1

Uniformisation de la vérification des liens

This commit is contained in:
2025-11-28 23:43:07 +01:00
parent 5e846aa4b4
commit 0260c1ab4e
7 changed files with 2145 additions and 1157 deletions

View File

@@ -6,6 +6,7 @@ const crypto = require("crypto");
const path = require("path");
const os = require("os");
const YAML = require("yaml");
const { buildUserAgent, checkUrl } = require("./lib/http");
const { getArchiveUrl, saveToArchive } = require("./lib/archive");
const { scrapePage } = require("./lib/puppeteer");
@@ -59,6 +60,14 @@ if (duplicateBundlePath) {
// Check URL accessibility and Archive.org availability
(async () => {
const userAgent = buildUserAgent();
const initialCheck = await checkUrl(url, { userAgent, timeoutMs: 8000 });
if (initialCheck.errorType || (typeof initialCheck.status === "number" && initialCheck.status >= 400)) {
console.warn(`⚠ Vérification HTTP avant scraping: ${initialCheck.errorType || initialCheck.status || "indéterminé"}`);
} else {
console.log(`🌐 Vérification HTTP avant scraping: ${initialCheck.status ?? "inconnue"}`);
}
let archiveUrl = await getArchiveUrl(url);
// If the URL is not archived, attempt to save it
@@ -106,13 +115,16 @@ if (duplicateBundlePath) {
// Scrape the page and capture a screenshot
console.log(`🔍 Scraping page and capturing screenshot...`);
const metadata = await scrapePage(url, tempScreenshotPath);
const metadata = await scrapePage(url, tempScreenshotPath, { userAgent });
// If Puppeteer failed, do not proceed
if (!metadata || !fs.existsSync(tempScreenshotPath)) {
console.error(`❌ Scraping failed. No bundle will be created.`);
process.exit(1);
}
if (!metadata.httpStatus && typeof initialCheck.status === "number") {
metadata.httpStatus = initialCheck.status;
}
// Create Hugo bundle only if scraping succeeded
console.log(`📦 Creating Hugo bundle for ${url}...`);

