chat_models.ts

import {
  GenerativeModel,
  GoogleGenerativeAI as GenerativeAI,
  FunctionDeclarationsTool as GoogleGenerativeAIFunctionDeclarationsTool,
  FunctionDeclaration as GenerativeAIFunctionDeclaration,
  type FunctionDeclarationSchema as GenerativeAIFunctionDeclarationSchema,
  GenerateContentRequest,
  SafetySetting,
  Part as GenerativeAIPart,
  ModelParams,
  RequestOptions,
  type CachedContent,
} from "@google/generative-ai";
import { CallbackManagerForLLMRun } from "@langchain/core/callbacks/manager";
import {
  AIMessageChunk,
  BaseMessage,
  UsageMetadata,
} from "@langchain/core/messages";
import { ChatGenerationChunk, ChatResult } from "@langchain/core/outputs";
import { getEnvironmentVariable } from "@langchain/core/utils/env";
import {
  BaseChatModel,
  type BaseChatModelCallOptions,
  type LangSmithParams,
  type BaseChatModelParams,
} from "@langchain/core/language_models/chat_models";
import { NewTokenIndices } from "@langchain/core/callbacks/base";
import {
  BaseLanguageModelInput,
  StructuredOutputMethodOptions,
} from "@langchain/core/language_models/base";
import {
  Runnable,
  RunnablePassthrough,
  RunnableSequence,
} from "@langchain/core/runnables";
import type { z } from "zod";
import { isZodSchema } from "@langchain/core/utils/types";
import { BaseLLMOutputParser } from "@langchain/core/output_parsers";
import { zodToGenerativeAIParameters } from "./utils/zod_to_genai_parameters.js";
import {
  convertBaseMessagesToContent,
  convertResponseContentToChatGenerationChunk,
  mapGenerateContentResultToChatResult,
} from "./utils/common.js";
import { GoogleGenerativeAIToolsOutputParser } from "./output_parsers.js";
import { GoogleGenerativeAIToolType } from "./types.js";
import { convertToolsToGenAI } from "./utils/tools.js";

interface TokenUsage {
  completionTokens?: number;
  promptTokens?: number;
  totalTokens?: number;
}

export type BaseMessageExamplePair = {
  input: BaseMessage;
  output: BaseMessage;
};

export interface GoogleGenerativeAIChatCallOptions
  extends BaseChatModelCallOptions {
  tools?: GoogleGenerativeAIToolType[];
  /**
   * Allowed functions to call when the mode is "any".
   * If empty, any one of the provided functions are called.
   */
  allowedFunctionNames?: string[];
  /**
   * Whether or not to include usage data, like token counts
   * in the streamed response chunks.
   * @default true
   */
  streamUsage?: boolean;
}

/**
 * An interface defining the input to the ChatGoogleGenerativeAI class.
 */
export interface GoogleGenerativeAIChatInput
  extends BaseChatModelParams,
    Pick<GoogleGenerativeAIChatCallOptions, "streamUsage"> {
  /**
   * @deprecated Use "model" instead.
   *
   * Model Name to use
   *
   * Alias for `model`
   *
   * Note: The format must follow the pattern - `{model}`
   */
  modelName?: string;
  /**
   * Model Name to use
   *
   * Note: The format must follow the pattern - `{model}`
   */
  model?: string;

  /**
   * Controls the randomness of the output.
   *
   * Values can range from [0.0,1.0], inclusive. A value closer to 1.0
   * will produce responses that are more varied and creative, while
   * a value closer to 0.0 will typically result in less surprising
   * responses from the model.
   *
   * Note: The default value varies by model
   */
  temperature?: number;

  /**
   * Maximum number of tokens to generate in the completion.
   */
  maxOutputTokens?: number;

  /**
   * Top-p changes how the model selects tokens for output.
   *
   * Tokens are selected from most probable to least until the sum
   * of their probabilities equals the top-p value.
   *
   * For example, if tokens A, B, and C have a probability of
   * .3, .2, and .1 and the top-p value is .5, then the model will
   * select either A or B as the next token (using temperature).
   *
   * Note: The default value varies by model
   */
  topP?: number;

