openai

type AWSBedrockReasoningContent struct {
	// See https://docs.aws.amazon.com/bedrock/latest/APIReference/API_runtime_ReasoningContentBlock.html for more information.
	ReasoningContent *awsbedrock.ReasoningContentBlock `json:"reasoningContent,omitzero"`
}

type AWSBedrockStreamReasoningContent struct {
	Text            string `json:"text,omitzero"`
	Signature       string `json:"signature,omitzero"`
	RedactedContent []byte `json:"redactedContent,omitzero"`
}

type Annotation struct {
	// Type is the type of the annotation. Always "url_citation" for web search.
	Type string `json:"type"`
	// URLCitation contains the citation details when type is "url_citation".
	URLCitation *URLCitation `json:"url_citation,omitempty"` //nolint:tagliatelle //follow openai api
}

type AnthropicVendorFields struct {
	// Thinking holds Anthropic thinking configuration options.
	//
	// https://docs.anthropic.com/en/api/messages#body-thinking
	Thinking *anthropic.ThinkingConfigParamUnion `json:"thinking,omitzero"`
}

type ChatCompletionAssistantMessageParam struct {
	// The role of the messages author, in this case `assistant`.
	Role string `json:"role"`
	// Data about a previous audio response from the model.
	// [Learn more](https://platform.openai.com/docs/guides/audio).
	Audio ChatCompletionAssistantMessageParamAudio `json:"audio,omitzero"`
	// The contents of the assistant message. Required unless `tool_calls` or
	// `function_call` is specified.
	Content StringOrAssistantRoleContentUnion `json:"content"`
	// An optional name for the participant. Provides the model information to
	// differentiate between participants of the same role.
	Name string `json:"name,omitempty"`
	// The refusal message by the assistant.
	Refusal string `json:"refusal,omitempty"`
	// The tool calls generated by the model, such as function calls.
	ToolCalls []ChatCompletionMessageToolCallParam `json:"tool_calls,omitempty"`
}

type ChatCompletionAssistantMessageParamAudio struct {
	// Unique identifier for a previous audio response from the model.
	ID string `json:"id"`
}

type ChatCompletionAssistantMessageParamContent struct {
	// The type of the content part.
	Type ChatCompletionAssistantMessageParamContentType `json:"type"`
	// The refusal message generated by the model.
	Refusal *string `json:"refusal,omitempty"`
	// The text content.
	Text *string `json:"text,omitempty"`

	// The signature for a thinking block.
	Signature       *string `json:"signature,omitempty"`
	RedactedContent []byte  `json:"redactedContent,omitempty"`
}

type ChatCompletionAssistantMessageParamContentType string

type ChatCompletionAudioFormat string

type ChatCompletionAudioParam struct {
	// Voice specifies the voice the model uses to respond. Supported voices are alloy, ash, ballad, coral, echo, fable, nova, onyx, sage, and shimmer.
	Voice ChatCompletionAudioVoice `json:"voice"`
	// Format specifies the output audio format. Must be one of wav, mp3, flac, opus, or pcm16.
	Format ChatCompletionAudioFormat `json:"format"`
}

type ChatCompletionAudioVoice string

type ChatCompletionChoicesFinishReason string

type ChatCompletionChoicesLogprobs struct {
	// A list of message content tokens with log probability information.
	Content []ChatCompletionTokenLogprob `json:"content,omitempty"`
	// A list of message refusal tokens with log probability information.
	Refusal []ChatCompletionTokenLogprob `json:"refusal,omitempty"`
}

type ChatCompletionChunkChoiceDeltaToolCall struct {
	Index int64 `json:"index"`
	// The ID of the tool call.
	ID *string `json:"id"`
	// The function that the model called.
	Function ChatCompletionMessageToolCallFunctionParam `json:"function"`
	// The type of the tool. Currently, only `function` is supported.
	Type ChatCompletionMessageToolCallType `json:"type,omitempty"`
}

type ChatCompletionContentPartFileFileParam struct {
	// The base64 encoded file data, used when passing the file to the model as a
	// string.
	FileData string `json:"file_data,omitzero"`
	// The ID of an uploaded file to use as input.
	FileID string `json:"file_id,omitzero"`
	// The name of the file, used when passing the file to the model as a string.
	Filename string `json:"filename,omitzero"`
}

type ChatCompletionContentPartFileParam struct {
	File ChatCompletionContentPartFileFileParam `json:"file,omitzero"`
	// The type of the content part. Always `file`.
	//
	// This field can be elided, and will marshal its zero value as "file".
	Type ChatCompletionContentPartFileType `json:"type"`
}

type ChatCompletionContentPartFileType string

type ChatCompletionContentPartImageImageURLDetail string

type ChatCompletionContentPartImageImageURLParam struct {
	// Either a URL of the image or the base64 encoded image data.
	URL string `json:"url"`
	// Specifies the detail level of the image. Learn more in the
	// [Vision guide](https://platform.openai.com/docs/guides/vision#low-or-high-fidelity-image-understanding).
	Detail ChatCompletionContentPartImageImageURLDetail `json:"detail,omitempty"`
}

type ChatCompletionContentPartImageParam struct {
	ImageURL ChatCompletionContentPartImageImageURLParam `json:"image_url"`
	// The type of the content part.
	Type ChatCompletionContentPartImageType `json:"type"`
}

type ChatCompletionContentPartImageType string

type ChatCompletionContentPartInputAudioInputAudioFormat string

type ChatCompletionContentPartInputAudioInputAudioParam struct {
	// Base64 encoded audio data.
	Data string `json:"data"`
	// The format of the encoded audio data. Currently supports "wav" and "mp3".
	Format ChatCompletionContentPartInputAudioInputAudioFormat `json:"format"`
}

