The plugin is exposing & returning post SEO fields but not returning field values pages or custom post types. Why can't figure out why.

Using 4.8.0 of the plugin Yoast version 15.3 WPGraphQL 1.1.3 gatsby-source-wordpress 4.0.1

Example Query:

  pages {
    edges {
      node {
        id
        title
        seo {
          canonical
          title
          metaDesc
          focuskw
          metaRobotsNoindex
          metaRobotsNofollow
          opengraphAuthor
          opengraphDescription
          opengraphTitle
          opengraphDescription
          opengraphImage {
            altText
            sourceUrl
            srcSet
          }
          opengraphUrl
          opengraphSiteName
          opengraphPublishedTime
          opengraphModifiedTime
          twitterTitle
          twitterDescription
          twitterImage {
            altText
            sourceUrl
            srcSet
          }
          breadcrumbs {
            url
            text
          }
          cornerstone
          schema {
            pageType
            articleType
          }
        }
      }
    }
  }
}```

Hi,

I have a page with both, the opengraph image and twitter image set, but both fields, opengraphImage and twitterImage show up as null in the API.

I have looked at the code and found the following problems: The value for YoastSEO()->meta->for_post($post->ID)->open_graph_images, which is used to retrieve the value for the open graph image, returns the images as an array or arrays. From the first image in the array, the url is taken and then put through wp's attachment_url_to_postid function to retrieve the id. Unfortunately, the URL in the array is a scaled Version of the original image, meaning that it has the dimensions of the thumbnail attached to the original filename (eg. my-image-1640x680.jpg). The attachment_url_to_postid function can't find the media attachment with this URL and returns 0, resulting in a null value for the API.

A similar problem exists for the twitter image, where the YoastSEO()->meta->for_post($post->ID)->twitter_image call returns a scaled version of the image, resulting in a return value of 0 from the attachment_url_to_postid function.

The problem with opengraphImage could possibly be overcome by using one of the other values in the image array. In my example, the array contains an id field, which could be used directly. However, I'm unsure if this value is always available.

I would also like to ask, wether it is necessary to return the images as MediaItems instead of as simple URLs. Yoast returns the scaled versions of the images, since they are optimized for the specific use-case. By expanding them back to MediaItems, the implementation that uses the information needs to re-decide which size to use.

I can offer to implement the changes if you like, but wanted to discuss the topic before creating a pull request.

Thank you for a great plugin!

Does this plugin still work with wp-graphql 0.3.2 (after the major 0.3.0 overhaul)? I did a quick test and it seems the graphql queries only return data when you manually edited the meta tags with the plugin. All autogenerated Yoast entries are empty strings

Firstly, great work on the plugin so far.

I had a suggestion around the social images (Open Graph and Twitter) that get returned. It would be great if we could return them as as a meditaItem. At the moment, they're just being returned as strings that point back to the WordPress instance, rather than Gatsby locally.

I had a go at getting this working but couldn't crack it. I figure we could use the type MediaItem rather than String but I just couldn't find enough documentation around what I actually needed to pass to it.

At the moment, I can see the plugin is using _yoast_wpseo_opengraph-image to grab the URL and I also see in the database that _yoast_wpseo_opengraph-image-id is available to get the ID of the media item. Just can't tie it all together. I'll submit a PR though if I get it figured out!

Yoast puts Json+ld automatically in the article generated by wordpress itself. It would be awesome if the same Json+ld object is automatically available in the post's seo.

Using %%excerpt%% as the default description on a custom post type don't return the generated value. In editor & WP frontend it works

in my GQL query i get

"seo": {
        "title": "Stable Baselines",
        "metaDesc": "%%excerpt%%",

Any ideas?

I'm not sure if this is specific to just the title tag field, but I've got an ampersand and bullet character in my title tags and they're showing up as & and • after fetching from GraphQL.

Looks like I'm getting encoded special characters for page. Thought this might have been fixed in the past but seems to be broken again.

Here's my query and what it returns.

Example shown is requesting titles but same issue occurs when requesting desc's Let me know if there's anymore more you want me to provide.

Cheers!

See this ticket as a main reference and for discussions: https://github.com/wp-graphql/wp-graphql/issues/1535

WordPress VIP has pretty high coding standards and everyone who will work with VIP in the future will face the same problems as our team right now. It would be great if we could get coding standards up and running with VIP's standards.

Ideally we could setup PHP_Codesniffer with all the configurations of them and let a bot check run through it before merging PRs.

We will attempt to setup everything and also fix major errors, so that we will be able to use WPGraphQL in the future. It would be great if some maintainers can share some of their attention on this, so we don't end up breaking stuff.

Also this is obviously open for discussion, we would just use that PR for our purposes anyways, but would greatly appreciate if it would make it into the main repos best practices.