  /**
   * Top-k changes how the model selects tokens for output.
   *
   * A top-k of 1 means the selected token is the most probable among
   * all tokens in the model’s vocabulary (also called greedy decoding),
   * while a top-k of 3 means that the next token is selected from
   * among the 3 most probable tokens (using temperature).
   *
   * Note: The default value varies by model
   */
  topK?: number;

  /**
   * The set of character sequences (up to 5) that will stop output generation.
   * If specified, the API will stop at the first appearance of a stop
   * sequence.
   *
   * Note: The stop sequence will not be included as part of the response.
   * Note: stopSequences is only supported for Gemini models
   */
  stopSequences?: string[];

  /**
   * A list of unique `SafetySetting` instances for blocking unsafe content. The API will block
   * any prompts and responses that fail to meet the thresholds set by these settings. If there
   * is no `SafetySetting` for a given `SafetyCategory` provided in the list, the API will use
   * the default safety setting for that category.
   */
  safetySettings?: SafetySetting[];

  /**
   * Google API key to use
   */
  apiKey?: string;

  /**
   * Google API version to use
   */
  apiVersion?: string;

  /**
   * Google API base URL to use
   */
  baseUrl?: string;

  /** Whether to stream the results or not */
  streaming?: boolean;

  /**
   * Whether or not to force the model to respond with JSON.
   * Available for `gemini-1.5` models and later.
   * @default false
   */
  json?: boolean;

  /**
   * Whether or not model supports system instructions.
   * The following models support system instructions:
   * - All Gemini 1.5 Pro model versions
   * - All Gemini 1.5 Flash model versions
   * - Gemini 1.0 Pro version gemini-1.0-pro-002
   */
  convertSystemMessageToHumanContent?: boolean | undefined;
}

