benchmark-app.js

const puppeteer = require('puppeteer');
const fs = require('fs');

// modified from https://github.com/epiviz/epiviz-chart/blob/master/performance/benchmarks.js
// const files = ["zeisel.matrix.mtx.gz"];
const datasets = ["zeisel-brain", 
    // "zilionis-lung",
    "paul-hsc", "ernst-sperm", "bacher-tcell",
    "bach-mammary"
];

const runs = 3;

async function track_metrics(datasets) {

    // headless: true to run in headless mode, otherwise this will 
    // open a browser and perform the actions. 
    // also requires shared array buffers to be enabled.
    let launchOptions = {
        headless: true,
        args: ['--start-maximized', "--enable-features=SharedArrayBuffer"]
    };

    const browser = await puppeteer.launch(launchOptions);

    for (let ds of datasets) {
        console.log("dataset: ", ds);

        for (let i = 0; i < runs; i++) {

            console.log("run: ", i);

            const page = await browser.newPage();
            var analysisTime;

            let path = `./data/${ds}`;
        
            // print all log messages; one of the logs is the analysis time
            // page.on('console', msg => { 
            //     console.log('PAGE LOG:', msg.text()) 
            // });
        
            // set viewport and user agent (just in case for nice viewing)
            await page.setViewport({ width: 1200, height: 800 });
        
            // whatever port serve says when you run
            await page.goto('http://localhost:3000');
        
            // App starts with the intro modal, continue to upload dataset
            await page.waitForSelector('#introSubmit');
            await page.waitForTimeout(100);
            await page.evaluate(() => document.getElementById('introSubmit').click());
        
            // Choose the 10x tab
            // await page.evaluate(() => document.getElementById('bp3-tab-title_undefined_tenx').click());
            // await page.waitForTimeout(100);
        
            // Choose the input element and upload a sample H5 file
            const mtx_input_sel = "#bp3-tab-panel_undefined_mtx > div > label:nth-child(1) > label > input[type=file]";
            // const h5_input_sel = "#bp3-tab-panel_undefined_tenx > div > label > label > input[type=file]"
            const gene_input_sel = "#bp3-tab-panel_undefined_mtx > div > label:nth-child(2) > label > input[type=file]";
            
            const inputUploadHandle = await page.$(mtx_input_sel);
            inputUploadHandle.uploadFile(`${path}/matrix.mtx.gz`);
        
            const geneUploadHandle = await page.$(gene_input_sel);
            geneUploadHandle.uploadFile(`${path}/features.tsv.gz`);
        
            await page.waitForTimeout(100);
        
            // Start the analysis
            await page.evaluate(() => document.getElementById('analysisSubmit').click());
        
            async function waitForEvent(page, eventName, seconds) {
                return Promise.race([
                    page.evaluate(eventName =>
                        new Promise((resolve) => {
                            document.addEventListener(eventName, (e) => {
                                resolve(JSON.stringify(e.detail));
                            }, { once: true });
                        }), eventName),
        
                    page.waitForTimeout(seconds * 1000)
                ]);
            }
        
            console.log("waiting for the kana-done event...");
            let resp = await waitForEvent(page, "kana-done", 50000);
            analysisTime = JSON.parse(resp)["total_time"];
            console.log("done!!!");
        
            console.log("capture session metrics");
            let metrics = await page.metrics();
            metrics["CustomTotalAnalysisTime"] = analysisTime;
            // console.log(JSON.stringify(metrics));
        
            fs.writeFileSync(`./app-results/${ds}_${i}_metrics.json`, JSON.stringify(metrics));
        
            console.log("capture screenshot before closing the browser");
            await page.screenshot({ path: `./app-results/${ds}_${i}_screenshot.png` });
            console.log(`screenshot stored in ./app-results/${ds}_${i}_screenshot.png`);
        
            // close the tab
            await page.close();
        }
    }

    await browser.close();
}


[1].forEach(async () => {
    await track_metrics(datasets);
    process.exit(0);
});