Mitch,

So you could tap into gathering node event then somehow figure out if the PDF is being used on a content node that is published. If content node not published then cancel the indexing event.

You may be able to determine media usage using nexu package https://our.umbraco.org/projects/backoffice-extensions/nexu/

Regards

Ismail

Thanks Ismail. Was hoping there would be a simple config setting I could use. I'll check out that Nexu package. Do you know if it has an API I can code to?

Thanks

Hi Mitch,

If you can't use my api directly you can always code against the normal Umbraco Relations API. So like Ismail set you can use it probably for you case.

Dave

Thank you so much guys. Very useful advice. I'll give it a go right now.

Mitch,

Let us know how you get on because this is very interesting use case and if it all works i am going to include note on it on the examine course i wrote under the pdf indexing exercise.

Regards

Ismail

I also have plans to extend the API with methods that can be used in your project just for this kinds of use cases.

But need to fix some other things first..;and of course find time :-)

Dave

So, I had a stab at it and it seems to work! Not much code and I'm sure it can be improved, but here it is...

public class PdfExamineEvents : ApplicationEventHandler
{
    private IRelationService _relationService;

    protected override void ApplicationStarted(UmbracoApplicationBase umbracoApplication, ApplicationContext applicationContext)
    {
        var helper = new UmbracoHelper(UmbracoContext.Current);

        ExamineManager.Instance.IndexProviderCollection["PDFIndexer"].NodeIndexing += (sender, e) => NodeIndexing(sender, e, helper);

        _relationService = applicationContext.Services.RelationService;
    }

    private void NodeIndexing(object sender, IndexingNodeEventArgs args, UmbracoHelper helper)
    {
        args.Cancel = !ShouldIndex(args.NodeId, helper);
    }

    private bool ShouldIndex(int nodeId, UmbracoHelper helper)
    {
        // Check if media item is PDF
        if (!IsPdf(nodeId, helper)) return false;

        // Check if if the PDF is in the Relations
        if (!_relationService.IsRelated(nodeId)) return false;

        // If any document with this PDF is published, add to index. If not, cancel
        var relations = _relationService.GetByChildId(nodeId);
        if (!AnyPublished(relations.Select(r => r.ParentId), helper)) return false;

        return true;
    }

    private static bool IsPdf(int nodeId, UmbracoHelper helper)
    {
        var mediaItem = helper.TypedMedia(nodeId);

        // Not sure if every IPublishedContent has this property
        return mediaItem.HasProperty("umbracoExtension") && mediaItem.GetPropertyValue<string>("umbracoExtension") == "pdf";
    }

    private static bool AnyPublished(IEnumerable<int> nodeIds, UmbracoHelper helper)
    {
        if (nodeIds.Select(id => helper.TypedContent(id)).Any(n => n != null)) return true;
        return false;
    }
}

Mitch,

One more thing you will need to handle content un publish and delete events. So if page a contains pdf b and page a is then unpublished you will need to tap into that event and remove pdf b from index.

Regards

Ismail

For the sake of completion, here is my code to remove a PDF from an index if the page it is on becomes unpublished...

private void ContentService_UnPublished(IPublishingStrategy sender, PublishEventArgs<IContent> e)
    {
        var pdfIndexer = ExamineManager.Instance.IndexProviderCollection["PDFIndexer"];

        foreach (var item in e.PublishedEntities)
        {
            // Get all relations where the current node is the parent
            var relations = _relationService.GetByParentId(item.Id).ToList();

            foreach (var relation in relations)
            {
                var mediaNode = _helper.TypedMedia(relation.ChildId);
                if (mediaNode == null) continue;

                if (mediaNode["umbracoExtension"].ToString() != "pdf") continue;

                // Get all relations for this PDF not including the current relation
                var otherRelations = _relationService.GetByChildId(mediaNode.Id).Where(rel => rel.ParentId != relation.ParentId);

                // If this PDF is used on other published pages, do nothing and leave PDF in index
                if (otherRelations.Any(i => _helper.TypedContent(i.ParentId) != null)) continue;

                pdfIndexer.DeleteFromIndex(mediaNode.Id.ToString());
            }
        }
    }

Any suggestions for improvements to this code are most welcome.

is working on a reply...