lundeen-bryan · September 10, 2023 02:28
diff --git a/get_dates.js b/get_dates.js
 /**
 * Date Extractor from HTML
 *
 * Version: 1.0.0
 *
 * Purpose:
 * This script is designed to extract dates present in an HTML document.
 * It uses multiple regular expressions to detect a variety of date formats.
 * After extracting all the unique dates, it presents them in a new popup window,
 * sorted in descending order from newest to oldest. If no dates are found,
 * a message is displayed in the popup indicating the absence of detectable dates.
 *
 * Creator: lundeen-bryan
 * Creation Date: 2023-09-09
 */
 (function () {
    // Function to format a Date object into DD-MMM-YYYY format
    function formatDate(d) {
        const monthNames = [
            "Jan",
            "Feb",
            "Mar",
            "Apr",
            "May",
            "Jun",
            "Jul",
            "Aug",
            "Sep",
            "Oct",
            "Nov",
            "Dec",
        ];
        return (
            ("0" + d.getDate()).slice(-2) +
            "-" +
            monthNames[d.getMonth()] +
            "-" +
            d.getFullYear()
        );
    }

    // Array to store unique date tags found in the document
    var tags = [];

    // Function to extract dates from the document based on given regex pattern
    function addMoreDates(reg) {
        var addTags = document.documentElement.innerHTML.match(reg);
        if (addTags) {
            addTags.forEach(function (newTag) {
                let parsedDate = Date.parse(newTag);
                if (!isNaN(parsedDate)) {
                    let formattedDate = formatDate(new Date(parsedDate));
                    // Add the date to the tags list if it's not already there
                    if (tags.indexOf(formattedDate) === -1) {
                        tags.push(formattedDate);
                    }
                }
            });
        }
    }

    // Array of Regex patterns to detect various date formats in the document
    const regexPatterns = [
        // ... (other regex patterns listed) ...
    ];

    // For each regex pattern, attempt to extract dates from the document
    regexPatterns.forEach(addMoreDates);

    // Sort the dates in descending order (newest to oldest)
    tags.sort(function (a, b) {
        return Date.parse(b) - Date.parse(a);
    });

    // If no dates were found, add a default message
    if (tags.length === 0) {
        tags = ["No dates detected in page."];
    }

    // Construct the popup content HTML
    var popupContent = "<html><head><title>Date List</title></head><body>";
    popupContent += "<h2>Dates in the HTML:</h2>";
    popupContent += "<ul>";
    tags.forEach(function (tag) {
        popupContent += "<li>" + tag + "</li>";
    });
    popupContent += "</ul></body></html>";

    // Open a new popup window and write the list of dates into it
    var popupWindow = window.open("", "DateListPopup", "width=400,height=400,scrollbars=1");
    popupWindow.document.write(popupContent);
    popupWindow.document.close();
 })();
	/**
	* Date Extractor from HTML
	*
	* Version: 1.0.0
	*
	* Purpose:
	* This script is designed to extract dates present in an HTML document.
	* It uses multiple regular expressions to detect a variety of date formats.
	* After extracting all the unique dates, it presents them in a new popup window,
	* sorted in descending order from newest to oldest. If no dates are found,
	* a message is displayed in the popup indicating the absence of detectable dates.
	*
	* Creator: lundeen-bryan
	* Creation Date: 2023-09-09
	*/
	(function () {
	// Function to format a Date object into DD-MMM-YYYY format
	function formatDate(d) {
	const monthNames = [
	"Jan",
	"Feb",
	"Mar",
	"Apr",
	"May",
	"Jun",
	"Jul",
	"Aug",
	"Sep",
	"Oct",
	"Nov",
	"Dec",
	];
	return (
	("0" + d.getDate()).slice(-2) +
	"-" +
	monthNames[d.getMonth()] +
	"-" +
	d.getFullYear()
	);
	}

	// Array to store unique date tags found in the document
	var tags = [];

	// Function to extract dates from the document based on given regex pattern
	function addMoreDates(reg) {
	var addTags = document.documentElement.innerHTML.match(reg);
	if (addTags) {
	addTags.forEach(function (newTag) {
	let parsedDate = Date.parse(newTag);
	if (!isNaN(parsedDate)) {
	let formattedDate = formatDate(new Date(parsedDate));
	// Add the date to the tags list if it's not already there
	if (tags.indexOf(formattedDate) === -1) {
	tags.push(formattedDate);
	}
	}
	});
	}
	}

	// Array of Regex patterns to detect various date formats in the document
	const regexPatterns = [
	// ... (other regex patterns listed) ...
	];

	// For each regex pattern, attempt to extract dates from the document
	regexPatterns.forEach(addMoreDates);

	// Sort the dates in descending order (newest to oldest)
	tags.sort(function (a, b) {
	return Date.parse(b) - Date.parse(a);
	});

	// If no dates were found, add a default message
	if (tags.length === 0) {
	tags = ["No dates detected in page."];
	}

	// Construct the popup content HTML
	var popupContent = "<html><head><title>Date List</title></head><body>";
	popupContent += "<h2>Dates in the HTML:</h2>";
	popupContent += "<ul>";
	tags.forEach(function (tag) {
	popupContent += "<li>" + tag + "</li>";
	});
	popupContent += "</ul></body></html>";

	// Open a new popup window and write the list of dates into it
	var popupWindow = window.open("", "DateListPopup", "width=400,height=400,scrollbars=1");
	popupWindow.document.write(popupContent);
	popupWindow.document.close();
	})();