langchain 0.0.147 → 0.0.149

package/dist/evaluation/embedding_distance/base.js

@@ -0,0 +1,156 @@
+import { distance, similarity } from "ml-distance";
+import { PairwiseStringEvaluator, StringEvaluator, } from "../base.js";
+import { OpenAIEmbeddings } from "../../embeddings/openai.js";
+/**
+ * Get the distance function for the given distance type.
+ * @param distance The distance type.
+ * @return The distance function.
+ */
+export function getDistanceCalculationFunction(distanceType) {
+    const distanceFunctions = {
+        cosine: (X, Y) => 1.0 - similarity.cosine(X, Y),
+        euclidean: distance.euclidean,
+        manhattan: distance.manhattan,
+        chebyshev: distance.chebyshev,
+    };
+    return distanceFunctions[distanceType];
+}
+/**
+ * Compute the score based on the distance metric.
+ * @param vectors The input vectors.
+ * @param distanceMetric The distance metric.
+ * @return The computed score.
+ */
+export function computeEvaluationScore(vectors, distanceMetric) {
+    const metricFunction = getDistanceCalculationFunction(distanceMetric);
+    return metricFunction(vectors[0], vectors[1]);
+}
+/**
+ * Use embedding distances to score semantic difference between
+ * a prediction and reference.
+ */
+export class EmbeddingDistanceEvalChain extends StringEvaluator {
+    constructor(fields) {
+        super();
+        Object.defineProperty(this, "requiresReference", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: true
+        });
+        Object.defineProperty(this, "requiresInput", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: false
+        });
+        Object.defineProperty(this, "outputKey", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: "score"
+        });
+        Object.defineProperty(this, "embedding", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: void 0
+        });
+        Object.defineProperty(this, "distanceMetric", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: "cosine"
+        });
+        this.embedding = fields?.embedding || new OpenAIEmbeddings();
+        this.distanceMetric = fields?.distanceMetric || "cosine";
+    }
+    _chainType() {
+        return `embedding_${this.distanceMetric}_distance`;
+    }
+    async _evaluateStrings(args, config) {
+        const result = await this.call(args, config);
+        return { [this.outputKey]: result[this.outputKey] };
+    }
+    get inputKeys() {
+        return ["reference", "prediction"];
+    }
+    get outputKeys() {
+        return [this.outputKey];
+    }
+    async _call(values, _runManager) {
+        const { prediction, reference } = values;
+        if (!this.embedding)
+            throw new Error("Embedding is undefined");
+        const vectors = await this.embedding.embedDocuments([
+            prediction,
+            reference,
+        ]);
+        const score = computeEvaluationScore(vectors, this.distanceMetric);
+        return { [this.outputKey]: score };
+    }
+}
+/**
+ * Use embedding distances to score semantic difference between two predictions.
+ */
+export class PairwiseEmbeddingDistanceEvalChain extends PairwiseStringEvaluator {
+    constructor(fields) {
+        super();
+        Object.defineProperty(this, "requiresReference", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: false
+        });
+        Object.defineProperty(this, "requiresInput", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: false
+        });
+        Object.defineProperty(this, "outputKey", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: "score"
+        });
+        Object.defineProperty(this, "embedding", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: void 0
+        });
+        Object.defineProperty(this, "distanceMetric", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: "cosine"
+        });
+        this.embedding = fields?.embedding || new OpenAIEmbeddings();
+        this.distanceMetric = fields?.distanceMetric || "cosine";
+    }
+    _chainType() {
+        return `pairwise_embedding_${this.distanceMetric}_distance`;
+    }
+    async _evaluateStringPairs(args, config) {
+        const result = await this.call(args, config);
+        return { [this.outputKey]: result[this.outputKey] };
+    }
+    get inputKeys() {
+        return ["prediction", "predictionB"];
+    }
+    get outputKeys() {
+        return [this.outputKey];
+    }
+    async _call(values, _runManager) {
+        const { prediction, predictionB } = values;
+        if (!this.embedding)
+            throw new Error("Embedding is undefined");
+        const vectors = await this.embedding.embedDocuments([
+            prediction,
+            predictionB,
+        ]);
+        const score = computeEvaluationScore(vectors, this.distanceMetric);
+        return { [this.outputKey]: score };
+    }
+}

