API errors on some regional filter requests

[steffilazerte]

@denislepage, @cjardine-bsc,

Lately the API has been returning an error when attempting to download data for IBAs

Endpoint: https://naturecounts.ca/api/data/list_collections (I also get the error on Sandbox)

'An error has occurred; java.lang.NullPointerException: Cannot read the array length because the return value of "org.bsc.monitoring.DataRequests2.getFilterShapes()" is null'

I've been trying with parameters "iba: IBA.AB001" or "iba: IBA.YK003" (I also tried just "iba: YK003", just in case but got the same error).

Also using bmdeVersion: BMDE2.00-min, if that's relevant and no collections.

Do I need to make updates to how naturecounts requests the data? Thanks!

EDIT: This also affects subnational regional filters

[denislepage]

Thanks! We’ll look into it. There were some changes done recently that would explain this.

[steffilazerte]

I also get this error with subnational codes

[denislepage]

Hi @steffilazerte

The problem should be fixed. Can you confirm on sandbox? I haven't deployed to live yet.

This request now works for me:

https://sandbox.naturecounts.ca/api/data/list_collections?filter={%22bmdeVersion%22:%22BMDE2.00-min%22,%22iba%22:[%22IBA.YK003%22]}

But still fails on the main site (as expected):

https://naturecounts.ca/api/data/list_collections?filter={%22bmdeVersion%22:%22BMDE2.00-min%22,%22iba%22:[%22IBA.YK003%22]}

[denislepage]

Oh, I overlooked the comment about subnational (or subnational2, = subNat2, ie. counties?) codes. subnational could refer to state/provinces (i.e. subnational1), but I don't think we ever use that name for that first level. The same fix should apply:

https://sandbox.naturecounts.ca/api/data/list_collections?filter={%22bmdeVersion%22:%22BMDE2.00-min%22,%22subNat2%22:[%22CA.ON.NF%22],%22species%22:[15770]}

Fails on main:
https://naturecounts.ca/api/data/list_collections?filter={"bmdeVersion":"BMDE2.00-min","subNat2":["CA.ON.NF"],"species":[15770]}

[steffilazerte]

Yup, it was subnational2.

Everything works on sandbox now, but there are some funny things returned.

1. Using the filter of IBA AB001 returns the IBA requested, but also observations with and IBA of "N/A". I think it should only return observations with the specific IBA, right?
2. There seems to only be one BCR now (BCR 10)... is that intentional? (see now BCRs matching changed #42 )

Or is this the nature of the sandbox data, and unrelated to the sandbox API?

[steffilazerte]

@denislepage, do you know when/if these last issues will be addressed?