n8n-nodes-lmstudio-embeddings/dist/nodes/LMStudioEmbeddings.node.js

n8n community node for LM Studio Embeddings API with encoding format selection

"use strict";
Object.defineProperty(exports, "__esModule", { value: true });
exports.LMStudioEmbeddings = void 0;
const openai_1 = require("@langchain/openai");
class LMStudioEmbeddings {
    constructor() {
        this.description = {
            displayName: 'LM Studio Embeddings',
            name: 'lmStudioEmbeddings',
            icon: 'file:lmstudio.svg',
            group: ['transform'],
            version: 1,
            description: 'Generate embeddings using LM Studio API',
            defaults: {
                name: 'LM Studio Embeddings',
            },
            inputs: [],
            outputs: ["ai_embedding" /* NodeConnectionType.AiEmbedding */],
            credentials: [
                {
                    name: 'lmStudioApi',
                    required: true,
                },
            ],
            properties: [
                {
                    displayName: 'Model',
                    name: 'model',
                    type: 'options',
                    typeOptions: {
                        loadOptionsMethod: 'getModels',
                    },
                    default: '',
                    description: 'The embedding model to use',
                    required: true,
                },
                {
                    displayName: 'Encoding Format',
                    name: 'encodingFormat',
                    type: 'options',
                    options: [
                        {
                            name: 'Float',
                            value: 'float',
                            description: 'Return embeddings as floating point numbers',
                        },
                        {
                            name: 'Base64',
                            value: 'base64',
                            description: 'Return embeddings encoded as base64',
                        },
                    ],
                    default: 'float',
                    description: 'The format for the embeddings',
                }
            ],
        };
        this.methods = {
            loadOptions: {
                async getModels() {
                    const credentials = await this.getCredentials('lmStudioApi');
                    try {
                        const response = await fetch(`${credentials.baseUrl}/models`, {
                            method: 'GET',
                            headers: {
                                'Content-Type': 'application/json',
                                ...(credentials.apiKey && { Authorization: `Bearer ${credentials.apiKey}` }),
                            },
                        });
                        if (!response.ok) {
                            throw new Error(`Failed to fetch models: ${response.status} ${response.statusText}`);
                        }
                        const result = await response.json();
                        const models = result.data || [];
                        return models.map((model) => ({
                            name: model.id,
                            value: model.id,
                        }));
                    }
                    catch (error) {
                        throw new Error(`Error fetching models from LM Studio: ${error instanceof Error ? error.message : 'Unknown error'}`);
                    }
                },
            },
        };
    }
    async supplyData(itemIndex) {
        // Создаем объект с методом embedQuery для vector store
        const embeddingProvider = {
            embedQuery: async (text) => {
                const credentials = await this.getCredentials('lmStudioApi');
                const model = this.getNodeParameter('model', itemIndex);
                const encodingFormat = this.getNodeParameter('encodingFormat', itemIndex);
                if (!text || !text.trim()) {
                    throw new Error('Text content is empty');
                }
                const embeddings = new openai_1.OpenAIEmbeddings({
                    configuration: {
                        baseURL: credentials.baseUrl,
                    },
                    apiKey: credentials.apiKey,
                    model: model,
                });
                // @ts-ignore
                const { data } = await embeddings.embeddingWithRetry({
                    model: model,
                    input: text,
                    encoding_format: encodingFormat,
                });
                return data[0].embedding;
            }
        };
        return {
            response: embeddingProvider,
        };
    }
}
exports.LMStudioEmbeddings = LMStudioEmbeddings;