package/dist/evaluation/embedding_distance/index.cjs

@@ -0,0 +1,17 @@
+"use strict";
+var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    var desc = Object.getOwnPropertyDescriptor(m, k);
+    if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
+      desc = { enumerable: true, get: function() { return m[k]; } };
+    }
+    Object.defineProperty(o, k2, desc);
+}) : (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    o[k2] = m[k];
+}));
+var __exportStar = (this && this.__exportStar) || function(m, exports) {
+    for (var p in m) if (p !== "default" && !Object.prototype.hasOwnProperty.call(exports, p)) __createBinding(exports, m, p);
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+__exportStar(require("./base.cjs"), exports);

package/dist/evaluation/embedding_distance/index.d.ts

@@ -0,0 +1 @@
+export * from "./base.js";

package/dist/evaluation/embedding_distance/index.js

@@ -0,0 +1 @@
+export * from "./base.js";

package/dist/evaluation/index.cjs

@@ -15,3 +15,9 @@ var __exportStar = (this && this.__exportStar) || function(m, exports) {
 };
 Object.defineProperty(exports, "__esModule", { value: true });
 __exportStar(require("./qa/index.cjs"), exports);
+__exportStar(require("./criteria/index.cjs"), exports);
+__exportStar(require("./agents/index.cjs"), exports);
+__exportStar(require("./embedding_distance/index.cjs"), exports);
+__exportStar(require("./comparison/index.cjs"), exports);
+__exportStar(require("./types.cjs"), exports);
+__exportStar(require("./loader.cjs"), exports);

package/dist/evaluation/index.d.ts

@@ -1 +1,7 @@
 export * from "./qa/index.js";
+export * from "./criteria/index.js";
+export * from "./agents/index.js";
+export * from "./embedding_distance/index.js";
+export * from "./comparison/index.js";
+export * from "./types.js";
+export * from "./loader.js";

package/dist/evaluation/index.js

@@ -1 +1,7 @@
 export * from "./qa/index.js";
+export * from "./criteria/index.js";
+export * from "./agents/index.js";
+export * from "./embedding_distance/index.js";
+export * from "./comparison/index.js";
+export * from "./types.js";
+export * from "./loader.js";

package/dist/evaluation/loader.cjs

@@ -0,0 +1,60 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.loadEvaluator = void 0;
+const index_js_1 = require("./criteria/index.cjs");
+const openai_js_1 = require("../chat_models/openai.cjs");
+const index_js_2 = require("./comparison/index.cjs");
+const index_js_3 = require("./embedding_distance/index.cjs");
+const index_js_4 = require("./agents/index.cjs");
+const base_js_1 = require("../chat_models/base.cjs");
+/**
+ * Load the requested evaluation chain specified by a string
+ * @param type The type of evaluator to load.
+ * @param options
+ *        - llm The language model to use for the evaluator.
+ *        - criteria The criteria to use for the evaluator.
+ *        - agentTools A list of tools available to the agent,for TrajectoryEvalChain.
+ */
+async function loadEvaluator(type, options) {
+    const { llm, chainOptions, criteria, agentTools } = options || {};
+    const llm_ = llm ??
+        new openai_js_1.ChatOpenAI({
+            modelName: "gpt-4",
+            temperature: 0.0,
+        });
+    let evaluator;
+    switch (type) {
+        case "criteria":
+            evaluator = await index_js_1.CriteriaEvalChain.fromLLM(llm_, criteria, chainOptions);
+            break;
+        case "labeled_criteria":
+            evaluator = await index_js_1.LabeledCriteriaEvalChain.fromLLM(llm_, criteria, chainOptions);
+            break;
+        case "pairwise_string":
+            evaluator = await index_js_2.PairwiseStringEvalChain.fromLLM(llm_, criteria, chainOptions);
+            break;
+        case "labeled_pairwise_string":
+            evaluator = await index_js_2.LabeledPairwiseStringEvalChain.fromLLM(llm_, criteria, chainOptions);
+            break;
+        case "trajectory":
+            // eslint-disable-next-line no-instanceof/no-instanceof
+            if (!(llm_ instanceof base_js_1.BaseChatModel)) {
+                throw new Error("LLM must be an instance of a base chat model.");
+            }
+            evaluator = await index_js_4.TrajectoryEvalChain.fromLLM(llm_, agentTools, chainOptions);
+            break;
+        case "embedding_distance":
+            evaluator = new index_js_3.EmbeddingDistanceEvalChain({
+                embedding: options?.embedding,
+                distanceMetric: options?.distanceMetric,
+            });
+            break;
+        case "pairwise_embedding_distance":
+            evaluator = new index_js_3.PairwiseEmbeddingDistanceEvalChain({});
+            break;
+        default:
+            throw new Error(`Unknown type: ${type}`);
+    }
+    return evaluator;
+}
+exports.loadEvaluator = loadEvaluator;

