\n      e2ee,\n      autoSubscribe: autoSubscribe == AutoSubscribe.SUBSCRIBE_ALL,\n      rtcConfig,\n      dynacast: false,\n    };\n\n    await this.#room.connect(this.#info.url, this.#info.token, opts);\n    this.#onConnect();\n\n    this.#room.remoteParticipants.forEach(this.onParticipantConnected);\n\n    if ([AutoSubscribe.AUDIO_ONLY, AutoSubscribe.VIDEO_ONLY].includes(autoSubscribe)) {\n      this.#room.remoteParticipants.forEach((p) => {\n        p.trackPublications.forEach((pub) => {\n          if (\n            (autoSubscribe === AutoSubscribe.AUDIO_ONLY && pub.kind === TrackKind.KIND_AUDIO) ||\n            (autoSubscribe === AutoSubscribe.VIDEO_ONLY && pub.kind === TrackKind.KIND_VIDEO)\n          ) {\n            pub.setSubscribed(true);\n          }\n        });\n      });\n    }\n  }\n\n  /**\n   * Gracefully shuts down the job, and runs all shutdown promises.\n   *\n   * @param reason - Optional reason for shutdown\n   */\n  shutdown(reason = '') {\n    this.#onShutdown(reason);\n  }\n\n  /** @internal */\n  onParticipantConnected(p: RemoteParticipant) {\n    for (const callback of this.#participantEntrypoints) {\n      if (this.#participantTasks[p.identity]?.callback == callback) {\n        this.#logger.warn(\n          'a participant has joined before a prior prticipant task matching the same identity has finished:',\n          p.identity,\n        );\n      }\n      const result = callback(this, p);\n      result.finally(() => delete this.#participantTasks[p.identity]);\n      this.#participantTasks[p.identity] = { callback, result };\n    }\n  }\n\n  /**\n   * Adds a promise to be awaited whenever a new participant joins the room.\n   *\n   * @throws {@link FunctionExistsError} if an entrypoint already exists\n   */\n  addParticipantEntrypoint(callback: (job: JobContext, p: RemoteParticipant) => Promise<void>) {\n    if (this.#participantEntrypoints.includes(callback)) {\n      throw new FunctionExistsError('entrypoints cannot be added more than once');\n    }\n\n    this.#participantEntrypoints.push(callback);\n  }\n}\n\nexport class JobProcess {\n  #pid = process.pid;\n  userData: { [id: string]: unknown } = {};\n\n  get pid(): number {\n    return this.#pid;\n  }\n}\n\n/**\n * A request sent by the server to spawn a new agent job.\n *\n * @remarks\n * For most applications, this is best left to the default, which simply accepts the job and\n * handles the logic inside the entrypoint function. This class is useful for vetting which\n * requests should fill idle processes and which should be outright rejected.\n */\nexport class JobRequest {\n  #job: proto.Job;\n  #onReject: () => Promise<void>;\n  #onAccept: (args: JobAcceptArguments) => Promise<void>;\n\n  /** @internal */\n  constructor(\n    job: proto.Job,\n    onReject: () => Promise<void>,\n    onAccept: (args: JobAcceptArguments) => Promise<void>,\n  ) {\n    this.#job = job;\n    this.#onReject = onReject;\n    this.#onAccept = onAccept;\n  }\n\n  /** @returns The ID of the job, set by the LiveKit server */\n  get id(): string {\n    return this.#job.id;\n  }\n\n  /** @see {@link https://www.npmjs.com/package/@livekit/protocol | @livekit/protocol} */\n  get job(): proto.Job {\n    return this.#job;\n  }\n\n  /** @see {@link https://www.npmjs.com/package/@livekit/protocol | @livekit/protocol} */\n  get room(): proto.Room | undefined {\n    return this.#job.room;\n  }\n\n  /** @see {@link https://www.npmjs.com/package/@livekit/protocol | @livekit/protocol} */\n  get publisher(): proto.ParticipantInfo | undefined {\n    return this.#job.participant;\n  }\n\n  /** @returns The agent's name, as set in {@link WorkerOptions} */\n  get agentName(): string {\n    return this.#job.agentName;\n  }\n\n  /** Rejects the job. */\n  async reject() {\n    await this.#onReject();\n  }\n\n  /** Accepts the job, launching it on an idle child process. */\n  async accept(name = '', identity = '', metadata = '', attributes?: { [key: string]: string }) {\n    if (identity === '') identity = 'agent-' + this.id;\n\n    this.#onAccept({ name, identity, metadata, attributes });\n  }\n}\n"],"mappings":"AAWA,SAAS,iBAAiB,WAAW,iBAAiB;AAEtD,SAAS,WAAW;AAGb,IAAK,gBAAL,kBAAKA,mBAAL;AACL,EAAAA,8BAAA;AACA,EAAAA,8BAAA;AACA,EAAAA,8BAAA;AACA,EAAAA,8BAAA;AAJU,SAAAA;AAAA,GAAA;AAsBL,MAAM,4BAA4B,MAAM;AAAA,EAC7C,YAAY,KAAc;AACxB,UAAM,GAAG;AACT,WAAO,eAAe,MAAM,WAAW,SAAS;AAAA,EAClD;AACF;AAGO,MAAM,WAAW;AAAA,EACtB;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA;AAAA,EAEA,oBAA6C,CAAC;AAAA,EAC9C,0BAAwF,CAAC;AAAA,EACzF,oBAKI,CAAC;AAAA,EACL;AAAA,EAEA,YACE,MACA,MACA,MACA,WACA,YACA;AACA,SAAK,QAAQ;AACb,SAAK,QAAQ;AACb,SAAK,QAAQ;AACb,SAAK,aAAa;AAClB,SAAK,cAAc;AACnB,SAAK,yBAAyB,KAAK,uBAAuB,KAAK,IAAI;AACnE,SAAK,MAAM,GAAG,UAAU,sBAAsB,KAAK,sBAAsB;AACzE,SAAK,UAAU,IAAI,EAAE,MAAM,EAAE,MAAM,KAAK,MAAM,CAAC;AAAA,EACjD;AAAA,EAEA,IAAI,OAAmB;AACrB,WAAO,KAAK;AAAA,EACd;AAAA,EAEA,IAAI,MAAiB;AACnB,WAAO,KAAK,MAAM;AAAA,EACpB;AAAA;AAAA,EAGA,IAAI,OAAa;AACf,WAAO,KAAK;AAAA,EACd;AAAA;AAAA,EAGA,IAAI,QAAsC;AACxC,WAAO,KAAK,MAAM;AAAA,EACpB;AAAA;AAAA,EAGA,oBAAoB,UAA+B;AACjD,SAAK,kBAAkB,KAAK,QAAQ;AAAA,EACtC;AAAA,EAEA,MAAM,mBAAmB,UAA+C;AACtE,QAAI,CAAC,KAAK,MAAM,aAAa;AAC3B,YAAM,IAAI,MAAM,uBAAuB;AAAA,IACzC;AAEA,eAAW,KAAK,KAAK,MAAM,mBAAmB,OAAO,GAAG;AACtD,WAAK,CAAC,YAAY,EAAE,aAAa,aAAa,EAAE,KAAK,QAAQ,gBAAgB,OAAO;AAClF,eAAO;AAAA,MACT;AAAA,IACF;AAEA,WAAO,IAAI,QAAQ,CAAC,SAAS,WAAW;AACtC,YAAM,yBAAyB,CAAC,gBAAmC;AACjE,aACG,CAAC,YAAY,YAAY,aAAa,aACvC,YAAY,KAAK,QAAQ,gBAAgB,OACzC;AACA,wBAAc;AACd,kBAAQ,WAAW;AAAA,QACrB;AAAA,MACF;AACA,YAAM,iBAAiB,MAAM;AAC3B,sBAAc;AACd,eAAO,IAAI,MAAM,iDAAiD,CAAC;AAAA,MACrE;AAEA,YAAM,gBAAgB,MAAM;AAC1B,aAAK,MAAM,IAAI,UAAU,sBAAsB,sBAAsB;AACrE,aAAK,MAAM,IAAI,UAAU,cAAc,cAAc;AAAA,MACvD;AAEA,WAAK,MAAM,GAAG,UAAU,sBAAsB,sBAAsB;AACpE,WAAK,MAAM,GAAG,UAAU,cAAc,cAAc;AAAA,IACtD,CAAC;AAAA,EACH;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,MAAM,QACJ,MACA,gBAA+B,uBAC/B,WACA;AACA,UAAM,OAAO;AAAA,MACX;AAAA,MACA,eAAe,iBAAiB;AAAA,MAChC;AAAA,MACA,UAAU;AAAA,IACZ;AAEA,UAAM,KAAK,MAAM,QAAQ,KAAK,MAAM,KAAK,KAAK,MAAM,OAAO,IAAI;AAC/D,SAAK,WAAW;AAEhB,SAAK,MAAM,mBAAmB,QAAQ,KAAK,sBAAsB;AAEjE,QAAI,CAAC,oBAA0B,kBAAwB,EAAE,SAAS,aAAa,GAAG;AAChF,WAAK,MAAM,mBAAmB,QAAQ,CAAC,MAAM;AAC3C,UAAE,kBAAkB,QAAQ,CAAC,QAAQ;AACnC,cACG,kBAAkB,sBAA4B,IAAI,SAAS,UAAU,cACrE,kBAAkB,sBAA4B,IAAI,SAAS,UAAU,YACtE;AACA,gBAAI,cAAc,IAAI;AAAA,UACxB;AAAA,QACF,CAAC;AAAA,MACH,CAAC;AAAA,IACH;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAS,SAAS,IAAI;AACpB,SAAK,YAAY,MAAM;AAAA,EACzB;AAAA;AAAA,EAGA,uBAAuB,GAAsB;AA9L/C;AA+LI,eAAW,YAAY,KAAK,yBAAyB;AACnD,YAAI,UAAK,kBAAkB,EAAE,QAAQ,MAAjC,mBAAoC,aAAY,UAAU;AAC5D,aAAK,QAAQ;AAAA,UACX;AAAA,UACA,EAAE;AAAA,QACJ;AAAA,MACF;AACA,YAAM,SAAS,SAAS,MAAM,CAAC;AAC/B,aAAO,QAAQ,MAAM,OAAO,KAAK,kBAAkB,EAAE,QAAQ,CAAC;AAC9D,WAAK,kBAAkB,EAAE,QAAQ,IAAI,EAAE,UAAU,OAAO;AAAA,IAC1D;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,yBAAyB,UAAoE;AAC3F,QAAI,KAAK,wBAAwB,SAAS,QAAQ,GAAG;AACnD,YAAM,IAAI,oBAAoB,4CAA4C;AAAA,IAC5E;AAEA,SAAK,wBAAwB,KAAK,QAAQ;AAAA,EAC5C;AACF;AAEO,MAAM,WAAW;AAAA,EACtB,OAAO,QAAQ;AAAA,EACf,WAAsC,CAAC;AAAA,EAEvC,IAAI,MAAc;AAChB,WAAO,KAAK;AAAA,EACd;AACF;AAUO,MAAM,WAAW;AAAA,EACtB;AAAA,EACA;AAAA,EACA;AAAA;AAAA,EAGA,YACE,KACA,UACA,UACA;AACA,SAAK,OAAO;AACZ,SAAK,YAAY;AACjB,SAAK,YAAY;AAAA,EACnB;AAAA;AAAA,EAGA,IAAI,KAAa;AACf,WAAO,KAAK,KAAK;AAAA,EACnB;AAAA;AAAA,EAGA,IAAI,MAAiB;AACnB,WAAO,KAAK;AAAA,EACd;AAAA;AAAA,EAGA,IAAI,OAA+B;AACjC,WAAO,KAAK,KAAK;AAAA,EACnB;AAAA;AAAA,EAGA,IAAI,YAA+C;AACjD,WAAO,KAAK,KAAK;AAAA,EACnB;AAAA;AAAA,EAGA,IAAI,YAAoB;AACtB,WAAO,KAAK,KAAK;AAAA,EACnB;AAAA;AAAA,EAGA,MAAM,SAAS;AACb,UAAM,KAAK,UAAU;AAAA,EACvB;AAAA;AAAA,EAGA,MAAM,OAAO,OAAO,IAAI,WAAW,IAAI,WAAW,IAAI,YAAwC;AAC5F,QAAI,aAAa,GAAI,YAAW,WAAW,KAAK;AAEhD,SAAK,UAAU,EAAE,MAAM,UAAU,UAAU,WAAW,CAAC;AAAA,EACzD;AACF;","names":["AutoSubscribe"]}