/**
 * Google Generative AI chat model integration.
 *
 * Setup:
 * Install `@langchain/google-genai` and set an environment variable named `GOOGLE_API_KEY`.
 *
 * ```bash
 * npm install @langchain/google-genai
 * export GOOGLE_API_KEY="your-api-key"
 * ```
 *
 * ## [Constructor args](https://api.js.langchain.com/classes/langchain_google_genai.ChatGoogleGenerativeAI.html#constructor)
 *
 * ## [Runtime args](https://api.js.langchain.com/interfaces/langchain_google_genai.GoogleGenerativeAIChatCallOptions.html)
 *
 * Runtime args can be passed as the second argument to any of the base runnable methods `.invoke`. `.stream`, `.batch`, etc.
 * They can also be passed via `.bind`, or the second arg in `.bindTools`, like shown in the examples below:
 *
 * ```typescript
 * // When calling `.bind`, call options should be passed via the first argument
 * const llmWithArgsBound = llm.bind({
 *   stop: ["\n"],
 *   tools: [...],
 * });
 *
 * // When calling `.bindTools`, call options should be passed via the second argument
 * const llmWithTools = llm.bindTools(
 *   [...],
 *   {
 *     stop: ["\n"],
 *   }
 * );
 * ```
 *
 * ## Examples
 *
 * <details open>
 * <summary><strong>Instantiate</strong></summary>
 *
 * ```typescript
 * import { ChatGoogleGenerativeAI } from '@langchain/google-genai';
 *
 * const llm = new ChatGoogleGenerativeAI({
 *   model: "gemini-1.5-flash",
 *   temperature: 0,
 *   maxRetries: 2,
 *   // apiKey: "...",
 *   // other params...
 * });
 * ```
 * </details>
 *
 * <br />
 *
 * <details>
 * <summary><strong>Invoking</strong></summary>
 *
 * ```typescript
 * const input = `Translate "I love programming" into French.`;
 *
 * // Models also accept a list of chat messages or a formatted prompt
 * const result = await llm.invoke(input);
 * console.log(result);
 * ```
 *
 * ```txt
 * AIMessage {
 *   "content": "There are a few ways to translate \"I love programming\" into French, depending on the level of formality and nuance you want to convey:\n\n**Formal:**\n\n* **J'aime la programmation.** (This is the most literal and formal translation.)\n\n**Informal:**\n\n* **J'adore programmer.** (This is a more enthusiastic and informal translation.)\n* **J'aime beaucoup programmer.** (This is a slightly less enthusiastic but still informal translation.)\n\n**More specific:**\n\n* **J'aime beaucoup coder.** (This specifically refers to writing code.)\n* **J'aime beaucoup développer des logiciels.** (This specifically refers to developing software.)\n\nThe best translation will depend on the context and your intended audience. \n",
 *   "response_metadata": {
 *     "finishReason": "STOP",
 *     "index": 0,
 *     "safetyRatings": [
 *       {
 *         "category": "HARM_CATEGORY_SEXUALLY_EXPLICIT",
 *         "probability": "NEGLIGIBLE"
 *       },
 *       {
 *         "category": "HARM_CATEGORY_HATE_SPEECH",
 *         "probability": "NEGLIGIBLE"
 *       },
 *       {
 *         "category": "HARM_CATEGORY_HARASSMENT",
 *         "probability": "NEGLIGIBLE"
 *       },
 *       {
 *         "category": "HARM_CATEGORY_DANGEROUS_CONTENT",
 *         "probability": "NEGLIGIBLE"
 *       }
 *     ]
 *   },
 *   "usage_metadata": {
 *     "input_tokens": 10,
 *     "output_tokens": 149,
 *     "total_tokens": 159
 *   }
 * }
 * ```
 * </details>
 *
 * <br />
 *
 * <details>
 * <summary><strong>Streaming Chunks</strong></summary>
 *
 * ```typescript
 * for await (const chunk of await llm.stream(input)) {
 *   console.log(chunk);
 * }
 * ```
 *
 * ```txt
 * AIMessageChunk {
 *   "content": "There",
 *   "response_metadata": {
 *     "index": 0
 *   }
 *   "usage_metadata": {
 *     "input_tokens": 10,
 *     "output_tokens": 1,
 *     "total_tokens": 11
 *   }
 * }
 * AIMessageChunk {
 *   "content": " are a few ways to translate \"I love programming\" into French, depending on",
 * }
 * AIMessageChunk {
 *   "content": " the level of formality and nuance you want to convey:\n\n**Formal:**\n\n",
 * }
 * AIMessageChunk {
 *   "content": "* **J'aime la programmation.** (This is the most literal and formal translation.)\n\n**Informal:**\n\n* **J'adore programmer.** (This",
 * }
 * AIMessageChunk {
 *   "content": " is a more enthusiastic and informal translation.)\n* **J'aime beaucoup programmer.** (This is a slightly less enthusiastic but still informal translation.)\n\n**More",
 * }
 * AIMessageChunk {
 *   "content": " specific:**\n\n* **J'aime beaucoup coder.** (This specifically refers to writing code.)\n* **J'aime beaucoup développer des logiciels.** (This specifically refers to developing software.)\n\nThe best translation will depend on the context and",
 * }
 * AIMessageChunk {
 *   "content": " your intended audience. \n",
 * }
 * ```
 * </details>
 *
 * <br />
 *
 * <details>
 * <summary><strong>Aggregate Streamed Chunks</strong></summary>
 *
 * ```typescript
 * import { AIMessageChunk } from '@langchain/core/messages';
 * import { concat } from '@langchain/core/utils/stream';
 *
 * const stream = await llm.stream(input);
 * let full: AIMessageChunk | undefined;
 * for await (const chunk of stream) {
 *   full = !full ? chunk : concat(full, chunk);
 * }
 * console.log(full);
 * ```
 *
 * ```txt
 * AIMessageChunk {