package/dist/evaluation/loader.d.ts

@@ -0,0 +1,27 @@
+import { BaseLanguageModel } from "../base_language/index.js";
+import { CriteriaLike } from "./criteria/index.js";
+import type { EvaluatorType } from "./types.js";
+import { StructuredTool } from "../tools/index.js";
+import { LLMEvalChainInput } from "./base.js";
+import { EmbeddingDistanceEvalChainInput } from "./embedding_distance/index.js";
+export type LoadEvaluatorOptions = EmbeddingDistanceEvalChainInput & {
+    llm?: BaseLanguageModel;
+    chainOptions?: Partial<Omit<LLMEvalChainInput, "llm">>;
+    /**
+     * The criteria to use for the evaluator.
+     */
+    criteria?: CriteriaLike;
+    /**
+     * A list of tools available to the agent, for TrajectoryEvalChain.
+     */
+    agentTools?: StructuredTool[];
+};
+/**
+ * Load the requested evaluation chain specified by a string
+ * @param type The type of evaluator to load.
+ * @param options
+ *        - llm The language model to use for the evaluator.
+ *        - criteria The criteria to use for the evaluator.
+ *        - agentTools A list of tools available to the agent,for TrajectoryEvalChain.
+ */
+export declare function loadEvaluator<T extends keyof EvaluatorType>(type: T, options?: LoadEvaluatorOptions): Promise<EvaluatorType[T]>;

package/dist/evaluation/loader.js

@@ -0,0 +1,56 @@
+import { CriteriaEvalChain, LabeledCriteriaEvalChain, } from "./criteria/index.js";
+import { ChatOpenAI } from "../chat_models/openai.js";
+import { LabeledPairwiseStringEvalChain, PairwiseStringEvalChain, } from "./comparison/index.js";
+import { EmbeddingDistanceEvalChain, PairwiseEmbeddingDistanceEvalChain, } from "./embedding_distance/index.js";
+import { TrajectoryEvalChain } from "./agents/index.js";
+import { BaseChatModel } from "../chat_models/base.js";
+/**
+ * Load the requested evaluation chain specified by a string
+ * @param type The type of evaluator to load.
+ * @param options
+ *        - llm The language model to use for the evaluator.
+ *        - criteria The criteria to use for the evaluator.
+ *        - agentTools A list of tools available to the agent,for TrajectoryEvalChain.
+ */
+export async function loadEvaluator(type, options) {
+    const { llm, chainOptions, criteria, agentTools } = options || {};
+    const llm_ = llm ??
+        new ChatOpenAI({
+            modelName: "gpt-4",
+            temperature: 0.0,
+        });
+    let evaluator;
+    switch (type) {
+        case "criteria":
+            evaluator = await CriteriaEvalChain.fromLLM(llm_, criteria, chainOptions);
+            break;
+        case "labeled_criteria":
+            evaluator = await LabeledCriteriaEvalChain.fromLLM(llm_, criteria, chainOptions);
+            break;
+        case "pairwise_string":
+            evaluator = await PairwiseStringEvalChain.fromLLM(llm_, criteria, chainOptions);
+            break;
+        case "labeled_pairwise_string":
+            evaluator = await LabeledPairwiseStringEvalChain.fromLLM(llm_, criteria, chainOptions);
+            break;
+        case "trajectory":
+            // eslint-disable-next-line no-instanceof/no-instanceof
+            if (!(llm_ instanceof BaseChatModel)) {
+                throw new Error("LLM must be an instance of a base chat model.");
+            }
+            evaluator = await TrajectoryEvalChain.fromLLM(llm_, agentTools, chainOptions);
+            break;
+        case "embedding_distance":
+            evaluator = new EmbeddingDistanceEvalChain({
+                embedding: options?.embedding,
+                distanceMetric: options?.distanceMetric,
+            });
+            break;
+        case "pairwise_embedding_distance":
+            evaluator = new PairwiseEmbeddingDistanceEvalChain({});
+            break;
+        default:
+            throw new Error(`Unknown type: ${type}`);
+    }
+    return evaluator;
+}

package/dist/evaluation/types.cjs

@@ -0,0 +1,2 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });

