seo 0.2.2 → 0.2.4

package/dist/index.js

@@ -1,5 +1,5 @@
 import { createRequire as __seoCreateRequire } from 'node:module'; const require = __seoCreateRequire(import.meta.url);
-import { n as load, t as version } from "./package-uOaBUbds.js";
+import { n as load, t as version } from "./package-D6bCOqrD.js";
 import { Headers, fetch as fetch$1 } from "undici";
 import { chmodSync, createWriteStream, existsSync, mkdirSync, readFileSync, renameSync, rmSync, statSync, writeFileSync } from "node:fs";
 import { homedir } from "node:os";
@@ -6694,6 +6694,33 @@ function sameDocument(left, right) {
 function bodyWordCount(value) {
 	return value.replace(/^---[\s\S]*?---\s*/u, "").replace(/```[\s\S]*?```/gu, " ").replace(/[#>*_`[\]()|:-]/gu, " ").split(/\s+/u).filter(Boolean).length;
 }
+function repeatedProseLines(markdown) {
+	const counts = /* @__PURE__ */ new Map();
+	let inFrontmatter = false;
+	let fence;
+	for (const [index, rawLine] of markdown.split(/\r?\n/u).entries()) {
+		const line = rawLine.trim();
+		if (index === 0 && line === "---") {
+			inFrontmatter = true;
+			continue;
+		}
+		if (inFrontmatter) {
+			if (line === "---") inFrontmatter = false;
+			continue;
+		}
+		if (!fence && (line.startsWith("```") || line.startsWith("~~~"))) {
+			fence = line.startsWith("```") ? "```" : "~~~";
+			continue;
+		}
+		if (fence) {
+			if (line.startsWith(fence)) fence = void 0;
+			continue;
+		}
+		if (line.length < 40 || /^(?:#{1,6}\s|[-+*]\s|\d+[.)]\s|>|\||`|\]\()/u.test(line)) continue;
+		counts.set(line, (counts.get(line) ?? 0) + 1);
+	}
+	return [...counts.values()].reduce((duplicates, count) => duplicates + Math.max(0, count - 1), 0);
+}
 function responseIsNoindex(response, body) {
 	if ((response.headers.get("x-robots-tag") ?? "").toLowerCase().includes("noindex")) return true;
 	return [...body.matchAll(/<meta\s+[^>]*>/giu)].some((match) => {
@@ -6707,7 +6734,6 @@ function markdownQuality(markdown, page) {
 	const introProbe = (page.contentSample?.replace(/\s+/gu, " ").trim())?.split(" ").slice(0, 10).join(" ");
 	const normalizedMarkdown = markdown.replace(/\s+/gu, " ");
 	const codeFences = markdown.match(/^```/gmu)?.length ?? 0;
-	const repeatedLines = [...markdown.matchAll(/^(.{12,})$/gmu)].map((match) => match[1]?.trim()).filter((line) => Boolean(line)).filter((line, index, lines) => lines.indexOf(line) !== index).length;
 	const wordCount = bodyWordCount(markdown);
 	const sourceWordCount = page.wordCount ?? 0;
 	return {
@@ -6722,7 +6748,7 @@ function markdownQuality(markdown, page) {
 		rawScriptTags: markdown.match(/<script\b/giu)?.length ?? 0,
 		rawStyleTags: markdown.match(/<style\b/giu)?.length ?? 0,
 		suspiciousConcatenations: markdown.match(/\b[a-z]{4,}[A-Z][A-Za-z]{3,}\b/gu)?.length ?? 0,
-		repeatedLines,
+		repeatedLines: repeatedProseLines(markdown),
 		sourceWordCount,
 		wordRetentionRatio: sourceWordCount > 0 ? Math.round(wordCount / sourceWordCount * 1e3) / 1e3 : null,
 		introductoryCopyRetained: introProbe ? normalizedMarkdown.includes(introProbe) : null,
@@ -7270,7 +7296,7 @@ function qualityFailureUrls(discovery) {
 	return discovery.markdownAlternates.pages.filter((page) => {
 		const quality = page.quality;
 		if (!quality) return true;
-		return !quality.frontmatterTitle || quality.h1Count !== 1 || !quality.codeFenceBalanced || quality.rawHtmlTags > 0 || quality.rawSvgTags > 0 || quality.rawScriptTags > 0 || quality.rawStyleTags > 0 || quality.navigationOnly || quality.repeatedLines > 0 || quality.wordRetentionRatio !== null && quality.wordRetentionRatio < .4 || quality.introductoryCopyRetained === false;
+		return !quality.frontmatterTitle || quality.h1Count !== 1 || !quality.codeFenceBalanced || quality.rawHtmlTags > 0 || quality.rawSvgTags > 0 || quality.rawScriptTags > 0 || quality.rawStyleTags > 0 || quality.navigationOnly || quality.repeatedLines > 0 || quality.wordRetentionRatio !== null && quality.wordRetentionRatio < .4;
 	}).map((page) => page.htmlUrl);
 }
 function profileChecks(discovery) {