type ChatCompletionContentPartInputAudioParam struct {
	InputAudio ChatCompletionContentPartInputAudioInputAudioParam `json:"input_audio"`
	// The type of the content part. Always `input_audio`.
	Type ChatCompletionContentPartInputAudioType `json:"type"`
}

type ChatCompletionContentPartInputAudioType string

type ChatCompletionContentPartRefusalParam struct {
	// The refusal message generated by the model.
	Refusal string `json:"refusal"`
	// The type of the content part.
	Type ChatCompletionContentPartRefusalType `json:"type"`
}

type ChatCompletionContentPartRefusalType string

type ChatCompletionContentPartTextParam struct {
	// The text content.
	Text string `json:"text"`
	// The type of the content part.
	Type string `json:"type"`
}

type ChatCompletionContentPartTextType string

type ChatCompletionContentPartUserUnionParam struct {
	OfText       *ChatCompletionContentPartTextParam       `json:",omitzero,inline"`
	OfInputAudio *ChatCompletionContentPartInputAudioParam `json:",omitzero,inline"`
	OfImageURL   *ChatCompletionContentPartImageParam      `json:",omitzero,inline"`
	OfFile       *ChatCompletionContentPartFileParam       `json:",omitzero,inline"`
}

type ChatCompletionDeveloperMessageParam struct {
	// Content: The contents of the developer message.
	Content ContentUnion `json:"content"`
	// The role of the messages author, in this case `developer`.
	Role string `json:"role"`
	// An optional name for the participant. Provides the model information to
	// differentiate between participants of the same role.
	Name string `json:"name,omitempty"`
}

type ChatCompletionMessageParamUnion struct {
	OfDeveloper *ChatCompletionDeveloperMessageParam `json:",omitzero,inline"`
	OfSystem    *ChatCompletionSystemMessageParam    `json:",omitzero,inline"`
	OfUser      *ChatCompletionUserMessageParam      `json:",omitzero,inline"`
	OfAssistant *ChatCompletionAssistantMessageParam `json:",omitzero,inline"`
	OfTool      *ChatCompletionToolMessageParam      `json:",omitzero,inline"`
}

type ChatCompletionMessageToolCallFunctionParam struct {
	// The arguments to call the function with, as generated by the model in JSON
	// format. Note that the model does not always generate valid JSON, and may
	// hallucinate parameters not defined by your function schema. Validate the
	// arguments in your code before calling your function.
	Arguments string `json:"arguments"`
	// The name of the function to call.
	Name string `json:"name"`
}

type ChatCompletionMessageToolCallParam struct {
	// The ID of the tool call.
	ID *string `json:"id"`
	// The function that the model called.
	Function ChatCompletionMessageToolCallFunctionParam `json:"function"`
	// The type of the tool. Currently, only `function` is supported.
	Type ChatCompletionMessageToolCallType `json:"type,omitempty"`
}

type ChatCompletionMessageToolCallType string

type ChatCompletionModality string

type ChatCompletionNamedToolChoice struct {
	// Type is the type of the tool. Currently, only `function` is supported.
	Type ToolType `json:"type"`
	// Function specifies the function to call.
	Function ChatCompletionNamedToolChoiceFunction `json:"function"`
}

type ChatCompletionNamedToolChoiceFunction struct {
	// Name is the name of the function to call.
	Name string `json:"name"`
}

type ChatCompletionRequest struct {
	// Messages: A list of messages comprising the conversation so far.
	// Depending on the model you use, different message types (modalities) are supported,
	// like text, images, and audio.
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-messages
	Messages []ChatCompletionMessageParamUnion `json:"messages"`

	// Model: ID of the model to use
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-model
	Model string `json:"model"`

	// FrequencyPenalty: Number between -2.0 and 2.0
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-frequency_penalty
	FrequencyPenalty *float32 `json:"frequency_penalty,omitempty"` //nolint:tagliatelle //follow openai api

	// LogitBias Modify the likelihood of specified tokens appearing in the completion.
	// It must be a token id string (specified by their token ID in the tokenizer), not a word string.
	// incorrect: `"logit_bias":{"You": 6}`, correct: `"logit_bias":{"1639": 6}`
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-logit_bias
	LogitBias map[string]int `json:"logit_bias,omitempty"` //nolint:tagliatelle //follow openai api

	// LogProbs indicates whether to return log probabilities of the output tokens or not.
	// If true, returns the log probabilities of each output token returned in the content of message.
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-logprobs
	LogProbs *bool `json:"logprobs,omitempty"`

	// TopLogProbs is an integer between 0 and 5 specifying the number of most likely tokens to return at each
	// token position, each with an associated log probability.
	// logprobs must be set to true if this parameter is used.
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-top_logprobs
	TopLogProbs *int `json:"top_logprobs,omitempty"` //nolint:tagliatelle //follow openai api

	// MaxTokens The maximum number of tokens that can be generated in the chat completion.
	// This value can be used to control costs for text generated via API.
	// This value is now deprecated in favor of max_completion_tokens, and is not compatible with o1 series models.
	// refs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-max_tokens
	MaxTokens *int64 `json:"max_tokens,omitempty"` //nolint:tagliatelle //follow openai api

	// MaxCompletionTokens is an Optional integer
	// An upper bound for the number of tokens that can be generated for a completion, including visible output tokens and reasoning tokens.
	// refs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-max_completion_tokens
	MaxCompletionTokens *int64 `json:"max_completion_tokens,omitempty"` //nolint:tagliatelle //follow openai api

	// N: LLM Gateway does not support multiple completions.
	// The only accepted value is 1.
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-n
	N *int `json:"n,omitempty"`

	// PresencePenalty Positive values penalize new tokens based on whether they appear in the text so far,
	// increasing the model's likelihood to talk about new topics.
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-presence_penalty
	PresencePenalty *float32 `json:"presence_penalty,omitempty"` //nolint:tagliatelle //follow openai api

