Fix issues with meta tags #1228

ybnd · 2021-06-14T08:11:00Z

References

Fixes Several minor SEO issues in HTML <meta> tags #1197

Description

Implement Google Scholar review suggestions for meta tags

Instructions for Reviewers

The following meta tags were changed:

citation_pdf_url: select Bitstreams ~ MIME type allowlist (if more than one option and no primary in ORIGINAL Bundle)
citation_abstract_html_url: use dc.identifier.uri if available, fall back to current URL
citation_publisher: use dc.publisher (except dissertations & tech reports, those use other tags)
citation_date → citation_publication_date
Remove og:title and og:description meta tags

Testing

Open some Item pages and inspect the head > meta in their HTML
- citation_abstract_html_url should not contain links to localhost, but should contain links ~ the Item's handle dc.identifier.uri is filled out. Otherwise, it should match the Item page URL
- citation_publisher should be present if dc.publisher is filled out
- citation_publication_date should be present
- The following meta tags should not be present: citation_date, og:title, og:description
For Items with one ORIGINAL Bitstream
- citation_pdf_url should contain a link to that Bitstream
For Items with more than one ORIGINAL Bitstream
- If a primary Bitstream is specified, citation_pdf_url should link to it
- Otherwise, citation_pdf_url should contain a link to the first ORIGINAL Bitstream with one of these MIME types

Checklist

My PR is small in size (e.g. less than 1,000 lines of code, not including comments & specs/tests), or I have provided reasons as to why that's not possible.
My PR passes TSLint validation using yarn run lint
My PR doesn't introduce circular dependencies
My PR includes TypeDoc comments for all new (or modified) public methods and classes. It also includes TypeDoc for large or complex private methods.
My PR passes all specs/tests and includes new/updated specs or tests based on the Code Testing Guide.
If my PR includes new, third-party dependencies (in package.json), I've made sure their licenses align with the DSpace BSD License based on the Licensing of Contributions documentation.

…eams

lgtm-com · 2021-06-14T08:18:50Z

This pull request introduces 1 alert when merging 64049fd into d253790 - view on LGTM.com

new alerts:

1 for Unused variable, import, function or class

lgtm-com · 2021-06-14T08:38:57Z

This pull request introduces 1 alert when merging 04b4f1c into d253790 - view on LGTM.com

new alerts:

1 for Unused variable, import, function or class

ybnd · 2021-06-14T08:46:16Z

@Art @tdonohue LGTM & unit test fix incoming

tdonohue

@ybnd and @artlowel : Overall, this fixes the issues you've described, but I've noticed an odd caching issue when testing with this PR. It seems the <meta> tags don't always refresh when you browse to a new page. Here's what I'm doing:

Start on a Community page, click on a Collection, then on an Item. Select "Inspect" in your browser and look at the <head> section.
- You'll see a <meta property='title'> with the Collection Name and a second with the Item name.
- You'll see a <meta property='citation_title>` with the Collection Name and a second with the Item name.
- All other <meta> tags for the Item will be there though, which is correct.
Click back to the Collection. Select "Inspect" in your browser and look at the <head> section.
- The Item <meta> tags (from step 1) are still visible. They seem to be cached, as they obviously shouldn't appear on a Collection.
Click on a different Item in the same Collection. Select "Inspect" in your browser and look at the <head> section.
- Now, <meta> tags are still odd...Some still appear from the first Item (from step 1) while others are correct for the Item you are on.

Overall, the behavior I'm seeing almost seems like a caching issue...like these <meta> tags are not being entirely "reset" (to empty) when you browse to a new page, and this results in very odd behavior. (NOTE This caching issue doesn't appear to be specific to this PR...as I can also sometimes get it to occur on demo7.dspace.org. So, if you'd rather treat this as a separate bug, we can do so. Besides this bug, the rest of the behavior of this PR seems correct.)

ybnd · 2021-06-28T10:07:52Z

@ybnd and @artlowel : Overall, this fixes the issues you've described, but I've noticed an odd caching issue when testing with this PR. It seems the <meta> tags don't always refresh when you browse to a new page

@tdonohue @artlowel After a bit of poking around on the current demo7.dspace.org I only ever saw the <title> meta tag change, and not even consistently.

