genkitx-azure-openai/lib/whisper.js

Genkit AI framework plugin for Azure OpenAI APIs.

"use strict";
var __defProp = Object.defineProperty;
var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
var __getOwnPropNames = Object.getOwnPropertyNames;
var __hasOwnProp = Object.prototype.hasOwnProperty;
var __export = (target, all) => {
  for (var name in all)
    __defProp(target, name, { get: all[name], enumerable: true });
};
var __copyProps = (to, from, except, desc) => {
  if (from && typeof from === "object" || typeof from === "function") {
    for (let key of __getOwnPropNames(from))
      if (!__hasOwnProp.call(to, key) && key !== except)
        __defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
  }
  return to;
};
var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
var whisper_exports = {};
__export(whisper_exports, {
  Whisper1ConfigSchema: () => Whisper1ConfigSchema,
  whisper1: () => whisper1,
  whisper1Model: () => whisper1Model
});
module.exports = __toCommonJS(whisper_exports);
var import_genkit = require("genkit");
var import_model = require("genkit/model");
const Whisper1ConfigSchema = import_genkit.GenerationCommonConfigSchema.extend({
  language: import_genkit.z.string().optional(),
  timestamp_granularities: import_genkit.z.array(import_genkit.z.enum(["word", "segment"])).optional(),
  response_format: import_genkit.z.enum(["json", "text", "srt", "verbose_json", "vtt"]).optional()
});
const whisper1 = (0, import_model.modelRef)({
  name: "azure-openai/whisper-1",
  info: {
    label: "OpenAI - Whisper",
    supports: {
      media: true,
      output: ["text", "json"],
      multiturn: false,
      systemRole: false,
      tools: false
    }
  },
  configSchema: Whisper1ConfigSchema
});
function toWhisper1Request(request) {
  const message = new import_genkit.Message(request.messages[0]);
  const media = message.media;
  if (!media?.url) {
    throw new Error("No media found in the request");
  }
  const mediaBuffer = Buffer.from(
    media.url.slice(media.url.indexOf(",") + 1),
    "base64"
  );
  const mediaFile = new File([mediaBuffer], "input", {
    type: media.contentType ?? media.url.slice("data:".length, media.url.indexOf(";"))
  });
  const options = {
    model: "whisper-1",
    file: mediaFile,
    prompt: message.text,
    temperature: request.config?.temperature,
    language: request.config?.language,
    timestamp_granularities: request.config?.timestamp_granularities
  };
  const outputFormat = request.output?.format;
  const customFormat = request.config?.response_format;
  if (outputFormat && customFormat) {
    if (outputFormat === "json" && customFormat !== "json" && customFormat !== "verbose_json") {
      throw new Error(
        `Custom response format ${customFormat} is not compatible with output format ${outputFormat}`
      );
    }
  }
  if (outputFormat === "media") {
    throw new Error(`Output format ${outputFormat} is not supported.`);
  }
  options.response_format = customFormat || outputFormat || "text";
  for (const k in options) {
    if (options[k] === void 0) {
      delete options[k];
    }
  }
  return options;
}
function toGenerateResponse(result) {
  return {
    candidates: [
      {
        index: 0,
        finishReason: "stop",
        message: {
          role: "model",
          content: [
            {
              text: typeof result === "string" ? result : result.text
            }
          ]
        }
      }
    ]
  };
}
function whisper1Model(ai, client) {
  return ai.defineModel(
    {
      name: whisper1.name,
      ...whisper1.info,
      configSchema: whisper1.configSchema
    },
    async (request) => {
      const result = await client.audio.transcriptions.create(
        toWhisper1Request(request)
      );
      return toGenerateResponse(result);
    }
  );
}
// Annotate the CommonJS export names for ESM import in node:
0 && (module.exports = {
  Whisper1ConfigSchema,
  whisper1,
  whisper1Model
});
//# sourceMappingURL=whisper.js.map