hfgo

type APIError = hferrors.APIError

type ChatChoice struct {
	// Required.
	FinishReason string `json:"finish_reason"`
	// Required.
	Index    int           `json:"index"`
	LogProbs *ChatLogProbs `json:"logprobs,omitempty"`
	// Required.
	Message ChatCompletionMessage `json:"message"`
}

type ChatCompletionMessage struct {
	Role string `json:"role"`
	// Content is present for text responses.
	Content *string `json:"content,omitempty"`
	// ToolCallID is set when returning tool-specific content.
	ToolCallID *string `json:"tool_call_id,omitempty"`
	// ToolCalls is present for tool call responses.
	ToolCalls []ChatToolCallOutput `json:"tool_calls,omitempty"`
}

type ChatFunctionCall struct {
	// Required.
	Name string `json:"name"`
	// Required.
	Arguments   string  `json:"arguments"`
	Description *string `json:"description,omitempty"`
}

type ChatFunctionDefinition struct {
	// Required.
	Name string `json:"name"`
	// A description of what the function does.
	Description *string `json:"description,omitempty"`
	// JSON schema describing function parameters.
	Parameters json.RawMessage `json:"parameters,omitempty"`
}

type ChatFunctionName struct {
	// Required.
	Name string `json:"name"`
}

type ChatImageURL struct {
	// Required.
	URL string `json:"url"`
}

type ChatJSONSchemaConfig struct {
	// The name of the response format.
	// Required.
	Name string `json:"name"`
	// A description of what the response format is for.
	Description *string `json:"description,omitempty"`
	// The schema for the response format as a JSON Schema object.
	// Learn how to build JSON schemas at https://json-schema.org/.
	Schema json.RawMessage `json:"schema,omitempty"`
	// Whether to enable strict schema adherence.
	Strict *bool `json:"strict,omitempty"`
}

type ChatLogProb struct {
	// Required.
	Token string `json:"token"`
	// Required.
	LogProb float64 `json:"logprob"`
	// Required.
	TopLogProbs []ChatTopLogProb `json:"top_logprobs"`
}

type ChatLogProbs struct {
	// Required.
	Content []ChatLogProb `json:"content"`
}

type ChatMessage struct {
	// Role of the message author (for example: system, user, assistant, tool).
	// Required.
	Role string `json:"role"`
	// Optional name for the participant.
	Name *string `json:"name,omitempty"`

	// Content may be a string or a list of content chunks.
	// Either Content or ToolCalls should be supplied.
	Content ChatMessageContent `json:"content"`

	// ToolCalls is used instead of Content when providing tool call messages.
	ToolCalls []ChatToolCall `json:"tool_calls,omitempty"`
}

type ChatMessageChunk struct {
	// Required when Type is "text".
	Text *string `json:"text,omitempty"`
	// Required when Type is "image_url".
	ImageURL *ChatImageURL `json:"image_url,omitempty"`
	// Required. Possible values: text, image_url.
	Type MessageChunkType `json:"type"`
}

type ChatMessageContent struct {

	// Text holds plain string content.
	Text *string `json:"-"`
	// Chunks holds structured content chunks.
	Chunks []ChatMessageChunk `json:"-"`
}

type ChatRequest struct {
	// Model to use for the chat completion.
	// Required.
	Model *string `json:"model,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"`

	// 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"`

	// The maximum number of tokens that can be generated in the chat completion.
	// Default: 1024. Minimum: 0.
	MaxTokens *int `json:"max_tokens,omitempty"`

	// A list of messages comprising the conversation so far.
	// Required.
	Messages []ChatMessage `json:"messages"`

	// 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"`

	// Response format configuration. Known types: text, json_schema, json_object.
	// Non-empty provider-specific values are also accepted.
	ResponseFormat *ChatResponseFormat `json:"response_format,omitempty"`

	// Seed for deterministic sampling.
	// Minimum: 0.
	Seed *int64 `json:"seed,omitempty"`

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

	// If true, generated tokens are returned as a stream using SSE.
	// For more information about streaming, see:
	// https://huggingface.co/docs/text-generation-inference/conceptual/streaming
	Stream *bool `json:"stream,omitempty"`

