feat: enhance OpenAI provider handling and add providerParams utility module

- Updated the `createBaseModel` function to handle OpenAI provider responses in strict mode.
- Modified `providerToAiSdkConfig` to include specific options for OpenAI when in strict mode.
- Introduced a new utility module `providerParams.ts` for managing provider-specific parameters, including OpenAI, Anthropic, and Gemini configurations.
- Added functions to retrieve service tiers, specific parameters, and reasoning efforts for various providers, improving overall provider management.

packages/aiCore/src/core/models/ProviderCreator.ts

@@ -65,10 +65,13 @@ export async function createBaseModel(
         `Creator function "${providerConfig.creatorFunctionName}" not found in the imported module for provider "${effectiveProviderId}"`
       )
     }
-
     // 创建provider实例
-    const provider = creatorFunction(options)
+    let provider = creatorFunction(options)
 
+    // 加一个特判
+    if (providerConfig.id === 'openai' && options.compatibility === 'strict') {
+      provider = provider.responses
+    }
     // 返回模型实例
     if (typeof provider === 'function') {
       let model: LanguageModelV1 = provider(modelId)

src/renderer/src/aiCore/index_new.ts

@@ -53,9 +53,16 @@ function providerToAiSdkConfig(provider: Provider): {
 
   const aiSdkProviderId = getAiSdkProviderId(actualProvider)
 
-  if (AiCore.isSupported(aiSdkProviderId) && aiSdkProviderId !== 'openai-compatible') {
-    const options = ProviderConfigFactory.fromProvider(aiSdkProviderId, actualProvider)
+  // 如果provider是openai，则使用strict模式并且默认responses api
+  const openaiResponseOptions =
+    aiSdkProviderId === 'openai'
+      ? {
+          compatibility: 'strict'
+        }
+      : undefined
 
+  if (AiCore.isSupported(aiSdkProviderId) && aiSdkProviderId !== 'openai-compatible') {
+    const options = ProviderConfigFactory.fromProvider(aiSdkProviderId, actualProvider, openaiResponseOptions)
     return {
       providerId: aiSdkProviderId as ProviderId,
       options
@@ -183,13 +190,13 @@ export default class ModernAiProvider {
 
     // 动态构建中间件数组
     const middlewares = buildAiSdkMiddlewares(middlewareConfig)
-    console.log('构建的中间件:', middlewares)
+    // console.log('构建的中间件:', middlewares)
 
     // 创建带有中间件的执行器
     if (middlewareConfig.onChunk) {
       // 流式处理 - 使用适配器
       const adapter = new AiSdkToChunkAdapter(middlewareConfig.onChunk)
-
+      console.log('最终params', params)
       const streamResult = await executor.streamText(
         modelId,
         params,