View File

@@ -2,17 +2,13 @@
const fs = require("fs");
const path = require("path");
const util = require("util");
const yaml = require("js-yaml");
const UserAgent = require("user-agents");
const { execFile } = require("child_process");
const { buildUserAgent, probeUrl, shouldRetry } = require("./lib/http");
const {
collectMarkdownLinksFromFile,
extractLinksFromText,
} = require("./lib/markdown_links");
const execFileAsync = util.promisify(execFile);
const SITE_ROOT = path.resolve(__dirname, "..");
const CONTENT_DIR = path.join(SITE_ROOT, "content");
const CONFIG_PATH = path.join(__dirname, "config", "config.json");
@@ -29,9 +25,8 @@ const DEFAULT_CONFIG = {
cacheTtlServerErrorDays: 1,
cacheTtlTimeoutDays: 7,
maxConcurrentHosts: 4,
maxRedirects: 5,
userAgent: null,
enableCookies: true,
cookieJar: path.join(__dirname, "cache", "curl_cookies.txt"),
};
function loadConfig() {
@@ -60,26 +55,24 @@ const CACHE_DIR = path.isAbsolute(settings.cacheDir)
const REPORT_PATH = path.isAbsolute(settings.cacheFile)
? settings.cacheFile
: path.join(CACHE_DIR, settings.cacheFile);
const COOKIE_JAR = settings.cookieJar
? path.isAbsolute(settings.cookieJar)
? settings.cookieJar
: path.resolve(SITE_ROOT, settings.cookieJar)
: path.join(CACHE_DIR, "curl_cookies.txt");
const HOST_DELAY_MS = Math.max(0, Number(settings.hostDelayMs) || 0);
const RETRY_DELAY_MS = Math.max(0, Number(settings.retryDelayMs) || 0);
const REQUEST_TIMEOUT_SECONDS = Math.max(1, Number(settings.requestTimeoutSeconds) || 5);
const REQUEST_TIMEOUT_MS = REQUEST_TIMEOUT_SECONDS * 1000;
const MAX_CONCURRENT_HOSTS = Math.max(
1,
Number.isFinite(Number(settings.maxConcurrentHosts))
? Number(settings.maxConcurrentHosts)
: DEFAULT_CONFIG.maxConcurrentHosts
);
const DEFAULT_USER_AGENT =
typeof settings.userAgent === "string" && settings.userAgent.trim()
? settings.userAgent.trim()
: new UserAgent().toString();
const ENABLE_COOKIES = settings.enableCookies !== false;
const MAX_REDIRECTS = Math.max(
0,
Number.isFinite(Number(settings.maxRedirects))
? Number(settings.maxRedirects)
: DEFAULT_CONFIG.maxRedirects
);
const DEFAULT_USER_AGENT = buildUserAgent(settings.userAgent);
const CACHE_TTL_SUCCESS_MS = daysToMs(
pickNumber(settings.cacheTtlSuccessDays, DEFAULT_CONFIG.cacheTtlSuccessDays)
@@ -95,12 +88,12 @@ const CACHE_TTL_TIMEOUT_MS = daysToMs(
);
fs.mkdirSync(CACHE_DIR, { recursive: true });
if (ENABLE_COOKIES) {
fs.mkdirSync(path.dirname(COOKIE_JAR), { recursive: true });
if (!fs.existsSync(COOKIE_JAR)) {
fs.closeSync(fs.openSync(COOKIE_JAR, "a"));
}
}
const BASE_HTTP_OPTIONS = {
userAgent: DEFAULT_USER_AGENT,
timeoutMs: REQUEST_TIMEOUT_MS,
maxRedirects: MAX_REDIRECTS,
};
function pickNumber(value, fallback) {
const parsed = Number(value);
@@ -536,59 +529,6 @@ function extractHost(url) {
}
}
async function curlRequest(url, method, hostHeader) {
const args = [
"--silent",
"--location",
"--fail",
"--max-time",
`${REQUEST_TIMEOUT_SECONDS}`,
"--output",
"/dev/null",
"--write-out",
"%{http_code}",
"--user-agent",
DEFAULT_USER_AGENT,
"--request",
method,
];
if (ENABLE_COOKIES) {
args.push("--cookie", COOKIE_JAR, "--cookie-jar", COOKIE_JAR);
}
if (hostHeader) {
args.push("-H", `Host: ${hostHeader}`);
}
args.push(url);
try {
const { stdout } = await execFileAsync("curl", args);
const status = parseInt(stdout.trim(), 10);
return {
status: Number.isNaN(status) ? null : status,
errorType: null,
method: method.toUpperCase(),
};
} catch (error) {
const rawStatus = error?.stdout?.toString().trim();
const status = rawStatus ? parseInt(rawStatus, 10) : null;
const errorCode = Number(error?.code);
const timeout = errorCode === 28 ? "timeout" : null;
return {
status: Number.isNaN(status) ? null : status,
errorType: timeout,
method: method.toUpperCase(),
};
}
}
function shouldRetryWithGet(result) {
if (!result) return true;
if (result.errorType) return true;
if (typeof result.status !== "number") return true;
return result.status >= 400;
}
function getTtlMs(entry) {
if (!entry) return 0;
if (entry.errorType === "timeout" || entry.status === 0 || entry.status === null) {
@@ -748,17 +688,16 @@ async function checkEntries(entriesToCheck, entries, snapshotMeta) {
if (host) {
await applyHostDelay(host);
}
const hostHeader = host || extractHost(entry.url);
let result = await curlRequest(entry.url, "HEAD", hostHeader);
let result = await probeUrl(entry.url, { ...BASE_HTTP_OPTIONS, method: "HEAD" });
recordHostCheck(host);
if (shouldRetryWithGet(result)) {
if (shouldRetry(result)) {
if (RETRY_DELAY_MS > 0) {
await delay(RETRY_DELAY_MS);
}
if (host) {
await applyHostDelay(host);
}
result = await curlRequest(entry.url, "GET", hostHeader);
result = await probeUrl(entry.url, { ...BASE_HTTP_OPTIONS, method: "GET" });
recordHostCheck(host);
}
updateEntryWithResult(entries[entry.url], result);