	// Stream options for SSE responses.
	StreamOptions *ChatStreamOptions `json:"stream_options,omitempty"`

	// Sampling temperature to use, between 0 and 2.
	// We generally recommend altering this or TopP but not both.
	Temperature *float64 `json:"temperature,omitempty"`

	// Tool choice behavior. Known values: auto, none, required, or
	// {"function":{"name":"..."}}. Meanings:
	// - auto: model can pick between generating a message or calling tools.
	// - none: model will not call any tool and instead generates a message.
	// - required: model must call one or more tools.
	// Non-empty provider-specific values are also accepted.
	ToolChoice *ChatToolChoice `json:"tool_choice,omitempty"`

	// A prompt to be appended before the tools.
	ToolPrompt *string `json:"tool_prompt,omitempty"`

	// 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.
	Tools []ChatTool `json:"tools,omitempty"`

	// Number of most likely tokens to return at each token position.
	// LogProbs must be true if this is set.
	// An integer between 0 and 5.
	TopLogProbs *int `json:"top_logprobs,omitempty"`

	// Nucleus sampling probability mass.
	// For example, 0.1 means only the tokens comprising the top 10% probability
	// mass are considered.
	TopP *float64 `json:"top_p,omitempty"`
}

type ChatResponse struct {
	// Required.
	ID string `json:"id"`
	// Unix timestamp in seconds.
	// Required.
	Created int64 `json:"created"`
	// Required.
	Model string `json:"model"`
	// Required.
	SystemFingerprint string `json:"system_fingerprint"`
	// Required.
	Choices []ChatChoice `json:"choices"`
	// Required.
	Usage ChatUsage `json:"usage"`
}

type ChatResponseFormat struct {
	// Known type values: text, json_schema, json_object.
	// Non-empty provider-specific values are also accepted.
	// For json_schema, JSONSchema is required.
	Type       ResponseFormatType    `json:"type"`
	JSONSchema *ChatJSONSchemaConfig `json:"json_schema,omitempty"`
}

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

type ChatStreamChoice struct {
	// Required.
	Delta        ChatStreamDelta `json:"delta"`
	FinishReason *string         `json:"finish_reason,omitempty"`
	// Required.
	Index    int           `json:"index"`
	LogProbs *ChatLogProbs `json:"logprobs,omitempty"`
}

type ChatStreamDelta struct {
	// Content is present for text deltas.
	Content *string `json:"content,omitempty"`
	// Role may be included with the first delta.
	Role *string `json:"role,omitempty"`
	// ToolCallID may be included for tool-specific content.
	ToolCallID *string `json:"tool_call_id,omitempty"`
	// ToolCalls is present for tool call deltas.
	ToolCalls []ChatStreamToolCall `json:"tool_calls,omitempty"`
}

type ChatStreamFunction struct {
	Name      string `json:"name"`
	Arguments string `json:"arguments"`
}

type ChatStreamOptions struct {
	// If set, an additional chunk is streamed before the [DONE] message
	// showing overall usage. 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 include a usage field with
	// a null value.
	IncludeUsage *bool `json:"include_usage,omitempty"`
}

type ChatStreamResponse struct {
	ID string `json:"id"`
	// Unix timestamp in seconds.
	Created           int64              `json:"created"`
	Model             string             `json:"model"`
	SystemFingerprint string             `json:"system_fingerprint"`
	Choices           []ChatStreamChoice `json:"choices"`
	Usage             *ChatUsage         `json:"usage,omitempty"`
}

type ChatStreamToolCall struct {
	ID       string             `json:"id,omitempty"`
	Type     string             `json:"type,omitempty"`
	Index    int                `json:"index"`
	Function ChatStreamFunction `json:"function"`
}

type ChatTool struct {
	// Tool type (currently only function is supported).
	Type     string                 `json:"type"`
	Function ChatFunctionDefinition `json:"function"`
}

type ChatToolCall struct {
	// Tool call ID.
	// Required.
	ID string `json:"id"`
	// Tool call type (for example: function).
	// Required.
	Type string `json:"type"`
	// Required.
	Function ChatFunctionCall `json:"function"`
}

type ChatToolCallOutput struct {
	// Required.
	ID string `json:"id"`
	// Required.
	Type string `json:"type"`
	// Required.
	Function ChatFunctionCall `json:"function"`
}