package/dist/evaluation/types.d.ts

@@ -0,0 +1,35 @@
+import { AgentTrajectoryEvaluator, LLMPairwiseStringEvaluator, LLMStringEvaluator, PairwiseStringEvaluator, StringEvaluator } from "./base.js";
+export interface EvaluatorType {
+    /**
+     * The criteria evaluator, which evaluates a model based on a
+     * custom set of criteria without any reference labels.
+     * */
+    criteria: LLMStringEvaluator;
+    /**
+     * The labeled criteria evaluator, which evaluates a model based on a
+     * custom set of criteria, with a reference label.
+     * */
+    labeled_criteria: LLMStringEvaluator;
+    /**
+     *  The pairwise string evaluator, which predicts the preferred prediction from
+     *  between two models.
+     */
+    pairwise_string: LLMPairwiseStringEvaluator;
+    /**
+     * The labeled pairwise string evaluator, which predicts the preferred prediction
+     * from between two models based on a ground truth reference label.
+     * */
+    labeled_pairwise_string: LLMPairwiseStringEvaluator;
+    /**
+     * The agent trajectory evaluator, which grades the agent's intermediate steps.
+     */
+    trajectory: AgentTrajectoryEvaluator;
+    /**
+     * Compare a prediction to a reference label using embedding distance.
+     * */
+    embedding_distance: StringEvaluator;
+    /**
+     * Compare two predictions using embedding distance.
+     * */
+    pairwise_embedding_distance: PairwiseStringEvaluator;
+}

package/dist/evaluation/types.js

@@ -0,0 +1 @@
+export {};

package/dist/experimental/multimodal_embeddings/googlevertexai.cjs

@@ -1,6 +1,7 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.GoogleVertexAIMultimodalEmbeddings = void 0;
+const google_auth_library_1 = require("google-auth-library");
 const base_js_1 = require("../../embeddings/base.cjs");
 const googlevertexai_connection_js_1 = require("../../util/googlevertexai-connection.cjs");
 /**
@@ -24,7 +25,10 @@ class GoogleVertexAIMultimodalEmbeddings extends base_js_1.Embeddings {
             value: void 0
         });
         this.model = fields?.model ?? this.model;
-        this.connection = new googlevertexai_connection_js_1.GoogleVertexAILLMConnection({ ...fields, ...this }, this.caller);
+        this.connection = new googlevertexai_connection_js_1.GoogleVertexAILLMConnection({ ...fields, ...this }, this.caller, new google_auth_library_1.GoogleAuth({
+            scopes: "https://www.googleapis.com/auth/cloud-platform",
+            ...fields?.authOptions,
+        }));
     }
     /**
      * Converts media (text or image) to an instance that can be used for

package/dist/experimental/multimodal_embeddings/googlevertexai.d.ts

@@ -1,11 +1,12 @@
 /// <reference types="node" resolution-mode="require"/>
+import { GoogleAuthOptions } from "google-auth-library";
 import { Embeddings, EmbeddingsParams } from "../../embeddings/base.js";
 import { GoogleVertexAIBaseLLMInput, GoogleVertexAIBasePrediction, GoogleVertexAILLMResponse } from "../../types/googlevertexai-types.js";
 /**
  * Parameters for the GoogleVertexAIMultimodalEmbeddings class, extending
  * both EmbeddingsParams and GoogleVertexAIConnectionParams.
  */
-export interface GoogleVertexAIMultimodalEmbeddingsParams extends EmbeddingsParams, GoogleVertexAIBaseLLMInput {
+export interface GoogleVertexAIMultimodalEmbeddingsParams extends EmbeddingsParams, GoogleVertexAIBaseLLMInput<GoogleAuthOptions> {
 }
 /**
  * An instance of media (text or image) that can be used for generating

package/dist/experimental/multimodal_embeddings/googlevertexai.js

@@ -1,3 +1,4 @@
+import { GoogleAuth } from "google-auth-library";
 import { Embeddings } from "../../embeddings/base.js";
 import { GoogleVertexAILLMConnection } from "../../util/googlevertexai-connection.js";
 /**
@@ -21,7 +22,10 @@ export class GoogleVertexAIMultimodalEmbeddings extends Embeddings {
             value: void 0
         });
         this.model = fields?.model ?? this.model;
-        this.connection = new GoogleVertexAILLMConnection({ ...fields, ...this }, this.caller);
+        this.connection = new GoogleVertexAILLMConnection({ ...fields, ...this }, this.caller, new GoogleAuth({
+            scopes: "https://www.googleapis.com/auth/cloud-platform",
+            ...fields?.authOptions,
+        }));
     }
     /**
      * Converts media (text or image) to an instance that can be used for

