[Cod-bugs] Authors matching in Optimade API

[Andrius Merkys]

Dear Mehmet,

On 2025-11-17 16:08, Mehmet Giritli - STFC UKRI wrote:
> I am trying to use Optimade API to search using author names. But I 
> noticed a behaviour which is a bit unexpected.
> 
> Consider this query:
> 
> authors.name HAS "A. Day"
> 
> The spec for HAS operator says that this should only match if there is 
> an exact match, i.e., there is an author which is equal to "A. Day", in 
> the record.
> 
> Here is COD's output:
> https://www.crystallography.net/cod/optimade/references?filter=authors.name%20HAS%20%22A.%20Day%22 <https://www.crystallography.net/cod/optimade/references?filter=authors.name%20HAS%20%22A.%20Day%22>
> 
> The single record above includes an author "A. Dayalan". However, I 
> think this should not be the case.  It seems like "HAS" implementation 
> behaves like what I would expect from "CONTAINS" implementation. But, 
> this means that there is no way to get only the records which contain 
> values we are explicit about.

You are right, here the COD's OPTIMADE implementation deviates slightly 
from the OPTIMADE specification. Given our backend, implementing support 
for such queries to the letter has proven to be too complicated, thus at 
the moment we have this deviation. I think it would be nice to at least 
emit a warning from our side.

> That said, I also noticed that CONTAINS is not implemented for 
> author.names at the moment:
> 
> https://www.crystallography.net/cod/optimade/references?filter=authors.name%20CONTAINS%20%22A.%20Day%22 <https://www.crystallography.net/cod/optimade/references?filter=authors.name%20CONTAINS%20%22A.%20Day%22>

In this case the filter string should be:

authors.name HAS CONTAINS "A. Day"

(Notice the addition of 'HAS'). 'CONTAINS' alone is an operator acting 
on strings, not on lists.

> To sum up my request, do you think it is possible to:
> 
>  1.
>     Change implementation of "author.names HAS" so that it allows only
>     exact matching.

This should be doable, it just needs additional work from our side.

>  2.
>     Implement "author.names CONTAINS" which is basically the current
>     implementation you have for HAS.

As explained, 'author.names CONTAINS' should be changed to 'author.names 
HAS CONTAINS', as 'author.names CONTAINS' should not work on any 
OPTIMADE implementation.

> I hope it made sense. Please let me know if there is a way to overcome 
> this situation currently (e.g., to exact match an author) in your API.

Thanks for contacting us regarding this issue.

Best wishes,
Andrius

-- 
Andrius Merkys
Vilnius University Institute of Biotechnology, Saulėtekio al. 7
LT-10257 Vilnius, Lithuania

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.