Contrib.social

Search API Attachments Not working

Open on Drupal.org →
Created on 12 August 2024, 3 months ago
Updated 30 August 2024, 3 months ago

Apologies for filing a support request ticket for debugging my own setup here, but I'm quite stumped, and perhaps if we find a resolution, this will help others.

I am attempting to set up Search API Attachments on my site, which we recently migrated to using Solr (from Elastic). This has previously worked under Elastic, but stopped at one point a while ago. We decided to move to Solr for a number of other reasons though.

Now I believe I have this set up correctly:

  • Processor is enabled
  • File fields set to "fulltext"
  • Solr extractor selected

But it doesn't seem to return the results for a test case of a unique string within an attached PDF document. Using the word "supercalifragilisticexpialidocious"

It doesn't appear to be indexing my file field content. Which is a "File" media asset reference. For good measure, I've tried adding both the media reference field and the file field on the media item to be indexed - as full text per the readme, and still nothing.
I've been trying to trace it down in xdebug, The best I can determine is that the processor isn't being added to this array:

            if ($property instanceof ProcessorPropertyInterface) {
              $processors_with_fields[$property->getProcessorId()] = TRUE;
            }

search_api/src/Item/Item.php

And the type for file fields here is always \Drupal\field\Entity\FieldConfig.

I believe I've followed all the steps in the Readme, but have I missed anything?

💬 Support request
Status

Closed: works as designed

Version

9.0

Component

Documentation

Created by

🇺🇸United States bburg Washington D.C.

Live updates comments and jobs are added and updated live.
Sign in to follow issues

Comments & Activities

  • Issue created by @bburg
  • Comment 3 months ago →
    🇺🇸United States bburg Washington D.C.

    So I figured out the issue, Search API Attachments adds a whole new field type to the fields you can add to be indexed, which appear as:

    "Search api attachments: {My Files Field}" with the machine name "saa_field_{files}" -- curly braces replacing what would be your specific field name. While I had the standard field added to the index, which of course just ended up indexing as the file ID. It's mentioned in this part of the documentation:

    "8. Go to admin/config/search/search-api/index/my_index/fields/add/nojs and:

    - in the General section, add the "Search api attachments: My pdfs" field.
    - in the Content section, add the "Title".
    - in the Content section, add the "Body".
    "

    Which didn't jump out at me on my first several read throughs. I added this to be indexed, and it worked as expected.

    Would it be unreasonable to request we update the documentation to call out that this is a new field, and not the standard file field?

  • Status changed to Closed: works as designed 3 months ago
  • Comment 3 months ago →
    🇫🇷France izus

    actually all these steps are documented in the README, please feel free to reopen and add the any documentation you find useful

contrib.social Blog FAQ Discussions
Production build 0.71.5 2024