Visualise geospatial data on dataset details page

[CharliePatterson]

Context:

• Displaying geospatial data on a map enables LPAs to visually check that the data they've provided is correct. A key finding from the Alpha phase was that this approach had much better responses that just displaying the data in a list or table format, as they'd be able to visually check the boundaries and locations were correct.
• This ticket covers the adding of the map to the page for geospatial datasets. There is a separate iteration to cover the adding of the summary/overview table: https://github.com/digital-land/submit/issues/407
• This page will be viewable to users when they click through to a live dataset that doesn't have any issues. In future there will be navigation that links together this page and the tasklist and issues pages, but this will be treated as a separate iteration.

Acceptance criteria:

• As a data provider, when I click through to a live dataset without any issues, I land on the dataset details page.
• As a data provider, when I'm viewing the details page of a dataset that contains geospatial data, I can see that data displayed on an interactive map.
• As a data provider, when I'm viewing the details page of a dataset that doesn't contain geospatial, I don't see a map.

The following datasets contain geospatial data:

• Article 4 direction area
• Conservation area
• Listed building outline
• Tree
• Tree preservation zone

Designs:

Designs in progress here: https://github.com/digital-land/submit/issues/311

Tasks:

• [ ] Create new page to contain dataset details
• [ ] Render map for geospatial datasets
• [ ] Speak to Owen around map tiles

[alextea]

[CharliePatterson]

Thanks @alextea 🙏

@DilwoarH clarification on how the navigation works for this first iteration - we'll update the navigation in this separate ticket: https://github.com/digital-land/submit/issues/408

[DilwoarH]

@CharliePatterson - there are some datasets which have too many entries that the datasette API returns an error (query taking too long).

We will need to look at using layers for these as suggested by @GeorgeGoodall.

[CharliePatterson]

@DilwoarH would this be resolved when we switch over to using the performance db?

[DilwoarH]

@DilwoarH would this be resolved when we switch over to using the performance db?

I don't think so as that would still use Datasette (which has the limitation).

The map layers work better long term as they are generated and cached - therefore won't require additional processing on our end. We need to investigate the following:

• Are the cached layers at a LPA level
• Which datasets are cached

Hope this makes sense