Assess the use of existing library to implement Google OneBox

[isedwards]

@nathanbaleeta wrote:

In response to 1.2.5, wagtail-schema.org library seems to be a ready solution out of the box to help standardize IoGT content as structured data for search engines like Google, Microsoft, Yahoo, Yandex etc to understand the semantics. It also provides JSON-LD tags for Wagtail sites which are consistent with Schema.org.

But it doesn't seem to offer a feature that can be enabled in the backend to allow content managers in the rich text editors to identify a mark-up for How-To, FAQ, and Video etc.

Therefore wagtail-schema.org assumes all the customization (basically embedding tags in templates) is handled by the developers leaving no room for manipulation by the content managers.

If the business requirement demands the content managers have to identify the JSON-LD tags for IoGT content on a rolling basis, then wagtail-schema.org may not fit the purpose, however if content managers can pre-determine the content which needs structuring and then let the developers do the heavy lifting, then this tool could suffice for this use case.

In response to 1.2.4, here is a great tool by Google for validating whether our JSON-LD tags in IoGT are working well: https://search.google.com/test/rich-results

PS: It may also be helpful to eventually follow and monitor how IoGT structured data is being used using Google search console.

More helpful resources: https://jsonld.com/

[isedwards]

I think wagtail-schema.org is the way to go. We should aim to get a working branch that demonstrates the out-of-the-box capabilities that it adds and then decide how to give content managers direct access to the capabilities (assuming they cannot use the provided template tags?).

Both releases of wagtail-schema.org on GitHub require wagtail 2.0, so we need to update to molo.core 8.0.0 before we can add this dependency.

@nathanbaleeta - Suggested steps:

• [x] Confirm why you have failing tests (we need tests to be passing before modifying the code)

Then, based on @michaelzhang608's work on dependencies, we then need to make the following changes to requirements.txt:

• [x] Pin molo.core to the last of the version 8 releases: molo.core==8.12.6
• [x] Pin molo.surveys and molo.polls to their latest versions: molo.surveys==6.10.5, molo.polls==6.4.0
• [x] Check requirements.txt in previous releases of molo.commenting and report which release(s) are suitable for use with molo.core 8 (and pin accordingly)
• [x] Disable molo.usermetadata (which does not support molo.core > 7.0.0). A seperate issue will be created to investigate
• [x] Add wagtail-schema.org==2.1.0 to requirements.txt and add wagtailschemaorg to INSTALLED_APPS in Django settings

[nathanbaleeta]

Found release molo.commenting==8.0.1 to be compatible with molo.core<9.0.0,>=8.0.0

Hint: https://pypi.org/project/molo.commenting/8.0.1/

[isedwards]

Is it working as expected? Can you give us an example of adding the extra information required by Google OneBox and how that is then rendered in the page?

[nathanbaleeta]

The installation part using the updated requirements.txt worked when I used pip3 version though with some warnings which could be addressed maybe in the near future. Below is a snapshot of the warnings to keep in the know:

However disabling molo.usermetadata dependency seems to cause other unprecedented bugs hence may not be able to verify whether the molo related updates are working as expected until this bug is resolved.

As you suggested, it would be good to create a separate issue to investigate this further. @isedwards

Below is the error it shows: ImportError: cannot import name 'urlresolvers' from 'django.core'

Stackoverflow gives some clue but still trying to figure out where to make the changes since what seems to be affected is a 3rd party library(django_comments) and not necessarily iogt code.