locator/index.js

135 lines
4.1 KiB
JavaScript

#!/usr/bin/env node
const fs = require('fs').promises;
const process = require('process');
const puppeteer = require('puppeteer');
async function main() {
if (process.argv[2] === undefined) {
console.error("This program expects an argument.");
console.error("USAGE: locator <path-to-HTML-file>");
process.exit(1);
}
// Path to the HTML file to analyse (given as relative path from current directory)
// We need the full path so that puppeteer is able to access it
const path = process.argv[2].startsWith('/') ? process.argv[2] : process.cwd() + '/' + process.argv[2];
// Check that the file exists
try {
await fs.access(path, fs.constants.F_OK);
} catch (e) {
console.error("No such file: " + path);
process.exit(1);
}
// Size of the rendering of the web page
const size = { width: 1280, height: 720 };
// Initialize browser
const browser = await puppeteer.launch();
const page = await browser.newPage();
await page.setViewport(size);
await page.goto("file://" + path);
// This will contain all the collected information
let info = [];
// Take a first screenshot
await page.screenshot({path: __dirname + '/' + 'screenshot1.png'});
// Edit the page to shrink elements in order to get better bounding boxes
for (let selector of ["h1", "h2", "h3", "h4", "h5", "h6", "a", "img", "p", "ul", "ol", "li"]) {
let query = selector;
let shouldCreateSpan = query !== "ul" && query !== "ol" && query != "img";
if (shouldCreateSpan) {
await page.evaluate(([query]) => {
for (let e of document.querySelectorAll(query)) {
if (e.children.length === 0) {
e.innerHTML = '<span>' + e.innerHTML + '</span>';
}
}
}, [query]);
}
}
// Take another screenshot and check the modification we made didn't change the layout of the page
await page.screenshot({path: __dirname + '/' + 'screenshot2.png'});
// Compare both screenshots
let file1 = await fs.readFile(__dirname + '/' + 'screenshot1.png');
let file2 = await fs.readFile(__dirname + '/' + 'screenshot2.png');
let filesAreSame = file1.map((x, i) => x === file2[i]).reduce((a, b) => a && b, true);
// Crash if they're different
if (!filesAreSame) {
throw new Error("Page edit changed the layout");
}
for (let selector of ["h1", "h2", "h3", "h4", "h5", "h6", "a", "img", "p", "ul", "ol", "li"]) {
let query = selector;
// Shrink the elements horizontally (to be able to get better bounding boxes)
let shouldCreateSpan = query !== "ul" && query !== "ol" && query != "img";
// Query the considered element
let parents = await page.$$(query);
let elements = await page.$$(query + (shouldCreateSpan ? ' > *:first-child' : ''));
for (let index = 0; index < elements.length; index ++) {
let parent = parents[index];
let element = elements[index];
let box = await element.boundingBox();
let classElement = shouldCreateSpan ? parent : element;
let classNameAttr = await classElement.getProperty('className');
let className = await classNameAttr.jsonValue();
// Scale the bounding box
box.x /= size.width;
box.width /= size.width;
box.y /= size.height;
box.height /= size.height;
// Give the selector as type
box.type = selector;
if (className !== "") {
box.class = className;
}
info.push(box);
}
}
// Sort the info by y and the x (top to bottom, then left to right)
info.sort((a, b) => {
if (a.y < b.y || (a.y == b.y && a.x < b.x)) {
return -1;
}
if (a.y > b.y || (a.y == b.y && a.x > b.x)) {
return 1;
}
return 0;
});
// Pretty print the output info
console.log(JSON.stringify(info, undefined, 4));
await browser.close();
}
main();