Looks like this distinctUntilKeyChanged filters out way more changes than it should, e.g. when going back/forward between a Collection & an Item I see dso.uuid change in that pipe but it doesn't pass that filter... not sure what that's about.

…oute

artlowel · 2021-06-30T12:27:49Z

@tdonohue and @ybnd I fixed the issue. The reason was not only the distinctUntilKeyChanged issue @ybnd mentioned, but the fact that the service responded both to the route changing, and calls for DSO page components initializing. With the second call often starting before the first had finished processing, which caused some tags that were just added to be immediately removed again, others added twice etc.

I fixed it by removing the calls on DSO pages, and only listening to route changes. That's possible now (but wasn't when the metadataservice was originally written) because we're using resolvers on those pages, and they all put the resolved object in the dso property in the route data.

I spent an additional 4 hours on this though, most of that not on the solution itself, but on the tests. They were basically integration tests in a unit test file. Didn't mock anything which caused them to break with the slightest change. I practically had to rewrite them from scratch, getting rid of anything that isn't strictly necessary for the test.

tdonohue · 2021-06-30T13:22:54Z

Thanks @artlowel . I'll add 4 hours to the original ticket and re-test this as soon as possible.

tdonohue

@artlowel & @ybnd : I retested today. With @artlowel 's updates, the behavior seems improved, however, I'm still seeing the citation_author field (and seemingly only that field) be cached between pages. Here's what I'm doing:

Running yarn start
Bring up UI. Immediately do inspect and open up the <head> section in a separate window
Now, in the UI, browse to an Item. I see the <meta> tags auto-refresh in my "inspect" window.
From that Item page, click on the Collection in breadcrumbs. Most <meta> tags auto-refresh, but the citation_author tags from Item still exist from the previous page.
Click on Community in Breadcrumbs. Again, most <meta> tags auto-refresh, but the citation_author tags still exist from Item.

Last time I tested, I saw this behavior across the <meta property="title"> and <meta property="citation_title"> tags...that bug seems fixed. But somehow the behavior still exists for citation_author. Is there something in the UI that could be caching the author names themselves? It's odd that this specific field is now the issue.

Overall, again, I'm in favor of this PR. I'd just really like to find a way to improve SEO by ensuring we always have accurate <meta> tags. If you both feel this final bug needs to be a separate ticket, we can split it out...but I'd still rate it as high priority for 7.0.

ybnd · 2021-07-02T06:49:03Z

@tdonohue Should be fixed now 👍

tdonohue

👍 Perfect! Thanks @ybnd and @artlowel ! This is now working perfectly for me & is ready to be merged. Thanks for tracking down what was going on with the meta tags!

tdonohue · 2021-07-02T15:00:41Z

Merging this immediately, as it was previously discussed as 1 approval in meetings (it includes a large number of refactored/updated specs), and I'd like to get it in place for final quick tests from Google Scholar.

[DSC-1446] added default file thumbnail Approved-by: Stefano Maffei

artlowel and others added 11 commits June 14, 2021 10:06

update citation_pdf_url logic

dde57b9

79768: Extract download route method for Bundles with multiple Bitstr…

9fe5a91

…eams

79768: Check against list of allowed mimetypes

1caba78

79768: Update & add unit tests

cb4446b

79768: Remove og:title & og:description

6f7b76e

79768: Rename citation_date to citation_publication_date

67f8ce7

79768: Fix tag: citation_abstract_html_url

304d8f7

79768: Fix tag: citation_publisher / citation_*_institution

81b76dd

79768: Rename method

6a4e563

79768: Update typedocs

2ed16aa

79768: Make relative URLs absolute

64049fd

artlowel assigned ybnd Jun 14, 2021

artlowel added bug component: SEO Search Engine Optimization e/4 Estimate in hours high priority testathon Reported by a tester during Community Testathon labels Jun 14, 2021

