@oh-my-pi/pi-utils 17.2.8 → 17.2.10

package/src/readability/readerable.ts

@@ -0,0 +1,51 @@
+/** Behavior-compatible reimplementation of @mozilla/readability's used surface. */
+
+import type { ReadabilityDocument, ReadabilityElement, ReadabilityNode } from "./types";
+
+const UNLIKELY =
+	/-ad-|ai2html|banner|breadcrumbs|combx|comment|community|cover-wrap|disqus|extra|footer|gdpr|header|legends|menu|related|remark|replies|rss|shoutbox|sidebar|skyscraper|social|sponsor|supplemental|ad-break|agegate|pagination|pager|popup|yom-remote/i;
+const POSSIBLE = /and|article|body|column|content|main|shadow/i;
+
+/** Options for the inexpensive readerability estimate. */
+export interface ReaderableOptions {
+	minContentLength?: number;
+	minScore?: number;
+	visibilityChecker?: (node: ReadabilityNode) => boolean;
+}
+
+/** Estimates whether a document contains enough prose for article extraction. */
+export function isProbablyReaderable(
+	document: ReadabilityDocument,
+	options: ReaderableOptions | ((node: ReadabilityNode) => boolean) = {},
+): boolean {
+	const resolved = typeof options === "function" ? { visibilityChecker: options } : options;
+	const minLength = resolved.minContentLength ?? 140;
+	const minScore = resolved.minScore ?? 20;
+	const visible =
+		resolved.visibilityChecker ??
+		((node: ReadabilityNode) => {
+			const element = node as ReadabilityElement;
+			const style = element.getAttribute("style")?.toLowerCase() ?? "";
+			return (
+				!element.hasAttribute("hidden") &&
+				element.getAttribute("aria-hidden") !== "true" &&
+				!/display\s*:\s*none/.test(style)
+			);
+		});
+	const candidates = new Set<ReadabilityElement>(Array.from(document.querySelectorAll("p, pre, article")));
+	for (const br of Array.from(document.querySelectorAll("div > br"))) {
+		if (br.parentNode) candidates.add(br.parentNode as ReadabilityElement);
+	}
+	let score = 0;
+	for (const node of candidates) {
+		if (!visible(node)) continue;
+		const label = `${node.className} ${node.id}`;
+		if (UNLIKELY.test(label) && !POSSIBLE.test(label)) continue;
+		if ((node.parentNode as ReadabilityElement | null)?.tagName === "LI" && node.tagName === "P") continue;
+		const length = (node.textContent ?? "").trim().length;
+		if (length < minLength) continue;
+		score += Math.sqrt(length - minLength);
+		if (score > minScore) return true;
+	}
+	return false;
+}

package/src/readability/types.ts

@@ -0,0 +1,72 @@
+/** Behavior-compatible reimplementation of @mozilla/readability's used surface. */
+
+/** Minimal standards-shaped DOM node accepted by the readability extractor. */
+export interface ReadabilityNode {
+	readonly nodeType: number;
+	readonly nodeName: string;
+	readonly tagName?: string;
+	readonly ownerDocument?: ReadabilityDocument | null;
+	parentNode: ReadabilityNode | null;
+	readonly children?: ArrayLike<ReadabilityElement>;
+	readonly childNodes: ArrayLike<ReadabilityNode>;
+	readonly firstChild: ReadabilityNode | null;
+	readonly textContent: string | null;
+	appendChild(node: ReadabilityNode): ReadabilityNode;
+	cloneNode(deep?: boolean): ReadabilityNode;
+	remove(): void;
+}
+
+/** Minimal standards-shaped DOM element accepted by the readability extractor. */
+export interface ReadabilityElement extends ReadabilityNode {
+	readonly children: ArrayLike<ReadabilityElement>;
+	readonly tagName: string;
+	id: string;
+	className: string;
+	innerHTML: string;
+	readonly attributes: ArrayLike<{ readonly name: string; readonly value: string }>;
+	getAttribute(name: string): string | null;
+	hasAttribute(name: string): boolean;
+	setAttribute(name: string, value: string): void;
+	removeAttribute(name: string): void;
+	getElementsByTagName(name: string): ArrayLike<ReadabilityElement>;
+	querySelector(selector: string): ReadabilityElement | null;
+	querySelectorAll(selector: string): ArrayLike<ReadabilityElement>;
+}
+
+/** Minimal standards-shaped document accepted by the readability extractor. */
+export interface ReadabilityDocument extends ReadabilityNode {
+	title: string;
+	readonly body: ReadabilityElement | null;
+	readonly documentElement: ReadabilityElement | null;
+	createElement(name: string): ReadabilityElement;
+	getElementsByTagName(name: string): ArrayLike<ReadabilityElement>;
+	querySelector(selector: string): ReadabilityElement | null;
+	querySelectorAll(selector: string): ArrayLike<ReadabilityElement>;
+}
+
+/** Configuration for article extraction. */
+export interface ReadabilityOptions<T = string> {
+	debug?: boolean;
+	maxElemsToParse?: number;
+	nbTopCandidates?: number;
+	charThreshold?: number;
+	classesToPreserve?: string[];
+	keepClasses?: boolean;
+	serializer?: (node: ReadabilityNode) => T;
+	disableJSONLD?: boolean;
+	allowedVideoRegex?: RegExp;
+}
+
+/** Extracted article and metadata. */
+export interface ReadabilityArticle<T = string> {
+	title: string | null | undefined;
+	content: T | null | undefined;
+	textContent: string | null | undefined;
+	length: number | null | undefined;
+	excerpt: string | null | undefined;
+	byline: string | null | undefined;
+	dir: string | null | undefined;
+	siteName: string | null | undefined;
+	lang: string | null | undefined;
+	publishedTime: string | null | undefined;
+}

package/src/readability.ts

@@ -0,0 +1,11 @@
+/** Behavior-compatible reimplementation of @mozilla/readability's used surface. */
+
+export { Readability } from "./readability/readability";
+export { isProbablyReaderable, type ReaderableOptions } from "./readability/readerable";
+export type {
+	ReadabilityArticle,
+	ReadabilityDocument,
+	ReadabilityElement,
+	ReadabilityNode,
+	ReadabilityOptions,
+} from "./readability/types";