 *   "content": "There are a few ways to translate \"I love programming\" into French, depending on the level of formality and nuance you want to convey:\n\n**Formal:**\n\n* **J'aime la programmation.** (This is the most literal and formal translation.)\n\n**Informal:**\n\n* **J'adore programmer.** (This is a more enthusiastic and informal translation.)\n* **J'aime beaucoup programmer.** (This is a slightly less enthusiastic but still informal translation.)\n\n**More specific:**\n\n* **J'aime beaucoup coder.** (This specifically refers to writing code.)\n* **J'aime beaucoup développer des logiciels.** (This specifically refers to developing software.)\n\nThe best translation will depend on the context and your intended audience. \n",
 *   "usage_metadata": {
 *     "input_tokens": 10,
 *     "output_tokens": 277,
 *     "total_tokens": 287
 *   }
 * }
 * ```
 * </details>
 *
 * <br />
 *
 * <details>
 * <summary><strong>Bind tools</strong></summary>
 *
 * ```typescript
 * import { z } from 'zod';
 *
 * const GetWeather = {
 *   name: "GetWeather",
 *   description: "Get the current weather in a given location",
 *   schema: z.object({
 *     location: z.string().describe("The city and state, e.g. San Francisco, CA")
 *   }),
 * }
 *
 * const GetPopulation = {
 *   name: "GetPopulation",
 *   description: "Get the current population in a given location",
 *   schema: z.object({
 *     location: z.string().describe("The city and state, e.g. San Francisco, CA")
 *   }),
 * }
 *
 * const llmWithTools = llm.bindTools([GetWeather, GetPopulation]);
 * const aiMsg = await llmWithTools.invoke(
 *   "Which city is hotter today and which is bigger: LA or NY?"
 * );
 * console.log(aiMsg.tool_calls);
 * ```
 *
 * ```txt
 * [
 *   {
 *     name: 'GetWeather',
 *     args: { location: 'Los Angeles, CA' },
 *     type: 'tool_call'
 *   },
 *   {
 *     name: 'GetWeather',
 *     args: { location: 'New York, NY' },
 *     type: 'tool_call'
 *   },
 *   {
 *     name: 'GetPopulation',
 *     args: { location: 'Los Angeles, CA' },
 *     type: 'tool_call'
 *   },
 *   {
 *     name: 'GetPopulation',
 *     args: { location: 'New York, NY' },
 *     type: 'tool_call'
 *   }
 * ]
 * ```
 * </details>
 *
 * <br />
 *
 * <details>
 * <summary><strong>Structured Output</strong></summary>
 *
 * ```typescript
 * const Joke = z.object({
 *   setup: z.string().describe("The setup of the joke"),
 *   punchline: z.string().describe("The punchline to the joke"),
 *   rating: z.number().optional().describe("How funny the joke is, from 1 to 10")
 * }).describe('Joke to tell user.');
 *
 * const structuredLlm = llm.withStructuredOutput(Joke, { name: "Joke" });
 * const jokeResult = await structuredLlm.invoke("Tell me a joke about cats");
 * console.log(jokeResult);
 * ```
 *
 * ```txt
 * {
 *   setup: "Why don\\'t cats play poker?",
 *   punchline: "Why don\\'t cats play poker? Because they always have an ace up their sleeve!"
 * }
 * ```
 * </details>
 *
 * <br />
 *
 * <details>
 * <summary><strong>Multimodal</strong></summary>
 *
 * ```typescript
 * import { HumanMessage } from '@langchain/core/messages';
 *
 * const imageUrl = "https://example.com/image.jpg";
 * const imageData = await fetch(imageUrl).then(res => res.arrayBuffer());
 * const base64Image = Buffer.from(imageData).toString('base64');
 *
 * const message = new HumanMessage({
 *   content: [
 *     { type: "text", text: "describe the weather in this image" },
 *     {
 *       type: "image_url",
 *       image_url: { url: `data:image/jpeg;base64,${base64Image}` },
 *     },
 *   ]
 * });
 *
 * const imageDescriptionAiMsg = await llm.invoke([message]);
 * console.log(imageDescriptionAiMsg.content);
 * ```
 *
 * ```txt
 * The weather in the image appears to be clear and sunny. The sky is mostly blue with a few scattered white clouds, indicating fair weather. The bright sunlight is casting shadows on the green, grassy hill, suggesting it is a pleasant day with good visibility. There are no signs of rain or stormy conditions.
 * ```
 * </details>
 *
 * <br />
 *
 * <details>
 * <summary><strong>Usage Metadata</strong></summary>
 *
 * ```typescript
 * const aiMsgForMetadata = await llm.invoke(input);
 * console.log(aiMsgForMetadata.usage_metadata);
 * ```
 *
 * ```txt
 * { input_tokens: 10, output_tokens: 149, total_tokens: 159 }
 * ```
 * </details>
 *
 * <br />
 *
 * <details>
 * <summary><strong>Response Metadata</strong></summary>
 *
 * ```typescript
 * const aiMsgForResponseMetadata = await llm.invoke(input);
 * console.log(aiMsgForResponseMetadata.response_metadata);
 * ```
 *
 * ```txt
 * {
 *   finishReason: 'STOP',
 *   index: 0,
 *   safetyRatings: [
 *     {
 *       category: 'HARM_CATEGORY_SEXUALLY_EXPLICIT',
 *       probability: 'NEGLIGIBLE'
 *     },
 *     {
 *       category: 'HARM_CATEGORY_HATE_SPEECH',
 *       probability: 'NEGLIGIBLE'
 *     },
 *     { category: 'HARM_CATEGORY_HARASSMENT', probability: 'NEGLIGIBLE' },
 *     {
 *       category: 'HARM_CATEGORY_DANGEROUS_CONTENT',
 *       probability: 'NEGLIGIBLE'
 *     }
 *   ]
 * }
 * ```
 * </details>
 *
 * <br />
 */
export class ChatGoogleGenerativeAI
  extends BaseChatModel<GoogleGenerativeAIChatCallOptions, AIMessageChunk>
  implements GoogleGenerativeAIChatInput
{
  static lc_name() {
    return "ChatGoogleGenerativeAI";
  }