artlowel added this to the 7.0 milestone Jun 14, 2021

Fix comment punctuation

04b4f1c

80084: Fix unit test & LGTM issues

34b117e

ybnd force-pushed the w2p-79768_fix-issues-with-meta-tags branch from aab8bd3 to 34b117e Compare June 14, 2021 08:53

tdonohue self-requested a review June 17, 2021 14:47

tdonohue requested changes Jun 21, 2021

View reviewed changes

fix issue where meta tags wouldn't be updated properly based on the r…

03e2e30

…oute

artlowel requested a review from tdonohue June 30, 2021 12:27

tdonohue removed the e/4 Estimate in hours label Jun 30, 2021

tdonohue requested changes Jun 30, 2021

View reviewed changes

move tagstore to ngrx

c86f163

tdonohue requested a review from atarix83 July 1, 2021 14:37

ybnd added 6 commits July 1, 2021 22:36

79768: Add unit tests for metaTagReducer

5040089

79768: Update & add MetadataService unit tests

fb8f28f

79768: Fix typo

8caa916

79768: Fix unused imports & lint issue

5ed41b3

Merge branch 'main' into w2p-79768_fix-issues-with-meta-tags

711abfc

79768: Fix followLink syntax

a91f16e

tdonohue self-requested a review July 2, 2021 14:04

tdonohue approved these changes Jul 2, 2021

View reviewed changes

tdonohue merged commit 1dba4c5 into DSpace:main Jul 2, 2021

tdonohue mentioned this pull request Jul 26, 2021

SEO + SSR fixes: Fix <meta> tag syntax and add a basic test that SSR is working #1282

Merged

tdonohue mentioned this pull request Jan 30, 2023

Items with many bitstreams cause a proxy error when trying to view an item (anonymously) or edit an item (as administrator) DSpace/DSpace#8648

Closed

4science-it pushed a commit to 4Science/dspace-angular that referenced this pull request Jan 26, 2024

Merged in DSC-1446-main (pull request DSpace#1228)

0a02da1

[DSC-1446] added default file thumbnail Approved-by: Stefano Maffei

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Fix issues with meta tags #1228

Fix issues with meta tags #1228

ybnd commented Jun 14, 2021

lgtm-com bot commented Jun 14, 2021

lgtm-com bot commented Jun 14, 2021

ybnd commented Jun 14, 2021

tdonohue left a comment •

edited

Loading

ybnd commented Jun 28, 2021 •

edited

Loading

artlowel commented Jun 30, 2021 •

edited

Loading

tdonohue commented Jun 30, 2021

tdonohue left a comment

ybnd commented Jul 2, 2021

tdonohue left a comment

tdonohue commented Jul 2, 2021

Fix issues with meta tags #1228

Fix issues with meta tags #1228

Conversation

ybnd commented Jun 14, 2021

References

Description

Instructions for Reviewers

Testing

Checklist

lgtm-com bot commented Jun 14, 2021

lgtm-com bot commented Jun 14, 2021

ybnd commented Jun 14, 2021

tdonohue left a comment • edited Loading

Choose a reason for hiding this comment

ybnd commented Jun 28, 2021 • edited Loading

artlowel commented Jun 30, 2021 • edited Loading

tdonohue commented Jun 30, 2021

tdonohue left a comment

Choose a reason for hiding this comment

ybnd commented Jul 2, 2021

tdonohue left a comment

Choose a reason for hiding this comment

tdonohue commented Jul 2, 2021

tdonohue left a comment •

edited

Loading

ybnd commented Jun 28, 2021 •

edited

Loading

artlowel commented Jun 30, 2021 •

edited

Loading