Describe the bug
Some ActivitySim GitHub Actions tests are failing. I don't think this is caused by any recent code commits from the development team, because:
- Tests are failing in the
Setup Miniforge step before doing anything related to ActivitySim, and the error message is about extracting a package (see screenshots below), and
- The error happens only in some, but not all tests, even though all tests have the same workflow steps for
Setup Miniforge. And every time there are different tests failing, e.g., production_semcog and estimation unit tests failed in this job; while standard 1-zone MTC failed in this job.
I think the workflow configuration for Setup Miniforge needs to be updated to make it stable. After some research, the most likely culprit is the libmamba solver that is current used for the workflow. Although it is faster than the default conda solver, it is less stable. Changing it to the classic conda solver seems to work in my test: https://github.com/wsp-sag/activitysim/actions/runs/13526804691/job/37799900361.
To Reproduce
Push a commit or open a pull request, you will probably see some tests failing in Setup Miniforge
Expected behavior
- GitHub Actions tests should not fail in
Setup Miniforge when creating python env.
- If one test fails in
Setup Miniforge, we expect all tests could fail at the same step.
Screenshots


Additional context
Add any other context about the problem here.
Describe the bug
Some ActivitySim GitHub Actions tests are failing. I don't think this is caused by any recent code commits from the development team, because:
Setup Miniforgestep before doing anything related to ActivitySim, and the error message is about extracting a package (see screenshots below), andSetup Miniforge. And every time there are different tests failing, e.g., production_semcog and estimation unit tests failed in this job; while standard 1-zone MTC failed in this job.I think the workflow configuration for Setup Miniforge needs to be updated to make it stable. After some research, the most likely culprit is the
libmambasolver that is current used for the workflow. Although it is faster than the default conda solver, it is less stable. Changing it to the classiccondasolver seems to work in my test: https://github.com/wsp-sag/activitysim/actions/runs/13526804691/job/37799900361.To Reproduce
Push a commit or open a pull request, you will probably see some tests failing in
Setup MiniforgeExpected behavior
Setup Miniforgewhen creating python env.Setup Miniforge, we expect all tests could fail at the same step.Screenshots

Additional context
Add any other context about the problem here.