Nulo/site-analyzer
1
0
Fork 0
Code Issues 5 Pull requests Projects 1 Releases Packages 1 Wiki Activity

Recurse directory instead of pulling stdin

Browse source
This commit is contained in:
Cat /dev/Nulo 2022-11-25 18:47:57 -03:00
parent 86885be850
commit d0bd26b959
1 changed files with 50 additions and 39 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

89
index.js
View file

@ -1,41 +1,39 @@
import { Parser } from "htmlparser2"; import { Parser } from "htmlparser2";
import { DomHandler } from "domhandler"; import { DomHandler } from "domhandler";
import { readdir, readFile } from "fs/promises";
import { join } from "path";
const noop = () => {}; const { argv } = process;
// const log = noop; const dirPath = argv[2] || ".";
const { log } = console;
const stdin = await readStdin(); function recurseElement(report, el) {
const rawHtml = stdin;
function recursive(el) {
if (el.name === "a") { if (el.name === "a") {
if (el.attribs.href) { if (el.attribs.href) {
checkUrl("link", el, el.attribs.href); checkUrl(report, "link", el, el.attribs.href);
} else { } else {
log("Link with no href:", getHtml(el)); report.warnings.push(`Link with no href: ${getHtml(el)}`);
} }
} }
if (["audio", "video", "img"].includes(el.name)) { if (["audio", "video", "img"].includes(el.name)) {
if (el.attribs.src) { if (el.attribs.src) {
checkUrl(el.name, el, el.attribs.src); checkUrl(report, el.name, el, el.attribs.src);
} else { } else {
log(`${el.name} with no src:`, getHtml(el)); report.warnings.push(`${el.name} with no src: ${getHtml(el)}`);
} }
} }
for (const child of el.children) { for (const child of el.children) {
if (child.type === "tag") { if (child.type === "tag") {
recursive(child); recurseElement(report, child);
} }
} }
} }
function checkUrl(type, el, url) { function checkUrl(report, type, el, url) {
if (isHttp(url)) { if (isHttp(url)) {
log(`HTTP/S ${type}:`, getHtml(el)); report.warnings.push(`HTTP/S ${type}: ${getHtml(el)}`);
} else if (isAbsolute(url)) { } else if (isAbsolute(url)) {
log(`Absolute ${type}:`, getHtml(el)); report.warnings.push(`Absolute ${type}: ${getHtml(el)}`);
} }
} }
@ -57,30 +55,43 @@ function getHtml(el) {
return text; return text;
} }
const handler = new DomHandler( function processFile(content) {
(error, dom) => {
if (error) {
// TODO: Handle error
} else {
console.time();
for (const el of dom) {
if (el.type === "tag") {
recursive(el);
}
}
console.timeEnd();
}
},
{ withEndIndices: true, withStartIndices: true }
);
const parser = new Parser(handler);
parser.parseComplete(rawHtml);
function readStdin() {
return new Promise((resolve, reject) => { return new Promise((resolve, reject) => {
let buffer = ""; const handler = new DomHandler(
process.stdin.resume(); (error, dom) => {
process.stdin.on("data", (d) => (buffer = buffer.concat(d.toString()))); if (error) {
process.stdin.on("close", () => resolve(buffer)); reject(error);
} else {
let report = {
warnings: [],
};
for (const el of dom) {
if (el.type === "tag") {
recurseElement(report, el);
}
}
resolve(report);
}
},
{ withEndIndices: true, withStartIndices: true }
);
const parser = new Parser(handler);
parser.parseComplete(content);
}); });
} }
async function recurseDirectory(path) {
const dir = await readdir(path, { withFileTypes: true });
for (const file of dir) {
const filePath = join(path, file.name);
if (file.isDirectory()) recurseDirectory(filePath);
else {
if (!file.name.endsWith(".html")) continue;
const content = await readFile(filePath, "utf-8");
console.time(filePath);
await processFile(content);
console.timeEnd(filePath);
}
}
}
await recurseDirectory(dirPath);