ARIA Landmark Auditor
Pricing
Pay per usage
Audit one public page for semantic HTML and ARIA landmark structure. Checks main, navigation, banner, contentinfo, repeated unlabeled landmarks, score, and accessibility recommendations.
Pricing
Pay per usage
Rating
0.0
(0)
Developer
Sanskar Jaiswal
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
10 days ago
Last modified
Categories
Share
node_modules/npm-debug.log*.apify/storage/FROM apify/actor-node:22
WORKDIR /usr/src/appCOPY package*.json ./RUN npm install --omit=devCOPY . ./CMD ["npm", "start"]{ "name": "aria-landmark-auditor", "version": "0.1.0", "description": "Apify Actor that audits public pages for ARIA and semantic landmark structure.", "type": "module", "private": true, "main": "src/main.js", "scripts": { "start": "node src/main.js", "test": "node --test test/*.test.js", "lint": "node --check src/main.js" }, "dependencies": { "apify": "^3.4.5" }, "devDependencies": {}}{ "actorSpecification": 1, "name": "aria-landmark-auditor", "title": "ARIA Landmark Auditor", "description": "Audit one public page for semantic HTML and ARIA landmark structure for accessibility and SEO QA.", "version": "0.1", "buildTag": "latest", "meta": { "templateId": "js-start" }, "input": "./input_schema.json", "output": "./output_schema.json", "storages": { "dataset": "./dataset_schema.json" }, "dockerfile": "./Dockerfile", "defaultRunOptions": { "build": "latest", "timeoutSecs": 60, "memoryMbytes": 512 }}{ "actorSpecification": 1, "fields": { "inputUrl": { "type": "string" }, "finalUrl": { "type": "string" }, "status": { "type": ["integer", "null"] }, "ok": { "type": "boolean" }, "landmarkCount": { "type": "integer" }, "mainCount": { "type": "integer" }, "navigationCount": { "type": "integer" }, "bannerCount": { "type": "integer" }, "contentinfoCount": { "type": "integer" }, "unlabeledRepeatedCount": { "type": "integer" }, "score": { "type": "integer" }, "grade": { "type": "string" }, "issues": { "type": "array" }, "recommendations": { "type": "array" }, "error": { "type": ["string", "null"] } }, "views": { "overview": { "title": "ARIA landmark audit", "transformation": { "fields": ["inputUrl", "finalUrl", "status", "ok", "landmarkCount", "mainCount", "navigationCount", "unlabeledRepeatedCount", "score", "grade", "issues"] }, "display": { "component": "table" } } }}{ "title": "ARIA Landmark Auditor Input", "description": "Enter one public HTTP or HTTPS page URL to audit semantic and ARIA landmarks.", "type": "object", "schemaVersion": 1, "properties": { "startUrl": { "title": "Page URL", "type": "string", "description": "Public HTTP or HTTPS page URL to audit. Credentials and private network targets are rejected.", "editor": "textfield", "prefill": "https://example.com" }, "timeoutSeconds": { "title": "Timeout seconds", "type": "integer", "description": "Request timeout for the page fetch.", "editor": "number", "default": 10, "minimum": 3, "maximum": 30 }, "maxLandmarks": { "title": "Max landmarks in output", "type": "integer", "description": "Maximum landmark details to include in the dataset item.", "editor": "number", "default": 100, "minimum": 1, "maximum": 300 } }, "required": ["startUrl"]}{ "actorOutputSchemaVersion": 1, "title": "ARIA Landmark Audit Output", "properties": { "results": { "type": "string", "title": "Audit results", "template": "{{links.apiDefaultDatasetUrl}}/items" } }}1import { Actor } from 'apify';2import dns from 'node:dns/promises';3import net from 'node:net';4import { fileURLToPath } from 'node:url';5