  lc_serializable = true;

  get lc_secrets(): { [key: string]: string } | undefined {
    return {
      apiKey: "GOOGLE_API_KEY",
    };
  }

  lc_namespace = ["langchain", "chat_models", "google_genai"];

  get lc_aliases() {
    return {
      apiKey: "google_api_key",
    };
  }

  modelName = "gemini-pro";

  model = "gemini-pro";

  temperature?: number; // default value chosen based on model

  maxOutputTokens?: number;

  topP?: number; // default value chosen based on model

  topK?: number; // default value chosen based on model

  stopSequences: string[] = [];

  safetySettings?: SafetySetting[];

  apiKey?: string;

  streaming = false;

  streamUsage = true;

  convertSystemMessageToHumanContent: boolean | undefined;

  private client: GenerativeModel;

  get _isMultimodalModel() {
    return (
      this.model.includes("vision") ||
      this.model.startsWith("gemini-1.5") ||
      this.model.startsWith("gemini-2")
    );
  }

  constructor(fields?: GoogleGenerativeAIChatInput) {
    super(fields ?? {});

    this.modelName =
      fields?.model?.replace(/^models\//, "") ??
      fields?.modelName?.replace(/^models\//, "") ??
      this.model;
    this.model = this.modelName;

    this.maxOutputTokens = fields?.maxOutputTokens ?? this.maxOutputTokens;

    if (this.maxOutputTokens && this.maxOutputTokens < 0) {
      throw new Error("`maxOutputTokens` must be a positive integer");
    }

    this.temperature = fields?.temperature ?? this.temperature;
    if (this.temperature && (this.temperature < 0 || this.temperature > 1)) {
      throw new Error("`temperature` must be in the range of [0.0,1.0]");
    }

    this.topP = fields?.topP ?? this.topP;
    if (this.topP && this.topP < 0) {
      throw new Error("`topP` must be a positive integer");
    }

    if (this.topP && this.topP > 1) {
      throw new Error("`topP` must be below 1.");
    }

    this.topK = fields?.topK ?? this.topK;
    if (this.topK && this.topK < 0) {
      throw new Error("`topK` must be a positive integer");
    }

    this.stopSequences = fields?.stopSequences ?? this.stopSequences;

    this.apiKey = fields?.apiKey ?? getEnvironmentVariable("GOOGLE_API_KEY");
    if (!this.apiKey) {
      throw new Error(
        "Please set an API key for Google GenerativeAI " +
          "in the environment variable GOOGLE_API_KEY " +
          "or in the `apiKey` field of the " +
          "ChatGoogleGenerativeAI constructor"
      );
    }

    this.safetySettings = fields?.safetySettings ?? this.safetySettings;
    if (this.safetySettings && this.safetySettings.length > 0) {
      const safetySettingsSet = new Set(
        this.safetySettings.map((s) => s.category)
      );
      if (safetySettingsSet.size !== this.safetySettings.length) {
        throw new Error(
          "The categories in `safetySettings` array must be unique"
        );
      }
    }

    this.streaming = fields?.streaming ?? this.streaming;

