providers

type BatchResult struct {
	Item   *PipelineItem
	Result *OCRResult
	Error  error
}

type Config struct {
	OCR            ProviderConfig       `yaml:"ocr"`
	Pipeline       PipelineConfig       `yaml:"pipeline"`
	RateLimit      RateLimitConfig      `yaml:"rate_limit"`
	TextCorrection TextCorrectionConfig `yaml:"text_correction"`

	// Provider-specific options that implement the ProviderOptionsInterface
	DoclingOptions *DoclingOptions `yaml:"docling_options,omitempty"`
	OpenAIOptions  *OpenAIOptions  `yaml:"openai_options,omitempty"`
	OllamaOptions  *OllamaOptions  `yaml:"ollama_options,omitempty"`
}

type Configurable interface {
	GetConfig() Config
}

type DoclingCliClient struct {
	Available  bool
	BinaryPath string
}

type DoclingClient struct {
	Client  *http.Client
	BaseURL string
}

type DoclingConvertDocumentResponse struct {
	Document       DoclingDocumentResponse `json:"document"`
	Status         string                  `json:"status"` // "success", "partial_success", "skipped", "failure"
	Errors         []any                   `json:"errors"`
	ProcessingTime float64                 `json:"processing_time"`
	Timings        map[string]any          `json:"timings"`
}

type DoclingDocumentResponse struct {
	Filename       string          `json:"filename"`
	MDContent      string          `json:"md_content"`
	JSONContent    json.RawMessage `json:"json_content"`
	HTMLContent    string          `json:"html_content"`
	TextContent    string          `json:"text_content"`
	DocTagsContent string          `json:"doctags_content"`
}

type DoclingHealthResponse struct {
	Status string `json:"status"`
}

type DoclingOptions struct {
	OCR       *bool `yaml:"do_ocr,omitempty"`
	Force_OCR *bool `yaml:"force_ocr,omitempty"`

	Pipeline string `yaml:"pipeline,omitempty"`

	PDF_Backend      string `yaml:"pdf_backend,omitempty"`
	Abort_On_Error   *bool  `yaml:"abort_on_error,omitempty"`
	Document_Timeout string `yaml:"document_timeout,omitempty"`
	Num_Threads      string `yaml:"num_threads,omitempty"`
	Device           string `yaml:"device,omitempty"`
	Verbose          string `yaml:"verbose,omitempty"`
}

type DoclingProcessPayload struct {
	FileBytes []byte
	Filename  string
	Options   map[string]string
}

type DoclingProvider struct {
	Client    *DoclingClient
	CliClient *DoclingCliClient
	Config    Config
}

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

type InputType int

type MistralOcrDimensions struct {
	DPI    int `json:"dpi"`
	Height int `json:"height"`
	Width  int `json:"width"`
}

type MistralOcrDocument struct {
	ImageURL     string `json:"image_url,omitempty"`
	DocumentURL  string `json:"document_url,omitempty"`
	DocumentName string `json:"document_name,omitempty"`
	Type         string `json:"type"`
}

type MistralOcrImage struct {
	ID           string `json:"id"`
	TopLeftX     int    `json:"top_left_x"`
	TopLeftY     int    `json:"top_left_y"`
	BottomRightX int    `json:"bottom_right_x"`
	BottomLeftX  int    `json:"bottom_left_x"`
	ImageBase64  string `json:"image_base64"`
}

type MistralOcrPage struct {
	Index      int                  `json:"index"`
	Markdown   string               `json:"markdown"`
	Images     []MistralOcrImage    `json:"images"`
	Dimensions MistralOcrDimensions `json:"dimensions"`
}

type MistralOcrResponse struct {
	Pages     []MistralOcrPage `json:"pages"`
	Model     string           `json:"model"`
	UsageInfo map[string]any   `json:"usage_info"`
}

type MistralOcrResquest struct {
	Model              string             `json:"model"`
	ID                 string             `json:"id,omitempty"`
	Document           MistralOcrDocument `json:"document"`
	Pages              []int              `json:"pages,omitempty"`
	IncludeImageBase64 bool               `json:"include_image_base64"`
	ImageLimit         int                `json:"image_limit,omitempty"`
	ImageMinSize       int                `json:"image_min_size,omitempty"`
}

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

type OCRImage struct {
	MistralImages []MistralOcrImage
}

type OCRInput struct {
	Type     InputType
	Image    image.Image
	PDFData  []byte
	Filename string
}

type OCRProvider interface {
	OCR(ctx context.Context, input OCRInput) (*OCRResult, error)
}

type OCRResult struct {
	Text     string
	Images   OCRImage
	Metadata map[string]string
}

type OllamaMessage struct {
	Role    string   `json:"role"`
	Content string   `json:"content"`
	Images  []string `json:"images,omitempty"` // Array of base64 encoded images
}

type OllamaOptions struct {
	MaxTokens   int64   `yaml:"max_tokens"`
	Temperature float64 `yaml:"temperature"`
}

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

type OllamaRequest struct {
	Model    string          `json:"model"`
	Messages []OllamaMessage `json:"messages"`
	Stream   bool            `json:"stream"`
	Options  map[string]any  `json:"options"`
}

type OllamaResponse struct {
	Model     string `json:"model"`
	CreatedAt string `json:"created_at"`
	Message   struct {
		Role    string `json:"role"`
		Content string `json:"content"`
	} `json:"message"`
	Done          bool  `json:"done"`
	TotalDuration int64 `json:"total_duration"`
}

type OpenAIOptions struct {
	MaxTokens   int64   `yaml:"max_tokens"`
	Temperature float64 `yaml:"temperature"`
}

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

type PDFCapable interface {
	SupportsPDF() bool
}

type PipelineConfig struct {
	DPI            float64 `yaml:"dpi"`
	Concurrency    int     `yaml:"concurrency"`
	OCRConcurrency int     `yaml:"ocr_concurrency"`
}

type PipelineItem struct {
	Input         *OCRInput
	OriginalIndex int
	PageIndex     int // -1 for non-PDF pages or single-page docs
}

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

type ProgressTrackerConfig struct {
	Total            int64
	UpdateFrequency  time.Duration
	MessageFormatter func(inProgress, completed, failed, total int64) string
	Prefix           string
}

type ProviderConfig struct {
	Provider     string `yaml:"provider"`
	Model        string `yaml:"model"`
	Prompt       string `yaml:"prompt"`
	AppendPrompt string `yaml:"append_prompt"`
	Language     string `yaml:"language"`
	Format       string `yaml:"format"`
	SupportsPDF  bool   `yaml:"supports_pdf"`
}

type ProviderOptionsInterface interface {
	// Apply merges default options with schema options and returns a config type the provider can use.
	Apply(defaults any, config Config) (any, error)
}

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

type RateLimitConfig struct {
	RPS   float64 `yaml:"rps"`
	Burst int     `yaml:"burst"`
}

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

type TesseractClient struct{}

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

type TextCorrectionConfig struct {
	Enabled   bool            `yaml:"enabled"`
	Provider  ProviderConfig  `yaml:"provider"`
	RateLimit RateLimitConfig `yaml:"rate_limit"`
}

type TextProcessor interface {
	Complete(ctx context.Context, text string) (string, error)
}