Types

Result

Extraction result containing text, metadata, and media information.

type Result struct {
    Text           string        `json:"text"`
    Title          string        `json:"title"`
    Images         []ImageInfo   `json:"images,omitempty"`
    Links          []LinkInfo    `json:"links,omitempty"`
    Videos         []VideoInfo   `json:"videos,omitempty"`
    Audios         []AudioInfo   `json:"audios,omitempty"`
    ProcessingTime time.Duration `json:"-"`       // Processing duration (excluded from standard serialization)
    WordCount      int           `json:"word_count"`
    ReadingTime    time.Duration `json:"-"`       // Estimated reading time (excluded from standard serialization)
}

MarshalJSON

Custom JSON serialization. ProcessingTime and ReadingTime have json:"-" tags (standard serialization skips them), but the custom MarshalJSON() method includes them as millisecond values.

func (r *Result) MarshalJSON() ([]byte, error)

WARNING

Result does not implement UnmarshalJSON. If you deserialize the output of MarshalJSON() back into a Result, duration fields such as ProcessingTime and ReadingTime will be lost — the JSON output key names (processing_time_ms, reading_time_ms) do not match the struct field names, so they cannot be restored.

This is intentional: the JSON format is designed for external consumption (e.g., API responses, logs, frontend display), not for bidirectional serialization.

ImageInfo

Image information.

type ImageInfo struct {
    URL          string `json:"url"`           // Image URL
    Alt          string `json:"alt"`           // Alternative text
    Title        string `json:"title"`         // Title
    Width        string `json:"width"`         // Width
    Height       string `json:"height"`        // Height
    IsDecorative bool   `json:"is_decorative"` // Whether the image is decorative
    Position     int    `json:"position"`      // Position in the document
}

Field Semantics

Field	Description
URL	The src attribute value of the image; only valid URLs (passing IsValidURL validation) are included — <img> with an invalid URL does not appear in the result
Alt	The original alt attribute; when empty, IsDecorative is true
Title	The original title attribute (not the page title)
Width/Height	The raw strings from the HTML attributes (e.g. "640", "50%"), not parsed into numbers — different pages may use inconsistent formats
IsDecorative	true when Alt is empty; can be used to identify and skip decorative images
Position	1-based index in the document; when PreserveImages = false, the entire Images slice is empty

Width/Height are string types

Width and Height are string types, not int. They preserve the original representation from the HTML source (which may include units, percentages, etc.). Parse them into numbers yourself when needed.

LinkInfo

Link information.

type LinkInfo struct {
    URL        string `json:"url"`         // Link URL
    Text       string `json:"text"`        // Link text
    Title      string `json:"title"`       // Link title
    IsExternal bool   `json:"is_external"` // Whether it is an external link (determined by whether the URL itself is an absolute external URL, not by comparison with BaseURL)
    IsNoFollow bool   `json:"is_nofollow"` // Whether the link is nofollow
    Position   int    `json:"position"`    // Position in the document
}

Field Semantics

Field	Description
URL	The href attribute value; only valid URLs (passing IsValidURL validation) are included — an <a> with an invalid URL still consumes a Position but is not added to the slice
Text	Concatenation of all text nodes inside the <a> tag (recursive GetTextContent)
Title	The original title attribute (not the link text)
IsExternal	Determined by whether the URL itself is an absolute external address, not by domain comparison with BaseURL — this differs from the internal/external link determination in ExtractAllLinks
IsNoFollow	true when the rel attribute contains nofollow (case-insensitive, ASCII-fold matching)
Position	1-based index in the document; invalid <a> (illegal or missing href) still consume index numbers but are not added to the slice, so Position values may be non-contiguous

VideoInfo

Video information.

type VideoInfo struct {
    URL      string `json:"url"`      // Video URL
    Type     string `json:"type"`     // Video type
    Poster   string `json:"poster"`   // Poster image URL
    Width    string `json:"width"`    // Width
    Height   string `json:"height"`   // Height
    Duration string `json:"duration"` // Duration
}

Type Field Value Rules

Type value	Meaning	When it occurs
"embed"	A video page referenced by an iframe	Embedded players such as YouTube, Vimeo, Youku, Bilibili
MIME type (e.g. "video/mp4")	Video file container	Attribute value from <source type="video/mp4">
Empty string	No type detected	<video src="..."> specifying the source directly with no <source> child element

Three sources of video extraction

Video extraction runs in three steps — raw HTML scan, DOM traversal, regex fallback — with deduplication at each step. See Media Extraction Guide.

AudioInfo

Audio information.

type AudioInfo struct {
    URL      string `json:"url"`      // Audio URL
    Type     string `json:"type"`     // Audio type
    Duration string `json:"duration"` // Duration
}

Type Field Value Rules

Type value	When it occurs
MIME type (e.g. "audio/mpeg")	Attribute value from <source type="audio/mpeg">
Empty string	<audio src="..."> specifying the source directly with no <source> child element

Duality of the .ogg extension

An OGG container can carry video or audio; a .ogg URL will appear in both Videos and Audios. The audio-only variant .oga appears only in Audios.

LinkResource

Link resource (used by link extraction API).

type LinkResource struct {
    URL   string // Link URL
    Title string // Link title
    Type  string // Link type
}

Statistics

Processing statistics.

type Statistics struct {
    TotalProcessed    int64         // Total processed count
    CacheHits         int64         // Cache hit count
    CacheMisses       int64         // Cache miss count
    ErrorCount        int64         // Error count
    AverageProcessTime time.Duration // Average processing time
}

BatchResult

Batch processing result.

type BatchResult struct {
    Results   []*Result // Extraction results, nil on failure or cancellation
    Errors    []error   // Failed errors
    Success   int       // Success count
    Failed    int       // Failure count
    Cancelled int       // Cancelled count
}

NodeAttr

HTML node attribute.

type NodeAttr struct {
    Key   string // Attribute name
    Value string // Attribute value
}