type ChatToolChoice struct {

	// Mode is a tool choice mode. Known values: auto, none, required.
	// Non-empty provider-specific values are also accepted.
	Mode *ToolChoiceMode `json:"-"`
	// Function selects a specific tool function by name.
	Function *ChatFunctionName `json:"-"`
}

type ChatTopLogProb struct {
	// Required.
	Token string `json:"token"`
	// Required.
	LogProb float64 `json:"logprob"`
}

type ChatUsage struct {
	// Required.
	CompletionTokens int `json:"completion_tokens"`
	// Required.
	PromptTokens int `json:"prompt_tokens"`
	// Required.
	TotalTokens int `json:"total_tokens"`
}

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

type FillMaskBatchRequest struct {
	// The inputs with masked tokens.
	// Required.
	Inputs []string `json:"inputs"`

	// Additional inference parameters for mask filling
	Parameters *FillMaskParameters `json:"parameters,omitempty"`
}

type FillMaskParameters struct {
	// When passed, the model will limit the scores to the passed targets instead of looking up
	// in the whole vocabulary. If the provided targets are not in the model vocab, they will be
	// tokenized and the first resulting token will be used (with a warning, and that might be
	// slower).
	Targets []string `json:"targets,omitempty"`

	// When passed, overrides the number of predictions to return.
	TopK *int `json:"top_k,omitempty"`
}

type FillMaskPrediction struct {
	// The input filled with the mask token prediction
	Sequence string `json:"sequence"`

	// The probability of the token prediction
	Score float64 `json:"score"`

	// The predicted token id (to replace the masked one).
	Token int `json:"token"`

	// The predicted token (to replace the masked one).
	TokenStr *string `json:"token_str"`
}

type FillMaskRequest struct {
	// The input text with masked tokens.
	// Required.
	Input string `json:"inputs"`

	// Additional inference parameters for mask filling
	Parameters *FillMaskParameters `json:"parameters,omitempty"`
}

type MessageChunkType string

type Option = request.Option

type QuestionAnswering struct {
	// The answer to the question.
	Answer string `json:"answer"`

	// The probability associated to the answer.
	Score float64 `json:"score"`

	// The character position in the input where the answer begins.
	Start int `json:"start"`

	// The character position in the input where the answer ends.
	End int `json:"end"`
}

type QuestionAnsweringInput struct {
	// The question to be answered.
	// Required.
	Question string `json:"question"`

	// The context to be used for answering the question.
	// Required.
	Context string `json:"context"`
}

type QuestionAnsweringParameters struct {
	// The number of answers to return (will be chosen by order of likelihood).
	// Note that less than top_k answers may be returned if there are not enough
	// options available within the context.
	TopK *int `json:"top_k,omitempty"`

	// If the context is too long to fit with the question for the model, it will
	// be split in several chunks with some overlap. This argument controls the
	// size of that overlap.
	DocStride *int `json:"doc_stride,omitempty"`

	// The maximum length of predicted answers (e.g., only answers with a shorter
	// length are considered).
	MaxAnswerLen *int `json:"max_answer_len,omitempty"`

	// The maximum length of the total sentence (context + question) in tokens of
	// each chunk passed to the model. The context will be split in several chunks
	// (using doc_stride as overlap) if needed.
	MaxSeqLen *int `json:"max_seq_len,omitempty"`

	// The maximum length of the question after tokenization. It will be truncated
	// if needed.
	MaxQuestionLen *int `json:"max_question_len,omitempty"`

	// Whether to accept impossible as an answer.
	HandleImpossibleAnswer *bool `json:"handle_impossible_answer,omitempty"`

	// Attempts to align the answer to real words. Improves quality on space
	// separated languages. Might hurt on non-space-separated languages (like
	// Japanese or Chinese).
	AlignToWords *bool `json:"align_to_words,omitempty"`
}

type QuestionAnsweringRequest struct {
	// The question and context pair to answer.
	// Required.
	Input QuestionAnsweringInput `json:"inputs"`

	// Additional inference parameters for question answering.
	Parameters *QuestionAnsweringParameters `json:"parameters,omitempty"`
}

type RawEvent struct {
	Data  []byte
	Event string
	ID    string
	Retry *time.Duration
}

