Papers

May 30

Google Scholar Scraper // Function to extract data from the page function scrapeData() { var citationsContainer = document.getElementById(“citations”); citationsContainer.innerHTML = “Loading…”; // URL of the Google Scholar profile var url = “https://scholar.google.com/citations?user=dvqkwNgAAAAJ&hl=id”; // Make a GET request to fetch the Read more

<style>
html, body, form, table, div, h1, h2, h3, h4, h5, h6, img, ol, ul, li, button {
    margin: 0px;
    padding: 0px;
    border: 0px none;
    color: #fff;
}

table {
    border-collapse: collapse;
    border-width: 0px;
    empty-cells: show;
}

body, td {
    font-size: 13px;
    font-family: Arial,sans-serif;
    line-height: 1.24;
}

.gsc_g_hist_wrp {
    padding-top: 10px !important;
}
</style>

<script>
function filterContent(content, url) {
    var contentText = '';
    var output = content.match(/<div class="gsc_rsb_s gsc_prf_pnl" id="gsc_rsb_cit" role="region" aria-labelledby="gsc_prf_t-cit">(.*)<\/div><div class="gsc_rsb_s gsc_prf_pnl" id="gsc_rsb_co" role="region" aria-labelledby="gsc_prf_t-ath">/is);
    
    // If the previous pattern didn't match, try another variant
    if (!output || !output[1]) {
        output = content.match(/<div class="gsc_rsb_s gsc_prf_pnl" id="gsc_rsb_cit" role="region" aria-labelledby="gsc_prf_t-cit">(.*)<\/div><div class="gsc_lcl" role="main" id="gsc_prf_w">/is);
    }
    
    contentText = output && output[1] ? output[1] : 'e1';

    var citationsMatch = contentText.match(/Citations<\/a><\/td><td class="gsc_rsb_std">(\d+)<\/td>/is);
    var citations = citationsMatch && citationsMatch[1] ? citationsMatch[1] : 'e2';

    var hindexMatch = contentText.match(/h-index<\/a><\/td><td class="gsc_rsb_std">(\d+)<\/td>/is);
    var hindex = hindexMatch && hindexMatch[1] ? hindexMatch[1] : 'e3';

    var contentText2Match = contentText.match(/<style>(.+)/is);
    var contentText2 = contentText2Match && contentText2Match[1] ? contentText2Match[1] : 'e4';

    contentText2 = '<style>' + contentText2;

    var parser = new DOMParser();
    var dom = parser.parseFromString(contentText2, 'text/html');

    var links = dom.getElementsByTagName('a');

    // Loop through each <a> tag and replace them with their text content
    for (var i = links.length - 1; i >= 0; i--) {
        var linkNode = links[i];
        var text = linkNode.textContent;
        var style = linkNode.getAttribute('style');
        var className = linkNode.getAttribute('class');
        var div = document.createElement('div');
        div.setAttribute('class', className);
        div.setAttribute('style', style);
        div.textContent = text;
        linkNode.parentNode.replaceChild(div, linkNode);
    }

    contentText2 = dom.documentElement.innerHTML;
    contentText2 = 'Citations according to <a href="' + url + '">Google Scholar</a>: ' +
        citations + ' (h-index: ' + hindex + ')' + contentText2;

    return contentText2;
}

function fetchData(id, lang) {
    if (id !== 'dvqkwNgAAAAJ&hl' && lang !== 'en') {
        var url = 'http://scholar.google.de/citations?user=' + id + '&hl=' + lang;
        var xhr = new XMLHttpRequest();
        xhr.open('GET', url, true);
        xhr.onreadystatechange = function () {
            if (xhr.readyState === 4 && xhr.status === 200) {
                var content = xhr.responseText;
                var filteredContent = filterContent(content, url);
                console.log(filteredContent);
            }
        };
        xhr.send();
    } else {
        console.log('ERROR: Cannot find Google Scholar Account');
    }
}

// Replace 'idValue' and 'langValue' with the actual values you want to use
fetchData('idValue', 'langValue');
</script>

Category: Papers

Papers

Recent Posts