Analyse if we do or can fix missing or invalid LDD at harvest time with sweeper

[tloubrieu-jpl]

💡 Description

For example when the product is loaded in the registry with harvest:

1. a LDD HTTP referenced in a product label might not be published yet .
   Or 2. The LDD might be accessible but invalid (invalidate dat, xml parsing issues)

This issues might be fixed later, I would like to know if sweeper can update the registry-ldd after harvest ran in these cases.

⚔️ Parent Epic / Related Tickets

No response

[alexdunnjpl]

@tloubrieu-jpl as a general statement given the following, we can put any post-hoc update in a sweeper, easy

• A document, once fixed with a given update, is considered fixed for good (and can be skipped for future processing)
• Sweepers has access to sufficient information to determine, given the content of a document, what the content should be
• If invalid data is present, sweepers has access to sufficient information to enumerate that invalid data at runtime

So this seems like a simple. Is the issue that records exist in registry-dd which are incomplete, or that records do were not created in registry-dd in the first place?

What is the delta between the current state and desired state after a given LDD is live and valid?

[tloubrieu-jpl]

Thanks @alexdunnjpl , the issue is that records are missing in registry-dd and would be added after the LDD is published or fixed.
Sorry I lost track of your latest development, but would that try to fix a product until all the LDD references in the original XML are solved and imported in registry-dd ?

[alexdunnjpl]

@tloubrieu-jpl negative - the recent development relates solely to ensuring that products' properties are all present in the index mappings. This prospective functionality would need to be a new sweeper.

[tloubrieu-jpl]

Thanks @alexdunnjpl , the. We can use that ticket to create this new sweeper.