src/renderer/src/aiCore/utils/providerParams.ts

@@ -0,0 +1,317 @@
+/**
+ * 提供商特定参数处理模块
+ * 从各个 API 客户端迁移的参数构建逻辑，用于 AI SDK 的 providerOptions
+ */
+
+import { Modality } from '@google/genai'
+import { DEFAULT_MAX_TOKENS } from '@renderer/config/constant'
+import {
+  findTokenLimit,
+  GEMINI_FLASH_MODEL_REGEX,
+  getOpenAIWebSearchParams,
+  isGeminiReasoningModel,
+  isOpenAIModel,
+  isOpenAIReasoningModel,
+  isReasoningModel,
+  isSupportedFlexServiceTier,
+  isSupportedReasoningEffortOpenAIModel
+} from '@renderer/config/models'
+import { getStoreSetting } from '@renderer/hooks/useSettings'
+import { getAssistantSettings } from '@renderer/services/AssistantService'
+import type { SettingsState } from '@renderer/store/settings'
+import { Assistant, EFFORT_RATIO, Model, OpenAIServiceTier } from '@renderer/types'
+
+// ===== OpenAI 相关参数 =====
+
+/**
+ * 获取 OpenAI 服务层级
+ * 从 BaseApiClient.getServiceTier 迁移
+ */
+export function getServiceTier(model: Model): OpenAIServiceTier | undefined {
+  if (!isOpenAIModel(model) || model.provider === 'github' || model.provider === 'copilot') {
+    return undefined
+  }
+
+  const openAI = getStoreSetting('openAI') as SettingsState['openAI']
+  let serviceTier = 'auto' as OpenAIServiceTier
+
+  if (openAI && openAI?.serviceTier === 'flex') {
+    if (isSupportedFlexServiceTier(model)) {
+      serviceTier = 'flex'
+    } else {
+      serviceTier = 'auto'
+    }
+  } else {
+    serviceTier = openAI.serviceTier
+  }
+
+  return serviceTier
+}
+
+/**
+ * 获取 OpenAI 提供商特定参数
+ * 从 OpenAIBaseClient.getProviderSpecificParameters 迁移
+ */
+export function getOpenAIProviderSpecificParameters(assistant: Assistant, model: Model) {
+  const { maxTokens } = getAssistantSettings(assistant)
+
+  if (model.provider === 'openrouter') {
+    if (model.id.includes('deepseek-r1')) {
+      return {
+        include_reasoning: true
+      }
+    }
+  }
+
+  if (isOpenAIReasoningModel(model)) {
+    return {
+      max_tokens: undefined,
+      max_completion_tokens: maxTokens
+    }
+  }
+
+  return {}
+}
+
+/**
+ * 获取 OpenAI 推理努力参数
+ * 从 OpenAIBaseClient.getReasoningEffort 迁移
+ */
+export function getOpenAIReasoningEffort(assistant: Assistant, model: Model) {
+  if (!isSupportedReasoningEffortOpenAIModel(model)) {
+    return {}
+  }
+
+  const openAI = getStoreSetting('openAI') as SettingsState['openAI']
+  const summaryText = openAI?.summaryText || 'off'
+
+  let summary: string | undefined = undefined
+
+  if (summaryText === 'off' || model.id.includes('o1-pro')) {
+    summary = undefined
+  } else {
+    summary = summaryText
+  }
+
+  const reasoningEffort = assistant?.settings?.reasoning_effort
+  if (!reasoningEffort) {
+    return {}
+  }
+
+  if (isSupportedReasoningEffortOpenAIModel(model)) {
+    return {
+      reasoning: {
+        effort: reasoningEffort,
+        summary: summary
+      }
+    }
+  }
+
+  return {}
+}
+
+// ===== Anthropic 相关参数 =====
+
+/**
+ * 获取 Anthropic 思维预算令牌配置
+ * 从 AnthropicAPIClient.getBudgetToken 迁移
+ */
+export function getAnthropicBudgetToken(assistant: Assistant, model: Model) {
+  if (!isReasoningModel(model)) {
+    return undefined
+  }
+
+  const { maxTokens } = getAssistantSettings(assistant)
+  const reasoningEffort = assistant?.settings?.reasoning_effort
+
+  if (reasoningEffort === undefined) {
+    return {
+      type: 'disabled'
+    }
+  }
+
+  const effortRatio = EFFORT_RATIO[reasoningEffort]
+  const tokenLimit = findTokenLimit(model.id)
+
+  if (!tokenLimit) {
+    return {
+      type: 'enabled',
+      budget_tokens: Math.max(1024, Math.floor((maxTokens || DEFAULT_MAX_TOKENS) * effortRatio))
+    }
+  }
+
+  const budgetTokens = Math.max(
+    1024,
+    Math.floor(
+      Math.min(
+        (tokenLimit.max - tokenLimit.min) * effortRatio + tokenLimit.min,
+        (maxTokens || DEFAULT_MAX_TOKENS) * effortRatio