package/dist/tokenize/basic/basic.cjs

@@ -34,8 +34,8 @@ __export(basic_exports, {
   tokenizeParagraphs: () => tokenizeParagraphs
 });
 module.exports = __toCommonJS(basic_exports);
-var tokenizer = __toESM(require("../index.cjs"), 1);
 var import_token_stream = require("../token_stream.cjs");
+var tokenizer = __toESM(require("../tokenizer.cjs"), 1);
 var import_hyphenator = require("./hyphenator.cjs");
 var import_paragraph = require("./paragraph.cjs");
 var import_sentence = require("./sentence.cjs");

package/dist/tokenize/basic/basic.cjs.map

@@ -1 +1 @@
-{"version":3,"sources":["../../../src/tokenize/basic/basic.ts"],"sourcesContent":["// SPDX-FileCopyrightText: 2024 LiveKit, Inc.\n//\n// SPDX-License-Identifier: Apache-2.0\nimport * as tokenizer from '../index.js';\nimport { BufferedSentenceStream, BufferedWordStream } from '../token_stream.js';\nimport { hyphenator } from './hyphenator.js';\nimport { splitParagraphs } from './paragraph.js';\nimport { splitSentences } from './sentence.js';\nimport { splitWords } from './word.js';\n\ninterface TokenizerOptions {\n  language: string;\n  minSentenceLength: number;\n  streamContextLength: number;\n}\n\nexport class SentenceTokenizer extends tokenizer.SentenceTokenizer {\n  #config: TokenizerOptions;\n\n  constructor(language = 'en-US', minSentenceLength = 20, streamContextLength = 10) {\n    super();\n    this.#config = {\n      language,\n      minSentenceLength,\n      streamContextLength,\n    };\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-unused-vars\n  tokenize(text: string, language?: string): string[] {\n    return splitSentences(text, this.#config.minSentenceLength).map((tok) => tok[0]);\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-unused-vars\n  stream(language?: string): tokenizer.SentenceStream {\n    return new BufferedSentenceStream(\n      (text: string) => splitSentences(text, this.#config.minSentenceLength),\n      this.#config.minSentenceLength,\n      this.#config.streamContextLength,\n    );\n  }\n}\n\nexport class WordTokenizer extends tokenizer.WordTokenizer {\n  #ignorePunctuation: boolean;\n\n  constructor(ignorePunctuation = true) {\n    super();\n    this.#ignorePunctuation = ignorePunctuation;\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-unused-vars\n  tokenize(text: string, language?: string): string[] {\n    return splitWords(text, this.#ignorePunctuation).map((tok) => tok[0]);\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-unused-vars\n  stream(language?: string): tokenizer.WordStream {\n    return new BufferedWordStream(\n      (text: string) => splitWords(text, this.#ignorePunctuation),\n      1,\n      1,\n    );\n  }\n}\n\nexport const hyphenateWord = (word: string): string[] => {\n  return hyphenator.hyphenateWord(word);\n};\n\nexport const tokenizeParagraphs = (text: string): string[] => {\n  return splitParagraphs(text).map((tok) => tok[0]);\n};\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAGA,gBAA2B;AAC3B,0BAA2D;AAC3D,wBAA2B;AAC3B,uBAAgC;AAChC,sBAA+B;AAC/B,kBAA2B;AAQpB,MAAM,0BAA0B,UAAU,kBAAkB;AAAA,EACjE;AAAA,EAEA,YAAY,WAAW,SAAS,oBAAoB,IAAI,sBAAsB,IAAI;AAChF,UAAM;AACN,SAAK,UAAU;AAAA,MACb;AAAA,MACA;AAAA,MACA;AAAA,IACF;AAAA,EACF;AAAA;AAAA,EAGA,SAAS,MAAc,UAA6B;AAClD,eAAO,gCAAe,MAAM,KAAK,QAAQ,iBAAiB,EAAE,IAAI,CAAC,QAAQ,IAAI,CAAC,CAAC;AAAA,EACjF;AAAA;AAAA,EAGA,OAAO,UAA6C;AAClD,WAAO,IAAI;AAAA,MACT,CAAC,aAAiB,gCAAe,MAAM,KAAK,QAAQ,iBAAiB;AAAA,MACrE,KAAK,QAAQ;AAAA,MACb,KAAK,QAAQ;AAAA,IACf;AAAA,EACF;AACF;AAEO,MAAM,sBAAsB,UAAU,cAAc;AAAA,EACzD;AAAA,EAEA,YAAY,oBAAoB,MAAM;AACpC,UAAM;AACN,SAAK,qBAAqB;AAAA,EAC5B;AAAA;AAAA,EAGA,SAAS,MAAc,UAA6B;AAClD,eAAO,wBAAW,MAAM,KAAK,kBAAkB,EAAE,IAAI,CAAC,QAAQ,IAAI,CAAC,CAAC;AAAA,EACtE;AAAA;AAAA,EAGA,OAAO,UAAyC;AAC9C,WAAO,IAAI;AAAA,MACT,CAAC,aAAiB,wBAAW,MAAM,KAAK,kBAAkB;AAAA,MAC1D;AAAA,MACA;AAAA,IACF;AAAA,EACF;AACF;AAEO,MAAM,gBAAgB,CAAC,SAA2B;AACvD,SAAO,6BAAW,cAAc,IAAI;AACtC;AAEO,MAAM,qBAAqB,CAAC,SAA2B;AAC5D,aAAO,kCAAgB,IAAI,EAAE,IAAI,CAAC,QAAQ,IAAI,CAAC,CAAC;AAClD;","names":[]}
+{"version":3,"sources":["../../../src/tokenize/basic/basic.ts"],"sourcesContent":["// SPDX-FileCopyrightText: 2024 LiveKit, Inc.\n//\n// SPDX-License-Identifier: Apache-2.0\nimport { BufferedSentenceStream, BufferedWordStream } from '../token_stream.js';\nimport * as tokenizer from '../tokenizer.js';\nimport { hyphenator } from './hyphenator.js';\nimport { splitParagraphs } from './paragraph.js';\nimport { splitSentences } from './sentence.js';\nimport { splitWords } from './word.js';\n\ninterface TokenizerOptions {\n  language: string;\n  minSentenceLength: number;\n  streamContextLength: number;\n}\n\nexport class SentenceTokenizer extends tokenizer.SentenceTokenizer {\n  #config: TokenizerOptions;\n\n  constructor(language = 'en-US', minSentenceLength = 20, streamContextLength = 10) {\n    super();\n    this.#config = {\n      language,\n      minSentenceLength,\n      streamContextLength,\n    };\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-unused-vars\n  tokenize(text: string, language?: string): string[] {\n    return splitSentences(text, this.#config.minSentenceLength).map((tok) => tok[0]);\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-unused-vars\n  stream(language?: string): tokenizer.SentenceStream {\n    return new BufferedSentenceStream(\n      (text: string) => splitSentences(text, this.#config.minSentenceLength),\n      this.#config.minSentenceLength,\n      this.#config.streamContextLength,\n    );\n  }\n}\n\nexport class WordTokenizer extends tokenizer.WordTokenizer {\n  #ignorePunctuation: boolean;\n\n  constructor(ignorePunctuation = true) {\n    super();\n    this.#ignorePunctuation = ignorePunctuation;\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-unused-vars\n  tokenize(text: string, language?: string): string[] {\n    return splitWords(text, this.#ignorePunctuation).map((tok) => tok[0]);\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-unused-vars\n  stream(language?: string): tokenizer.WordStream {\n    return new BufferedWordStream(\n      (text: string) => splitWords(text, this.#ignorePunctuation),\n      1,\n      1,\n    );\n  }\n}\n\nexport const hyphenateWord = (word: string): string[] => {\n  return hyphenator.hyphenateWord(word);\n};\n\nexport const tokenizeParagraphs = (text: string): string[] => {\n  return splitParagraphs(text).map((tok) => tok[0]);\n};\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAGA,0BAA2D;AAC3D,gBAA2B;AAC3B,wBAA2B;AAC3B,uBAAgC;AAChC,sBAA+B;AAC/B,kBAA2B;AAQpB,MAAM,0BAA0B,UAAU,kBAAkB;AAAA,EACjE;AAAA,EAEA,YAAY,WAAW,SAAS,oBAAoB,IAAI,sBAAsB,IAAI;AAChF,UAAM;AACN,SAAK,UAAU;AAAA,MACb;AAAA,MACA;AAAA,MACA;AAAA,IACF;AAAA,EACF;AAAA;AAAA,EAGA,SAAS,MAAc,UAA6B;AAClD,eAAO,gCAAe,MAAM,KAAK,QAAQ,iBAAiB,EAAE,IAAI,CAAC,QAAQ,IAAI,CAAC,CAAC;AAAA,EACjF;AAAA;AAAA,EAGA,OAAO,UAA6C;AAClD,WAAO,IAAI;AAAA,MACT,CAAC,aAAiB,gCAAe,MAAM,KAAK,QAAQ,iBAAiB;AAAA,MACrE,KAAK,QAAQ;AAAA,MACb,KAAK,QAAQ;AAAA,IACf;AAAA,EACF;AACF;AAEO,MAAM,sBAAsB,UAAU,cAAc;AAAA,EACzD;AAAA,EAEA,YAAY,oBAAoB,MAAM;AACpC,UAAM;AACN,SAAK,qBAAqB;AAAA,EAC5B;AAAA;AAAA,EAGA,SAAS,MAAc,UAA6B;AAClD,eAAO,wBAAW,MAAM,KAAK,kBAAkB,EAAE,IAAI,CAAC,QAAQ,IAAI,CAAC,CAAC;AAAA,EACtE;AAAA;AAAA,EAGA,OAAO,UAAyC;AAC9C,WAAO,IAAI;AAAA,MACT,CAAC,aAAiB,wBAAW,MAAM,KAAK,kBAAkB;AAAA,MAC1D;AAAA,MACA;AAAA,IACF;AAAA,EACF;AACF;AAEO,MAAM,gBAAgB,CAAC,SAA2B;AACvD,SAAO,6BAAW,cAAc,IAAI;AACtC;AAEO,MAAM,qBAAqB,CAAC,SAA2B;AAC5D,aAAO,kCAAgB,IAAI,EAAE,IAAI,CAAC,QAAQ,IAAI,CAAC,CAAC;AAClD;","names":[]}