	// An object specifying the format that the model must output.
	//
	// Setting to `{ "type": "json_schema", "json_schema": {...} }` enables Structured
	// Outputs which ensures the model will match your supplied JSON schema. Learn more
	// in the
	// [Structured Outputs guide](https://platform.openai.com/docs/guides/structured-outputs).
	//
	// Setting to `{ "type": "json_object" }` enables the older JSON mode, which
	// ensures the message the model generates is valid JSON. Using `json_schema` is
	// preferred for models that support it.
	ResponseFormat *ChatCompletionResponseFormatUnion `json:"response_format,omitempty"` //nolint:tagliatelle //follow openai api

	// Seed: This feature is in Beta. If specified, our system will make a best effort to
	// sample deterministically, such that repeated requests with the same `seed` and
	// parameters should return the same result. Determinism is not guaranteed, and you
	// should refer to the `system_fingerprint` response parameter to monitor changes
	// in the backend.
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-seed
	Seed *int `json:"seed,omitempty"`

	// Constrains effort on reasoning for
	// [reasoning models](https://platform.openai.com/docs/guides/reasoning). Currently
	// supported values are `minimal`, `low`, `medium`, and `high`. Reducing reasoning
	// effort can result in faster responses and fewer tokens used on reasoning in a
	// response.
	//
	// Any of "minimal", "low", "medium", "high".
	ReasoningEffort openai.ReasoningEffort `json:"reasoning_effort,omitzero"`

	// ServiceTier:string or null - Defaults to auto
	// Specifies the processing type used for serving the request.
	// If set to 'auto', then the request will be processed with the service tier configured in the Project settings. Unless otherwise configured, the Project will use 'default'.
	// If set to 'default', then the request will be processed with the standard pricing and performance for the selected model.
	// If set to 'flex' or 'priority', then the request will be processed with the corresponding service tier.
	// When the service_tier parameter is set, the response body will include the service_tier value based on the processing mode actually used to serve the request.
	// This response value may be different from the value set in the parameter.
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-service_tier
	// Any of "auto", "default", "flex", "scale", "priority".
	ServiceTier openai.ChatCompletionNewParamsServiceTier `json:"service_tier,omitzero"`

	// Constrains the verbosity of the model's response. Lower values will result in
	// more concise responses, while higher values will result in more verbose
	// responses. Currently supported values are `low`, `medium`, and `high`.
	//
	// Any of "low", "medium", "high".
	Verbosity openai.ChatCompletionNewParamsVerbosity `json:"verbosity,omitzero"`

	// Stop string / array / null Defaults to null
	// Up to 4 sequences where the API will stop generating further tokens.
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-stop
	Stop openai.ChatCompletionNewParamsStopUnion `json:"stop,omitzero"`

	// Stream: If set, partial message deltas will be sent, like in ChatGPT.
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-stream
	Stream bool `json:"stream,omitempty"`

	// StreamOptions for streaming response. Only set this when you set stream: true.
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-stream_options
	StreamOptions *StreamOptions `json:"stream_options,omitempty"` //nolint:tagliatelle //follow openai api

	// Temperature What sampling temperature to use, between 0 and 2.
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-temperature
	Temperature *float64 `json:"temperature,omitempty"`

	// TopP An alternative to sampling with temperature, called nucleus sampling,
	// where the model considers the results of the tokens with top_p probability mass.
	// So 0.1 means only the tokens comprising the top 10% probability mass are considered.
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-top_p
	TopP *float64 `json:"top_p,omitempty"` //nolint:tagliatelle //follow openai api

	// Tools provide a list of tool definitions to be used by the LLM.
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-response_format
	Tools []Tool `json:"tools,omitempty"`

	// ToolChoice controls which (if any) tool is called by the model.
	// `none` means the model will not call any tool and instead generates a message.
	// `auto` means the model can pick between generating a message or calling one or more tools.
	// `required` means the model must call one or more tools.
	// Specifying a particular tool via `{"type": "function", "function": {"name": "my_function"}}` forces the model to call that tool.
	// `none` is the default when no tools are present. `auto` is the default if tools are present.
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-tool_choice
	ToolChoice *ChatCompletionToolChoiceUnion `json:"tool_choice,omitempty"` //nolint:tagliatelle //follow openai api

	// ParallelToolCalls enables multiple tools to be returned by the model.
	// Docs: https://platform.openai.com/docs/guides/function-calling/parallel-function-calling
	ParallelToolCalls *bool `json:"parallel_tool_calls,omitempty"` //nolint:tagliatelle //follow openai api

	// User: A unique identifier representing your end-user, which can help OpenAI to monitor and detect abuse.
	// Docs: https://platform.openai.com/docs/api-reference/chat/create#chat-create-user
	User string `json:"user,omitempty"`

	// Modalities specifies the output types that you would like the model to generate. Most models are capable of generating text, which is the default. The gpt-4o-audio-preview model can also generate audio.
	Modalities []ChatCompletionModality `json:"modalities,omitempty"`

	// Audio specifies parameters for audio output. Required when audio output is requested with modalities: ["audio"].
	Audio *ChatCompletionAudioParam `json:"audio,omitempty"`

	// PredictionContent provides configuration for a Predicted Output, which can greatly improve response times when large parts of the model response are known ahead of time.
	PredictionContent *PredictionContent `json:"prediction,omitempty"`

	// WebSearchOptions configures web search tool for models that support it.
	// This tool searches the web for relevant results to use in a response.
	// Docs: https://platform.openai.com/docs/guides/tools-web-search?api-mode=chat
	WebSearchOptions *WebSearchOptions `json:"web_search_options,omitempty"` //nolint:tagliatelle //follow openai api