References:

• https://wpvip.com/documentation/how-to-install-php-code-sniffer-for-wordpress-vip/
• https://github.com/Automattic/VIP-Coding-Standards/tree/master/WordPress-VIP-Go

Added the following Yoast data to the Schema:

SEO - opengraphUrl, opengraphSiteName, opengraphPublishedTime, opengraphModifiedTime.

SEOSchema - siteName, wordpressSiteName, siteUrl.

Steps to repro:

1. wordpress site with yoast 14.0.3
2. install wordpress plugin wpgraphql 0.8.4
3. install wordpress plugin wp-graphql-yoasy-seo 3.3.0
4. using some graphql editor (eg GraphIQL) execute the following query:

{
  posts(first: 10) {
    edges {
      node {
        id
        title
        seo {
          title
          metaDesc
          focuskw
          metaKeywords
          metaRobotsNoindex
          metaRobotsNofollow
          opengraphTitle
          opengraphDescription
          opengraphImage {
            altText
            sourceUrl
            srcSet
          }
          twitterTitle
          twitterDescription
          twitterImage {
            altText
            sourceUrl
            srcSet
        }
        }
      }
    }
  }
}

expected: Title and meta description reflect each individual post

actual: Title and meta description match the most recently generated post

This PR complementary of #85 makes the following improvements to the archive SEO data by using Yoast SEO surfaces.

• Show encoded title instead of Yoast SEO variables.
• Add the metaRobotsNofollow value even if it is not configurable from the Search Appearance to simplify integrations.
• Return a string instead of a boolean for the metaRobotsNoindex and metaRobotsNofollow values to keep it consistent across individual nodes.

While the API is well-documented for current pages, I didn't find a way to get the global values for individual pages. I'm also wondering if the archiveLink field should be renamed canonical.

Before

After

Hello, I'm new to this plugin and WPGraphQL and I would like to know if it's correct that the fields are empty with asPreview or if I rather need to change something.

This is my query:

const GET_POST_BY_ID = gql`
	query PostBySlug($id: ID!, $idType: PostIdType = URI, $asPreview: Boolean = false) {
		post(id: $id, idType: $idType, asPreview: $asPreview) {
			id
			slug
			isPreview
			title(format: RENDERED)
			content(format: RENDERED)
			featuredImage {
				node {
					uri
					altText
				}
			}
		seo {
			title
			metaDesc
			metaKeywords
			breadcrumbs {
				text
				url
			}
		}
		}
	}
`

And this is the method I use to run the query:

async function getSinglePost(id, asPreview = false) {
	const idType = asPreview ? 'DATABASE_ID' : 'URI';
	const { data } = await client.query({
		query: GET_POST_BY_ID,
		variables: {
			id,
			idType,
			asPreview,
		},
	});
	return data?.post || {};
}

When the variable asPreview is set to true, some fields related to Yoast SEO are empty. In particular:

seo {
	title
	metaDesc
	metaKeywords
	breadcrumbs {
		text
		url
	}
}

Am I wrong something or is it correct that it behaves like this?

There are cases that we need to query if there is a redirection set for a specific URI. For example, in case of "not found error" you may want to redirect the user to correct page instead of showing a 404 page.

Currently, we can get all redirects from the SEO (Yoast config) node but it is not feasible to get a full list of redirects and filter on client side. It would be perfect if we can query the redirect with a where param, use the query result to redirect to correct page or to throw a 404 error.

I would like to know your thoughts and possible development ideas on this.

When running a query for twitterImage?.sourceUrl() I get the following error (when no image has been added to the Twitter preview/Twitter Image section in WordPress):

[GQtyError: Internal server error] {
  graphQLErrors: [
    {
      debugMessage: 'Cannot return null for non-nullable field "MediaItem.id".',
      message: 'Internal server error',
      extensions: [Object],
      locations: [Array],
      path: [Array]
    }
  ]
}

Doesn't seem to be an issue with opengraphImage?.sourceUrl()

Hello,

Thanks for the great plugin. I am new to gatsbyjs and started my first project but can't able to solve the problem. I am getting "property="og:type" content="website"" on Post instead of "property="og:type" content="article".

Apart from that, on the homepage no Yoast metadata showing in Gatsby

Please help me

Waiting for your reply.

I posted this on the Cloudinary's WordPress Plugin repo as well. I wasn't sure which side could fix this.

I know there are other issues already created for Cloudinary URLs populating correctly for WPGraphQL, but I think there's a subtle difference in what I'm facing (not pulling images from the content or raw HTML).

I'm using the WPGraphQL Yoast SEO Addon to query SEO data from the Yoast general settings and individual Post/Page Yoast settings.

When I set up a Facebook social image in the Yoast general settings, I get a WordPress-based URL instead of a Cloudinary URL.

In addition to that, when I set the Facebook or Twitter images on a specific Post/Page, I get null for the opengraphImage result and a WordPress URL for the twitterImage result.