package/dist/tokenize/basic/basic.d.ts

@@ -1,4 +1,4 @@
-import * as tokenizer from '../index.js';
+import * as tokenizer from '../tokenizer.js';
 export declare class SentenceTokenizer extends tokenizer.SentenceTokenizer {
     #private;
     constructor(language?: string, minSentenceLength?: number, streamContextLength?: number);

package/dist/tokenize/basic/basic.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"basic.d.ts","sourceRoot":"","sources":["../../../src/tokenize/basic/basic.ts"],"names":[],"mappings":"AAGA,OAAO,KAAK,SAAS,MAAM,aAAa,CAAC;AAazC,qBAAa,iBAAkB,SAAQ,SAAS,CAAC,iBAAiB;;gBAGpD,QAAQ,SAAU,EAAE,iBAAiB,SAAK,EAAE,mBAAmB,SAAK;IAUhF,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,QAAQ,CAAC,EAAE,MAAM,GAAG,MAAM,EAAE;IAKnD,MAAM,CAAC,QAAQ,CAAC,EAAE,MAAM,GAAG,SAAS,CAAC,cAAc;CAOpD;AAED,qBAAa,aAAc,SAAQ,SAAS,CAAC,aAAa;;gBAG5C,iBAAiB,UAAO;IAMpC,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,QAAQ,CAAC,EAAE,MAAM,GAAG,MAAM,EAAE;IAKnD,MAAM,CAAC,QAAQ,CAAC,EAAE,MAAM,GAAG,SAAS,CAAC,UAAU;CAOhD;AAED,eAAO,MAAM,aAAa,SAAU,MAAM,KAAG,MAAM,EAElD,CAAC;AAEF,eAAO,MAAM,kBAAkB,SAAU,MAAM,KAAG,MAAM,EAEvD,CAAC"}
+{"version":3,"file":"basic.d.ts","sourceRoot":"","sources":["../../../src/tokenize/basic/basic.ts"],"names":[],"mappings":"AAIA,OAAO,KAAK,SAAS,MAAM,iBAAiB,CAAC;AAY7C,qBAAa,iBAAkB,SAAQ,SAAS,CAAC,iBAAiB;;gBAGpD,QAAQ,SAAU,EAAE,iBAAiB,SAAK,EAAE,mBAAmB,SAAK;IAUhF,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,QAAQ,CAAC,EAAE,MAAM,GAAG,MAAM,EAAE;IAKnD,MAAM,CAAC,QAAQ,CAAC,EAAE,MAAM,GAAG,SAAS,CAAC,cAAc;CAOpD;AAED,qBAAa,aAAc,SAAQ,SAAS,CAAC,aAAa;;gBAG5C,iBAAiB,UAAO;IAMpC,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,QAAQ,CAAC,EAAE,MAAM,GAAG,MAAM,EAAE;IAKnD,MAAM,CAAC,QAAQ,CAAC,EAAE,MAAM,GAAG,SAAS,CAAC,UAAU;CAOhD;AAED,eAAO,MAAM,aAAa,SAAU,MAAM,KAAG,MAAM,EAElD,CAAC;AAEF,eAAO,MAAM,kBAAkB,SAAU,MAAM,KAAG,MAAM,EAEvD,CAAC"}

package/dist/tokenize/basic/basic.js

@@ -1,5 +1,5 @@
-import * as tokenizer from "../index.js";
 import { BufferedSentenceStream, BufferedWordStream } from "../token_stream.js";
+import * as tokenizer from "../tokenizer.js";
 import { hyphenator } from "./hyphenator.js";
 import { splitParagraphs } from "./paragraph.js";
 import { splitSentences } from "./sentence.js";

