Documentation
¶
Overview ¶
Package pagecapture captures normalized, reusable evidence about web pages — rendered HTML, a screenshot, visible text, and structural signals (title, language, links, headings) — using a w3pilot browser session. It is domain-neutral: accessibility evaluation, functional/journey testing, i18n checks, and security review all consume the same PageEvidence rather than each driving the browser themselves.
Index ¶
- Constants
- type CrawlOptions
- type Heading
- type Link
- type Options
- type PageEvidence
- func Capture(ctx context.Context, pilot *w3pilot.Pilot, pageURL string, opts Options) (*PageEvidence, error)
- func CaptureCurrent(ctx context.Context, pilot *w3pilot.Pilot, pageURL string, screenshot bool) *PageEvidence
- func CaptureURL(ctx context.Context, pageURL string, opts Options) (*PageEvidence, error)
- type SettleMode
- type SiteEvidence
Constants ¶
const ( MaxHTMLBytes = 200_000 MaxTextBytes = 40_000 )
Bounds on captured text to keep evidence bundles a sane size.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type CrawlOptions ¶
type CrawlOptions struct {
Options
// MaxPages caps how many pages to capture (including the start page).
MaxPages int
// SameHostOnly restricts the crawl to the start URL's host.
SameHostOnly bool
}
CrawlOptions configure a site capture.
func DefaultCrawlOptions ¶
func DefaultCrawlOptions() CrawlOptions
DefaultCrawlOptions returns defaults: up to 5 same-host pages.
type Options ¶
type Options struct {
// LoadWait bounds how long to wait for the page to settle. The wait is
// best-effort: capture proceeds against the rendered DOM regardless.
LoadWait time.Duration
// MinSettle is the minimum time to wait before accepting a stable DOM.
// SPAs render in phases (shell, then async data) with quiet gaps between;
// this floor prevents settling in an early gap before data arrives.
// Defaults to 2.5s when zero.
MinSettle time.Duration
// Settle selects the readiness strategy (default SettleDOMStable).
Settle SettleMode
// Screenshot toggles screenshot capture (the most expensive part).
Screenshot bool
}
Options configure a capture.
func DefaultOptions ¶
func DefaultOptions() Options
DefaultOptions returns sensible defaults: DOM-stability settle, screenshot on.
type PageEvidence ¶
type PageEvidence struct {
URL string `json:"url"`
Title string `json:"title"`
Language string `json:"language,omitempty"` // <html lang>
// HTML is the rendered outer HTML, bounded to MaxHTMLBytes.
HTML string `json:"html,omitempty"`
// Text is the visible body text, bounded to MaxTextBytes.
Text string `json:"text,omitempty"`
// ScreenshotPNG is a full-page screenshot; base64-encoded when serialized.
ScreenshotPNG []byte `json:"screenshotPng,omitempty"`
Links []Link `json:"links,omitempty"`
Headings []Heading `json:"headings,omitempty"`
// Truncated reports whether HTML/Text were clipped to their bounds.
Truncated bool `json:"truncated,omitempty"`
// LoadWarning is set when the page-load wait did not fully settle.
LoadWarning string `json:"loadWarning,omitempty"`
CapturedAt time.Time `json:"capturedAt"`
}
PageEvidence is normalized evidence for a single page.
func Capture ¶
func Capture(ctx context.Context, pilot *w3pilot.Pilot, pageURL string, opts Options) (*PageEvidence, error)
Capture navigates to pageURL, waits for it to settle, and captures evidence.
func CaptureCurrent ¶
func CaptureCurrent(ctx context.Context, pilot *w3pilot.Pilot, pageURL string, screenshot bool) *PageEvidence
CaptureCurrent captures evidence from the page the Pilot is *already* on, without navigating or settling. Use it when the caller has already loaded and settled the page (e.g. an audit engine that just ran against it), to avoid a redundant navigation. pageURL labels the evidence.
func CaptureURL ¶
CaptureURL launches a headless browser, captures one page, and closes it. Convenience for callers that don't already hold a Pilot.
func (PageEvidence) HasScreenshot ¶
func (p PageEvidence) HasScreenshot() bool
HasScreenshot reports whether a screenshot was captured (used by consumers to decide whether visual criteria can be judged).
type SettleMode ¶
type SettleMode int
SettleMode selects how Capture waits for a page to be ready.
const ( // SettleDOMStable (default) polls the DOM until it stops changing and has // real content. This is the robust choice for client-rendered SPAs, whose // "load" event fires before content exists and whose network may never idle. SettleDOMStable SettleMode = iota // SettleLoad waits only for the "load" event. Fine for static pages. SettleLoad // SettleNone captures immediately after navigation. SettleNone )
type SiteEvidence ¶
type SiteEvidence struct {
StartURL string `json:"startUrl"`
Pages []PageEvidence `json:"pages"`
}
SiteEvidence is evidence captured across multiple pages of one site.
func CaptureSite ¶
func CaptureSite(ctx context.Context, pilot *w3pilot.Pilot, startURL string, opts CrawlOptions) (*SiteEvidence, error)
CaptureSite crawls from startURL, breadth-first, capturing evidence for each page up to MaxPages. It reuses one Pilot for all pages.
Source Files