With symmetry, Hydra throws minor warning

Given this input:

/opt/.miniconda/envs/protein_rfdiffusion/bin/python /home/tristin/downloads/large_language_models/RFdiffusion/scripts/run_inference.py --config-name=symmetry inference.output_prefix="/home/tristin/documents/core/rfdiff_out/backbone" inference.num_designs=1 'contigmap.contigs=[50-50/0 50-50/0 50-50]' inference.symmetry="c3" 'potentials.guiding_potentials=["type:olig_contacts,weight_intra:1,weight_inter:0.1"]' potentials.olig_intra_all=True potentials.olig_inter_all=True potentials.guide_scale=2.0 potentials.guide_decay="quadratic"

I get this minor Warning:

Exception: /opt/.miniconda/envs/protein_rfdiffusion/lib/python3.9/site-packages/hydra/_internal/defaults_list.py:251: UserWarning: In 'symmetry': Defaults list is missing _self_. See https://hydra.cc/docs/1.2/upgrades/1.0_to_1.1/default_composition_order for more information warnings.warn(msg, UserWarning)

This appears easily solved by modifying config/inference/symmetry.yaml

From:

defaults:
  - base

To:

defaults:
  - base
  - _self_

My understanding is this manages the order the settings are loaded in, but I'm not familiar enough with RFdiffusion to be sure I'm setting things up right. Best guess for an alternative set up would be to put _self_ before base. Suppose I could move the entire block too, but I suspect that'd just break things. Regardless, this fix works for me.

Note: I believe most users can ignore this warning; however, when setting up a pipeline it may present a minor issue.

Open config/inference/symmetry.yaml and compare its defaults list with the Hydra composition-order guidance linked in the issue. Run the provided symmetry inference command to confirm the warning, then verify that the warning is gone while the symmetry configuration still loads correctly.