6const USER_AGENT = 'AriaLandmarkAuditor/0.1 (+https://apify.com)';7const MAX_BODY_BYTES = 2_000_000;8const LANDMARK_ROLES = new Set(['banner', 'complementary', 'contentinfo', 'form', 'main', 'navigation', 'region', 'search']);9const IMPLICIT_ROLES = { main: 'main', nav: 'navigation', aside: 'complementary', header: 'banner', footer: 'contentinfo' };10const REPEATED_ROLES_NEED_LABELS = new Set(['navigation', 'complementary', 'form', 'region', 'search']);11
12function isPrivateIPv4(ip) {13 const parts = ip.split('.').map(Number);14 if (parts.length !== 4 || parts.some((n) => Number.isNaN(n))) return false;15 const [a, b] = parts;16 return a === 10 || (a === 172 && b >= 16 && b <= 31) || (a === 192 && b === 168) || a === 127 || a === 0 || (a === 169 && b === 254);17}18
19function isPrivateIPv6(ip) {20 const normalized = ip.toLowerCase();21 return normalized === '::1' || normalized.startsWith('fc') || normalized.startsWith('fd') || normalized.startsWith('fe80:');22}23
24export async function normalizeAndValidateUrl(rawUrl) {25 if (!rawUrl || typeof rawUrl !== 'string') throw new Error('startUrl is required');26 if (/^[a-z][a-z0-9+.-]*:/i.test(rawUrl) && !/^https?:\/\//i.test(rawUrl)) throw new Error('Only HTTP and HTTPS URLs are supported');27
28 const withScheme = /^https?:\/\//i.test(rawUrl) ? rawUrl : `https://${rawUrl}`;29 const url = new URL(withScheme);30 if (!['http:', 'https:'].includes(url.protocol)) throw new Error('Only HTTP and HTTPS URLs are supported');31 if (!url.hostname || url.username || url.password) throw new Error('URL must be public and must not include credentials');32
33 const literalType = net.isIP(url.hostname);34 if (literalType === 4 && isPrivateIPv4(url.hostname)) throw new Error('Private IPv4 targets are blocked');35 if (literalType === 6 && isPrivateIPv6(url.hostname)) throw new Error('Private IPv6 targets are blocked');36
37 const records = literalType ? [{ address: url.hostname, family: literalType }] : await dns.lookup(url.hostname, { all: true });38 for (const record of records) {39 if (record.family === 4 && isPrivateIPv4(record.address)) throw new Error('DNS resolves to a private IPv4 address; blocked for SSRF safety');40 if (record.family === 6 && isPrivateIPv6(record.address)) throw new Error('DNS resolves to a private IPv6 address; blocked for SSRF safety');41 }42 return url;43}44
45function clampInteger(value, fallback, min, max) {46 const parsed = Number(value);47 if (!Number.isFinite(parsed)) return fallback;48 return Math.min(Math.max(Math.trunc(parsed), min), max);49}50
51function getAttr(tag, name) {52 const match = tag.match(new RegExp(`\\s${name}\\s*=\\s*("([^"]*)"|'([^']*)'|([^\\s>]+))`, 'i'));53 return match ? (match[2] ?? match[3] ?? match[4] ?? '').trim() : null;54}55
56function hasAttr(tag, name) {57 return new RegExp(`\\s${name}(\\s|=|>|/)`, 'i').test(tag);58}59
60export function parseLandmarks(html, maxLandmarks = 100) {61 const landmarks = [];62 for (const match of html.matchAll(/<([a-z][a-z0-9-]*)\b[^>]*>/gi)) {63 const tag = match[0];64 const tagName = match[1].toLowerCase();65 if (tag.startsWith('</') || hasAttr(tag, 'hidden') || getAttr(tag, 'aria-hidden') === 'true') continue;66
67 const explicitRole = (getAttr(tag, 'role') || '').toLowerCase().split(/\s+/).find((role) => LANDMARK_ROLES.has(role));68 const role = explicitRole || IMPLICIT_ROLES[tagName];69 if (!role) continue;70
71 const label = getAttr(tag, 'aria-label') || getAttr(tag, 'aria-labelledby') || null;72 landmarks.push({73 tag: tagName,74 role,75 explicitRole: explicitRole || null,76 label,77 id: getAttr(tag, 'id') || null,78 className: getAttr(tag, 'class') || null,79 });80 }81
