/** Pure HTML extraction for site onboarding crawls. No network. */
import { type SocialLink } from "./social.ts";
export type ExtractedImage = {
    src: string;
    alt: string;
    width?: number;
    height?: number;
    inChrome?: boolean;
    fromJsonLdLogo?: boolean;
    fromOg?: boolean;
    aboveFold?: boolean;
    pageUrl?: string;
    nearbyHeading?: string;
};
export type ExtractedCta = {
    href: string;
    label: string;
    kind: "tel" | "mailto" | "internal" | "external" | "other";
};
export type ExtractedTestimonial = {
    quote: string;
    attribution?: string;
};
export type ExtractedFaq = {
    question: string;
    answer: string;
};
export type ExtractedPage = {
    aliasOf?: string;
    url: string;
    title: string;
    description: string;
    canonical?: string;
    lang?: string;
    h1: string[];
    headings: string[];
    outline: string[];
    paragraphs: string[];
    lists: string[];
    testimonials: ExtractedTestimonial[];
    faqs: ExtractedFaq[];
    links: {
        href: string;
        label: string;
        relative?: boolean;
    }[];
    ctas: ExtractedCta[];
    images: ExtractedImage[];
    phones: string[];
    emails: string[];
    social: SocialLink[];
    jsonLd: unknown[];
    audit: PageAuditHints;
};
export type PageAuditHints = {
    h1Count: number;
    missingAlts: number;
    unlabeledInputs: number;
    hasCanonical: boolean;
    hasLang: boolean;
    hasJsonLd: boolean;
    titleLength: number;
    descriptionLength: number;
};
export declare function shouldSkipPath(pathname: string): boolean;
/** Drop runaway query-string variants while keeping unique content URLs. */
export declare function shouldSkipQueryUrl(href: string): boolean;
export declare function stripNoise(html: string): string;
export declare function absolutize(href: string, base: string): string | null;
/** Resolve `<base href>` against the document URL; fall back to the page URL. */
export declare function documentBase(html: string, pageUrl: string): string;
/** Path-relative href (about, ../x) as opposed to root-relative, absolute, or hash. */
export declare function isPathRelativeHref(href: string): boolean;
export declare function sameOrigin(a: string, b: string): boolean;
/** Treat www and apex as the same site for crawl bounds. */
export declare function sameSite(a: string, b: string): boolean;
export declare function isLikelyTrackingImage(img: ExtractedImage): boolean;
export declare function extractPhones(text: string): string[];
export declare function extractEmails(text: string): string[];
export declare function parseJsonLd(html: string): unknown[];
export declare function businessFromJsonLd(nodes: unknown[]): {
    name?: string;
    phone?: string;
    email?: string;
    address?: string;
    city?: string;
    hours?: string[];
    url?: string;
    logoUrl?: string;
};
export declare function jsonLdServices(nodes: unknown[]): {
    name: string;
    blurb: string;
}[];
/** Builders often put their prose in divs. Keep leaf prose without duplicating
 * parent containers, headings, controls or existing semantic paragraphs. */
export declare function extractNarrativeParagraphs(html: string): string[];
export declare function extractPage(html: string, pageUrl: string): ExtractedPage;
export declare function parseSitemapLocs(xml: string): string[];
export declare function inferPrimaryCta(ctas: ExtractedCta[]): ExtractedCta | undefined;
/** Slugify a page URL into a safe copy filename stem. */
export declare function pageCopySlug(pageUrl: string, index: number): string;