openai

type AssistantMessage[T string | []AssistantMessageContentPart] struct {

	// The content of the message.
	Content T

	// An optional name for the participant.
	// Provides the model information to differentiate between participants of the same role.
	Name string

	// The refusal message generated by the model, if any.
	Refusal string

	// The tool calls generated by the model, such as function calls.
	ToolCalls []ToolCall

	// Data about a previous audio response from the model.
	Audio *AudioRef
}

type AssistantMessageContentPart interface {
	ContentPart
	// contains filtered or unexported methods
}

type Audio struct {

	// The raw audio data.
	Data []byte `json:"data"`

	// The format of the audio data, such as "wav" or "mp3".
	// The format must be a valid audio format supported by the model.
	Format string `json:"format"`
}

type AudioContentPart struct {
	// The audio information.
	Audio Audio
}

type AudioOutput struct {

	// Unique identifier for this audio response.
	Id string `json:"id"`

	// The time at which this audio content will no longer be accessible on the server for use in multi-turn conversations.
	ExpiresAt time.Time `json:"expires_at"`

	// The raw audio data, in the format specified in the request.
	Data []byte `json:"data"`

	// Transcript of the audio generated by the model.
	Transcript string `json:"transcript"`
}

type AudioParameters struct {

	// The voice the model should use for audio output, such as "ash" or "ballad".
	// See the model's documentation for a list of all supported voices.
	Voice string `json:"voice"`

	// The format of the audio data, such as "wav" or "mp3".
	// See the model's documentation for a list of all supported formats.
	Format string `json:"format"`
}

type AudioRef struct {

	// Unique identifier for a previous audio response from the model.
	Id string `json:"id"`
}

type ChatModel struct {
	// contains filtered or unexported fields
}

type ChatModelInput struct {

	// The name of the model to use for the chat.
	//
	// Must be the exact string expected by the model provider.
	// For example, "gpt-3.5-turbo".
	Model string `json:"model"`

	// The list of messages to send to the chat model.
	Messages []RequestMessage `json:"messages"`

	// Output types that you want the model to generate.
	// Text modality is implied if no modalities are specified.
	Modalities []Modality `json:"modalities,omitempty"`

	// Parameters for audio output.
	// Required when audio output is requested in the modalities field.
	Audio *AudioParameters `json:"audio,omitempty"`

	// 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.
	FrequencyPenalty float64 `json:"frequency_penalty,omitempty"`

	// Modifies the likelihood of specified tokens appearing in the completion.
	//
	// Accepts an object that maps tokens (specified by their token ID in the tokenizer) to an associated bias value from -100 to 100.
	// Mathematically, the bias is added to the logits generated by the model prior to sampling. The exact effect will vary per model,
	// but values between -1 and 1 should decrease or increase likelihood of selection; values like -100 or 100 should result in a ban
	// or exclusive selection of the relevant token.
	LogitBias map[string]float64 `json:"logit_bias,omitempty"`

	// 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.
	Logprobs bool `json:"logprobs,omitempty"`

	// An integer between 0 and 20 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.
	TopLogprobs int `json:"top_logprobs,omitempty"`

	// The maximum number of tokens to generate in the chat completion.
	//
	// The default (0) is equivalent to 4096.
	//
	// Deprecated: Use the MaxCompletionTokens parameter instead, unless the model specifically requires passing "max_tokens".
	MaxTokens int `json:"max_tokens,omitempty"`

	// The maximum number of tokens to generate in the chat completion.
	//
	// The default (0) is equivalent to 4096.
	MaxCompletionTokens int `json:"max_completion_tokens,omitempty"`

	// The number of completions to generate for each prompt.
	//
	// The default (0) is equivalent to 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.
	PresencePenalty float64 `json:"presence_penalty,omitempty"`

	// Specifies the requested format for the response.
	//  - ResponseFormatText requests a plain text string.
	//  - ResponseFormatJson requests a JSON object.
	//  - ResponseFormatJsonSchema requests a JSON object that conforms to the provided JSON schema.
	//
	// The default is ResponseFormatText.
	ResponseFormat ResponseFormat `json:"response_format"`

	// If specified, the model 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 use the SystemFingerprint response
	// parameter to monitor changes in the backend.
	//
	// The default (0) is equivalent to a random seed.
	Seed int `json:"seed,omitempty"`

	// Specifies the latency tier to use for processing the request.
	// This is relevant for customers subscribed to the scale tier service of the model hosting platform.
	//
	// - If set to 'ServiceTierAuto', and the Project is Scale tier enabled, the system will utilize scale tier credits until they are exhausted.
	// - If set to 'ServiceTierAuto', and the Project is not Scale tier enabled, the request will be processed using the default service tier with a lower uptime SLA and no latency guarantee.
	// - If set to 'ServiceTierDefault', the request will be processed using the default service tier with a lower uptime SLA and no latency guarantee.
	// - When not set, the default behavior is 'ServiceTierAuto'.
	//
	// When this parameter is set, the response `serviceTier` property will indicate the service tier utilized.
	ServiceTier ServiceTier `json:"service_tier,omitempty"`