82 const counts = Object.fromEntries([...LANDMARK_ROLES].map((role) => [role, landmarks.filter((item) => item.role === role).length]));83 const unlabeledRepeated = landmarks.filter((item) => REPEATED_ROLES_NEED_LABELS.has(item.role) && counts[item.role] > 1 && !item.label);84 return {85 landmarkCount: landmarks.length,86 mainCount: counts.main,87 navigationCount: counts.navigation,88 bannerCount: counts.banner,89 contentinfoCount: counts.contentinfo,90 complementaryCount: counts.complementary,91 searchCount: counts.search,92 formLandmarkCount: counts.form,93 regionCount: counts.region,94 unlabeledRepeatedCount: unlabeledRepeated.length,95 landmarks: landmarks.slice(0, maxLandmarks),96 unlabeledRepeatedLandmarks: unlabeledRepeated.slice(0, maxLandmarks),97 };98}99
100function scoreLandmarks({ parsed, status, error }) {101 const issues = [];102 const recommendations = [];103 let score = 100;104
105 if (error) return { score: 0, grade: 'F', issues: [error], recommendations: ['Verify the URL is public, reachable, and returns HTML.'] };106 if (!status || status >= 400) {107 score -= 40;108 issues.push(`HTTP status is ${status || 'unknown'}`);109 recommendations.push('Audit a live 2xx HTML page.');110 }111 if (parsed.landmarkCount === 0) {112 score -= 40;113 issues.push('No semantic or ARIA landmarks found');114 recommendations.push('Add semantic landmarks such as main, nav, header, footer, and aside.');115 }116 if (parsed.mainCount === 0) {117 score -= 25;118 issues.push('No main landmark found');119 recommendations.push('Add one main element or role="main" landmark for primary content.');120 } else if (parsed.mainCount > 1) {121 score -= 25;122 issues.push(`${parsed.mainCount} main landmarks found`);123 recommendations.push('Keep exactly one main landmark on the page.');124 }125 if (parsed.unlabeledRepeatedCount > 0) {126 score -= Math.min(30, parsed.unlabeledRepeatedCount * 10);127 issues.push(`${parsed.unlabeledRepeatedCount} repeated landmark(s) lack accessible labels`);128 recommendations.push('Add aria-label or aria-labelledby to repeated navigation, search, form, complementary, or region landmarks.');129 }130 if (parsed.bannerCount > 1) {131 score -= 10;132 issues.push(`${parsed.bannerCount} banner landmarks found`);133 recommendations.push('Use one top-level header/banner landmark unless nested context requires otherwise.');134 }135 if (parsed.contentinfoCount > 1) {136 score -= 10;137 issues.push(`${parsed.contentinfoCount} contentinfo landmarks found`);138 recommendations.push('Use one top-level footer/contentinfo landmark unless nested context requires otherwise.');139 }140
141 const bounded = Math.max(0, score);142 const grade = bounded >= 90 ? 'A' : bounded >= 75 ? 'B' : bounded >= 60 ? 'C' : bounded >= 45 ? 'D' : 'F';143 return { score: bounded, grade, issues: [...new Set(issues)], recommendations: [...new Set(recommendations)] };144}145
146async function fetchHtml(url, timeoutSeconds) {147 await normalizeAndValidateUrl(url.href);148 const controller = new AbortController();149 const timeout = setTimeout(() => controller.abort(), timeoutSeconds * 1000);150 try {151 const response = await fetch(url, { redirect: 'manual', signal: controller.signal, headers: { 'user-agent': USER_AGENT, accept: 'text/html,*/*;q=0.1' } });152 const location = response.headers.get('location');153 if (location && response.status >= 300 && response.status < 400) {154 const next = new URL(location, url.href);155 await normalizeAndValidateUrl(next.href);156 return fetchHtml(next, timeoutSeconds);157 }158 return { status: response.status, finalUrl: url.href, contentType: response.headers.get('content-type') || '', body: (await response.text()).slice(0, MAX_BODY_BYTES) };159 } finally {160 clearTimeout(timeout);161 }162}163
