Scraped url missing the protocol and domain

[GrahamHotchkiss]

Reads "url": "/sitelink-api/v1/sites/1" rather than "url": ""https://apps.snh.gov.uk/sitelink-api/v1/sites/1"

[GrahamHotchkiss]

@MikeCoats - so I see this is a deliberate choice https://github.com/Scottish-Natural-Heritage/sitelink-api-mirror/blob/136378d9607d8f1bfac6fbccf43aa6eab86a13dd/mirror.mjs#L158-L160 + https://github.com/Scottish-Natural-Heritage/sitelink-api-mirror/blob/136378d9607d8f1bfac6fbccf43aa6eab86a13dd/mirror.mjs#L205-L206

Looks like it is trying to create a relative path but not seeing anything in the blaim history to enlighten me any further.

[MikeCoats]

It allows you to host locally or in any cloud environment without having to think about the domain you're serving from.

• If you host locally, you'd need http://localhost:8080/sitelink-api/v1/sites/1
• If you serve from the S3 bucket you'd need http://naturescot-sitelink-mirror.s3-website.eu-west-2.amazonaws.com/sitelink-api/v1/sites/1
• If you served from the cloudfront endpoint, you'd need https://d3ovo1aozvg78c.cloudfront.net/sitelink-api/v1/sites/1
• If you served from the CNAME, you'd need https://sitelink-api.nature.scot/sitelink-api/v1/sites/1

By using just /sitelink-api/v1/sites/1 you get away from trying to be too clever or baking in an artefact of how you deploy the mirror.

[GrahamHotchkiss]

But the application actually uses the 'url' as a URL, we published it as such and this is a material change from the original.