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

	// A number between 0.0 and 2.0 that controls the sampling temperature.
	//
	// Higher values like 0.8 will make the output more random, while lower values
	// like 0.2 will make it more focused and deterministic.
	//
	// We generally recommend altering this or TopP but not both.
	//
	// The default value is 1.0.
	Temperature float64 `json:"temperature"`

	// An alternative to sampling with temperature, called nucleus sampling, where the model
	// considers the results of the tokens with TopP probability mass.
	//
	// For example, 0.1 means only the tokens comprising the top 10% probability mass are considered.
	//
	// We generally recommend altering this or Temperature but not both.
	//
	// The default value is 1.0.
	TopP float64 `json:"top_p"`

	// A list of tools the model may call. Currently, only functions are supported as a tool.
	// Use this to provide a list of functions the model may generate JSON inputs for.
	// A max of 128 functions are supported.
	Tools []Tool `json:"tools,omitempty"`

	// Controls which (if any) tool is called by the model.
	//  - ToolChoiceNone means the model will not call any tool and instead generates a message.
	//  - ToolChoiceAuto means the model can pick between generating a message or calling one or more tools.
	//  - ToolChoiceRequired means the model must call one or more tools.
	//  - ToolChoiceFunction(name) forces the model to call a specific tool.
	//
	// The default is ToolChoiceAuto when tools are present, and ToolChoiceNone otherwise.
	ToolChoice ToolChoice `json:"tool_choice,omitempty"`

	// Whether to enable parallel function calling during tool use.
	//
	// The default is true.
	ParallelToolCalls bool `json:"parallel_tool_calls"`

	// The user ID to associate with the request, as described in the [documentation].
	// If not specified, the request will be anonymous.
	//
	// [documentation]: https://platform.openai.com/docs/guides/safety-best-practices/end-user-ids
	User string `json:"user,omitempty"`
}

type ChatModelOutput struct {

	// A unique identifier for the chat completion.
	Id string `json:"id"`

	// The name of the output object type returned by the API.
	// This will always be "chat.completion".
	Object string `json:"object"`

	// A list of chat completion choices. Can be more than one if n is greater than 1 in the input options.
	Choices []Choice `json:"choices"`

	// The timestamp of when the chat completion was created.
	Created time.Time `json:"created"`

	// The name of the model used to generate the chat.
	// In most cases, this will match the requested model field in the input.
	Model string `json:"model"`

	// The service tier used for processing the request.
	//
	// This field is only included if the ServiceTier parameter is specified in the request.
	ServiceTier ServiceTier `json:"service_tier"`

	// This fingerprint represents the OpenAI backend configuration that the model runs with.
	//
	// Can be used in conjunction with the Seed request parameter to understand when backend changes
	// have been made that might impact determinism.
	SystemFingerprint string `json:"system_fingerprint"`

	// The usage statistics for the request.
	Usage Usage `json:"usage"`
}

type Choice struct {

	// The reason the model stopped generating tokens.
	//
	// Possible values are:
	//  - "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 content filters
	//  - "tool_calls" if the model called a tool
	FinishReason string `json:"finish_reason"`

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

	// A message generated by the model.
	Message CompletionMessage `json:"message"`

	// Log probability information for the choice.
	Logprobs Logprobs `json:"logprobs"`
}

type CompletionMessage struct {

	// The content of the message.
	Content string `json:"content"`

	// The refusal message generated by the model, if any.
	Refusal string `json:"refusal,omitempty"`

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

	// The audio output generated by the model, if any.
	// Used only when audio output is requested in the modalities field, and when the model and host support audio output.
	Audio *AudioOutput `json:"audio,omitempty"`
}

type ContentPart interface {

	// The type of the content part.
	Type() string
	// contains filtered or unexported methods
}

type DeveloperContentPart = SystemContentPart

type DeveloperMessage[T string | []DeveloperContentPart] struct {

	// The content of the message.
	Content T

	// An optional name for the participant.
	// Provides the model information to differentiate between participants of the same role.
	Name string
}

type Embedding struct {

	// The name of the output object type returned by the API.
	// This will always be "embedding".
	Object string `json:"object"`

	// The index of the input text that corresponds to this embedding.
	// Used when requesting embeddings for multiple texts.
	Index int `json:"index"`

	// The vector embedding of the input text.
	Embedding []float32 `json:"embedding"`
}

type EmbeddingsModel struct {
	// contains filtered or unexported fields
}

type EmbeddingsModelInput struct {

	// The name of the model to use to generate the embeddings.
	//
	// Must be the exact string expected by the model provider.
	// For example, "text-embedding-3-small".
	Model string `json:"model"`

	// The input content to vectorize.
	Input any `json:"input"`

	// The format for the output embeddings.
	// The default ("") is equivalent to [EncodingFormatFloat], which is currently the only supported format.
	EncodingFormat EncodingFormat `json:"encoding_format,omitempty"`