package/dist/tokenize/basic/basic.js.map

@@ -1 +1 @@
-{"version":3,"sources":["../../../src/tokenize/basic/basic.ts"],"sourcesContent":["// SPDX-FileCopyrightText: 2024 LiveKit, Inc.\n//\n// SPDX-License-Identifier: Apache-2.0\nimport * as tokenizer from '../index.js';\nimport { BufferedSentenceStream, BufferedWordStream } from '../token_stream.js';\nimport { hyphenator } from './hyphenator.js';\nimport { splitParagraphs } from './paragraph.js';\nimport { splitSentences } from './sentence.js';\nimport { splitWords } from './word.js';\n\ninterface TokenizerOptions {\n  language: string;\n  minSentenceLength: number;\n  streamContextLength: number;\n}\n\nexport class SentenceTokenizer extends tokenizer.SentenceTokenizer {\n  #config: TokenizerOptions;\n\n  constructor(language = 'en-US', minSentenceLength = 20, streamContextLength = 10) {\n    super();\n    this.#config = {\n      language,\n      minSentenceLength,\n      streamContextLength,\n    };\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-unused-vars\n  tokenize(text: string, language?: string): string[] {\n    return splitSentences(text, this.#config.minSentenceLength).map((tok) => tok[0]);\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-unused-vars\n  stream(language?: string): tokenizer.SentenceStream {\n    return new BufferedSentenceStream(\n      (text: string) => splitSentences(text, this.#config.minSentenceLength),\n      this.#config.minSentenceLength,\n      this.#config.streamContextLength,\n    );\n  }\n}\n\nexport class WordTokenizer extends tokenizer.WordTokenizer {\n  #ignorePunctuation: boolean;\n\n  constructor(ignorePunctuation = true) {\n    super();\n    this.#ignorePunctuation = ignorePunctuation;\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-unused-vars\n  tokenize(text: string, language?: string): string[] {\n    return splitWords(text, this.#ignorePunctuation).map((tok) => tok[0]);\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-unused-vars\n  stream(language?: string): tokenizer.WordStream {\n    return new BufferedWordStream(\n      (text: string) => splitWords(text, this.#ignorePunctuation),\n      1,\n      1,\n    );\n  }\n}\n\nexport const hyphenateWord = (word: string): string[] => {\n  return hyphenator.hyphenateWord(word);\n};\n\nexport const tokenizeParagraphs = (text: string): string[] => {\n  return splitParagraphs(text).map((tok) => tok[0]);\n};\n"],"mappings":"AAGA,YAAY,eAAe;AAC3B,SAAS,wBAAwB,0BAA0B;AAC3D,SAAS,kBAAkB;AAC3B,SAAS,uBAAuB;AAChC,SAAS,sBAAsB;AAC/B,SAAS,kBAAkB;AAQpB,MAAM,0BAA0B,UAAU,kBAAkB;AAAA,EACjE;AAAA,EAEA,YAAY,WAAW,SAAS,oBAAoB,IAAI,sBAAsB,IAAI;AAChF,UAAM;AACN,SAAK,UAAU;AAAA,MACb;AAAA,MACA;AAAA,MACA;AAAA,IACF;AAAA,EACF;AAAA;AAAA,EAGA,SAAS,MAAc,UAA6B;AAClD,WAAO,eAAe,MAAM,KAAK,QAAQ,iBAAiB,EAAE,IAAI,CAAC,QAAQ,IAAI,CAAC,CAAC;AAAA,EACjF;AAAA;AAAA,EAGA,OAAO,UAA6C;AAClD,WAAO,IAAI;AAAA,MACT,CAAC,SAAiB,eAAe,MAAM,KAAK,QAAQ,iBAAiB;AAAA,MACrE,KAAK,QAAQ;AAAA,MACb,KAAK,QAAQ;AAAA,IACf;AAAA,EACF;AACF;AAEO,MAAM,sBAAsB,UAAU,cAAc;AAAA,EACzD;AAAA,EAEA,YAAY,oBAAoB,MAAM;AACpC,UAAM;AACN,SAAK,qBAAqB;AAAA,EAC5B;AAAA;AAAA,EAGA,SAAS,MAAc,UAA6B;AAClD,WAAO,WAAW,MAAM,KAAK,kBAAkB,EAAE,IAAI,CAAC,QAAQ,IAAI,CAAC,CAAC;AAAA,EACtE;AAAA;AAAA,EAGA,OAAO,UAAyC;AAC9C,WAAO,IAAI;AAAA,MACT,CAAC,SAAiB,WAAW,MAAM,KAAK,kBAAkB;AAAA,MAC1D;AAAA,MACA;AAAA,IACF;AAAA,EACF;AACF;AAEO,MAAM,gBAAgB,CAAC,SAA2B;AACvD,SAAO,WAAW,cAAc,IAAI;AACtC;AAEO,MAAM,qBAAqB,CAAC,SAA2B;AAC5D,SAAO,gBAAgB,IAAI,EAAE,IAAI,CAAC,QAAQ,IAAI,CAAC,CAAC;AAClD;","names":[]}
+{"version":3,"sources":["../../../src/tokenize/basic/basic.ts"],"sourcesContent":["// SPDX-FileCopyrightText: 2024 LiveKit, Inc.\n//\n// SPDX-License-Identifier: Apache-2.0\nimport { BufferedSentenceStream, BufferedWordStream } from '../token_stream.js';\nimport * as tokenizer from '../tokenizer.js';\nimport { hyphenator } from './hyphenator.js';\nimport { splitParagraphs } from './paragraph.js';\nimport { splitSentences } from './sentence.js';\nimport { splitWords } from './word.js';\n\ninterface TokenizerOptions {\n  language: string;\n  minSentenceLength: number;\n  streamContextLength: number;\n}\n\nexport class SentenceTokenizer extends tokenizer.SentenceTokenizer {\n  #config: TokenizerOptions;\n\n  constructor(language = 'en-US', minSentenceLength = 20, streamContextLength = 10) {\n    super();\n    this.#config = {\n      language,\n      minSentenceLength,\n      streamContextLength,\n    };\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-unused-vars\n  tokenize(text: string, language?: string): string[] {\n    return splitSentences(text, this.#config.minSentenceLength).map((tok) => tok[0]);\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-unused-vars\n  stream(language?: string): tokenizer.SentenceStream {\n    return new BufferedSentenceStream(\n      (text: string) => splitSentences(text, this.#config.minSentenceLength),\n      this.#config.minSentenceLength,\n      this.#config.streamContextLength,\n    );\n  }\n}\n\nexport class WordTokenizer extends tokenizer.WordTokenizer {\n  #ignorePunctuation: boolean;\n\n  constructor(ignorePunctuation = true) {\n    super();\n    this.#ignorePunctuation = ignorePunctuation;\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-unused-vars\n  tokenize(text: string, language?: string): string[] {\n    return splitWords(text, this.#ignorePunctuation).map((tok) => tok[0]);\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-unused-vars\n  stream(language?: string): tokenizer.WordStream {\n    return new BufferedWordStream(\n      (text: string) => splitWords(text, this.#ignorePunctuation),\n      1,\n      1,\n    );\n  }\n}\n\nexport const hyphenateWord = (word: string): string[] => {\n  return hyphenator.hyphenateWord(word);\n};\n\nexport const tokenizeParagraphs = (text: string): string[] => {\n  return splitParagraphs(text).map((tok) => tok[0]);\n};\n"],"mappings":"AAGA,SAAS,wBAAwB,0BAA0B;AAC3D,YAAY,eAAe;AAC3B,SAAS,kBAAkB;AAC3B,SAAS,uBAAuB;AAChC,SAAS,sBAAsB;AAC/B,SAAS,kBAAkB;AAQpB,MAAM,0BAA0B,UAAU,kBAAkB;AAAA,EACjE;AAAA,EAEA,YAAY,WAAW,SAAS,oBAAoB,IAAI,sBAAsB,IAAI;AAChF,UAAM;AACN,SAAK,UAAU;AAAA,MACb;AAAA,MACA;AAAA,MACA;AAAA,IACF;AAAA,EACF;AAAA;AAAA,EAGA,SAAS,MAAc,UAA6B;AAClD,WAAO,eAAe,MAAM,KAAK,QAAQ,iBAAiB,EAAE,IAAI,CAAC,QAAQ,IAAI,CAAC,CAAC;AAAA,EACjF;AAAA;AAAA,EAGA,OAAO,UAA6C;AAClD,WAAO,IAAI;AAAA,MACT,CAAC,SAAiB,eAAe,MAAM,KAAK,QAAQ,iBAAiB;AAAA,MACrE,KAAK,QAAQ;AAAA,MACb,KAAK,QAAQ;AAAA,IACf;AAAA,EACF;AACF;AAEO,MAAM,sBAAsB,UAAU,cAAc;AAAA,EACzD;AAAA,EAEA,YAAY,oBAAoB,MAAM;AACpC,UAAM;AACN,SAAK,qBAAqB;AAAA,EAC5B;AAAA;AAAA,EAGA,SAAS,MAAc,UAA6B;AAClD,WAAO,WAAW,MAAM,KAAK,kBAAkB,EAAE,IAAI,CAAC,QAAQ,IAAI,CAAC,CAAC;AAAA,EACtE;AAAA;AAAA,EAGA,OAAO,UAAyC;AAC9C,WAAO,IAAI;AAAA,MACT,CAAC,SAAiB,WAAW,MAAM,KAAK,kBAAkB;AAAA,MAC1D;AAAA,MACA;AAAA,IACF;AAAA,EACF;AACF;AAEO,MAAM,gBAAgB,CAAC,SAA2B;AACvD,SAAO,WAAW,cAAc,IAAI;AACtC;AAEO,MAAM,qBAAqB,CAAC,SAA2B;AAC5D,SAAO,gBAAgB,IAAI,EAAE,IAAI,CAAC,QAAQ,IAAI,CAAC,CAAC;AAClD;","names":[]}

package/dist/tokenize/basic/sentence.cjs

@@ -33,19 +33,25 @@ const splitSentences = (text, minLength = 20) => {
   text = text.replaceAll("\n", " ");
   text = text.replaceAll(prefixes, "$1<prd>");
   text = text.replaceAll(websites, "<prd>$2");
-  text = text.replaceAll(new RegExp(`${digits}[.]${digits}`, "g"), "$1<prd>$2");
+  text = text.replaceAll(new RegExp(`${digits.source}[.]${digits.source}`, "g"), "$1<prd>$2");
   text = text.replaceAll(dots, (match) => "<prd>".repeat(match.length));
   text = text.replaceAll("Ph.D.", "Ph<prd>D<prd>");
-  text = text.replaceAll(new RegExp(`s${alphabets}[.]`, "g"), "$1<prd>");
-  text = text.replaceAll(new RegExp(`${acronyms} ${starters}`, "g"), "$1<stop> $2");
+  text = text.replaceAll(new RegExp(`\\s${alphabets.source}[.] `, "g"), " $1<prd> ");
+  text = text.replaceAll(new RegExp(`${acronyms.source} ${starters.source}`, "g"), "$1<stop> $2");
   text = text.replaceAll(
-    new RegExp(`${alphabets}[.]${alphabets}[.]${alphabets}[.]`, "g"),
+    new RegExp(`${alphabets.source}[.]${alphabets.source}[.]${alphabets.source}[.]`, "g"),
     "$1<prd>$2<prd>$3<prd>"
   );
-  text = text.replaceAll(new RegExp(`${alphabets}[.]${alphabets}[.]`, "g"), "$1<prd>$2<prd>");
-  text = text.replaceAll(new RegExp(` ${suffixes}[.] ${starters}`, "g"), "$1<stop> $2");
-  text = text.replaceAll(new RegExp(` ${suffixes}[.]`, "g"), "$1<prd>");
-  text = text.replaceAll(new RegExp(` ${alphabets}[.]`, "g"), "$1<prd>");
+  text = text.replaceAll(
+    new RegExp(`${alphabets.source}[.]${alphabets.source}[.]`, "g"),
+    "$1<prd>$2<prd>"
+  );
+  text = text.replaceAll(
+    new RegExp(` ${suffixes.source}[.] ${starters.source}`, "g"),
+    "$1<stop> $2"
+  );
+  text = text.replaceAll(new RegExp(` ${suffixes.source}[.]`, "g"), "$1<prd>");
+  text = text.replaceAll(new RegExp(` ${alphabets.source}[.]`, "g"), "$1<prd>");
   text = text.replaceAll(".\u201D", "\u201D.");
   text = text.replaceAll('."', '".');
   text = text.replaceAll('!"', '"!');

