Description

• CoreTool* in ai package

  PR: fix (ai): remove deprecated CoreTool* types #5768

  packages/ai/core/generate-text/tool-call.ts:

    19   */
    20  // TODO remove in v5
    21  export type CoreToolCallUnion<TOOLS extends ToolSet> = ToolCallUnion<ToolSet>;
  

  packages/ai/core/generate-text/tool-result.ts:

    31   */
    32  // TODO remove in v5
    33  export type CoreToolResultUnion<TOOLS extends ToolSet> = ToolResultUnion<TOOLS>;
  

  packages/ai/core/tool/tool.ts:

    108   */
    109  // TODO remove in v5
    110  export type CoreTool<
  

  packages/ai/core/types/language-model.ts:

    61   */
    62  // TODO remove in v5
    63  export type CoreToolChoice<TOOLS extends Record<string, unknown>> =
  

• Special handling of temperature in ai

  PR: refactor (ai): remove obsolete handling of temperature when it is undefined #5769

  packages/ai/core/middleware/default-settings-middleware.ts:

  TODO: verify if temperature now defaults to undefined

    31          // special case for temperature 0
    32          // TODO remove when temperature defaults to undefined
    33          temperature:
  

• experimental_wrapLanguageModel in ai

  PR: fix (ai): remove deprecated experimental_wrapLanguageModel #5771

  packages/ai/core/middleware/wrap-language-model.ts:

    101   */
    102  // TODO remove in v5
    103  export const experimental_wrapLanguageModel = wrapLanguageModel;
  

• options.throwErrorForEmptyVectors for cosineSimilarity() function

  PR: fix (ai): remove deprecated options.throwErrorForEmptyVectors from cosineSimilarity() #5772

  The options argument can be removed entirely.

  packages/ai/core/util/cosine-similarity.ts:

    20    vector2: number[],
    21    // TODO remove throw option in 5.0
    22    options?: {
  

• max_completion_tokens in openai package

  Moved to max completion/output token calculations #5792

  packages/openai/src/openai-chat-language-model.ts:

    190        // openai specific settings:
    19 :       // TODO remove in next major version; we auto-map maxOutputTokens now
    192        max_completion_tokens: openaiOptions.maxCompletionTokens,
  

• rawCall in language-model

  PR: fix (provider): remove obsolete .rawCall from doStream() and doGenerate() results (now there is request) #5774

  First step would be to copy the v1 folder to v2 and remove rawCall in there, then export v1 explicitly so that dependends can be migrated one-by-one.

  packages/provider/src/language-model/v1/language-model-v1.ts:

    250       */
    25 :     // TODO remove in v2 (there is now request)
    252      rawCall: {
  

• Experimental_LanguageModelV2Middleware in language-model-middleware

  PR: fix (provider): remove deprecated Experimental_LanguageModelV2Middleware #5775

  packages/provider/src/language-model-middleware/v2/language-model-v2-middleware.ts:

    65   */
    66  // TODO remove in v5
    67  export type Experimental_LanguageModelV2Middleware = LanguageModelV2Middleware;
  

• CoreTool* in provider-utils

  PR: fix (provider-utils): remove deprecated CoreToolCall and CoreToolResult types #5776

  packages/provider-utils/src/types/tool-call.ts:

    24   */
    25  // TODO remove in v5
    26  export type CoreToolCall<NAME extends string, ARGS> = ToolCall<NAME, ARGS>;
  

  packages/provider-utils/src/types/tool-result.ts:

    29   */
    30  // TODO remove in v5
    31  export type CoreToolResult<NAME extends string, ARGS, RESULT> = ToolResult<