list_surveys() returns journal DOI (not Zenodo DOI) for MixIT, breaking download_survey()

[sbfnk-bot]

list_surveys() returns the wrong url for the MixIT survey — its journal DOI (https://doi.org/10.1038/s43856-026-01543-4) instead of its Zenodo dataset DOI (https://doi.org/10.5281/zenodo.17579537). download_survey() then correctly refuses the non-Zenodo DOI, so the survey can't be downloaded.

Reproduce

library(contactsurveys)
sl <- list_surveys()
mix <- sl[grepl("MixIT", sl$title), ]
mix$url
#> "https://doi.org/10.1038/s43856-026-01543-4"   # journal DOI

download_survey(mix$url, tempdir())
#> ✖ DOI '10.1038/s43856-026-01543-4' does not seem to be a Zenodo DOI
#> Error: Request failed after 4 attempts.

The Zenodo dataset itself exists and downloads fine via the correct DOI:

f <- download_survey("https://doi.org/10.5281/zenodo.17579537", tempdir())
#> 8 files, md5-verified

(socialmixr::load_survey(f) then loads it cleanly: 4979 participants, 23718 contacts.)

Cause

The MixIT Zenodo record (10.5281/zenodo.17579537, in the social_contact_data community) carries the journal DOI as a related/alternate identifier, and list_surveys() seems to pick that up as the record's url rather than the record's own Zenodo DOI. Every other survey lists a 10.5281/zenodo.* DOI.

Impact

This surfaces downstream as the monthly socialmixr survey-check failing for MixIT (epiforecasts/socialmixr#351, and #341 before it) with an opaque invalid 'file' argument.

Suggested fix

list_surveys() should return the record's own Zenodo DOI (the 10.5281/zenodo.* concept/version DOI) as url, rather than a related/alternate journal DOI.

[lwillem]

Fixing this issue would be great to make the package more robust. However, I also reached out to the ZENODO repo maintainer(s) on this.

[LLucchini]

Thanks for reaching out.
I have now removed the repeated publication DOI from the "Alternate identifiers" field.
Let me know if this solves the issue.