package/dist/tokenize/basic/sentence.cjs.map

@@ -1 +1 @@
-{"version":3,"sources":["../../../src/tokenize/basic/sentence.ts"],"sourcesContent":["// SPDX-FileCopyrightText: 2024 LiveKit, Inc.\n//\n// SPDX-License-Identifier: Apache-2.0\n\n/**\n * Split the text into sentences.\n */\nexport const splitSentences = (text: string, minLength = 20): [string, number, number][] => {\n  const alphabets = /([A-Za-z])/g;\n  const prefixes = /(Mr|St|Mrs|Ms|Dr)[.]/g;\n  const suffixes = /(Inc|Ltd|Jr|Sr|Co)/g;\n  const starters =\n    /(Mr|Mrs|Ms|Dr|Prof|Capt|Cpt|Lt|He\\s|She\\s|It\\s|They\\s|Their\\s|Our\\s|We\\s|But\\s|However\\s|That\\s|This\\s|Wherever)/g;\n  const acronyms = /([A-Z][.][A-Z][.](?:[A-Z][.])?)/g;\n  const websites = /[.](com|net|org|io|gov|edu|me)/g;\n  const digits = /([0-9])/g;\n  const dots = /\\.{2,}/g;\n\n  text = text.replaceAll('\\n', ' ');\n  text = text.replaceAll(prefixes, '$1<prd>');\n  text = text.replaceAll(websites, '<prd>$2');\n  text = text.replaceAll(new RegExp(`${digits}[.]${digits}`, 'g'), '$1<prd>$2');\n  text = text.replaceAll(dots, (match) => '<prd>'.repeat(match.length));\n  text = text.replaceAll('Ph.D.', 'Ph<prd>D<prd>');\n  text = text.replaceAll(new RegExp(`\\s${alphabets}[.]`, 'g'), '$1<prd>');\n  text = text.replaceAll(new RegExp(`${acronyms} ${starters}`, 'g'), '$1<stop> $2');\n  text = text.replaceAll(\n    new RegExp(`${alphabets}[.]${alphabets}[.]${alphabets}[.]`, 'g'),\n    '$1<prd>$2<prd>$3<prd>',\n  );\n  text = text.replaceAll(new RegExp(`${alphabets}[.]${alphabets}[.]`, 'g'), '$1<prd>$2<prd>');\n  text = text.replaceAll(new RegExp(` ${suffixes}[.] ${starters}`, 'g'), '$1<stop> $2');\n  text = text.replaceAll(new RegExp(` ${suffixes}[.]`, 'g'), '$1<prd>');\n  text = text.replaceAll(new RegExp(` ${alphabets}[.]`, 'g'), '$1<prd>');\n  text = text.replaceAll('.”', '”.');\n  text = text.replaceAll('.\"', '\".');\n  text = text.replaceAll('!\"', '\"!');\n  text = text.replaceAll('?\"', '\"?');\n  text = text.replaceAll('.', '.<stop>');\n  text = text.replaceAll('?', '?<stop>');\n  text = text.replaceAll('!', '!<stop>');\n  text = text.replaceAll('<prd>', '.');\n\n  const split = text.split('<stop>');\n  text = text.replaceAll('<stop>', '');\n\n  const sentences: [string, number, number][] = [];\n  let buf = '';\n  let start = 0;\n  let end = 0;\n  for (const match of split) {\n    const sentence = match.trim();\n    if (!sentence) continue;\n\n    buf += ' ' + sentence;\n    end += match.length;\n    if (buf.length > minLength) {\n      sentences.push([buf.slice(1), start, end]);\n      start = end;\n      buf = '';\n    }\n  }\n\n  if (buf) {\n    sentences.push([buf.slice(1), start, text.length - 1]);\n  }\n\n  return sentences;\n};\n"],"mappings":";;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;AAAA;AAOO,MAAM,iBAAiB,CAAC,MAAc,YAAY,OAAmC;AAC1F,QAAM,YAAY;AAClB,QAAM,WAAW;AACjB,QAAM,WAAW;AACjB,QAAM,WACJ;AACF,QAAM,WAAW;AACjB,QAAM,WAAW;AACjB,QAAM,SAAS;AACf,QAAM,OAAO;AAEb,SAAO,KAAK,WAAW,MAAM,GAAG;AAChC,SAAO,KAAK,WAAW,UAAU,SAAS;AAC1C,SAAO,KAAK,WAAW,UAAU,SAAS;AAC1C,SAAO,KAAK,WAAW,IAAI,OAAO,GAAG,MAAM,MAAM,MAAM,IAAI,GAAG,GAAG,WAAW;AAC5E,SAAO,KAAK,WAAW,MAAM,CAAC,UAAU,QAAQ,OAAO,MAAM,MAAM,CAAC;AACpE,SAAO,KAAK,WAAW,SAAS,eAAe;AAC/C,SAAO,KAAK,WAAW,IAAI,OAAO,IAAK,SAAS,OAAO,GAAG,GAAG,SAAS;AACtE,SAAO,KAAK,WAAW,IAAI,OAAO,GAAG,QAAQ,IAAI,QAAQ,IAAI,GAAG,GAAG,aAAa;AAChF,SAAO,KAAK;AAAA,IACV,IAAI,OAAO,GAAG,SAAS,MAAM,SAAS,MAAM,SAAS,OAAO,GAAG;AAAA,IAC/D;AAAA,EACF;AACA,SAAO,KAAK,WAAW,IAAI,OAAO,GAAG,SAAS,MAAM,SAAS,OAAO,GAAG,GAAG,gBAAgB;AAC1F,SAAO,KAAK,WAAW,IAAI,OAAO,IAAI,QAAQ,OAAO,QAAQ,IAAI,GAAG,GAAG,aAAa;AACpF,SAAO,KAAK,WAAW,IAAI,OAAO,IAAI,QAAQ,OAAO,GAAG,GAAG,SAAS;AACpE,SAAO,KAAK,WAAW,IAAI,OAAO,IAAI,SAAS,OAAO,GAAG,GAAG,SAAS;AACrE,SAAO,KAAK,WAAW,WAAM,SAAI;AACjC,SAAO,KAAK,WAAW,MAAM,IAAI;AACjC,SAAO,KAAK,WAAW,MAAM,IAAI;AACjC,SAAO,KAAK,WAAW,MAAM,IAAI;AACjC,SAAO,KAAK,WAAW,KAAK,SAAS;AACrC,SAAO,KAAK,WAAW,KAAK,SAAS;AACrC,SAAO,KAAK,WAAW,KAAK,SAAS;AACrC,SAAO,KAAK,WAAW,SAAS,GAAG;AAEnC,QAAM,QAAQ,KAAK,MAAM,QAAQ;AACjC,SAAO,KAAK,WAAW,UAAU,EAAE;AAEnC,QAAM,YAAwC,CAAC;AAC/C,MAAI,MAAM;AACV,MAAI,QAAQ;AACZ,MAAI,MAAM;AACV,aAAW,SAAS,OAAO;AACzB,UAAM,WAAW,MAAM,KAAK;AAC5B,QAAI,CAAC,SAAU;AAEf,WAAO,MAAM;AACb,WAAO,MAAM;AACb,QAAI,IAAI,SAAS,WAAW;AAC1B,gBAAU,KAAK,CAAC,IAAI,MAAM,CAAC,GAAG,OAAO,GAAG,CAAC;AACzC,cAAQ;AACR,YAAM;AAAA,IACR;AAAA,EACF;AAEA,MAAI,KAAK;AACP,cAAU,KAAK,CAAC,IAAI,MAAM,CAAC,GAAG,OAAO,KAAK,SAAS,CAAC,CAAC;AAAA,EACvD;AAEA,SAAO;AACT;","names":[]}
+{"version":3,"sources":["../../../src/tokenize/basic/sentence.ts"],"sourcesContent":["// SPDX-FileCopyrightText: 2024 LiveKit, Inc.\n//\n// SPDX-License-Identifier: Apache-2.0\n\n/**\n * Split the text into sentences.\n */\nexport const splitSentences = (text: string, minLength = 20): [string, number, number][] => {\n  const alphabets = /([A-Za-z])/g;\n  const prefixes = /(Mr|St|Mrs|Ms|Dr)[.]/g;\n  const suffixes = /(Inc|Ltd|Jr|Sr|Co)/g;\n  const starters =\n    /(Mr|Mrs|Ms|Dr|Prof|Capt|Cpt|Lt|He\\s|She\\s|It\\s|They\\s|Their\\s|Our\\s|We\\s|But\\s|However\\s|That\\s|This\\s|Wherever)/g;\n  const acronyms = /([A-Z][.][A-Z][.](?:[A-Z][.])?)/g;\n  const websites = /[.](com|net|org|io|gov|edu|me)/g;\n  const digits = /([0-9])/g;\n  const dots = /\\.{2,}/g;\n\n  text = text.replaceAll('\\n', ' ');\n  text = text.replaceAll(prefixes, '$1<prd>');\n  text = text.replaceAll(websites, '<prd>$2');\n  text = text.replaceAll(new RegExp(`${digits.source}[.]${digits.source}`, 'g'), '$1<prd>$2');\n  text = text.replaceAll(dots, (match) => '<prd>'.repeat(match.length));\n  text = text.replaceAll('Ph.D.', 'Ph<prd>D<prd>');\n  text = text.replaceAll(new RegExp(`\\\\s${alphabets.source}[.] `, 'g'), ' $1<prd> ');\n  text = text.replaceAll(new RegExp(`${acronyms.source} ${starters.source}`, 'g'), '$1<stop> $2');\n  text = text.replaceAll(\n    new RegExp(`${alphabets.source}[.]${alphabets.source}[.]${alphabets.source}[.]`, 'g'),\n    '$1<prd>$2<prd>$3<prd>',\n  );\n  text = text.replaceAll(\n    new RegExp(`${alphabets.source}[.]${alphabets.source}[.]`, 'g'),\n    '$1<prd>$2<prd>',\n  );\n  text = text.replaceAll(\n    new RegExp(` ${suffixes.source}[.] ${starters.source}`, 'g'),\n    '$1<stop> $2',\n  );\n  text = text.replaceAll(new RegExp(` ${suffixes.source}[.]`, 'g'), '$1<prd>');\n  text = text.replaceAll(new RegExp(` ${alphabets.source}[.]`, 'g'), '$1<prd>');\n  text = text.replaceAll('.”', '”.');\n  text = text.replaceAll('.\"', '\".');\n  text = text.replaceAll('!\"', '\"!');\n  text = text.replaceAll('?\"', '\"?');\n  text = text.replaceAll('.', '.<stop>');\n  text = text.replaceAll('?', '?<stop>');\n  text = text.replaceAll('!', '!<stop>');\n  text = text.replaceAll('<prd>', '.');\n\n  const split = text.split('<stop>');\n  text = text.replaceAll('<stop>', '');\n\n  const sentences: [string, number, number][] = [];\n  let buf = '';\n  let start = 0;\n  let end = 0;\n  for (const match of split) {\n    const sentence = match.trim();\n    if (!sentence) continue;\n\n    buf += ' ' + sentence;\n    end += match.length;\n    if (buf.length > minLength) {\n      sentences.push([buf.slice(1), start, end]);\n      start = end;\n      buf = '';\n    }\n  }\n\n  if (buf) {\n    sentences.push([buf.slice(1), start, text.length - 1]);\n  }\n\n  return sentences;\n};\n"],"mappings":";;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;AAAA;AAOO,MAAM,iBAAiB,CAAC,MAAc,YAAY,OAAmC;AAC1F,QAAM,YAAY;AAClB,QAAM,WAAW;AACjB,QAAM,WAAW;AACjB,QAAM,WACJ;AACF,QAAM,WAAW;AACjB,QAAM,WAAW;AACjB,QAAM,SAAS;AACf,QAAM,OAAO;AAEb,SAAO,KAAK,WAAW,MAAM,GAAG;AAChC,SAAO,KAAK,WAAW,UAAU,SAAS;AAC1C,SAAO,KAAK,WAAW,UAAU,SAAS;AAC1C,SAAO,KAAK,WAAW,IAAI,OAAO,GAAG,OAAO,MAAM,MAAM,OAAO,MAAM,IAAI,GAAG,GAAG,WAAW;AAC1F,SAAO,KAAK,WAAW,MAAM,CAAC,UAAU,QAAQ,OAAO,MAAM,MAAM,CAAC;AACpE,SAAO,KAAK,WAAW,SAAS,eAAe;AAC/C,SAAO,KAAK,WAAW,IAAI,OAAO,MAAM,UAAU,MAAM,QAAQ,GAAG,GAAG,WAAW;AACjF,SAAO,KAAK,WAAW,IAAI,OAAO,GAAG,SAAS,MAAM,IAAI,SAAS,MAAM,IAAI,GAAG,GAAG,aAAa;AAC9F,SAAO,KAAK;AAAA,IACV,IAAI,OAAO,GAAG,UAAU,MAAM,MAAM,UAAU,MAAM,MAAM,UAAU,MAAM,OAAO,GAAG;AAAA,IACpF;AAAA,EACF;AACA,SAAO,KAAK;AAAA,IACV,IAAI,OAAO,GAAG,UAAU,MAAM,MAAM,UAAU,MAAM,OAAO,GAAG;AAAA,IAC9D;AAAA,EACF;AACA,SAAO,KAAK;AAAA,IACV,IAAI,OAAO,IAAI,SAAS,MAAM,OAAO,SAAS,MAAM,IAAI,GAAG;AAAA,IAC3D;AAAA,EACF;AACA,SAAO,KAAK,WAAW,IAAI,OAAO,IAAI,SAAS,MAAM,OAAO,GAAG,GAAG,SAAS;AAC3E,SAAO,KAAK,WAAW,IAAI,OAAO,IAAI,UAAU,MAAM,OAAO,GAAG,GAAG,SAAS;AAC5E,SAAO,KAAK,WAAW,WAAM,SAAI;AACjC,SAAO,KAAK,WAAW,MAAM,IAAI;AACjC,SAAO,KAAK,WAAW,MAAM,IAAI;AACjC,SAAO,KAAK,WAAW,MAAM,IAAI;AACjC,SAAO,KAAK,WAAW,KAAK,SAAS;AACrC,SAAO,KAAK,WAAW,KAAK,SAAS;AACrC,SAAO,KAAK,WAAW,KAAK,SAAS;AACrC,SAAO,KAAK,WAAW,SAAS,GAAG;AAEnC,QAAM,QAAQ,KAAK,MAAM,QAAQ;AACjC,SAAO,KAAK,WAAW,UAAU,EAAE;AAEnC,QAAM,YAAwC,CAAC;AAC/C,MAAI,MAAM;AACV,MAAI,QAAQ;AACZ,MAAI,MAAM;AACV,aAAW,SAAS,OAAO;AACzB,UAAM,WAAW,MAAM,KAAK;AAC5B,QAAI,CAAC,SAAU;AAEf,WAAO,MAAM;AACb,WAAO,MAAM;AACb,QAAI,IAAI,SAAS,WAAW;AAC1B,gBAAU,KAAK,CAAC,IAAI,MAAM,CAAC,GAAG,OAAO,GAAG,CAAC;AACzC,cAAQ;AACR,YAAM;AAAA,IACR;AAAA,EACF;AAEA,MAAI,KAAK;AACP,cAAU,KAAK,CAAC,IAAI,MAAM,CAAC,GAAG,OAAO,KAAK,SAAS,CAAC,CAAC;AAAA,EACvD;AAEA,SAAO;AACT;","names":[]}