133
tools/lib/http.js Normal file
View File

@@ -0,0 +1,133 @@
const { fetch } = require("undici");
const UserAgent = require("user-agents");
const DEFAULT_ACCEPT =
"text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,*/*;q=0.8";
const DEFAULT_ACCEPT_LANGUAGE = "fr-FR,fr;q=0.9,en;q=0.7";
const DEFAULT_TIMEOUT_MS = 5000;
const DEFAULT_MAX_REDIRECTS = 5;
function buildUserAgent(preferred) {
if (typeof preferred === "string" && preferred.trim()) {
return preferred.trim();
}
const ua = new UserAgent();
return ua.toString();
}
async function fetchWithRedirects(targetUrl, options, maxRedirects) {
let currentUrl = targetUrl;
let response = null;
let redirects = 0;
while (redirects <= maxRedirects) {
response = await fetch(currentUrl, { ...options, redirect: "manual" });
const location = response.headers.get("location");
if (
response.status >= 300 &&
response.status < 400 &&
location &&
redirects < maxRedirects
) {
if (response.body && typeof response.body.cancel === "function") {
try {
await response.body.cancel();
} catch (_) {
// Ignore cancellation errors; we're moving to the next hop.
}
}
currentUrl = new URL(location, currentUrl).toString();
redirects += 1;
continue;
}
break;
}
return response;
}
async function probeUrl(url, options = {}) {
const method = typeof options.method === "string" ? options.method.toUpperCase() : "HEAD";
const timeoutMs = Number.isFinite(options.timeoutMs) ? options.timeoutMs : DEFAULT_TIMEOUT_MS;
const maxRedirects = Number.isFinite(options.maxRedirects)
? options.maxRedirects
: DEFAULT_MAX_REDIRECTS;
const userAgent = buildUserAgent(options.userAgent);
const headers = {
"user-agent": userAgent,
"accept-language": DEFAULT_ACCEPT_LANGUAGE,
"accept": DEFAULT_ACCEPT,
...(options.headers || {}),
};
const controller = new AbortController();
const timer = setTimeout(() => controller.abort(), timeoutMs);
try {
const response = await fetchWithRedirects(
url,
{
method,
headers,
signal: controller.signal,
},
maxRedirects
);
const status = response ? response.status : null;
const finalUrl = response?.url || url;
if (response?.body && typeof response.body.cancel === "function") {
try {
await response.body.cancel();
} catch (_) {
// Ignore cancellation errors; the status is all we needed.
}
}
return {
status,
finalUrl,
method,
errorType: null,
};
} catch (error) {
if (error.name === "AbortError") {
return {
status: null,
finalUrl: url,
method,
errorType: "timeout",
};
}
return {
status: null,
finalUrl: url,
method,
errorType: "network",
message: error.message,
};
} finally {
clearTimeout(timer);
}
}
function shouldRetry(result) {
if (!result) return true;
if (result.errorType) return true;
if (typeof result.status !== "number") return true;
return result.status >= 400;
}
async function checkUrl(url, options = {}) {
const firstMethod = options.firstMethod || "HEAD";
let result = await probeUrl(url, { ...options, method: firstMethod });
if (options.retryWithGet !== false && shouldRetry(result)) {
result = await probeUrl(url, { ...options, method: "GET" });
}
return result;
}
module.exports = {
buildUserAgent,
checkUrl,
probeUrl,
shouldRetry,
};

View File

@@ -1,6 +1,6 @@
const puppeteer = require("puppeteer-extra");
const StealthPlugin = require("puppeteer-extra-plugin-stealth");
const UserAgent = require("user-agents");
const { buildUserAgent } = require("./http");
puppeteer.use(StealthPlugin());
@@ -8,9 +8,11 @@ puppeteer.use(StealthPlugin());
* Scrape a webpage to extract metadata and take a screenshot.
* @param {string} url - The URL of the page to scrape.
* @param {string} screenshotPath - Path where the screenshot should be saved.
* @param {object} options
* @param {string} [options.userAgent] - Optional user agent to use for the session.
* @returns {Promise<object>} - Metadata including title, description, keywords, language, and HTTP status.
*/
async function scrapePage(url, screenshotPath) {
async function scrapePage(url, screenshotPath, options = {}) {
console.log(`🔍 Scraping: ${url}`);
const browser = await puppeteer.launch({
@@ -35,9 +37,8 @@ async function scrapePage(url, screenshotPath) {
const page = await browser.newPage();
// Generate a fresh, realistic user-agent
const userAgent = new UserAgent();
await page.setUserAgent(userAgent.toString());
const userAgent = buildUserAgent(options.userAgent);
await page.setUserAgent(userAgent);
// Add headers to simulate a real browser
await page.setExtraHTTPHeaders({