Use breadcrumbs based on organisation, if no parent linked

[unoduetre]

What

The Content Types Discovery team's audit uncovered that corporate_information pages lack a breadcrumb that connects back to the organisation.

Currently corporate information pages without the linked parent use taxon based breadcrumbs. An example would be the following page: https://www.gov.uk/government/organisations/government-actuarys-department/about

This PR fixes that, and if there is no parent linked, it uses the first linked organisation to generate breadcrumbs.

The logic is the following:

1. if parent is linked to, use it to generate breadcrumbs;
2. otherwise, if organisations are linked to, use the first one to generate breadcrumbs;
3. otherwise, if taxons are linked to, use them to generate breadcrumbs;
4. otherwise use the default breadcrumbs linking to homepage.

Why

So that users can navigate back to the organisation to which this corporate information relates.

Trello ticket

Testing

https://government-frontend-pr-3347.herokuapp.com/government/organisations/government-actuarys-department/about
https://government-frontend-pr-3347.herokuapp.com/government/organisations/department-for-business-energy-and-industrial-strategy/about/personal-information-charter
https://government-frontend-pr-3347.herokuapp.com/government/organisations/standards-and-testing-agency/about
https://government-frontend-pr-3347.herokuapp.com/government/organisations/hm-courts-and-tribunals-service/about

Visual Changes

Before

After

[MartinJJones]

I've had a quick look and have one thought on the updated breadcrumbs, using the page below as an example:

www.gov.uk/government/organisations/government-actuarys-department/about

The breadcrumbs are updated from Home > Government to Home > Government Actuary's Department which definitely feels like a nice a improvement.

It does appear that this page also has a grandparent, "Departments, agencies and public bodies", I think we should check with a content designer if this should also be included in the breadcrumbs as well, for example:

Home > Departments, agencies and public bodies > Government Actuary's Department

[unoduetre]

I've had a quick look and have one thought on the updated breadcrumbs, using the page below as an example:

www.gov.uk/government/organisations/government-actuarys-department/about

The breadcrumbs are updated from Home > Government to Home > Government Actuary's Department which definitely feels like a nice a improvement.

It does appear that this page also has a grandparent, "Departments, agencies and public bodies", I think we should check with a content designer if this should also be included in the breadcrumbs as well, for example:

Home > Departments, agencies and public bodies > Government Actuary's Department

Unfortunately, I can't find the information about the higher levels in the content item, so getting this information would require either adding things to the content item (probably with the publishing team's help) or doing a separate call to the api to retrieve that, which seems like an overkill. If we want/can modify the content item in the first place, we can add a parent link instead. This would resolve the issue, as the logic prioritises parent links. Also some examples of corporate information pages in publishing-api have parent links, so there shouldn't be a problem with adding it. So maybe we can ask the publishing team to add the parent link and in the meantime we can use this solution?

[unoduetre]

Or maybe we could use something like Home > Organisations > Government Actuary's Department
as in the Government Actuary's Department's page the breadcrumbs are Home > Organisations?

[unoduetre]

Or maybe we could use something like Home > Organisations > Government Actuary's Department as in the Government Actuary's Department's page the breadcrumbs are Home > Organisations?

This change has now been applied.

[unoduetre]

This is still worth considering: https://www.gov.uk/government/organisations/hm-courts-and-tribunals-service/about vs https://government-frontend-pr-3347.herokuapp.com/government/organisations/hm-courts-and-tribunals-service/about

[MartinJJones]

Looks good to me from a frontend perspective

@hannako please can you help review the backend changes

[hannako]

Thanks for adding all these tests Matt 🏅

I've requested a few tiny refactors. But my main question is on what valid corporate information pages look like

[hannako]

Minor:

I think we should follow the existing pattern in this file of just reading from the links. This method is just a boolean ie we don't need to do anything with the data. So unless I'm missing something, instantiating a new object feels unnecessary.

content_item.dig("links", "organisations").present?

[unoduetre]

This must match line 26 of lib/govuk_publishing_components/presenters/content_breadcrumbs_based_on_organisations.rb
So if this is to be changed, the other line needs to be changed as well.
It wouldn't be good, if we duplicated the code, as this wouldn't ensure the code (the list of methods) is always matching. Duplicating would easily lead to errors when it's changed in one place but not the other.

This solution calls the same, already existing method #related_organisations in ContentItem, which is a wrapper for raw content item data.
(Also this method contains additional logic which ensures the organisation we are interested in is always of the document type "organisation".
I know currently all organisations in links -> organisations are of "organisation" document type, so the additional logic is redundant right now.)

If we still wanted to change that, we'd need to create a method somewhere which would be called by both `lib/govuk_publishing_components/presenters/contextual_navigation.rb and lib/govuk_publishing_components/presenters/content_breadcrumbs_based_on_organisations.rb which would ensure that exactly the same list of organisations is used in both cases.

[hannako]

Thanks for addressing the notes, and for adding all the tests ⭐