	// GCPVertexAIVendorFields configures the GCP VertexAI specific fields during schema translation.
	*GCPVertexAIVendorFields `json:",inline,omitempty"`

	// AnthropicVendorFields configures the Anthropic specific fields during schema translation.
	*AnthropicVendorFields `json:",inline,omitempty"`

	// GuidedChoice: The output will be exactly one of the choices.
	GuidedChoice []string `json:"guided_choice,omitzero"`

	// GuidedRegex: The output will follow the regex pattern.
	GuidedRegex string `json:"guided_regex,omitzero"`

	// GuidedJSON: The output will follow the JSON schema.
	GuidedJSON json.RawMessage `json:"guided_json,omitzero"`
}

type ChatCompletionResponse struct {
	// ID is a unique identifier for the chat completion.
	ID string `json:"id,omitempty"`
	// Choices are described in the OpenAI API documentation:
	// https://platform.openai.com/docs/api-reference/chat/object#chat/object-choices
	Choices []ChatCompletionResponseChoice `json:"choices,omitempty"`

	// Created is the Unix timestamp (in seconds) of when the chat completion was created.
	Created JSONUNIXTime `json:"created,omitzero"`

	// Model is the model used for the chat completion.
	Model string `json:"model,omitempty"`

	// ServiceTier is the service tier used for the completion.
	ServiceTier string `json:"service_tier,omitempty"`

	// SystemFingerprint represents the backend configuration that the model runs with.
	SystemFingerprint string `json:"system_fingerprint,omitempty"`

	// Object is always "chat.completion" for completions.
	// https://platform.openai.com/docs/api-reference/chat/object#chat/object-object
	Object string `json:"object,omitempty"`

	// Usage is described in the OpenAI API documentation:
	// https://platform.openai.com/docs/api-reference/chat/object#chat/object-usage
	Usage Usage `json:"usage,omitzero"`

	// Obfuscation are random characters that normalize payload sizes as a
	// mitigation to certain side-channel attacks.
	// https://platform.openai.com/docs/api-reference/responses/get#responses_get-include_obfuscation
	Obfuscation string `json:"obfuscation,omitempty"`
}

type ChatCompletionResponseChoice struct {
	// The reason the model stopped generating tokens. This will be `stop` if the model
	// hit a natural stop point or a provided stop sequence, `length` if the maximum
	// number of tokens specified in the request was reached, `content_filter` if
	// content was omitted due to a flag from our content filters, `tool_calls` if the
	// model called a tool, or `function_call` (deprecated) if the model called a
	// function.
	FinishReason ChatCompletionChoicesFinishReason `json:"finish_reason"`
	// The index of the choice in the list of choices.
	Index int64 `json:"index"`
	// Log probability information for the choice.
	Logprobs ChatCompletionChoicesLogprobs `json:"logprobs,omitzero"`
	// Message is described in the OpenAI API documentation:
	// https://platform.openai.com/docs/api-reference/chat/object#chat/object-choices
	Message ChatCompletionResponseChoiceMessage `json:"message,omitempty"`
}

type ChatCompletionResponseChoiceMessage struct {
	// The contents of the message.
	Content *string `json:"content,omitempty"`

	// The role of the author of this message.
	Role string `json:"role,omitempty"`

	// The tool calls generated by the model, such as function calls.
	ToolCalls []ChatCompletionMessageToolCallParam `json:"tool_calls,omitempty"`

	// Annotations for the message, when applicable, as when using the web search tool.
	Annotations *[]Annotation `json:"annotations,omitempty"`

	// Audio is the audio response generated by the model, if applicable.
	Audio *ChatCompletionResponseChoiceMessageAudio `json:"audio,omitempty"`

	// ReasoningContent is used to hold any non-standard fields from the backend which supports reasoning,
	// like "reasoningContent" from AWS Bedrock.
	ReasoningContent *ReasoningContentUnion `json:"reasoning_content,omitempty"`

	// SafetyRatings contains safety ratings copied from the GCP Vertex AI response as-is.
	// List of ratings for the safety of a response candidate. There is at most one rating per category.
	// https://cloud.google.com/vertex-ai/generative-ai/docs/reference/rest/v1/GenerateContentResponse#SafetyRating
	SafetyRatings []*genai.SafetyRating `json:"safety_ratings,omitempty"`
}

type ChatCompletionResponseChoiceMessageAudio struct {
	Data       string `json:"data"`
	ExpiresAt  int64  `json:"expires_at"`
	ID         string `json:"id"`
	Transcript string `json:"transcript"`
}

type ChatCompletionResponseChunk struct {
	// ID is a unique identifier for the chat completion chunk.
	ID string `json:"id,omitempty"`
	// Choices are described in the OpenAI API documentation:
	// https://platform.openai.com/docs/api-reference/chat/streaming#chat/streaming-choices
	Choices []ChatCompletionResponseChunkChoice `json:"choices,omitempty"`

	// Created is the Unix timestamp (in seconds) of when the chat completion was created.
	Created JSONUNIXTime `json:"created,omitzero"`

	// Model is the model used for the chat completion.
	Model string `json:"model,omitempty"`

	// ServiceTier is the service tier used for the completion.
	ServiceTier string `json:"service_tier,omitempty"`

	// SystemFingerprint represents the backend configuration that the model runs with.
	SystemFingerprint string `json:"system_fingerprint,omitempty"`

	// Object is always "chat.completion.chunk" for completions.
	// https://platform.openai.com/docs/api-reference/chat/streaming#chat/streaming-object
	Object string `json:"object,omitempty"`

	// Usage is described in the OpenAI API documentation:
	// https://platform.openai.com/docs/api-reference/chat/streaming#chat/streaming-usage
	Usage *Usage `json:"usage,omitempty"`