    this.client = new GenerativeAI(this.apiKey).getGenerativeModel(
      {
        model: this.model,
        safetySettings: this.safetySettings as SafetySetting[],
        generationConfig: {
          candidateCount: 1,
          stopSequences: this.stopSequences,
          maxOutputTokens: this.maxOutputTokens,
          temperature: this.temperature,
          topP: this.topP,
          topK: this.topK,
          ...(fields?.json ? { responseMimeType: "application/json" } : {}),
        },
      },
      {
        apiVersion: fields?.apiVersion,
        baseUrl: fields?.baseUrl,
      }
    );
    this.streamUsage = fields?.streamUsage ?? this.streamUsage;
  }

  useCachedContent(
    cachedContent: CachedContent,
    modelParams?: ModelParams,
    requestOptions?: RequestOptions
  ): void {
    if (!this.apiKey) return;
    this.client = new GenerativeAI(
      this.apiKey
    ).getGenerativeModelFromCachedContent(
      cachedContent,
      modelParams,
      requestOptions
    );
  }

  get useSystemInstruction(): boolean {
    return typeof this.convertSystemMessageToHumanContent === "boolean"
      ? !this.convertSystemMessageToHumanContent
      : this.computeUseSystemInstruction;
  }

  get computeUseSystemInstruction(): boolean {
    // This works on models from April 2024 and later
    //   Vertex AI: gemini-1.5-pro and gemini-1.0-002 and later
    //   AI Studio: gemini-1.5-pro-latest
    if (this.modelName === "gemini-1.0-pro-001") {
      return false;
    } else if (this.modelName.startsWith("gemini-pro-vision")) {
      return false;
    } else if (this.modelName.startsWith("gemini-1.0-pro-vision")) {
      return false;
    } else if (this.modelName === "gemini-pro") {
      // on AI Studio gemini-pro is still pointing at gemini-1.0-pro-001
      return false;
    }
    return true;
  }

  getLsParams(options: this["ParsedCallOptions"]): LangSmithParams {
    return {
      ls_provider: "google_genai",
      ls_model_name: this.model,
      ls_model_type: "chat",
      ls_temperature: this.client.generationConfig.temperature,
      ls_max_tokens: this.client.generationConfig.maxOutputTokens,
      ls_stop: options.stop,
    };
  }

  _combineLLMOutput() {
    return [];
  }

  _llmType() {
    return "googlegenerativeai";
  }

  override bindTools(
    tools: GoogleGenerativeAIToolType[],
    kwargs?: Partial<GoogleGenerativeAIChatCallOptions>
  ): Runnable<
    BaseLanguageModelInput,
    AIMessageChunk,
    GoogleGenerativeAIChatCallOptions
  > {
    return this.bind({ tools: convertToolsToGenAI(tools)?.tools, ...kwargs });
  }

  invocationParams(
    options?: this["ParsedCallOptions"]
  ): Omit<GenerateContentRequest, "contents"> {
    const toolsAndConfig = options?.tools?.length
      ? convertToolsToGenAI(options.tools, {
          toolChoice: options.tool_choice,
          allowedFunctionNames: options.allowedFunctionNames,
        })
      : undefined;

    return {
      ...(toolsAndConfig?.tools ? { tools: toolsAndConfig.tools } : {}),
      ...(toolsAndConfig?.toolConfig
        ? { toolConfig: toolsAndConfig.toolConfig }
        : {}),
    };
  }

  async _generate(
    messages: BaseMessage[],
    options: this["ParsedCallOptions"],
    runManager?: CallbackManagerForLLMRun
  ): Promise<ChatResult> {
    const prompt = convertBaseMessagesToContent(
      messages,
      this._isMultimodalModel,
      this.useSystemInstruction
    );
    let actualPrompt = prompt;
    if (prompt[0].role === "system") {
      const [systemInstruction] = prompt;
      this.client.systemInstruction = systemInstruction;
      actualPrompt = prompt.slice(1);
    }
    const parameters = this.invocationParams(options);

