update infio api

src/core/rag/embedding.ts

@@ -4,7 +4,7 @@ import { URL } from 'url'
 import { GoogleGenerativeAI } from '@google/generative-ai'
 import { OpenAI } from 'openai'
 
-import { ALIBABA_QWEN_BASE_URL, OPENAI_BASE_URL, SILICONFLOW_BASE_URL } from "../../constants"
+import { ALIBABA_QWEN_BASE_URL, INFIO_BASE_URL, OPENAI_BASE_URL, SILICONFLOW_BASE_URL } from "../../constants"
 import { EmbeddingModel } from '../../types/embedding'
 import { ApiProvider } from '../../types/llm/model'
 import { InfioSettings } from '../../types/settings'
@@ -20,6 +20,42 @@ export const getEmbeddingModel = (
 	settings: InfioSettings,
 ): EmbeddingModel => {
 	switch (settings.embeddingModelProvider) {
+		case ApiProvider.Infio: {
+			const openai = new OpenAI({
+				apiKey: settings.infioProvider.apiKey,
+				baseURL: INFIO_BASE_URL,
+				dangerouslyAllowBrowser: true,
+			})
+			const modelInfo = GetEmbeddingModelInfo(settings.embeddingModelProvider, settings.embeddingModelId)
+			return {
+				id: settings.embeddingModelId,
+				dimension: modelInfo.dimensions,
+				getEmbedding: async (text: string) => {
+					try {
+						if (!openai.apiKey) {
+							throw new LLMAPIKeyNotSetException(
+								'OpenAI API key is missing. Please set it in settings menu.',
+							)
+						}
+						const embedding = await openai.embeddings.create({
+							model: settings.embeddingModelId,
+							input: text,
+						})
+						return embedding.data[0].embedding
+					} catch (error) {
+						if (
+							error.status === 429 &&
+							error.message.toLowerCase().includes('rate limit')
+						) {
+							throw new LLMRateLimitExceededException(
+								'OpenAI API rate limit exceeded. Please try again later.',
+							)
+						}
+						throw error
+					}
+				},
+			}
+		}
 		case ApiProvider.OpenAI: {
 			const baseURL = settings.openaiProvider.useCustomUrl ? settings.openaiProvider.baseUrl : OPENAI_BASE_URL
 			const openai = new OpenAI({

src/utils/api.ts

@@ -188,6 +188,26 @@ async function fetchInfioModels(apiKey?: string): Promise<Record<string, ModelIn
 	}
 }
 
+export const infioEmbeddingModels = {
+	"openai/text-embedding-3-small": {
+		dimensions: 1536,
+		description: "Increased performance over 2nd generation ada embedding model"
+	},
+	"gemini/gemini-embedding-exp-03-07": {
+		dimensions: 1024,
+		description: "Most capable 2nd generation embedding model, replacing 16 first generation models"
+	},
+	"deepseek/embedding-large-text": {
+		dimensions: 1024,
+		description: "Most capable embedding model for both English and non-English tasks"
+	},
+	"deepseek/embedding-text": {
+		dimensions: 512,
+		description: "Most capable embedding model for both English and non-English tasks"
+	}
+} as const satisfies Record<string, EmbeddingModelInfo>
+
+
 // OpenRouter
 // https://openrouter.ai/models?order=newest&supported_parameters=tools
 export const openRouterDefaultModelId = "anthropic/claude-sonnet-4" // will always exist in openRouterModels
@@ -1593,6 +1613,7 @@ export const GetAllProviders = (): ApiProvider[] => {
 
 export const GetEmbeddingProviders = (): ApiProvider[] => {
 	return [
+		ApiProvider.Infio,
 		ApiProvider.OpenAI,
 		ApiProvider.SiliconFlow,
 		ApiProvider.Google,
@@ -1684,6 +1705,8 @@ export const GetProviderModelIds = async (provider: ApiProvider, settings?: Infi
 // Get all embedding models for a provider
 export const GetEmbeddingProviderModels = (provider: ApiProvider): Record<string, EmbeddingModelInfo> => {
 	switch (provider) {
+		case ApiProvider.Infio:
+			return infioEmbeddingModels
 		case ApiProvider.Google:
 			return geminiEmbeddingModels
 		case ApiProvider.SiliconFlow: