@langchain/openai
Version:
OpenAI integrations for LangChain.js
196 lines (195 loc) • 7.24 kB
JavaScript
"use strict";
Object.defineProperty(exports, "__esModule", { value: true });
exports.OpenAIEmbeddings = void 0;
const openai_1 = require("openai");
const env_1 = require("@langchain/core/utils/env");
const embeddings_1 = require("@langchain/core/embeddings");
const chunk_array_1 = require("@langchain/core/utils/chunk_array");
const azure_js_1 = require("./utils/azure.cjs");
const openai_js_1 = require("./utils/openai.cjs");
/**
* Class for generating embeddings using the OpenAI API.
*
* To use with Azure, import the `AzureOpenAIEmbeddings` class.
*
* @example
* ```typescript
* // Embed a query using OpenAIEmbeddings to generate embeddings for a given text
* const model = new OpenAIEmbeddings();
* const res = await model.embedQuery(
* "What would be a good company name for a company that makes colorful socks?",
* );
* console.log({ res });
*
* ```
*/
class OpenAIEmbeddings extends embeddings_1.Embeddings {
constructor(fields) {
const fieldsWithDefaults = { maxConcurrency: 2, ...fields };
super(fieldsWithDefaults);
Object.defineProperty(this, "model", {
enumerable: true,
configurable: true,
writable: true,
value: "text-embedding-ada-002"
});
/** @deprecated Use "model" instead */
Object.defineProperty(this, "modelName", {
enumerable: true,
configurable: true,
writable: true,
value: void 0
});
Object.defineProperty(this, "batchSize", {
enumerable: true,
configurable: true,
writable: true,
value: 512
});
// TODO: Update to `false` on next minor release (see: https://github.com/langchain-ai/langchainjs/pull/3612)
Object.defineProperty(this, "stripNewLines", {
enumerable: true,
configurable: true,
writable: true,
value: true
});
/**
* The number of dimensions the resulting output embeddings should have.
* Only supported in `text-embedding-3` and later models.
*/
Object.defineProperty(this, "dimensions", {
enumerable: true,
configurable: true,
writable: true,
value: void 0
});
Object.defineProperty(this, "timeout", {
enumerable: true,
configurable: true,
writable: true,
value: void 0
});
Object.defineProperty(this, "organization", {
enumerable: true,
configurable: true,
writable: true,
value: void 0
});
Object.defineProperty(this, "client", {
enumerable: true,
configurable: true,
writable: true,
value: void 0
});
Object.defineProperty(this, "clientConfig", {
enumerable: true,
configurable: true,
writable: true,
value: void 0
});
const apiKey = fieldsWithDefaults?.apiKey ??
fieldsWithDefaults?.openAIApiKey ??
(0, env_1.getEnvironmentVariable)("OPENAI_API_KEY");
this.organization =
fieldsWithDefaults?.configuration?.organization ??
(0, env_1.getEnvironmentVariable)("OPENAI_ORGANIZATION");
this.model =
fieldsWithDefaults?.model ?? fieldsWithDefaults?.modelName ?? this.model;
this.modelName = this.model;
this.batchSize = fieldsWithDefaults?.batchSize ?? this.batchSize;
this.stripNewLines =
fieldsWithDefaults?.stripNewLines ?? this.stripNewLines;
this.timeout = fieldsWithDefaults?.timeout;
this.dimensions = fieldsWithDefaults?.dimensions;
this.clientConfig = {
apiKey,
organization: this.organization,
dangerouslyAllowBrowser: true,
...fields?.configuration,
};
}
/**
* Method to generate embeddings for an array of documents. Splits the
* documents into batches and makes requests to the OpenAI API to generate
* embeddings.
* @param texts Array of documents to generate embeddings for.
* @returns Promise that resolves to a 2D array of embeddings for each document.
*/
async embedDocuments(texts) {
const batches = (0, chunk_array_1.chunkArray)(this.stripNewLines ? texts.map((t) => t.replace(/\n/g, " ")) : texts, this.batchSize);
const batchRequests = batches.map((batch) => {
const params = {
model: this.model,
input: batch,
};
if (this.dimensions) {
params.dimensions = this.dimensions;
}
return this.embeddingWithRetry(params);
});
const batchResponses = await Promise.all(batchRequests);
const embeddings = [];
for (let i = 0; i < batchResponses.length; i += 1) {
const batch = batches[i];
const { data: batchResponse } = batchResponses[i];
for (let j = 0; j < batch.length; j += 1) {
embeddings.push(batchResponse[j].embedding);
}
}
return embeddings;
}
/**
* Method to generate an embedding for a single document. Calls the
* embeddingWithRetry method with the document as the input.
* @param text Document to generate an embedding for.
* @returns Promise that resolves to an embedding for the document.
*/
async embedQuery(text) {
const params = {
model: this.model,
input: this.stripNewLines ? text.replace(/\n/g, " ") : text,
};
if (this.dimensions) {
params.dimensions = this.dimensions;
}
const { data } = await this.embeddingWithRetry(params);
return data[0].embedding;
}
/**
* Private method to make a request to the OpenAI API to generate
* embeddings. Handles the retry logic and returns the response from the
* API.
* @param request Request to send to the OpenAI API.
* @returns Promise that resolves to the response from the API.
*/
async embeddingWithRetry(request) {
if (!this.client) {
const openAIEndpointConfig = {
baseURL: this.clientConfig.baseURL,
};
const endpoint = (0, azure_js_1.getEndpoint)(openAIEndpointConfig);
const params = {
...this.clientConfig,
baseURL: endpoint,
timeout: this.timeout,
maxRetries: 0,
};
if (!params.baseURL) {
delete params.baseURL;
}
this.client = new openai_1.OpenAI(params);
}
const requestOptions = {};
return this.caller.call(async () => {
try {
const res = await this.client.embeddings.create(request, requestOptions);
return res;
}
catch (e) {
const error = (0, openai_js_1.wrapOpenAIClientError)(e);
throw error;
}
});
}
}
exports.OpenAIEmbeddings = OpenAIEmbeddings;