package/dist/llms/bedrock.cjs

@@ -112,6 +112,12 @@ class Bedrock extends base_js_1.LLM {
             writable: true,
             value: void 0
         });
+        Object.defineProperty(this, "endpointUrl", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: void 0
+        });
         Object.defineProperty(this, "codec", {
             enumerable: true,
             configurable: true,
@@ -132,6 +138,7 @@ class Bedrock extends base_js_1.LLM {
         this.temperature = fields?.temperature ?? this.temperature;
         this.maxTokens = fields?.maxTokens ?? this.maxTokens;
         this.fetchFn = fields?.fetchFn ?? fetch;
+        this.endpointUrl = fields?.endpointUrl;
     }
     /** Call out to Bedrock service model.
       Arguments:
@@ -154,7 +161,8 @@ class Bedrock extends base_js_1.LLM {
         const provider = this.model.split(".")[0];
         const service = "bedrock";
         const inputBody = BedrockLLMInputOutputAdapter.prepareInput(provider, prompt, this.maxTokens, this.temperature);
-        const url = new URL(`https://${service}.${this.region}.amazonaws.com/model/${this.model}/invoke-with-response-stream`);
+        const endpointUrl = this.endpointUrl ?? `${service}.${this.region}.amazonaws.com`;
+        const url = new URL(`https://${endpointUrl}/model/${this.model}/invoke-with-response-stream`);
         const request = new protocol_http_1.HttpRequest({
             hostname: url.hostname,
             path: url.pathname,

package/dist/llms/bedrock.d.ts

@@ -29,6 +29,8 @@ export interface BedrockInput {
     maxTokens?: number;
     /** A custom fetch function for low-level access to AWS API. Defaults to fetch() */
     fetchFn?: typeof fetch;
+    /** Override the default endpoint url */
+    endpointUrl?: string;
 }
 /**
  * A type of Large Language Model (LLM) that interacts with the Bedrock
@@ -46,6 +48,7 @@ export declare class Bedrock extends LLM implements BedrockInput {
     temperature?: number | undefined;
     maxTokens?: number | undefined;
     fetchFn: typeof fetch;
+    endpointUrl?: string;
     codec: EventStreamCodec;
     get lc_secrets(): {
         [key: string]: string;

package/dist/llms/bedrock.js

@@ -109,6 +109,12 @@ export class Bedrock extends LLM {
             writable: true,
             value: void 0
         });
+        Object.defineProperty(this, "endpointUrl", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: void 0
+        });
         Object.defineProperty(this, "codec", {
             enumerable: true,
             configurable: true,
@@ -129,6 +135,7 @@ export class Bedrock extends LLM {
         this.temperature = fields?.temperature ?? this.temperature;
         this.maxTokens = fields?.maxTokens ?? this.maxTokens;
         this.fetchFn = fields?.fetchFn ?? fetch;
+        this.endpointUrl = fields?.endpointUrl;
     }
     /** Call out to Bedrock service model.
       Arguments:
@@ -151,7 +158,8 @@ export class Bedrock extends LLM {
         const provider = this.model.split(".")[0];
         const service = "bedrock";
         const inputBody = BedrockLLMInputOutputAdapter.prepareInput(provider, prompt, this.maxTokens, this.temperature);
-        const url = new URL(`https://${service}.${this.region}.amazonaws.com/model/${this.model}/invoke-with-response-stream`);
+        const endpointUrl = this.endpointUrl ?? `${service}.${this.region}.amazonaws.com`;
+        const url = new URL(`https://${endpointUrl}/model/${this.model}/invoke-with-response-stream`);
         const request = new HttpRequest({
             hostname: url.hostname,
             path: url.pathname,

