Update studies from v1 to v2

[iaindillingham]

Simon noticed^1 that many studies are using research-action v1. This version notifies the following users if a restricted dataset is referenced:

https://github.com/opensafely-core/research-action/blob/c1c04e8f8de82ea8764d99bcc9c0a65a3abba65d/action.sh#L52

The latest version is v2. This version notifies the following users:

https://github.com/opensafely-core/research-action/blob/dfd59edbcd71109c0f2e517d4b3a797155c9abca/action.yml#L9

Actions

• [ ] Open and merge PRs on the affected studies. These PRs should update .github/workflows/test_runner.yaml

Affected studies

covid-ve-change-over-time
data-notebooks-andrea
lone_households
MH_pandemic
NICE-audit-BP
NICE-audit-lipids
non-specific-immunity-research
oral-anticoagulant-covid
os-demo-research
OS_OC_v001-research
os-sch-children-2021
OS_VidGroupClinic_v001-research
output-explorer-test-repo
PaCa_Enzyme_Rx
Pancreatic_cancer
pathology-comparators-short-report
pincer-measures-change-detection
pincer-measures
population-outcomes-burden-research
post-admission-admissions-research
post-covid-bristol
post-covid-cvd-methods
post-covid-diabetes
post-covid-kidney-outcomes
Post_covid_nice_compliance
post-covid-outcomes-research
post-covid-pre-vaccinated-cardiovascular
post-covid-unvaccinated
post-covid-vaccinated
PostOpCovid
post-vax-ae-attendance
post-vax-outcomes-report
pregnancy-vaccine-safety
primary-care-covid-codes-research
primis-codelist-prevalence
principle-inhaler-uptake
PSA2021
qof-utilities
rapid-reports
raw-data-plausibility-checks
rct-test
reinfection
release-process-test
renal-short-data-report
representativeness
restoration-observatory-emis-research
SA_carehomerx
school-age-children-and-covid2
second-line-diabetes-covid-research
SGTF-617-research
SGTF-CFR-research
SGTF-Omi-research
sotrovimab-and-molnupiravir
sro-gastro
sro-measures-demo
SRO-Measures
SRO-pulse-oximetry
SRO-PULSE-OXIMETRY-UPD
SRO-smr
sro-template
suicide-deaths-pcc
surg-covid-safely
surgery-research
T1DM_covid_research
test-age-distribution
test-point-and-click
tncc_ve
Uptake-of-NHS-home-interventions-during-COVID-19
vaccine-characteristics-research
vaccine-effectiveness-hospital-admissions-validation
vaccine-neuro-cohort
vaccine-rd-research
vaccine_status

[bloodearnest]

Action tags are designed to work differently - the current version should be tagged v2 - its a moving tag.

When it was updated to add Liam, it HEAD was tagged v2 (as per README.md, its not a breaking changed). To bring it up to date we need to run make tag-release, which will tag the current HEAD as v2.

We can also tag v2.1 if we want, but we should update v2 to point to v2.1. This is the convention for action tags: v2 points to the latest of v2.x.y.

[iaindillingham]

Ah, that's really useful! Thank you, @bloodearnest.

[bloodearnest]

We could also quickly fix the v1 action this way too.

If we branch from the v1 tag, change the names, then tag that branch as the new v1, we'll get the correct notifications w/o having to update all the repos.

Of course, we should probably do the work to update all the repos using v1 anyway, but the above is quick fix.

[iaindillingham]

I've run make tag-release, so if a study is run in CI, points to research-action v2 (in .github/workflows/test_runner.yaml), and references a restricted dataset, then the agreed people will be notified.

In today's stand-up, we agreed to defer work on updating studies from v1 to v2 until we've created some guidelines for updating studies. We will likely need to update studies as part of the current Codespaces initiative.