extract-image-info.js

var fs = require('fs');
var cheerio = require('cheerio');

function extractInfo(article) {
    var prettyLink = article.find("h2 a").attr("href");
    var matches = article.find(".permalink a").attr("href").match(/p\/([0-9]+)/);
    var postID = matches[1];
    var title = article.find("h2").text();
    var date = article.find(".date").attr("data-date");
    return { prettyLink: prettyLink,
             postID: postID,
             title: title,
             date: date};
}

function collectLinks(article) {
    var results = [];
    article.find(".body a").each(function() {
        var link = $(this);
        if (link.attr('href') && (link.attr('href').match(/sachachua/)
            || !link.attr('href').match(/^http/))) {
            results.push({href: link.attr('href'),
                          text: link.text()});
        }
    });
    return results;
}

function formatEntry(data) {
    var s = "";
    for (var i = 0; i < data.links.length; i++) {
        var matches;
        if (s) { s += "\n"; }
        if ((matches = data.links[i].href.match(/uploads\/[0-9]+\/[0-9]+\/(.*?)\.(?:png|jpg)/))) {
            s += matches[1] + "\t" + data.prettyLink;
        }
    }
    return s;
}

function parseFile(filename) {
    $ = cheerio.load(fs.readFileSync(filename));
    $('article').each(function() {
        var data = extractInfo($(this));
        data.links = collectLinks($(this));
        var output = formatEntry(data);
        if (output) {
            console.log(output);
        }
    });
        
}

for (var i = 2; i < process.argv.length; i++) {
    parseFile(process.argv[i]);
}