	// Obfuscation are random characters that normalize payload sizes as a
	// mitigation to certain side-channel attacks.
	// https://platform.openai.com/docs/api-reference/responses/get#responses_get-include_obfuscation
	Obfuscation string `json:"obfuscation,omitempty"`
}

type ChatCompletionResponseChunkChoice struct {
	Index        int64                                   `json:"index"`
	Delta        *ChatCompletionResponseChunkChoiceDelta `json:"delta,omitzero"`
	Logprobs     *ChatCompletionChoicesLogprobs          `json:"logprobs,omitzero"`
	FinishReason ChatCompletionChoicesFinishReason       `json:"finish_reason,omitempty"`
}

type ChatCompletionResponseChunkChoiceDelta struct {
	Content          *string                                  `json:"content,omitempty"`
	Role             string                                   `json:"role,omitempty"`
	ToolCalls        []ChatCompletionChunkChoiceDeltaToolCall `json:"tool_calls,omitempty"`
	Annotations      *[]Annotation                            `json:"annotations,omitempty"`
	ReasoningContent *AWSBedrockStreamReasoningContent        `json:"reasoning_content,omitempty"`
}

type ChatCompletionResponseFormatJSONObjectParam struct {
	// The type of response format being defined. Always `json_object`.
	Type ChatCompletionResponseFormatType `json:"type"`
}

type ChatCompletionResponseFormatJSONSchema struct {
	// Structured Outputs configuration options, including a JSON Schema.
	JSONSchema ChatCompletionResponseFormatJSONSchemaJSONSchema `json:"json_schema,omitzero"`
	// The type of response format being defined. Always `json_schema`.
	//
	// This field can be elided, and will marshal its zero value as "json_schema".
	Type ChatCompletionResponseFormatType `json:"type"`
}

type ChatCompletionResponseFormatJSONSchemaJSONSchema struct {
	// The name of the response format. Must be a-z, A-Z, 0-9, or contain underscores
	// and dashes, with a maximum length of 64.
	Name string `json:"name"`
	// A description of what the response format is for, used by the model to determine
	// how to respond in the format.
	Description string `json:"description,omitempty"`
	// The schema for the response format, described as a JSON Schema object. Learn how
	// to build JSON schemas [here](https://json-schema.org/).
	Schema json.RawMessage `json:"schema"`
	// Whether to enable strict schema adherence when generating the output. If set to
	// true, the model will always follow the exact schema defined in the `schema`
	// field. Only a subset of JSON Schema is supported when `strict` is `true`. To
	// learn more, read the
	// [Structured Outputs guide](https://platform.openai.com/docs/guides/structured-outputs).
	Strict bool `json:"strict,omitempty"`
}

type ChatCompletionResponseFormatTextParam struct {
	// The type of response format being defined. Always `text`.
	Type ChatCompletionResponseFormatType `json:"type"`
}

type ChatCompletionResponseFormatType string

type ChatCompletionResponseFormatUnion struct {
	OfText       *ChatCompletionResponseFormatTextParam       `json:",omitempty,inline"`
	OfJSONSchema *ChatCompletionResponseFormatJSONSchema      `json:",omitempty,inline"`
	OfJSONObject *ChatCompletionResponseFormatJSONObjectParam `json:",omitempty,inline"`
}

type ChatCompletionSystemMessageParam struct {
	// Content: The contents of the system message.
	Content ContentUnion `json:"content"`
	// The role of the messages author, in this case `system`.
	Role string `json:"role"`
	// An optional name for the participant. Provides the model information to
	// differentiate between participants of the same role.
	Name string `json:"name,omitempty"`
}

type ChatCompletionTokenLogprob struct {
	// The token.
	Token string `json:"token"`
	// A list of integers representing the UTF-8 bytes representation of the token.
	// Useful in instances where characters are represented by multiple tokens and
	// their byte representations must be combined to generate the correct text
	// representation. Can be `null` if there is no bytes representation for the token.
	Bytes []int64 `json:"bytes,omitempty"`
	// The log probability of this token, if it is within the top 20 most likely
	// tokens. Otherwise, the value `-9999.0` is used to signify that the token is very
	// unlikely.
	Logprob float64 `json:"logprob"`
	// List of the most likely tokens and their log probability, at this token
	// position. In rare cases, there may be fewer than the number of requested
	// `top_logprobs` returned.
	TopLogprobs []ChatCompletionTokenLogprobTopLogprob `json:"top_logprobs"`
}

type ChatCompletionTokenLogprobTopLogprob struct {
	// The token.
	Token string `json:"token"`
	// A list of integers representing the UTF-8 bytes representation of the token.
	// Useful in instances where characters are represented by multiple tokens and
	// their byte representations must be combined to generate the correct text
	// representation. Can be `null` if there is no bytes representation for the token.
	Bytes []int64 `json:"bytes,omitempty"`
	// The log probability of this token, if it is within the top 20 most likely
	// tokens. Otherwise, the value `-9999.0` is used to signify that the token is very
	// unlikely.
	Logprob float64 `json:"logprob"`
}

type ChatCompletionToolChoiceUnion struct {
	Value any
}

type ChatCompletionToolMessageParam struct {
	// Content: The contents of the tool message.
	Content ContentUnion `json:"content"`
	// The role of the messages author, in this case `tool`.
	Role string `json:"role"`
	// Tool call that this message is responding to.
	ToolCallID string `json:"tool_call_id"`
}

type ChatCompletionUserMessageParam struct {
	// The contents of the user message.
	Content StringOrUserRoleContentUnion `json:"content"`
	// The role of the messages author, in this case `user`.
	Role string `json:"role"`
	// An optional name for the participant. Provides the model information to
	// differentiate between participants of the same role.
	Name string `json:"name,omitempty"`
}

type CompletionChoice struct {
	// The generated text completion.
	Text string `json:"text"`