package/dist/tokenize/basic/sentence.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"sentence.d.ts","sourceRoot":"","sources":["../../../src/tokenize/basic/sentence.ts"],"names":[],"mappings":"AAIA;;GAEG;AACH,eAAO,MAAM,cAAc,SAAU,MAAM,yBAAmB,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,CAAC,EA6DrF,CAAC"}
+{"version":3,"file":"sentence.d.ts","sourceRoot":"","sources":["../../../src/tokenize/basic/sentence.ts"],"names":[],"mappings":"AAIA;;GAEG;AACH,eAAO,MAAM,cAAc,SAAU,MAAM,yBAAmB,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,CAAC,EAmErF,CAAC"}

package/dist/tokenize/basic/sentence.js

@@ -10,19 +10,25 @@ const splitSentences = (text, minLength = 20) => {
   text = text.replaceAll("\n", " ");
   text = text.replaceAll(prefixes, "$1<prd>");
   text = text.replaceAll(websites, "<prd>$2");
-  text = text.replaceAll(new RegExp(`${digits}[.]${digits}`, "g"), "$1<prd>$2");
+  text = text.replaceAll(new RegExp(`${digits.source}[.]${digits.source}`, "g"), "$1<prd>$2");
   text = text.replaceAll(dots, (match) => "<prd>".repeat(match.length));
   text = text.replaceAll("Ph.D.", "Ph<prd>D<prd>");
-  text = text.replaceAll(new RegExp(`s${alphabets}[.]`, "g"), "$1<prd>");
-  text = text.replaceAll(new RegExp(`${acronyms} ${starters}`, "g"), "$1<stop> $2");
+  text = text.replaceAll(new RegExp(`\\s${alphabets.source}[.] `, "g"), " $1<prd> ");
+  text = text.replaceAll(new RegExp(`${acronyms.source} ${starters.source}`, "g"), "$1<stop> $2");
   text = text.replaceAll(
-    new RegExp(`${alphabets}[.]${alphabets}[.]${alphabets}[.]`, "g"),
+    new RegExp(`${alphabets.source}[.]${alphabets.source}[.]${alphabets.source}[.]`, "g"),
     "$1<prd>$2<prd>$3<prd>"
   );
-  text = text.replaceAll(new RegExp(`${alphabets}[.]${alphabets}[.]`, "g"), "$1<prd>$2<prd>");
-  text = text.replaceAll(new RegExp(` ${suffixes}[.] ${starters}`, "g"), "$1<stop> $2");
-  text = text.replaceAll(new RegExp(` ${suffixes}[.]`, "g"), "$1<prd>");
-  text = text.replaceAll(new RegExp(` ${alphabets}[.]`, "g"), "$1<prd>");
+  text = text.replaceAll(
+    new RegExp(`${alphabets.source}[.]${alphabets.source}[.]`, "g"),
+    "$1<prd>$2<prd>"
+  );
+  text = text.replaceAll(
+    new RegExp(` ${suffixes.source}[.] ${starters.source}`, "g"),
+    "$1<stop> $2"
+  );
+  text = text.replaceAll(new RegExp(` ${suffixes.source}[.]`, "g"), "$1<prd>");
+  text = text.replaceAll(new RegExp(` ${alphabets.source}[.]`, "g"), "$1<prd>");
   text = text.replaceAll(".\u201D", "\u201D.");
   text = text.replaceAll('."', '".');
   text = text.replaceAll('!"', '"!');