package/dist/llms/{googlevertexai.js → googlevertexai/common.cjs}

@@ -1,22 +1,26 @@
-import { BaseLLM } from "./base.js";
-import { GoogleVertexAILLMConnection } from "../util/googlevertexai-connection.js";
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.BaseGoogleVertexAI = void 0;
+const base_js_1 = require("../base.cjs");
 /**
- * Enables calls to the Google Cloud's Vertex AI API to access
- * Large Language Models.
- *
- * To use, you will need to have one of the following authentication
- * methods in place:
- * - You are logged into an account permitted to the Google Cloud project
- *   using Vertex AI.
- * - You are running this on a machine using a service account permitted to
- *   the Google Cloud project using Vertex AI.
- * - The `GOOGLE_APPLICATION_CREDENTIALS` environment variable is set to the
- *   path of a credentials file for a service account permitted to the
- *   Google Cloud project using Vertex AI.
+ * Base class for Google Vertex AI LLMs.
+ * Implemented subclasses must provide a GoogleVertexAILLMConnection
+ * with an appropriate auth client.
  */
-export class GoogleVertexAI extends BaseLLM {
+class BaseGoogleVertexAI extends base_js_1.BaseLLM {
+    get lc_aliases() {
+        return {
+            model: "model_name",
+        };
+    }
     constructor(fields) {
         super(fields ?? {});
+        Object.defineProperty(this, "lc_serializable", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: true
+        });
         Object.defineProperty(this, "model", {
             enumerable: true,
             configurable: true,
@@ -65,10 +69,9 @@ export class GoogleVertexAI extends BaseLLM {
         this.maxOutputTokens = fields?.maxOutputTokens ?? this.maxOutputTokens;
         this.topP = fields?.topP ?? this.topP;
         this.topK = fields?.topK ?? this.topK;
-        this.connection = new GoogleVertexAILLMConnection({ ...fields, ...this }, this.caller);
     }
     _llmType() {
-        return "googlevertexai";
+        return "vertexai";
     }
     async _generate(prompts, options) {
         const generations = await Promise.all(prompts.map((prompt) => this._generatePrompt(prompt, options)));
@@ -129,3 +132,4 @@ export class GoogleVertexAI extends BaseLLM {
         return result?.data?.predictions[0];
     }
 }
+exports.BaseGoogleVertexAI = BaseGoogleVertexAI;

package/dist/llms/{googlevertexai.d.ts → googlevertexai/common.d.ts}

@@ -1,11 +1,8 @@
-import { BaseLLM } from "./base.js";
-import { Generation, LLMResult } from "../schema/index.js";
-import { GoogleVertexAIBaseLLMInput, GoogleVertexAIBasePrediction, GoogleVertexAILLMResponse } from "../types/googlevertexai-types.js";
-/**
- * Interface representing the input to the Google Vertex AI model.
- */
-export interface GoogleVertexAITextInput extends GoogleVertexAIBaseLLMInput {
-}
+import { BaseLLM } from "../base.js";
+import { Generation, LLMResult } from "../../schema/index.js";
+import { GoogleVertexAILLMConnection } from "../../util/googlevertexai-connection.js";
+import { GoogleVertexAIBaseLLMInput, GoogleVertexAIBasePrediction, GoogleVertexAILLMResponse } from "../../types/googlevertexai-types.js";
+import { BaseLanguageModelCallOptions } from "../../base_language/index.js";
 /**
  * Interface representing the instance of text input to the Google Vertex
  * AI model.
@@ -32,27 +29,20 @@ interface TextPrediction extends GoogleVertexAIBasePrediction {
     content: string;
 }
 /**
- * Enables calls to the Google Cloud's Vertex AI API to access
- * Large Language Models.
- *
- * To use, you will need to have one of the following authentication
- * methods in place:
- * - You are logged into an account permitted to the Google Cloud project
- *   using Vertex AI.
- * - You are running this on a machine using a service account permitted to
- *   the Google Cloud project using Vertex AI.
- * - The `GOOGLE_APPLICATION_CREDENTIALS` environment variable is set to the
- *   path of a credentials file for a service account permitted to the
- *   Google Cloud project using Vertex AI.
+ * Base class for Google Vertex AI LLMs.
+ * Implemented subclasses must provide a GoogleVertexAILLMConnection
+ * with an appropriate auth client.
  */
-export declare class GoogleVertexAI extends BaseLLM implements GoogleVertexAITextInput {
+export declare class BaseGoogleVertexAI<AuthOptions> extends BaseLLM implements GoogleVertexAIBaseLLMInput<AuthOptions> {
+    lc_serializable: boolean;
     model: string;
     temperature: number;
     maxOutputTokens: number;
     topP: number;
     topK: number;
-    private connection;
-    constructor(fields?: GoogleVertexAITextInput);
+    protected connection: GoogleVertexAILLMConnection<BaseLanguageModelCallOptions, GoogleVertexAILLMInstance, TextPrediction, AuthOptions>;
+    get lc_aliases(): Record<string, string>;
+    constructor(fields?: GoogleVertexAIBaseLLMInput<AuthOptions>);
     _llmType(): string;
     _generate(prompts: string[], options: this["ParsedCallOptions"]): Promise<LLMResult>;
     _generatePrompt(prompt: string, options: this["ParsedCallOptions"]): Promise<Generation[]>;