+      )
+    )
+  )
+
+  return {
+    type: 'enabled',
+    budget_tokens: budgetTokens
+  }
+}
+
+// ===== Gemini 相关参数 =====
+
+/**
+ * 获取 Gemini 思维配置
+ * 从 GeminiAPIClient.getBudgetToken 迁移
+ */
+export function getGeminiBudgetToken(assistant: Assistant, model: Model) {
+  if (isGeminiReasoningModel(model)) {
+    const reasoningEffort = assistant?.settings?.reasoning_effort
+
+    // 如果 thinking_budget 是 undefined，不思考
+    if (reasoningEffort === undefined) {
+      return {
+        thinkingConfig: {
+          includeThoughts: false,
+          ...(GEMINI_FLASH_MODEL_REGEX.test(model.id) ? { thinkingBudget: 0 } : {})
+        }
+      }
+    }
+
+    const effortRatio = EFFORT_RATIO[reasoningEffort]
+
+    if (effortRatio > 1) {
+      return {
+        thinkingConfig: {
+          includeThoughts: true
+        }
+      }
+    }
+
+    const tokenLimit = findTokenLimit(model.id)
+    const { min = 0, max = 0 } = tokenLimit || {}
+
+    // 计算 budgetTokens，确保不低于 min
+    const budget = Math.floor((max - min) * effortRatio + min)
+
+    return {
+      thinkingConfig: {
+        ...(budget > 0 ? { thinkingBudget: budget } : {}),
+        includeThoughts: true
+      }
+    }
+  }
+
+  return {}
+}
+
+/**
+ * 获取 Gemini 安全设置
+ * 从 GeminiAPIClient.getSafetySettings 迁移
+ */
+export function getGeminiSafetySettings() {
+  return [
+    {
+      category: 'HARM_CATEGORY_HARASSMENT',
+      threshold: 'BLOCK_NONE'
+    },
+    {
+      category: 'HARM_CATEGORY_HATE_SPEECH',
+      threshold: 'BLOCK_NONE'
+    },
+    {
+      category: 'HARM_CATEGORY_SEXUALLY_EXPLICIT',
+      threshold: 'BLOCK_NONE'
+    },
+    {
+      category: 'HARM_CATEGORY_DANGEROUS_CONTENT',
+      threshold: 'BLOCK_NONE'
+    }
+  ]
+}
+
+/**
+ * 获取 Gemini 图像生成参数
+ * 从 GeminiAPIClient.getGenerateImageParameter 迁移
+ */
+export function getGeminiGenerateImageParameter() {
+  return {
+    systemInstruction: undefined,
+    responseModalities: [Modality.TEXT, Modality.IMAGE],
+    responseMimeType: 'text/plain'
+  }
+}
+
+// ===== Web 搜索相关参数 =====
+
+/**
+ * 获取 OpenAI 的 Web 搜索 providerOptions
+ * 复用现有的 getOpenAIWebSearchParams 函数
+ */
+export function getOpenAIWebSearchProviderOptions(model: Model, enableWebSearch: boolean): Record<string, any> {
+  return getOpenAIWebSearchParams(model, enableWebSearch)
+}
+
+/**
+ * 获取 Anthropic 的 Web 搜索 providerOptions
+ */
+export function getAnthropicWebSearchProviderOptions(model: Model, enableWebSearch: boolean): Record<string, any> {
+  if (!enableWebSearch) {
+    return {}
+  }
+
+  // Anthropic 通过 tools 实现，但在 AI SDK 中应该通过 providerOptions 统一
+  return {
+    webSearch: {
+      enabled: true,
+      toolType: 'web_search_20250305',
+      maxUses: 5
+    }
+  }
+}
+
+/**
+ * 获取 Gemini 的 Web 搜索 providerOptions
+ */
+export function getGeminiWebSearchProviderOptions(model: Model, enableWebSearch: boolean): Record<string, any> {
+  if (!enableWebSearch) {
+    return {}
+  }
+
+  // Gemini 通过 googleSearch 工具实现，但在 AI SDK 中应该通过 providerOptions 统一
+  return {
+    webSearch: {
+      enabled: true,
+      toolType: 'googleSearch'
+    }
+  }
+}
+
+// ===== 通用辅助函数 =====
+
+/**
+ * 获取自定义参数
+ * 从 BaseApiClient.getCustomParameters 迁移
+ */
+export function getCustomParameters(assistant: Assistant) {
+  return (
+    assistant?.settings?.customParameters?.reduce((acc, param) => {
+      if (!param.name?.trim()) {
+        return acc
+      }
+      if (param.type === 'json') {
+        const value = param.value as string
+        if (value === 'undefined') {
+          return { ...acc, [param.name]: undefined }
+        }
+        try {
+          return { ...acc, [param.name]: JSON.parse(value) }
+        } catch {
+          return { ...acc, [param.name]: value }
+        }
+      }
+      return {
+        ...acc,
+        [param.name]: param.value
+      }
+    }, {}) || {}
+  )
+}