locator/index.js

#!/usr/bin/env node

const fs = require('fs').promises;
const process = require('process');
const puppeteer = require('puppeteer');

async function main() {

    // 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 = [];

    for (let selector of ["h1", "h2", "h3", "h4", "h5", "h6", "a", "img", "p", "ul", "ol", "li"]) {

        let query = selector;
        if (query === "p") {
            // Skip paragraphs that have children (which means that they are not text paragraphs)
            query += ":not(:has(*))";
        }

        // Query the considered element
        let elements = await page.$$(query);

        for (let element of elements) {
            let box = await element.boundingBox();

            // 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;

            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();