Broken URL for dataset_get

[fontikar]

Oh dear. During the rename process, we’ve broken load_taxonomic_resources

stable_resources <- load_taxonomic_resources(stable_or_current_data = "stable")
Error in vapply(url, register_, character(1L), registries = registries,  :                                    
  values must be length 1,
 but FUN(X[[1]]) result is length 2
...done

This URL is not found: https://github.com/traitecoevo/APCalign/releases/download/
but https://github.com/traitecoevo/APCalign/releases still exists! 🤷‍♀️ I'm not sure what to do here, maybe GitHub just needs a day to create the download page after renaming?

[fontikar]

GH documentation for reference https://docs.github.com/en/repositories/releasing-projects-on-github/linking-to-releases

[fontikar]

So strange because Copy Link for apc.parquet you get: https://github.com/traitecoevo/APCalign/releases/download/0.0.2.9000/apc.parquet

[fontikar]

Broken on master after rename. going to branch from master (stable-resources) to:

• write a test to capture this
• try debug issue

[fontikar]

Error comes down to contentid::register()

apc_hash <- contentid::register(url)
Error in vapply(url, register_, character(1L), registries = registries,  :                                    
  values must be length 1,
 but FUN(X[[1]]) result is length 2

[fontikar]

Interesting that contentid::register works for apni URL (https://github.com/traitecoevo/APCalign/releases/download/0.0.2.9000/apni.parquet) but not for the APC parquet

[fontikar]

Was able to register content ids with:

• csv APC from website and saving locally as parquet so its not the parquet...
• parquet download from GitHub releases

 APC <-
+     readr::read_csv(
+         "https://biodiversity.org.au/nsl/services/export/taxonCsv",
+         n_max = 110000,
+         col_types =
+             readr::cols(
+                 .default = readr::col_character(),
+                 proParte = readr::col_logical(),
+                 taxonRankSortOrder = readr::col_double(),
+                 created = readr::col_datetime(format = ""),
+                 modified = readr::col_datetime(format = "")
+             )
+ )
                                                                                                              
> arrow::write_parquet(APC, "ignore/apc.parquet")
> contentid::register("ignore/apc.parquet")
[1] "hash://sha256/8f7c2de1b19051081bbf1919f876f20b83614cab063ddde172723c8734eab420" 
contentid::register("ignore/gh_apc.parquet")
[1] "hash://sha256/bc369b29b9c5c3e600135d8c9aa49a09cfb83e905be7d0b9607543b89a5bd0b4"      

[fontikar]

A hash is created for the APCalign release url
registry.csv

But searching for the hash here gives the link to ausflora!!!

[fontikar]

Awaiting for GitHub to update servers test again in a day or two