Get external page urls on a multi-tenant multi-language Episerver site

Sometimes it seems as if the hardest thing in Episerver is to retreive a proper external url of a PageData or ContentReference item. This especially occurs on multi-language or multi-tenant sites. 

The url repository of Episerver may sometimes be complex, and the api's may be changing. What I typically need is a stable method that will be able to resolve the external url of content links no matter which site setup I got. The methods I use have been developed over years and the one I use mostly at the moment seem to do it all.

TLDR; here it is

/// <summary>
/// Resolve the absolute external url of a ContentReference
/// http://www.herlitz.nu/2016/11/29/get-external-page-urls-on-a-multi-tenant-multi-language-episerver-site/
/// </summary>
/// <param name="contentLink"></param>
/// <returns></returns>
public static class PageDataExtensions
{
    public static string ExternalUrl(this ContentReference contentLink)
    {
        var locator = ServiceLocator.Current.GetInstance<IContentLoader>();
        var pd = locator.Get(contentLink);

        var urlString = UrlResolver.Current.GetUrl(contentLink, pd.LanguageBranch, new VirtualPathArguments { ForceCanonical = true });
        if (string.IsNullOrEmpty(urlString))
        {
            return urlString;
        }

        var uri = new Uri(urlString, UriKind.RelativeOrAbsolute);
        if (uri.IsAbsoluteUri)
        {
            return urlString;
        }

        // optional
        //if (!uri.IsAbsoluteUri && HttpContext.Current != null)
        //{
        //    return new Uri(HttpContext.Current.Request.Url, uri).ToString();
        //}

        if (!uri.IsAbsoluteUri)
        {
            var siteDefinitionResolover = ServiceLocator.Current.GetInstance<SiteDefinitionResolver>();
            SiteDefinition siteDefinition = siteDefinitionResolover.GetDefinitionForContent(
                contentLink: contentLink,
                fallbackToWildcardMapped: true,
                fallbackToEmpty: true);

            return new Uri(siteDefinition.SiteUrl, uri).ToString();
        }

        return urlString;
    }
}

In a bit more detail

First I construct a PageData object from the contentLink which is required to distinguish the different language branches

var locator = ServiceLocator.Current.GetInstance<IContentLoader>();
var pd = locator.Get(contentLink);

Next up I use the UrlResolver singleton to tech the url for the page, this is where the language specific url is resolved. If no url is detected this is a good place to exit, handle the null return in your call

var urlString = UrlResolver.Current.GetUrl(contentLink, pd.LanguageBranch, new VirtualPathArguments { ForceCanonical = true });
if (string.IsNullOrEmpty(urlString))
{
    return urlString;
}

Check if the urlString was absolute, if it was we are done

var uri = new Uri(urlString, UriKind.RelativeOrAbsolute);
if (uri.IsAbsoluteUri)
{
    return urlString;
}

If the url still is relative we are probably missing httpcontext, possibly the method is run by a service like an indexer or another timer job. To ensure we construct a correct absolute url we need to fetch the site url from the SiteDefinition which we can construct from the contentLink using the SiteDefinitionResolver

if (!uri.IsAbsoluteUri)
{
    var siteDefinitionResolover = ServiceLocator.Current.GetInstance<SiteDefinitionResolver>();
    SiteDefinition siteDefinition = siteDefinitionResolover.GetDefinitionForContent(
        contentLink: contentLink, 
        fallbackToWildcardMapped: true,
        fallbackToEmpty: true);

    return new Uri(siteDefinition.SiteUrl, uri).ToString();
}

Thats it, happy urling.