    // Handle streaming
    if (this.streaming) {
      const tokenUsage: TokenUsage = {};
      const stream = this._streamResponseChunks(messages, options, runManager);
      const finalChunks: Record<number, ChatGenerationChunk> = {};

      for await (const chunk of stream) {
        const index =
          (chunk.generationInfo as NewTokenIndices)?.completion ?? 0;
        if (finalChunks[index] === undefined) {
          finalChunks[index] = chunk;
        } else {
          finalChunks[index] = finalChunks[index].concat(chunk);
        }
      }
      const generations = Object.entries(finalChunks)
        .sort(([aKey], [bKey]) => parseInt(aKey, 10) - parseInt(bKey, 10))
        .map(([_, value]) => value);

      return { generations, llmOutput: { estimatedTokenUsage: tokenUsage } };
    }

    const res = await this.completionWithRetry({
      ...parameters,
      contents: actualPrompt,
    });

    let usageMetadata: UsageMetadata | undefined;
    if ("usageMetadata" in res.response) {
      const genAIUsageMetadata = res.response.usageMetadata as {
        promptTokenCount: number | undefined;
        candidatesTokenCount: number | undefined;
        totalTokenCount: number | undefined;
      };
      usageMetadata = {
        input_tokens: genAIUsageMetadata.promptTokenCount ?? 0,
        output_tokens: genAIUsageMetadata.candidatesTokenCount ?? 0,
        total_tokens: genAIUsageMetadata.totalTokenCount ?? 0,
      };
    }

    const generationResult = mapGenerateContentResultToChatResult(
      res.response,
      {
        usageMetadata,
      }
    );
    await runManager?.handleLLMNewToken(
      generationResult.generations[0].text ?? ""
    );
    return generationResult;
  }

  async *_streamResponseChunks(
    messages: BaseMessage[],
    options: this["ParsedCallOptions"],
    runManager?: CallbackManagerForLLMRun
  ): AsyncGenerator<ChatGenerationChunk> {
    const prompt = convertBaseMessagesToContent(
      messages,
      this._isMultimodalModel,
      this.useSystemInstruction
    );
    let actualPrompt = prompt;
    if (prompt[0].role === "system") {
      const [systemInstruction] = prompt;
      this.client.systemInstruction = systemInstruction;
      actualPrompt = prompt.slice(1);
    }
    const parameters = this.invocationParams(options);
    const request = {
      ...parameters,
      contents: actualPrompt,
    };
    const stream = await this.caller.callWithOptions(
      { signal: options?.signal },
      async () => {
        const { stream } = await this.client.generateContentStream(request);
        return stream;
      }
    );

    let usageMetadata: UsageMetadata | undefined;
    let index = 0;
    for await (const response of stream) {
      if (
        "usageMetadata" in response &&
        this.streamUsage !== false &&
        options.streamUsage !== false
      ) {
        const genAIUsageMetadata = response.usageMetadata as {
          promptTokenCount: number;
          candidatesTokenCount: number;
          totalTokenCount: number;
        };
        if (!usageMetadata) {
          usageMetadata = {
            input_tokens: genAIUsageMetadata.promptTokenCount,
            output_tokens: genAIUsageMetadata.candidatesTokenCount,
            total_tokens: genAIUsageMetadata.totalTokenCount,
          };
        } else {
          // Under the hood, LangChain combines the prompt tokens. Google returns the updated
          // total each time, so we need to find the difference between the tokens.
          const outputTokenDiff =
            genAIUsageMetadata.candidatesTokenCount -
            usageMetadata.output_tokens;
          usageMetadata = {
            input_tokens: 0,
            output_tokens: outputTokenDiff,
            total_tokens: outputTokenDiff,
          };
        }
      }

      const chunk = convertResponseContentToChatGenerationChunk(response, {
        usageMetadata,
        index,
      });
      index += 1;
      if (!chunk) {
        continue;
      }

      yield chunk;
      await runManager?.handleLLMNewToken(chunk.text ?? "");
    }
  }

  async completionWithRetry(
    request: string | GenerateContentRequest | (string | GenerativeAIPart)[],
    options?: this["ParsedCallOptions"]
  ) {
    return this.caller.callWithOptions(
      { signal: options?.signal },
      async () => {
        try {
          return await this.client.generateContent(request);
          // eslint-disable-next-line @typescript-eslint/no-explicit-any
        } catch (e: any) {
          // TODO: Improve error handling
          if (e.message?.includes("400 Bad Request")) {
            e.status = 400;
          }
          throw e;
        }
      }
    );
  }