	// The index of the choice in the list of choices.
	Index *int `json:"index,omitempty"`

	// Log probability information for the choice.
	Logprobs *CompletionLogprobs `json:"logprobs,omitzero"`

	// The reason the model stopped generating tokens.
	// This will be "stop" if the model hit a natural stop point or a provided stop
	// sequence, "length" if the maximum number of tokens specified in the request
	// was reached, or "content_filter" if content was omitted due to a flag from
	// our content filters.
	FinishReason string `json:"finish_reason,omitzero"`
}

type CompletionLogprobs struct {
	// The tokens chosen by the model.
	Tokens []string `json:"tokens,omitzero"`

	// The log probability of each token.
	TokenLogprobs []float64 `json:"token_logprobs,omitzero"`

	// A list of the top log probabilities for each token position.
	TopLogprobs []map[string]float64 `json:"top_logprobs,omitzero"`

	// The character offset from the start of the returned text for each token.
	TextOffset []int `json:"text_offset,omitzero"`
}

type CompletionRequest struct {
	// ID of the model to use. You can use the List models API to see all of your
	// available models, or see the Model overview for descriptions of them.
	Model string `json:"model"`

	// The prompt(s) to generate completions for, encoded as a string, array of
	// strings, array of tokens, or array of token arrays.
	// Note that <|endoftext|> is the document separator that the model sees during
	// training, so if a prompt is not specified the model will generate as if from
	// the beginning of a new document.
	Prompt PromptUnion `json:"prompt"`

	// Generates `best_of` completions server-side and returns the "best" (the one
	// with the highest log probability per token). Results cannot be streamed.
	// When used with `n`, `best_of` controls the number of candidate completions
	// and `n` specifies how many to return – `best_of` must be greater than `n`.
	// Minimum: 0, Maximum: 20, Default: 1
	BestOf *int `json:"best_of,omitempty"`

	// Echo back the prompt in addition to the completion.
	// Default: false
	Echo bool `json:"echo,omitzero"`

	// Number between -2.0 and 2.0. Positive values penalize new tokens based on
	// their existing frequency in the text so far, decreasing the model's
	// likelihood to repeat the same line verbatim.
	// Minimum: -2, Maximum: 2, Default: 0
	FrequencyPenalty *float64 `json:"frequency_penalty,omitempty"`

	// Modify the likelihood of specified tokens appearing in the completion.
	// Accepts a JSON object that maps tokens (specified by their token ID in the
	// GPT tokenizer) to an associated bias value from -100 to 100.
	LogitBias map[string]int `json:"logit_bias,omitempty"`

	// Include the log probabilities on the `logprobs` most likely output tokens,
	// as well the chosen tokens. The maximum value for `logprobs` is 5.
	// Minimum: 0, Maximum: 5, Default: null
	Logprobs *int `json:"logprobs,omitempty"`

	// The maximum number of tokens that can be generated in the completion.
	// The token count of your prompt plus `max_tokens` cannot exceed the model's
	// context length.
	// Minimum: 0, Default: 16
	MaxTokens *int `json:"max_tokens,omitempty"`

	// How many completions to generate for each prompt.
	// Minimum: 1, Maximum: 128, Default: 1
	N *int `json:"n,omitempty"`

	// Number between -2.0 and 2.0. Positive values penalize new tokens based on
	// whether they appear in the text so far, increasing the model's likelihood
	// to talk about new topics.
	// Minimum: -2, Maximum: 2, Default: 0
	PresencePenalty *float64 `json:"presence_penalty,omitempty"`

	// If specified, our system will make a best effort to sample deterministically,
	// such that repeated requests with the same `seed` and parameters should return
	// the same result. Determinism is not guaranteed.
	Seed *int64 `json:"seed,omitempty"`

	// Up to 4 sequences where the API will stop generating further tokens.
	Stop any `json:"stop,omitempty"`

	// Whether to stream back partial progress. If set, tokens will be sent as
	// data-only server-sent events as they become available, with the stream
	// terminated by a `data: [DONE]` message.
	// Default: false
	Stream bool `json:"stream,omitzero"`

	// Options for streaming response. Only set this when you set stream: true.
	StreamOptions *StreamOptions `json:"stream_options,omitempty"`

	// The suffix that comes after a completion of inserted text.
	// This parameter is only supported for `gpt-3.5-turbo-instruct`.
	// Default: null
	Suffix string `json:"suffix,omitzero"`

	// What sampling temperature to use, between 0 and 2. Higher values like 0.8
	// will make the output more random, while lower values like 0.2 will make it
	// more focused and deterministic.
	// Minimum: 0, Maximum: 2, Default: 1
	Temperature *float64 `json:"temperature,omitempty"`

	// An alternative to sampling with temperature, called nucleus sampling, where
	// the model considers the results of the tokens with top_p probability mass.
	// So 0.1 means only the tokens comprising the top 10% probability mass are
	// considered. We generally recommend altering this or temperature but not both.
	// Minimum: 0, Maximum: 1, Default: 1
	TopP *float64 `json:"top_p,omitempty"`

	// A unique identifier representing your end-user, which can help OpenAI to
	// monitor and detect abuse.
	User string `json:"user,omitzero"`
}

type CompletionResponse struct {
	// A unique identifier for the completion.
	ID string `json:"id"`

	// The object type, which is always "text_completion".
	Object string `json:"object"`

	// The Unix timestamp (in seconds) of when the completion was created.
	Created JSONUNIXTime `json:"created,omitzero"`

	// The model used for completion.
	Model string `json:"model"`

	// This fingerprint represents the backend configuration that the model runs with.
	SystemFingerprint string `json:"system_fingerprint,omitzero"`

	// The list of completion choices generated by the model.
	Choices []CompletionChoice `json:"choices"`