164export async function auditAriaLandmarks(input) {165 const startUrl = await normalizeAndValidateUrl(input.startUrl);166 const timeoutSeconds = clampInteger(input.timeoutSeconds, 10, 3, 30);167 const maxLandmarks = clampInteger(input.maxLandmarks, 100, 1, 300);168 const checkedAt = new Date().toISOString();169 let fetched = null;170 let error = null;171
172 try { fetched = await fetchHtml(startUrl, timeoutSeconds); } catch (caught) { error = caught.message; }173
174 const parsed = parseLandmarks(fetched?.body || '', maxLandmarks);175 const scored = scoreLandmarks({ parsed, status: fetched?.status, error });176 return {177 inputUrl: input.startUrl,178 normalizedInputUrl: startUrl.href,179 finalUrl: fetched?.finalUrl || startUrl.href,180 status: fetched?.status || null,181 ok: !error && fetched?.status >= 200 && fetched?.status < 400 && scored.score >= 90,182 checkedAt,183 ...parsed,184 score: scored.score,185 grade: scored.grade,186 issues: scored.issues,187 recommendations: scored.recommendations,188 error,189 };190}191
192const isExecutedDirectly = process.argv[1] && fileURLToPath(import.meta.url) === process.argv[1];193
194if (process.env.NODE_ENV !== 'test' && isExecutedDirectly) {195 await Actor.init();196 try {197 const input = await Actor.getInput();198 const result = await auditAriaLandmarks(input || {});199 await Actor.pushData(result);200 await Actor.setValue('OUTPUT', result);201 Actor.log.info('ARIA landmark audit complete', { finalUrl: result.finalUrl, score: result.score, grade: result.grade });202 } finally {203 await Actor.exit();204 }205}1import assert from 'node:assert/strict';2import { test } from 'node:test';3import { auditAriaLandmarks, normalizeAndValidateUrl, parseLandmarks } from '../src/main.js';4
5test('blocks non-public and credentialed targets', async () => {6 await assert.rejects(() => normalizeAndValidateUrl('ftp://example.com'), /Only HTTP/);7 await assert.rejects(() => normalizeAndValidateUrl('https://user:pass@example.com'), /credentials/);8 await assert.rejects(() => normalizeAndValidateUrl('http://127.0.0.1'), /Private IPv4/);9});10
11test('parses semantic and explicit landmarks', () => {12 const parsed = parseLandmarks(`13 <header>Top</header>14 <nav aria-label="Primary"></nav>15 <nav></nav>16 <main id="content"></main>17 <aside role="complementary" aria-label="Related"></aside>18 <div role="search"></div>19 <footer></footer>20 `);21 assert.equal(parsed.landmarkCount, 7);22 assert.equal(parsed.mainCount, 1);23 assert.equal(parsed.navigationCount, 2);24 assert.equal(parsed.bannerCount, 1);25 assert.equal(parsed.contentinfoCount, 1);26 assert.equal(parsed.searchCount, 1);27 assert.equal(parsed.unlabeledRepeatedCount, 1);28});29
30test('audits missing and duplicate main landmarks', async () => {31 const originalFetch = globalThis.fetch;32 globalThis.fetch = async () => new Response('<header></header><nav></nav><nav></nav><main></main><main></main><footer></footer>', { status: 200, headers: { 'content-type': 'text/html' } });33
34 try {35 const result = await auditAriaLandmarks({ startUrl: 'https://example.com/page' });36 assert.equal(result.mainCount, 2);37 assert.equal(result.unlabeledRepeatedCount, 2);38 assert.equal(result.ok, false);39 assert.match(result.issues.join(' '), /main landmarks/);40 assert.equal(result.grade, 'D');41 } finally {42 globalThis.fetch = originalFetch;43 }44});