package/dist/tokenize/basic/sentence.js.map

@@ -1 +1 @@
-{"version":3,"sources":["../../../src/tokenize/basic/sentence.ts"],"sourcesContent":["// SPDX-FileCopyrightText: 2024 LiveKit, Inc.\n//\n// SPDX-License-Identifier: Apache-2.0\n\n/**\n * Split the text into sentences.\n */\nexport const splitSentences = (text: string, minLength = 20): [string, number, number][] => {\n  const alphabets = /([A-Za-z])/g;\n  const prefixes = /(Mr|St|Mrs|Ms|Dr)[.]/g;\n  const suffixes = /(Inc|Ltd|Jr|Sr|Co)/g;\n  const starters =\n    /(Mr|Mrs|Ms|Dr|Prof|Capt|Cpt|Lt|He\\s|She\\s|It\\s|They\\s|Their\\s|Our\\s|We\\s|But\\s|However\\s|That\\s|This\\s|Wherever)/g;\n  const acronyms = /([A-Z][.][A-Z][.](?:[A-Z][.])?)/g;\n  const websites = /[.](com|net|org|io|gov|edu|me)/g;\n  const digits = /([0-9])/g;\n  const dots = /\\.{2,}/g;\n\n  text = text.replaceAll('\\n', ' ');\n  text = text.replaceAll(prefixes, '$1<prd>');\n  text = text.replaceAll(websites, '<prd>$2');\n  text = text.replaceAll(new RegExp(`${digits}[.]${digits}`, 'g'), '$1<prd>$2');\n  text = text.replaceAll(dots, (match) => '<prd>'.repeat(match.length));\n  text = text.replaceAll('Ph.D.', 'Ph<prd>D<prd>');\n  text = text.replaceAll(new RegExp(`\\s${alphabets}[.]`, 'g'), '$1<prd>');\n  text = text.replaceAll(new RegExp(`${acronyms} ${starters}`, 'g'), '$1<stop> $2');\n  text = text.replaceAll(\n    new RegExp(`${alphabets}[.]${alphabets}[.]${alphabets}[.]`, 'g'),\n    '$1<prd>$2<prd>$3<prd>',\n  );\n  text = text.replaceAll(new RegExp(`${alphabets}[.]${alphabets}[.]`, 'g'), '$1<prd>$2<prd>');\n  text = text.replaceAll(new RegExp(` ${suffixes}[.] ${starters}`, 'g'), '$1<stop> $2');\n  text = text.replaceAll(new RegExp(` ${suffixes}[.]`, 'g'), '$1<prd>');\n  text = text.replaceAll(new RegExp(` ${alphabets}[.]`, 'g'), '$1<prd>');\n  text = text.replaceAll('.”', '”.');\n  text = text.replaceAll('.\"', '\".');\n  text = text.replaceAll('!\"', '\"!');\n  text = text.replaceAll('?\"', '\"?');\n  text = text.replaceAll('.', '.<stop>');\n  text = text.replaceAll('?', '?<stop>');\n  text = text.replaceAll('!', '!<stop>');\n  text = text.replaceAll('<prd>', '.');\n\n  const split = text.split('<stop>');\n  text = text.replaceAll('<stop>', '');\n\n  const sentences: [string, number, number][] = [];\n  let buf = '';\n  let start = 0;\n  let end = 0;\n  for (const match of split) {\n    const sentence = match.trim();\n    if (!sentence) continue;\n\n    buf += ' ' + sentence;\n    end += match.length;\n    if (buf.length > minLength) {\n      sentences.push([buf.slice(1), start, end]);\n      start = end;\n      buf = '';\n    }\n  }\n\n  if (buf) {\n    sentences.push([buf.slice(1), start, text.length - 1]);\n  }\n\n  return sentences;\n};\n"],"mappings":"AAOO,MAAM,iBAAiB,CAAC,MAAc,YAAY,OAAmC;AAC1F,QAAM,YAAY;AAClB,QAAM,WAAW;AACjB,QAAM,WAAW;AACjB,QAAM,WACJ;AACF,QAAM,WAAW;AACjB,QAAM,WAAW;AACjB,QAAM,SAAS;AACf,QAAM,OAAO;AAEb,SAAO,KAAK,WAAW,MAAM,GAAG;AAChC,SAAO,KAAK,WAAW,UAAU,SAAS;AAC1C,SAAO,KAAK,WAAW,UAAU,SAAS;AAC1C,SAAO,KAAK,WAAW,IAAI,OAAO,GAAG,MAAM,MAAM,MAAM,IAAI,GAAG,GAAG,WAAW;AAC5E,SAAO,KAAK,WAAW,MAAM,CAAC,UAAU,QAAQ,OAAO,MAAM,MAAM,CAAC;AACpE,SAAO,KAAK,WAAW,SAAS,eAAe;AAC/C,SAAO,KAAK,WAAW,IAAI,OAAO,IAAK,SAAS,OAAO,GAAG,GAAG,SAAS;AACtE,SAAO,KAAK,WAAW,IAAI,OAAO,GAAG,QAAQ,IAAI,QAAQ,IAAI,GAAG,GAAG,aAAa;AAChF,SAAO,KAAK;AAAA,IACV,IAAI,OAAO,GAAG,SAAS,MAAM,SAAS,MAAM,SAAS,OAAO,GAAG;AAAA,IAC/D;AAAA,EACF;AACA,SAAO,KAAK,WAAW,IAAI,OAAO,GAAG,SAAS,MAAM,SAAS,OAAO,GAAG,GAAG,gBAAgB;AAC1F,SAAO,KAAK,WAAW,IAAI,OAAO,IAAI,QAAQ,OAAO,QAAQ,IAAI,GAAG,GAAG,aAAa;AACpF,SAAO,KAAK,WAAW,IAAI,OAAO,IAAI,QAAQ,OAAO,GAAG,GAAG,SAAS;AACpE,SAAO,KAAK,WAAW,IAAI,OAAO,IAAI,SAAS,OAAO,GAAG,GAAG,SAAS;AACrE,SAAO,KAAK,WAAW,WAAM,SAAI;AACjC,SAAO,KAAK,WAAW,MAAM,IAAI;AACjC,SAAO,KAAK,WAAW,MAAM,IAAI;AACjC,SAAO,KAAK,WAAW,MAAM,IAAI;AACjC,SAAO,KAAK,WAAW,KAAK,SAAS;AACrC,SAAO,KAAK,WAAW,KAAK,SAAS;AACrC,SAAO,KAAK,WAAW,KAAK,SAAS;AACrC,SAAO,KAAK,WAAW,SAAS,GAAG;AAEnC,QAAM,QAAQ,KAAK,MAAM,QAAQ;AACjC,SAAO,KAAK,WAAW,UAAU,EAAE;AAEnC,QAAM,YAAwC,CAAC;AAC/C,MAAI,MAAM;AACV,MAAI,QAAQ;AACZ,MAAI,MAAM;AACV,aAAW,SAAS,OAAO;AACzB,UAAM,WAAW,MAAM,KAAK;AAC5B,QAAI,CAAC,SAAU;AAEf,WAAO,MAAM;AACb,WAAO,MAAM;AACb,QAAI,IAAI,SAAS,WAAW;AAC1B,gBAAU,KAAK,CAAC,IAAI,MAAM,CAAC,GAAG,OAAO,GAAG,CAAC;AACzC,cAAQ;AACR,YAAM;AAAA,IACR;AAAA,EACF;AAEA,MAAI,KAAK;AACP,cAAU,KAAK,CAAC,IAAI,MAAM,CAAC,GAAG,OAAO,KAAK,SAAS,CAAC,CAAC;AAAA,EACvD;AAEA,SAAO;AACT;","names":[]}
+{"version":3,"sources":["../../../src/tokenize/basic/sentence.ts"],"sourcesContent":["// SPDX-FileCopyrightText: 2024 LiveKit, Inc.\n//\n// SPDX-License-Identifier: Apache-2.0\n\n/**\n * Split the text into sentences.\n */\nexport const splitSentences = (text: string, minLength = 20): [string, number, number][] => {\n  const alphabets = /([A-Za-z])/g;\n  const prefixes = /(Mr|St|Mrs|Ms|Dr)[.]/g;\n  const suffixes = /(Inc|Ltd|Jr|Sr|Co)/g;\n  const starters =\n    /(Mr|Mrs|Ms|Dr|Prof|Capt|Cpt|Lt|He\\s|She\\s|It\\s|They\\s|Their\\s|Our\\s|We\\s|But\\s|However\\s|That\\s|This\\s|Wherever)/g;\n  const acronyms = /([A-Z][.][A-Z][.](?:[A-Z][.])?)/g;\n  const websites = /[.](com|net|org|io|gov|edu|me)/g;\n  const digits = /([0-9])/g;\n  const dots = /\\.{2,}/g;\n\n  text = text.replaceAll('\\n', ' ');\n  text = text.replaceAll(prefixes, '$1<prd>');\n  text = text.replaceAll(websites, '<prd>$2');\n  text = text.replaceAll(new RegExp(`${digits.source}[.]${digits.source}`, 'g'), '$1<prd>$2');\n  text = text.replaceAll(dots, (match) => '<prd>'.repeat(match.length));\n  text = text.replaceAll('Ph.D.', 'Ph<prd>D<prd>');\n  text = text.replaceAll(new RegExp(`\\\\s${alphabets.source}[.] `, 'g'), ' $1<prd> ');\n  text = text.replaceAll(new RegExp(`${acronyms.source} ${starters.source}`, 'g'), '$1<stop> $2');\n  text = text.replaceAll(\n    new RegExp(`${alphabets.source}[.]${alphabets.source}[.]${alphabets.source}[.]`, 'g'),\n    '$1<prd>$2<prd>$3<prd>',\n  );\n  text = text.replaceAll(\n    new RegExp(`${alphabets.source}[.]${alphabets.source}[.]`, 'g'),\n    '$1<prd>$2<prd>',\n  );\n  text = text.replaceAll(\n    new RegExp(` ${suffixes.source}[.] ${starters.source}`, 'g'),\n    '$1<stop> $2',\n  );\n  text = text.replaceAll(new RegExp(` ${suffixes.source}[.]`, 'g'), '$1<prd>');\n  text = text.replaceAll(new RegExp(` ${alphabets.source}[.]`, 'g'), '$1<prd>');\n  text = text.replaceAll('.”', '”.');\n  text = text.replaceAll('.\"', '\".');\n  text = text.replaceAll('!\"', '\"!');\n  text = text.replaceAll('?\"', '\"?');\n  text = text.replaceAll('.', '.<stop>');\n  text = text.replaceAll('?', '?<stop>');\n  text = text.replaceAll('!', '!<stop>');\n  text = text.replaceAll('<prd>', '.');\n\n  const split = text.split('<stop>');\n  text = text.replaceAll('<stop>', '');\n\n  const sentences: [string, number, number][] = [];\n  let buf = '';\n  let start = 0;\n  let end = 0;\n  for (const match of split) {\n    const sentence = match.trim();\n    if (!sentence) continue;\n\n    buf += ' ' + sentence;\n    end += match.length;\n    if (buf.length > minLength) {\n      sentences.push([buf.slice(1), start, end]);\n      start = end;\n      buf = '';\n    }\n  }\n\n  if (buf) {\n    sentences.push([buf.slice(1), start, text.length - 1]);\n  }\n\n  return sentences;\n};\n"],"mappings":"AAOO,MAAM,iBAAiB,CAAC,MAAc,YAAY,OAAmC;AAC1F,QAAM,YAAY;AAClB,QAAM,WAAW;AACjB,QAAM,WAAW;AACjB,QAAM,WACJ;AACF,QAAM,WAAW;AACjB,QAAM,WAAW;AACjB,QAAM,SAAS;AACf,QAAM,OAAO;AAEb,SAAO,KAAK,WAAW,MAAM,GAAG;AAChC,SAAO,KAAK,WAAW,UAAU,SAAS;AAC1C,SAAO,KAAK,WAAW,UAAU,SAAS;AAC1C,SAAO,KAAK,WAAW,IAAI,OAAO,GAAG,OAAO,MAAM,MAAM,OAAO,MAAM,IAAI,GAAG,GAAG,WAAW;AAC1F,SAAO,KAAK,WAAW,MAAM,CAAC,UAAU,QAAQ,OAAO,MAAM,MAAM,CAAC;AACpE,SAAO,KAAK,WAAW,SAAS,eAAe;AAC/C,SAAO,KAAK,WAAW,IAAI,OAAO,MAAM,UAAU,MAAM,QAAQ,GAAG,GAAG,WAAW;AACjF,SAAO,KAAK,WAAW,IAAI,OAAO,GAAG,SAAS,MAAM,IAAI,SAAS,MAAM,IAAI,GAAG,GAAG,aAAa;AAC9F,SAAO,KAAK;AAAA,IACV,IAAI,OAAO,GAAG,UAAU,MAAM,MAAM,UAAU,MAAM,MAAM,UAAU,MAAM,OAAO,GAAG;AAAA,IACpF;AAAA,EACF;AACA,SAAO,KAAK;AAAA,IACV,IAAI,OAAO,GAAG,UAAU,MAAM,MAAM,UAAU,MAAM,OAAO,GAAG;AAAA,IAC9D;AAAA,EACF;AACA,SAAO,KAAK;AAAA,IACV,IAAI,OAAO,IAAI,SAAS,MAAM,OAAO,SAAS,MAAM,IAAI,GAAG;AAAA,IAC3D;AAAA,EACF;AACA,SAAO,KAAK,WAAW,IAAI,OAAO,IAAI,SAAS,MAAM,OAAO,GAAG,GAAG,SAAS;AAC3E,SAAO,KAAK,WAAW,IAAI,OAAO,IAAI,UAAU,MAAM,OAAO,GAAG,GAAG,SAAS;AAC5E,SAAO,KAAK,WAAW,WAAM,SAAI;AACjC,SAAO,KAAK,WAAW,MAAM,IAAI;AACjC,SAAO,KAAK,WAAW,MAAM,IAAI;AACjC,SAAO,KAAK,WAAW,MAAM,IAAI;AACjC,SAAO,KAAK,WAAW,KAAK,SAAS;AACrC,SAAO,KAAK,WAAW,KAAK,SAAS;AACrC,SAAO,KAAK,WAAW,KAAK,SAAS;AACrC,SAAO,KAAK,WAAW,SAAS,GAAG;AAEnC,QAAM,QAAQ,KAAK,MAAM,QAAQ;AACjC,SAAO,KAAK,WAAW,UAAU,EAAE;AAEnC,QAAM,YAAwC,CAAC;AAC/C,MAAI,MAAM;AACV,MAAI,QAAQ;AACZ,MAAI,MAAM;AACV,aAAW,SAAS,OAAO;AACzB,UAAM,WAAW,MAAM,KAAK;AAC5B,QAAI,CAAC,SAAU;AAEf,WAAO,MAAM;AACb,WAAO,MAAM;AACb,QAAI,IAAI,SAAS,WAAW;AAC1B,gBAAU,KAAK,CAAC,IAAI,MAAM,CAAC,GAAG,OAAO,GAAG,CAAC;AACzC,cAAQ;AACR,YAAM;AAAA,IACR;AAAA,EACF;AAEA,MAAI,KAAK;AACP,cAAU,KAAK,CAAC,IAAI,MAAM,CAAC,GAAG,OAAO,KAAK,SAAS,CAAC,CAAC;AAAA,EACvD;AAEA,SAAO;AACT;","names":[]}