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

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

type ResponseFormatType string

type SDKError = hferrors.SDKError

type SDKErrorKind = hferrors.SDKErrorKind

type Summarization struct {
	// The summarized text.
	SummaryText string `json:"summary_text"`
}

type SummarizationBatchRequest struct {
	// The texts to summarize.
	// Required.
	Inputs []string `json:"inputs"`

	// Additional inference parameters for summarization.
	Parameters *SummarizationParameters `json:"parameters,omitempty"`
}

type SummarizationParameters struct {
	// Whether to clean up the potential extra spaces in the text output.
	CleanUpTokenizationSpaces *bool `json:"clean_up_tokenization_spaces,omitempty"`

	// The truncation strategy to use.
	Truncation *string `json:"truncation,omitempty"`

	// GenerateParameters provides additional parametrization of the text
	// generation algorithm (e.g. "max_new_tokens", "temperature", "top_k").
	//
	// It is exposed because it is part of the upstream inference schema, but the
	// set of valid arguments is not documented by Hugging Face and may depend on
	// the model being used. Invalid or unsupported arguments can be rejected by
	// the API.
	GenerateParameters map[string]any `json:"generate_parameters,omitempty"`
}

type SummarizationRequest struct {
	// The text to summarize.
	// Required.
	Input string `json:"inputs"`

	// Additional inference parameters for summarization.
	Parameters *SummarizationParameters `json:"parameters,omitempty"`
}

type TableQuestionAnswer struct {
	// The answer to the question given the table. If there is an aggregator,
	// the answer will be preceded by "AGGREGATOR >".
	Answer string `json:"answer"`

	// Coordinates of the cells of the answers.
	Coordinates [][]int `json:"coordinates"`

	// List of strings made up of the answer cell values.
	Cells []string `json:"cells"`

	// If the model has an aggregator, this returns the aggregator.
	Aggregator *string `json:"aggregator,omitempty"`
}

type TableQuestionAnsweringInput struct {
	// The question to be answered about the table.
	// Required.
	Question string `json:"question"`

	// The table to serve as context for the questions.
	// Each key is a column name, and the value is a list of cell values for that column.
	// Required.
	Table map[string][]string `json:"table"`
}

type TableQuestionAnsweringParameters struct {
	// Activates and controls padding.
	Padding *string `json:"padding,omitempty"`

	// Whether to do inference sequentially or as a batch. Batching is faster,
	// but models like SQA require the inference to be done sequentially to
	// extract relations within sequences, given their conversational nature.
	Sequential *bool `json:"sequential,omitempty"`

	// Activates and controls truncation.
	Truncation *bool `json:"truncation,omitempty"`
}

type TableQuestionAnsweringRequest struct {
	// The question and table pair to answer.
	// Required.
	Input TableQuestionAnsweringInput `json:"inputs"`

	// Additional inference parameters for table question answering.
	Parameters *TableQuestionAnsweringParameters `json:"parameters,omitempty"`
}

type TextClassification struct {
	// The predicted class label.
	Label string `json:"label"`

	// The corresponding probability.
	Score float64 `json:"score"`
}

type TextClassificationBatchRequest struct {
	// The texts to classify.
	// Required.
	Inputs []string `json:"inputs"`

	// Additional inference parameters for text classification
	Parameters *TextClassificationParameters `json:"parameters,omitempty"`
}

type TextClassificationParameters struct {
	// Possible values: sigmoid, softmax, none.
	FunctionToApply *string `json:"function_to_apply,omitempty"`

	// When specified, limits the output to the top K most probable classes.
	TopK *int `json:"top_k,omitempty"`
}

type TextClassificationRequest struct {
	// The text to classify.
	// Required.
	Input string `json:"inputs"`

	// Additional inference parameters for text classification
	Parameters *TextClassificationParameters `json:"parameters,omitempty"`
}

type TokenClassification struct {
	// The predicted label for a group of one or more tokens.
	// Present when an aggregation strategy other than "none" is used.
	EntityGroup *string `json:"entity_group,omitempty"`

	// The predicted label for a single token.
	// Present when aggregation_strategy is "none".
	Entity *string `json:"entity,omitempty"`