	// The maximum number of dimensions for the output embeddings.
	// The default (0) indicates that the model's default number of dimensions will be used.
	Dimensions int `json:"dimensions,omitempty"`

	// The user ID to associate with the request, as described in the [documentation].
	// If not specified, the request will be anonymous.
	//
	// [documentation]: https://platform.openai.com/docs/guides/safety-best-practices/end-user-ids
	User string `json:"user,omitempty"`
}

type EmbeddingsModelOutput struct {

	// The name of the output object type returned by the API.
	// This will always be "list".
	Object string `json:"object"`

	// The name of the model used to generate the embeddings.
	// In most cases, this will match the requested model field in the input.
	Model string `json:"model"`

	// The usage statistics for the request.
	Usage Usage `json:"usage"`

	// The output vector embeddings data.
	Data []Embedding `json:"data"`
}

type EncodingFormat string

type FunctionCall struct {

	// The name of the function to call.
	Name string `json:"name"`

	// The arguments to call the function with, as generated by the model in JSON format.
	Arguments string `json:"arguments"`
}

type FunctionDefinition struct {
	// 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"`

	// An optional description of what the function does, used by the model to choose when and how to call the function.
	Description string `json:"description,omitempty"`

	// Whether to enable strict schema adherence when generating the function call.
	// If set to true, the model will follow the exact schema defined in the parameters field.
	//
	// The default is false.
	//
	// See https://platform.openai.com/docs/guides/function-calling
	//
	// NOTE:
	// In order to guarantee strict schema adherence, disable parallel function calls
	// by setting ParallelToolCalls to false.
	//
	// See https://platform.openai.com/docs/guides/function-calling/parallel-function-calling-and-structured-outputs
	Strict bool `json:"strict,omitempty"`

	// The parameters the functions accepts, described as a JSON Schema object.
	//
	// See https://platform.openai.com/docs/guides/function-calling
	Parameters utils.RawJsonString `json:"parameters,omitempty"`
}

type Image struct {

	// The URL of the image.
	Url string `json:"url"`

	// An optional detail string for the image.
	// Can be set to "low", "high", or "auto".
	// The default is "auto".
	Detail string `json:"detail,omitempty"`
}

type ImageContentPart struct {
	// The image information.
	Image Image
}

type Logprobs struct {
	// A list of message content tokens with log probability information.
	Content []LogprobsContent `json:"content"`
}

type LogprobsContent struct {
	LogprobsContentObject

	// 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 TopLogprobs returned.
	TopLogprobs []LogprobsContentObject `json:"top_logprobs"`
}

type LogprobsContentObject struct {
	// The token.
	Token string `json:"token"`

	// 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"`

	// 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 nil if there is no bytes representation for the token.
	Bytes []byte `json:"bytes"`
}

type Modality string

type RefusalContentPart struct {
	// The refusal message generated by the model.
	Refusal string
}

type RequestMessage interface {
	json.Marshaler

	// The role of the author of this message.
	Role() string
}

type ResponseFormat struct {

	// The type of response format.
	Type string `json:"type"`

	// The JSON schema to use for the response format.
	JsonSchema utils.RawJsonString `json:"json_schema,omitempty"`
}

type ServiceTier string

type SystemContentPart interface {
	ContentPart
	// contains filtered or unexported methods
}

type SystemMessage[T string | []SystemContentPart] struct {
	// The content of the message.
	Content T

	// An optional name for the participant.
	// Provides the model information to differentiate between participants of the same role.
	Name string
}

type TextContentPart struct {
	// The text string.
	Text string
}

type Tool struct {

	// The type of the tool. Currently, only `"function"` is supported.
	Type string `json:"type"`

	// The definition of the function.
	Function FunctionDefinition `json:"function"`
	// contains filtered or unexported fields
}

type ToolCall struct {

	// The ID of the tool call.
	Id string `json:"id"`

	// The type of the tool. Currently, only `"function"` is supported.
	Type string `json:"type"`

	// The function that the model called.
	Function FunctionCall `json:"function"`
}

type ToolChoice struct {

	// The type of tool to call.
	Type string `json:"type"`

	// The function to call.
	Function struct {
		// The name of the function to call.
		Name string `json:"name"`
	} `json:"function,omitempty"`
}

type ToolMessage[T string | []ToolMessageContentPart] struct {

	// The content of the message.
	Content T

	// The tool call that this message is responding to.
	ToolCallId string
}

type ToolMessageContentPart interface {
	ContentPart
	// contains filtered or unexported methods
}

type Usage struct {

	// The number of completion tokens used.
	CompletionTokens int `json:"completion_tokens"`

	// The number of prompt tokens used.
	PromptTokens int `json:"prompt_tokens"`

	// The total number of tokens used.
	TotalTokens int `json:"total_tokens"`
}

type UserMessage[T string | []UserMessageContentPart] struct {

	// The content of the message.
	Content T

	// An optional name for the participant.
	// Provides the model information to differentiate between participants of the same role.
	Name string
}

type UserMessageContentPart interface {
	ContentPart
	// contains filtered or unexported methods
}