This may be a lower priority issue since this is an older version of activitysim, but reporting here for posterity. When trying to run activitysim 1.2.2 with e.g. test_prototype_mtc (or sandag-abm3-example too), it just fails silently and no output (not even logs) are produced.

The only bigger issue here is that the current docs asks users to install with python=3.9, which would default to activitysim 1.2.2. Having the primary example then failing when new users are following the log creates a point of confusion.

If there's not an obvious reason to why this is happening, and I assume it's not worth anyone's time to figure this out for an old version of activitysim, perhaps the best solution is to either 1. update the docs and/or 2. take down 1.2.2 from conda-forge?

To Reproduce
Steps to reproduce the behavior:

1. Follow the docs https://activitysim.github.io/activitysim/v1.3.1/users-guide/modelsetup.html and install activitysim with mamba create -n asim python=3.9 activitysim -c conda-forge --override-channels
2. Then follow the instructions to run test_prototype_mtc at https://activitysim.github.io/activitysim/v1.3.1/users-guide/run_primary_example.html
3. After issuing activitysim run -c configs -o output -d data, everything finishes in like a second and no output (screen or logs or anything else) is produced at all.