	// Usage statistics for the completion request.
	Usage *Usage `json:"usage,omitzero"`
}

type CompletionTokensDetails struct {
	// Text input tokens present in the prompt.
	TextTokens int `json:"text_tokens,omitzero"`

	// When using Predicted Outputs, the number of tokens in the prediction that appeared in the completion.
	AcceptedPredictionTokens int `json:"accepted_prediction_tokens,omitzero"`
	// Audio input tokens generated by the model.
	AudioTokens int `json:"audio_tokens,omitzero"`
	// Tokens generated by the model for reasoning.
	ReasoningTokens int `json:"reasoning_tokens,omitzero"`
	// When using Predicted Outputs, the number of tokens in the prediction that did not appear in the completion.
	// However, like reasoning tokens, these tokens are still counted in the total completion tokens
	// for purposes of billing, output, and context window limits.
	RejectedPredictionTokens int `json:"rejected_prediction_tokens,omitzero"`
}

type ContentUnion struct {
	Value any
}

type Embedding struct {
	// Object: The object type, which is always "embedding".
	Object string `json:"object"`

	// Embedding: The embedding vector, which can be a list of floats or a string.
	// The length of vector depends on the model as listed in the embedding guide.
	Embedding EmbeddingUnion `json:"embedding"`

	// Index: The index of the embedding in the list of embeddings.
	Index int `json:"index"`
}

type EmbeddingRequest struct {
	// Input: Input text to embed, encoded as a string or array of tokens.
	// To embed multiple inputs in a single request, pass an array of strings or array of token arrays.
	// The input must not exceed the max input tokens for the model (8192 tokens for text-embedding-ada-002),
	// cannot be an empty string, and any array must be 2048 dimensions or less.
	// Docs: https://platform.openai.com/docs/api-reference/embeddings/create#embeddings-create-input
	Input EmbeddingRequestInput `json:"input"`

	// Model: ID of the model to use.
	// Docs: https://platform.openai.com/docs/api-reference/embeddings/create#embeddings-create-model
	Model string `json:"model"`

	// EncodingFormat: The format to return the embeddings in. Can be either float or base64.
	// Docs: https://platform.openai.com/docs/api-reference/embeddings/create#embeddings-create-encoding_format
	EncodingFormat *string `json:"encoding_format,omitempty"` //nolint:tagliatelle //follow openai api

	// Dimensions: The number of dimensions the resulting output embeddings should have.
	// Only supported in text-embedding-3 and later models.
	// Docs: https://platform.openai.com/docs/api-reference/embeddings/create#embeddings-create-dimensions
	Dimensions *int `json:"dimensions,omitempty"`

	// User: A unique identifier representing your end-user, which can help OpenAI to monitor and detect abuse.
	// Docs: https://platform.openai.com/docs/api-reference/embeddings/create#embeddings-create-user
	User *string `json:"user,omitempty"`
}

type EmbeddingRequestInput struct {
	Value any
}

type EmbeddingResponse struct {
	// Object: The object type, which is always "list".
	// https://platform.openai.com/docs/api-reference/embeddings/object#embeddings/object-object
	Object string `json:"object"`

	// Data: The list of embeddings generated by the model.
	// https://platform.openai.com/docs/api-reference/embeddings/object#embeddings/object-data
	Data []Embedding `json:"data"`

	// Model: The name of the model used to generate the embedding.
	// https://platform.openai.com/docs/api-reference/embeddings/object#embeddings/object-model
	Model string `json:"model"`

	// Usage: The usage information for the request.
	// https://platform.openai.com/docs/api-reference/embeddings/object#embeddings/object-usage
	Usage EmbeddingUsage `json:"usage"`
}

type EmbeddingUnion struct {
	Value any
}

type EmbeddingUsage struct {
	// PromptTokens: The number of tokens used by the prompt.
	PromptTokens int `json:"prompt_tokens"` //nolint:tagliatelle //follow openai api

	// TotalTokens: The total number of tokens used by the request.
	TotalTokens int `json:"total_tokens"` //nolint:tagliatelle //follow openai api
}

type Error struct {
	// The unique ID of the server event.
	EventID *string `json:"event_id,omitempty"`
	// The event type, must be error.
	Type string `json:"type"`
	// Details of the error.
	Error ErrorType `json:"error"`
}

type ErrorType struct {
	// The type of error (e.g., "invalid_request_error", "server_error").
	Type string `json:"type"`
	// Error code, if any.
	Code *string `json:"code,omitempty"`
	// A human-readable error message.
	Message string `json:"message,omitempty"`
	// Parameter related to the error, if any.
	Param *string `json:"param,omitempty"`
	// The event_id of the client event that caused the error, if applicable.
	EventID *string `json:"event_id,omitempty"`
}

type FunctionDefinition struct {
	// Name is the name of the function to be called. Must be a-z, A-Z, 0-9, or contain underscores and dashes, with a maximum length of 64.
	Name string `json:"name"`
	// Description is a description of what the function does, used by the model to choose when and how to call the function.
	Description string `json:"description,omitempty"`
	Strict      bool   `json:"strict,omitempty"`
	// Parameters is an object describing the function.
	// You can pass json.RawMessage to describe the schema,
	// or you can pass in a struct which serializes to the proper JSON schema.
	// The jsonschema package is provided for convenience, but you should
	// consider another specialized library if you require more complex schemas.
	Parameters any `json:"parameters"`
}

type GCPVertexAIGenerationConfig struct {
	// ThinkingConfig holds Gemini thinking configuration options.
	//
	// https://cloud.google.com/vertex-ai/docs/reference/rest/v1/GenerationConfig#ThinkingConfig
	ThinkingConfig *genai.ThinkingConfig `json:"thinkingConfig,omitzero"`
}

type GCPVertexAIVendorFields struct {
	// GenerationConfig holds Gemini generation configuration options.
	// Currently only a subset of the options are supported.
	//
	// https://cloud.google.com/vertex-ai/docs/reference/rest/v1/GenerationConfig
	GenerationConfig *GCPVertexAIGenerationConfig `json:"generationConfig,omitzero"`

