Ian,

You will need to implement gathering node data event, then in that test for the content type and then test if the picker field has been selected then from the id get the media item.

Next you will need to extract the content of the pdf then inject that extracted content into examine field. For the pdf extraction you could install examine pdf indexer and use the api from that to extract the contents of the pdf or you could write something yourself with 3rd party pdf libraries to extract the content.

So its not out of the box but with a bit of code is it doable.

So I should have posted this nearly 12 months ago, when I fixed it at the time, but I have finally got around to it!

Nik and Ismail's replies really helped.

In my case, I wanted to index the name of each of the documents added by the content author to a node type in Umbraco. So Nik's advice was most helpful in my case, but if I ever need to index the content aswell, then I will be able to thanks to Ismail's help.

I'll post the code here in case it's ever useful to anyone else.

Add a class that inherits from IApplicationEventHandler:

public class AppEvents : IApplicationEventHandler
{

    public void OnApplicationStarted(UmbracoApplicationBase umbracoApplication, ApplicationContext applicationContext)
    {
        var helper = new UmbracoHelper(UmbracoContext.Current);
        ExamineManager.Instance.IndexProviderCollection["ExternalIndexer"].GatheringNodeData
                             += (sender, e) => ExamineEvents.GatheringContentData(sender, e, helper);
    }
}

Now below I will create the static class that was referenced in the ApplicationEventHandler above.

When creating the search index, the code needs to check if the content author has added one or more media items ("ItemDocuments") inside a document type called AccordionTabs. If some documents do exist, then loop through the media items, getting their GUID, retrieving them as TypedMedia to allow me to add the media item's name to the index.

I also had to put a line in there to try to get the media items as UDI documents, or to retrieve those documents that use an integer instead as their id. Some documents seemed to be stored as Umbraco.Core.Udi but some others were not. I'm not sure why, but adding in this check meant no documents were missed from the index.

public static class ExamineEvents
{
    public static void GatheringContentData(object sender, IndexingNodeDataEventArgs e, UmbracoHelper helper)
    {
        if (e.IndexType == "content")
        {
            var content = helper.TypedContent(e.NodeId);

            // Custom indexing logic

            try
            {

                if (content != null && content.ContentType.Alias == "AccordionTab" && content.HasProperty("AccordionTabs") && content.HasValue("AccordionTabs"))
                {

                    IEnumerable<IPublishedContent> tabs = (IEnumerable<IPublishedContent>)content.GetProperty("AccordionTabs").Value;

                    if (tabs != null)
                    {
                        var fieldDocuments = new StringBuilder();

                        foreach (var t in tabs)
                        {

                            if (t != null && t.HasProperty("ItemDocuments") && t.HasValue("ItemDocuments"))
                            {
                                var ipublishDocs = t.GetPropertyValue<IEnumerable<IPublishedContent>>("ItemDocuments");

                                if (ipublishDocs == null)
                                {
                                    var udiDocs = (IEnumerable<Umbraco.Core.Udi>)t.GetProperty("ItemDocuments").Value;

                                    foreach (Umbraco.Core.GuidUdi d in udiDocs)
                                    {

                                        string id = d.Guid.ToString();

                                        var m = helper.TypedMedia(id);

                                        if (m != null)
                                        {
                                            fieldDocuments.AppendLine(m.Name);
                                        }

                                    }
                                }
                                else
                                {

                                    foreach (var d in ipublishDocs)
                                    {

                                        var m = helper.TypedMedia(d.Id);

                                        if (m != null)
                                        {
                                            fieldDocuments.AppendLine(m.Name);
                                        }

                                    }
                                }

                            }

                        }

                        e.Fields.Add("websiteDocuments", fieldDocuments.ToString());

                    }


                }
            }
            catch (Exception err)
            {
                throw err;
            }

        }

    }
}

is working on a reply...