  withStructuredOutput<
    // eslint-disable-next-line @typescript-eslint/no-explicit-any
    RunOutput extends Record<string, any> = Record<string, any>
  >(
    outputSchema:
      | z.ZodType<RunOutput>
      // eslint-disable-next-line @typescript-eslint/no-explicit-any
      | Record<string, any>,
    config?: StructuredOutputMethodOptions<false>
  ): Runnable<BaseLanguageModelInput, RunOutput>;

  withStructuredOutput<
    // eslint-disable-next-line @typescript-eslint/no-explicit-any
    RunOutput extends Record<string, any> = Record<string, any>
  >(
    outputSchema:
      | z.ZodType<RunOutput>
      // eslint-disable-next-line @typescript-eslint/no-explicit-any
      | Record<string, any>,
    config?: StructuredOutputMethodOptions<true>
  ): Runnable<BaseLanguageModelInput, { raw: BaseMessage; parsed: RunOutput }>;

  withStructuredOutput<
    // eslint-disable-next-line @typescript-eslint/no-explicit-any
    RunOutput extends Record<string, any> = Record<string, any>
  >(
    outputSchema:
      | z.ZodType<RunOutput>
      // eslint-disable-next-line @typescript-eslint/no-explicit-any
      | Record<string, any>,
    config?: StructuredOutputMethodOptions<boolean>
  ):
    | Runnable<BaseLanguageModelInput, RunOutput>
    | Runnable<
        BaseLanguageModelInput,
        { raw: BaseMessage; parsed: RunOutput }
      > {
    // eslint-disable-next-line @typescript-eslint/no-explicit-any
    const schema: z.ZodType<RunOutput> | Record<string, any> = outputSchema;
    const name = config?.name;
    const method = config?.method;
    const includeRaw = config?.includeRaw;
    if (method === "jsonMode") {
      throw new Error(
        `ChatGoogleGenerativeAI only supports "functionCalling" as a method.`
      );
    }

    let functionName = name ?? "extract";
    let outputParser: BaseLLMOutputParser<RunOutput>;
    let tools: GoogleGenerativeAIFunctionDeclarationsTool[];
    if (isZodSchema(schema)) {
      const jsonSchema = zodToGenerativeAIParameters(schema);
      tools = [
        {
          functionDeclarations: [
            {
              name: functionName,
              description:
                jsonSchema.description ?? "A function available to call.",
              parameters: jsonSchema as GenerativeAIFunctionDeclarationSchema,
            },
          ],
        },
      ];
      outputParser = new GoogleGenerativeAIToolsOutputParser<
        z.infer<typeof schema>
      >({
        returnSingle: true,
        keyName: functionName,
        zodSchema: schema,
      });
    } else {
      let geminiFunctionDefinition: GenerativeAIFunctionDeclaration;
      if (
        typeof schema.name === "string" &&
        typeof schema.parameters === "object" &&
        schema.parameters != null
      ) {
        geminiFunctionDefinition = schema as GenerativeAIFunctionDeclaration;
        functionName = schema.name;
      } else {
        geminiFunctionDefinition = {
          name: functionName,
          description: schema.description ?? "",
          parameters: schema as GenerativeAIFunctionDeclarationSchema,
        };
      }
      tools = [
        {
          functionDeclarations: [geminiFunctionDefinition],
        },
      ];
      outputParser = new GoogleGenerativeAIToolsOutputParser<RunOutput>({
        returnSingle: true,
        keyName: functionName,
      });
    }
    const llm = this.bind({
      tools,
      tool_choice: functionName,
    });

    if (!includeRaw) {
      return llm.pipe(outputParser).withConfig({
        runName: "ChatGoogleGenerativeAIStructuredOutput",
      }) as Runnable<BaseLanguageModelInput, RunOutput>;
    }

    const parserAssign = RunnablePassthrough.assign({
      // eslint-disable-next-line @typescript-eslint/no-explicit-any
      parsed: (input: any, config) => outputParser.invoke(input.raw, config),
    });
    const parserNone = RunnablePassthrough.assign({
      parsed: () => null,
    });
    const parsedWithFallback = parserAssign.withFallbacks({
      fallbacks: [parserNone],
    });
    return RunnableSequence.from<
      BaseLanguageModelInput,
      { raw: BaseMessage; parsed: RunOutput }
    >([
      {
        raw: llm,
      },
      parsedWithFallback,
    ]).withConfig({
      runName: "StructuredOutputRunnable",
    });
  }
}