	// SafetySettings: Safety settings in the request to block unsafe content in the response.
	//
	// https://cloud.google.com/vertex-ai/docs/reference/rest/v1/SafetySetting
	SafetySettings []*genai.SafetySetting `json:"safetySettings,omitzero"`
}

type JSONUNIXTime time.Time

type LogProb struct {
	Token   string  `json:"token"`
	LogProb float64 `json:"logprob"`
	Bytes   []byte  `json:"bytes,omitempty"` // Omitting the field if it is null.
	// TopLogProbs is a list of the most likely tokens and their log probability, at this token position.
	// In rare cases, there may be fewer than the number of requested top_logprobs returned.
	TopLogProbs []TopLogProbs `json:"top_logprobs"` //nolint:tagliatelle //follow openai api
}

type LogProbs struct {
	// Content is a list of message content tokens with log probability information.
	Content []LogProb `json:"content"`
}

type Model struct {
	// ID is the model identifier, which can be referenced in the API endpoints.
	ID string `json:"id"`
	// Created is the Unix timestamp (in seconds) when the model was created.
	Created JSONUNIXTime `json:"created"`
	// Object is the object type, which is always "model".
	Object string `json:"object"`
	// OwnedBy is the organization that owns the model.
	OwnedBy string `json:"owned_by"`
}

type ModelList struct {
	// Data is a list of models.
	Data []Model `json:"data"`
	// Object is the object type, which is always "list".
	Object string `json:"object"`
}

type PredictionContent struct {
	// Type: The type of the predicted content you want to provide.
	// This type is currently always "content".
	Type PredictionContentType `json:"type"`

	// Content: The content that should be matched when generating a model response.
	// If generated tokens would match this content, the entire model response
	// can be returned much more quickly.
	Content ContentUnion `json:"content"`
}

type PredictionContentType string

type PromptTokensDetails struct {
	// Text input tokens present in the prompt.
	TextTokens int `json:"text_tokens,omitzero"`

	// Audio input tokens present in the prompt.
	AudioTokens int `json:"audio_tokens,omitzero"`
	// Cached tokens present in the prompt.
	CachedTokens int `json:"cached_tokens,omitzero"`
}

type PromptUnion struct {
	Value interface{}
}

type ReasoningContentUnion struct {
	Value any
}

type StreamOptions struct {
	// If set, an additional chunk will be streamed before the data: [DONE] message.
	// The usage field on this chunk shows the token usage statistics for the entire request,
	// and the choices field will always be an empty array.
	// All other chunks will also include a usage field, but with a null value.
	IncludeUsage bool `json:"include_usage,omitempty"` //nolint:tagliatelle //follow openai api
}

type StringOrAssistantRoleContentUnion struct {
	Value any
}

type StringOrUserRoleContentUnion struct {
	Value any
}

type Tool struct {
	Type     ToolType            `json:"type"`
	Function *FunctionDefinition `json:"function,omitempty"`
}

type ToolChoiceType string

type ToolType string

type TopLogProbs struct {
	Token   string  `json:"token"`
	LogProb float64 `json:"logprob"`
	Bytes   []byte  `json:"bytes,omitempty"`
}

type URLCitation struct {
	// EndIndex is the index of the last character of the URL citation in the message.
	EndIndex int `json:"end_index"`
	// StartIndex is the index of the first character of the URL citation in the message.
	StartIndex int `json:"start_index"`
	// URL is the URL of the web resource.
	URL string `json:"url"`
	// Title is the title of the web resource.
	Title string `json:"title"`
}

type Usage struct {
	// Number of tokens in the prompt.
	PromptTokens int `json:"prompt_tokens,omitempty"` //nolint:tagliatelle //follow openai api

	// Number of tokens in the generated completion.
	CompletionTokens int `json:"completion_tokens,omitempty"` //nolint:tagliatelle //follow openai api

	// Total number of tokens used in the request (prompt + completion).
	TotalTokens int `json:"total_tokens,omitempty"` //nolint:tagliatelle //follow openai api

	// CompletionTokensDetails: Breakdown of tokens used in a completion.
	// Only populated for /v1/chat/completions endpoint, not for /v1/completions.
	CompletionTokensDetails *CompletionTokensDetails `json:"completion_tokens_details,omitempty"` //nolint:tagliatelle //follow openai api

	// PromptTokensDetails: Breakdown of tokens used in the prompt.
	// Only populated for /v1/chat/completions endpoint, not for /v1/completions.
	PromptTokensDetails *PromptTokensDetails `json:"prompt_tokens_details,omitempty"` //nolint:tagliatelle //follow openai api
}

type WebSearchContextSize string

type WebSearchLocation struct {
	// City is the approximate city name.
	City string `json:"city,omitempty"`
	// Region is the approximate region or state.
	Region string `json:"region,omitempty"`
	// Country is the approximate country.
	Country string `json:"country,omitempty"`
}

type WebSearchOptions struct {
	// UserLocation provides approximate location parameters for the search.
	UserLocation *WebSearchUserLocation `json:"user_location,omitempty"` //nolint:tagliatelle //follow openai api
	// SearchContextSize controls how much context to include from search results.
	SearchContextSize WebSearchContextSize `json:"search_context_size,omitempty"` //nolint:tagliatelle //follow openai api
}

type WebSearchUserLocation struct {
	// Type is the type of location approximation. Always "approximate".
	Type string `json:"type"`
	// Approximate contains the approximate location details.
	Approximate WebSearchLocation `json:"approximate"`
}