package/dist/tokenize/tokenizer.test.cjs

@@ -0,0 +1,220 @@
+"use strict";
+var import_vitest = require("vitest");
+var import_basic = require("./basic/index.cjs");
+var import_paragraph = require("./basic/paragraph.cjs");
+const TEXT = "Hi! LiveKit is a platform for live audio and video applications and services. R.T.C stands for Real-Time Communication... again R.T.C. Mr. Theo is testing the sentence tokenizer. This is a test. Another test. A short sentence. A longer sentence that is longer than the previous sentence. f(x) = x * 2.54 + 42. Hey! Hi! Hello! ";
+const EXPECTED_MIN_20 = [
+  "Hi! LiveKit is a platform for live audio and video applications and services.",
+  "R.T.C stands for Real-Time Communication... again R.T.C.",
+  "Mr. Theo is testing the sentence tokenizer.",
+  "This is a test. Another test.",
+  "A short sentence. A longer sentence that is longer than the previous sentence.",
+  "f(x) = x * 2.54 + 42.",
+  "Hey! Hi! Hello!"
+];
+const WORDS_TEXT = "This is a test. Blabla another test! multiple consecutive spaces:     done";
+const WORDS_EXPECTED = [
+  "This",
+  "is",
+  "a",
+  "test",
+  "Blabla",
+  "another",
+  "test",
+  "multiple",
+  "consecutive",
+  "spaces",
+  "done"
+];
+const WORDS_PUNCT_TEXT = 'This is <phoneme alphabet="cmu-arpabet" ph="AE K CH UW AH L IY">actually</phoneme> tricky to handle.';
+const WORDS_PUNCT_EXPECTED = [
+  "This",
+  "is",
+  "<phoneme",
+  'alphabet="cmu-arpabet"',
+  'ph="AE',
+  "K",
+  "CH",
+  "UW",
+  "AH",
+  "L",
+  'IY">actually</phoneme>',
+  "tricky",
+  "to",
+  "handle."
+];
+const HYPHENATOR_TEXT = ["Segment", "expected", "communication", "window", "welcome", "bedroom"];
+const HYPHENATOR_EXPECTED = [
+  ["Seg", "ment"],
+  ["ex", "pect", "ed"],
+  ["com", "mu", "ni", "ca", "tion"],
+  ["win", "dow"],
+  ["wel", "come"],
+  ["bed", "room"]
+];
+const PARAGRAPH_TEST_CASES = [
+  ["Single paragraph.", [["Single paragraph.", 0, 17]]],
+  [
+    "Paragraph 1.\n\nParagraph 2.",
+    [
+      ["Paragraph 1.", 0, 12],
+      ["Paragraph 2.", 14, 26]
+    ]
+  ],
+  [
+    "Para 1.\n\nPara 2.\n\nPara 3.",
+    [
+      ["Para 1.", 0, 7],
+      ["Para 2.", 9, 16],
+      ["Para 3.", 18, 25]
+    ]
+  ],
+  ["\n\nParagraph with leading newlines.", [["Paragraph with leading newlines.", 2, 34]]],
+  ["Paragraph with trailing newlines.\n\n", [["Paragraph with trailing newlines.", 0, 33]]],
+  [
+    "\n\n  Paragraph with leading and trailing spaces.  \n\n",
+    [["Paragraph with leading and trailing spaces.", 4, 47]]
+  ],
+  [
+    "Para 1.\n\n\n\nPara 2.",
+    // Multiple newlines between paragraphs
+    [
+      ["Para 1.", 0, 7],
+      ["Para 2.", 11, 18]
+    ]
+  ],
+  [
+    "Para 1.\n \n \nPara 2.",
+    // Newlines with spaces between paragraphs
+    [
+      ["Para 1.", 0, 7],
+      ["Para 2.", 12, 19]
+    ]
+  ],
+  [
+    "",
+    // Empty string
+    []
+  ],
+  [
+    "\n\n\n",
+    // Only newlines
+    []
+  ],
+  [
+    "Line 1\nLine 2\nLine 3",
+    // Single paragraph with newlines
+    [["Line 1\nLine 2\nLine 3", 0, 20]]
+  ]
+];
+(0, import_vitest.describe)("tokenizer", () => {
+  (0, import_vitest.describe)("SentenceTokenizer", () => {
+    const tokenizer = new import_basic.SentenceTokenizer();
+    (0, import_vitest.it)("should tokenize sentences correctly", () => {
+      (0, import_vitest.expect)(tokenizer.tokenize(TEXT).every((x, i) => EXPECTED_MIN_20[i] === x)).toBeTruthy();
+    });
+    (0, import_vitest.it)("should stream tokenize sentences correctly", async () => {
+      const pattern = [1, 2, 4];
+      let text = TEXT;
+      const chunks = [];
+      const patternIter = Array(Math.ceil(text.length / pattern.reduce((sum, num) => sum + num, 0))).fill(pattern).flat()[Symbol.iterator]();
+      for (const size of patternIter) {
+        if (!text) break;
+        chunks.push(text.slice(void 0, size));
+        text = text.slice(size);
+      }
+      const stream = tokenizer.stream();
+      for (const chunk of chunks) {
+        stream.pushText(chunk);
+      }
+      stream.endInput();
+      stream.close();
+      for (const x of EXPECTED_MIN_20) {
+        await stream.next().then((value) => {
+          if (value.value) {
+            (0, import_vitest.expect)(value.value.token).toStrictEqual(x);
+          }
+        });
+      }
+    });
+  });
+  (0, import_vitest.describe)("WordTokenizer", () => {
+    const tokenizer = new import_basic.WordTokenizer();
+    (0, import_vitest.it)("should tokenize words correctly", () => {
+      (0, import_vitest.expect)(tokenizer.tokenize(WORDS_TEXT).every((x, i) => WORDS_EXPECTED[i] === x)).toBeTruthy();
+    });
+    (0, import_vitest.it)("should stream tokenize words correctly", async () => {
+      const pattern = [1, 2, 4];
+      let text = WORDS_TEXT;
+      const chunks = [];
+      const patternIter = Array(Math.ceil(text.length / pattern.reduce((sum, num) => sum + num, 0))).fill(pattern).flat()[Symbol.iterator]();
+      for (const size of patternIter) {
+        if (!text) break;
+        chunks.push(text.slice(void 0, size));
+        text = text.slice(size);
+      }
+      const stream = tokenizer.stream();
+      for (const chunk of chunks) {
+        stream.pushText(chunk);
+      }
+      stream.endInput();
+      stream.close();
+      for (const x of WORDS_EXPECTED) {
+        await stream.next().then((value) => {
+          if (value.value) {
+            (0, import_vitest.expect)(value.value.token).toStrictEqual(x);
+          }
+        });
+      }
+    });
+    (0, import_vitest.describe)("punctuation handling", () => {
+      const tokenizerPunct = new import_basic.WordTokenizer(false);
+      (0, import_vitest.it)("should tokenize words correctly", () => {
+        (0, import_vitest.expect)(
+          tokenizerPunct.tokenize(WORDS_PUNCT_TEXT).every((x, i) => WORDS_PUNCT_EXPECTED[i] === x)
+        ).toBeTruthy();
+      });
+      (0, import_vitest.it)("should stream tokenize words correctly", async () => {
+        const pattern = [1, 2, 4];
+        let text = WORDS_PUNCT_TEXT;
+        const chunks = [];
+        const patternIter = Array(
+          Math.ceil(text.length / pattern.reduce((sum, num) => sum + num, 0))
+        ).fill(pattern).flat()[Symbol.iterator]();
+        for (const size of patternIter) {
+          if (!text) break;
+          chunks.push(text.slice(void 0, size));
+          text = text.slice(size);
+        }
+        const stream = tokenizerPunct.stream();
+        for (const chunk of chunks) {
+          stream.pushText(chunk);
+        }
+        stream.endInput();
+        stream.close();
+        for (const x of WORDS_PUNCT_EXPECTED) {
+          await stream.next().then((value) => {
+            if (value.value) {
+              (0, import_vitest.expect)(value.value.token).toStrictEqual(x);
+            }
+          });
+        }
+      });
+    });
+  });
+  (0, import_vitest.describe)("hyphenateWord", () => {
+    (0, import_vitest.it)("should hyphenate correctly", () => {
+      HYPHENATOR_TEXT.forEach((x, i) => {
+        (0, import_vitest.expect)((0, import_basic.hyphenateWord)(x)).toStrictEqual(HYPHENATOR_EXPECTED[i]);
+      });
+    });
+  });
+  (0, import_vitest.describe)("splitParagraphs", () => {
+    (0, import_vitest.it)("should tokenize paragraphs correctly", () => {
+      PARAGRAPH_TEST_CASES.forEach(([a, b]) => {
+        (0, import_vitest.expect)((0, import_paragraph.splitParagraphs)(a)).toStrictEqual(b);
+      });
+    });
+  });
+});
+//# sourceMappingURL=tokenizer.test.cjs.map