	// The associated score / probability.
	Score float64 `json:"score"`

	// The corresponding text.
	Word string `json:"word"`

	// The character position in the input where this group begins.
	Start int `json:"start"`

	// The character position in the input where this group ends.
	End int `json:"end"`
}

type TokenClassificationBatchRequest struct {
	// The texts to classify tokens from.
	// Required.
	Inputs []string `json:"inputs"`

	// Additional inference parameters for token classification.
	Parameters *TokenClassificationParameters `json:"parameters,omitempty"`
}

type TokenClassificationParameters struct {
	// A list of labels to ignore in the classification results.
	IgnoreLabels []string `json:"ignore_labels,omitempty"`

	// The number of overlapping tokens between chunks when splitting the input text.
	Stride *int `json:"stride,omitempty"`

	// The strategy used to fuse tokens based on model predictions.
	// Possible values: "none", "simple", "first", "average", "max".
	AggregationStrategy *string `json:"aggregation_strategy,omitempty"`
}

type TokenClassificationRequest struct {
	// The text to classify tokens from.
	// Required.
	Input string `json:"inputs"`

	// Additional inference parameters for token classification.
	Parameters *TokenClassificationParameters `json:"parameters,omitempty"`
}

type ToolChoiceMode string

type Translation struct {
	// The translated text.
	TranslationText string `json:"translation_text"`
}

type TranslationBatchRequest struct {
	// The texts to translate.
	// Required.
	Inputs []string `json:"inputs"`

	// Additional inference parameters for translation.
	Parameters *TranslationParameters `json:"parameters,omitempty"`
}

type TranslationParameters struct {
	// Whether to clean up the potential extra spaces in the text output.
	CleanUpTokenizationSpaces *bool `json:"clean_up_tokenization_spaces,omitempty"`

	// The source language of the text. Required for models that can
	// translate from multiple languages.
	SrcLang *string `json:"src_lang,omitempty"`

	// The target language to translate to. Required for models that can
	// translate to multiple languages.
	TgtLang *string `json:"tgt_lang,omitempty"`

	// The truncation strategy to use.
	Truncation *string `json:"truncation,omitempty"`

	// GenerateParameters provides additional parametrization of the text
	// generation algorithm (e.g. "max_new_tokens", "temperature", "top_k").
	//
	// It is exposed because it is part of the upstream inference schema, but the
	// set of valid arguments is not documented by Hugging Face and may depend on
	// the model being used. Invalid or unsupported arguments can be rejected by
	// the API.
	GenerateParameters map[string]any `json:"generate_parameters,omitempty"`
}

type TranslationRequest struct {
	// The text to translate.
	// Required.
	Input string `json:"inputs"`

	// Additional inference parameters for translation.
	Parameters *TranslationParameters `json:"parameters,omitempty"`
}

type ZeroShotTextClassification struct {
	// The predicted class label.
	Label string `json:"label"`

	// The corresponding probability.
	Score float64 `json:"score"`
}

type ZeroShotTextClassificationBatchRequest struct {
	// The texts to classify.
	// Required.
	Inputs []string `json:"inputs"`

	// Additional inference parameters for zero-shot text classification.
	// Required.
	Parameters *ZeroShotTextClassificationParameters `json:"parameters,omitempty"`
}

type ZeroShotTextClassificationParameters struct {
	// The set of possible class labels to classify the text into.
	// Required.
	CandidateLabels []string `json:"candidate_labels,omitempty"`

	// The sentence used in conjunction with candidate_labels to attempt
	// the text classification by replacing the placeholder with the
	// candidate labels.
	HypothesisTemplate *string `json:"hypothesis_template,omitempty"`

	// Whether multiple candidate labels can be true. If false, the scores
	// are normalized such that the sum of the label likelihoods for each
	// sequence is 1. If true, the labels are considered independent and
	// probabilities are normalized for each candidate.
	MultiLabel *bool `json:"multi_label,omitempty"`
}

type ZeroShotTextClassificationRequest struct {
	// The text to classify.
	// Required.
	Input string `json:"inputs"`

	// Additional inference parameters for zero-shot text classification.
	// Required.
	Parameters *ZeroShotTextClassificationParameters `json:"parameters,omitempty"`
}