ci: added git-cliff changelog automation

[shamykyzer]

● Closes #41.

Adds git-cliff changelog automation. NEWS.md entries are generated from
Conventional-Commits PR titles, run locally by the release manager before tagging.

• cliff.toml: git-cliff config (groups by type, filters to V* tags, links (#NNN) to PRs, skips noise).
• CONTRIBUTING.md: documents the PR-title convention and release command.
• .Rbuildignore: excludes cliff.toml from the package tarball.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (51810ca) to head (75e47e9).

Additional details and impacted files

@@            Coverage Diff            @@
##              main       #49   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            4         4           
  Lines          115       115           
=========================================
  Hits           115       115           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[shamykyzer]

Hello @rpreen, could you have a look at this and let me know what you think?

[rpreen]

I just took a quick look at this and I think the config needs tweaking - if you look at the toml in the Python ACRO repo I made some changes before it was merged - in particular there needs to be a default fall through with category 'other' to handle the cases where someone has merged/committed to the main branch without using conventional commits.

[shamykyzer]

Hi @rpreen, could you please have a look at my changes? I can't seem to pass the macOS CI check either..

[JessUWE]

@shamykyzer Can you look through our stand-up notes and see if any of the suggestions would help.

[shamykyzer]

Again, sorry @JessUWE for not checking the stand-up notes, this should be good to merge now.

I switched to reticulate's own Python resolver instead of blindly picking the first python3 in PATH, which was failing on conda runners.

Before:

python <- get_python()

After:

python <- reticulate::virtualenv_starter()
if (is.null(python)) python <- get_python()

[jim-smith]

Again, sorry @JessUWE for not checking the stand-up notes, this should be good to merge now.

I switched to reticulate's own Python resolver instead of blindly picking the first python3 in PATH, which was failing on conda runners.

Before:

python <- get_python()

After:

python <- reticulate::virtualenv_starter()
if (is.null(python)) python <- get_python()

@shamykyzer need to check this logic (using reticulate's resolver), because 'it's hard to install the R version" is a really common complaint from TREs.

We need to make sure that we pick up the version of python that has acro installed - this will/should be in a virtual environment of its own, and that env should be pushed to the front of PATH - but all the relevant info should be in the installation / help guides

[shamykyzer]

As discussed in the meeting, this PR should be ready for merge now

[jim-smith]

@rpreen sorry -this is a 2-3 minute ask, related to the work you did on making the R-python setup more robust.

@shamykyzer had put in a fix that made the conda macOS runner work - see above- it's currently reverted but we could put it back im. Could you have a look and see what you think.

[rpreen]

the macOS runner failure is another issue not caused by anything here