Is there a way to combine Inpainting and Fold Conditioning?

Hello! Thank you for all the work and advancements into RFdiffusion!

I am interested in specifying the secondary structure of a new inpainted section of an input pdb. Essentially, I want to keep parts of the input pdb fixed, add connections between the fixed parts, and be able to specify those connections to be beta sheets or alpha helices or etc.

I did try a few runs with provided _ss.pt and _adj.pt files to correspond to the secondary structure I would like the inpainted section to have, but I would get an Assertion Error:

Error executing job with overrides: ['inference.output_prefix=./outputs/guide_test', 'inference.input_pdb=inputs/1sp1_KT_relaxed1.pdb', 'contigmap.contigs=[A1-7/2-2/A8-30]', 'inference.num_designs=2', 'scaffoldguided.scaffoldguided=True', 'scaffoldguided.target_pdb=False', 'scaffoldguided.scaffold_dir=../helper_scripts/E2/']
Traceback (most recent call last):
File "/scratch/ml1628/Klab/RFdiffusion/Zn_hinge/../scripts/run_inference.py", line 84, in main
x_init, seq_init = sampler.sample_init()
File "/cache/home/ml1628/RFdiffusion/rfdiffusion/inference/model_runners.py", line 851, in sample_init
assert L_mapped==self.adj.shape[0]
AssertionError

Before I go down the rabbit hole of trying to get this to work, I wanted to ask: is this possible? Any and all advice would be greatly appreciated!

Reproduce the reported command, then inspect scripts/run_inference.py and rfdiffusion/inference/model_runners.py around sample_init(), especially the assertion comparing L_mapped with self.adj.shape[0]. Compare the contigmap layout with the supplied _ss.pt and _adj.pt files in the scaffoldguided directory. Done means establishing whether combined inpainting and fold conditioning is supported and, if so, documenting the required inputs or reproducer.