README update and minor environment fix

I updated the README to clarify the language, update the fixed issues with the apptainer commands, and remove repeated information.

I changed the name of the conda environment in environment.yml because "base" cannot be used for a conda environment.

Author: rclune

README.md

@@ -13,48 +13,29 @@ This repository contains the code for the [associated preprint](https://www.bior
 }
 ```
 
-You can clone this repository using git, then inference can be run when you have downloaded model weights and the necessary apptainer to run (see below).
-
-To download model weights:
-  * [click here for the best RNA-design weights](https://files.ipd.uw.edu/pub/2025_RFDpoly/train_session2024-07-08_1720455712_BFF_3.00.pt)
-  * [click here for the best general multi-polymer weights](https://files.ipd.uw.edu/pub/2025_RFDpoly/train_session2024-06-27_1719522052_BFF_7.00.pt)
-
-This software runs using apptainers. To download the associated apptainer `.sif` file, [click here](https://files.ipd.uw.edu/pub/2025_RFDpoly/SE3nv.sif)
-Equivalent setup instructions for conda are in progress.
-
-Once you have these files, you can follow the [design tutorial](https://github.com/RosettaCommons/RFDpoly/blob/main/RFDpoly_tutorial.pdf) to see various design tasks and the associated documentation/syntax.
-
-Contact [email protected] if you have trouble accessing files or hit any bugs.
-
-
+## Installation and setup
 
+0. Clone the repository 
+```bash
+git clone [email protected]:RosettaCommons/RFDpoly.git
+```
 
+1. Set environment variables and create directories 
 
-## Installation and setup
+Set relevant paths for installation and then create the directories the paths point to:
 
-1. Set relevant paths for installation and navigate to directory for setup (change accordingly given your filesystem and choice of storage location):
-```Bash
-# Where you choose to install things (CHANGE THESE):
-RFDPOLY_DIR=~/git/RFDpoly_paper_version
-WEIGHTS_DIR=~/git/RFDpoly_paper_version/weights
-ENV_DIR=~/git/RFDpoly_paper_version/exec
+```bash
+# Feel free to change the exact names of the directories
+RFDPOLY_DIR=/path/to/RFDpoly
+WEIGHTS_DIR=/path/to/RFDpoly/weights
 
-# Wherever you want to run some design jobs (ALSO CHANGE THIS):
-DESIGN_DIR=~/git/RFDpoly_paper_version/design_jobs
+# path for where to put the Apptainer .sif file
+ENV_DIR=/path/to/RFDpoly/exec
 
 # Make the directories
 mkdir -p $RFDPOLY_DIR
 mkdir -p $WEIGHTS_DIR
 mkdir -p $ENV_DIR
-mkdir -p $DESIGN_DIR
-
-```
-
-2. Navigate to the directory where you want to set up the repo, and clone it:
-```Bash
-cd $RFDPOLY_DIR
-git clone [email protected]:RosettaCommons/RFDpoly.git
-
 ```
 
 3. Navigate to the directory where you want to save model weights, download them, and set path variable for weights of your choice:
@@ -68,64 +49,74 @@ curl -O https://files.ipd.uw.edu/pub/2025_RFDpoly/train_session2024-06-27_171952
 curl -O https://files.ipd.uw.edu/pub/2025_RFDpoly/train_session2024-07-08_1720455712_BFF_3.00.pt
 
 # Set weights path (replace with your choice of weights):
-MODEL_WEIGHTS_PATH=$WEIGHTS_DIR/train_session2024-07-08_1720455712_BFF_3.00.pt
-export RFDPOLY_CKPT_PATH="$MODEL_WEIGHTS_PATH"
-
+export RFDPOLY_CKPT_PATH=\$WEIGHTS_DIR/train_session2024-07-08_1720455712_BFF_3.00.pt
 ```
 
 4. Setting up the environment:
-Navigate to the directory where you want to apptainer file, download it, and set path variable:
+Navigate to the directory where you want to [Apptainer](https://apptainer.org/) file, download it, and set path variable:
 ```Bash
 cd $ENV_DIR
 curl -O https://files.ipd.uw.edu/pub/2025_RFDpoly/SE3nv.sif
 APPTAINER_PATH=$ENV_DIR/SE3nv.sif
-
 ```
+Downloading the `.sif` file may take several minutes. 
+
+<!--Equivalent python environments can be set up using [conda](https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html), or an environment manager of your choice. Environment files are provided in `rf_diffusion/environment` and can be used as follows:
+```bash
+apptainer build SE3nv.sif SE
+```-->
 
-Equivalent python environments can be set up using conda, or an environment manager of your choice.
-All dependencies and packages are specified in the files found in `rf_diffusion/environment` (see `environment.yml` and `SE3nv.spec`) 
+## Basic Use Case Example: Testing Your Setup
+
+In this example, RFDpoly will create a single design
+with three chains:
+1. A DNA chain with 33 segments
+2. An RNA chain with 33 segments
+3. A protein chain with 75 residues
+
+Make a directory to store the outputs of the demo and store 
+it in the `DESIGN_DIR` environment variable:
+```bash
+export DESIGN_DIR=/path/to/your/output/directory
+```
 
-## Testing that everything works (the *DEMO*):
-Change directory to run from within the RFDpoly directory useful for filepath searches if you are not providing an input pdb), and run example inference script.
-Be sure to specify your desired output directory with `$DESIGN_DIR`.
 ```Bash
 cd $DESIGN_DIR/
 
-$APPTAINER_PATH $RFDPOLY_DIR/RFDpoly/rf_diffusion/run_inference.py --config-name=multi_polymer \
+apptainer run --nv $APPTAINER_PATH $RFDPOLY_DIR/rf_diffusion/run_inference.py --config-name=multi_polymer \
 diffuser.T=50 \
-inference.ckpt_path=$MODEL_WEIGHTS_PATH \
+inference.ckpt_path=$RFDPOLY_CKPT_PATH \
 inference.num_designs=1 \
 contigmap.contigs=[\'33\ 33\ 75\'] \
 contigmap.polymer_chains=[\'dna\',\'rna\',\'protein\'] \
 inference.output_prefix=$DESIGN_DIR/test_outputs/basic_uncond_test01
-
 ```
 The initial run will take a little while to precompute the IGSO3 cache, but subsequent runs will be more direct and quick.
 
 
 if that throws errors, then try:
-```
-$APPTAINER_PATH $RFDPOLY_DIR/RFDpoly/rf_diffusion/run_inference.py --config-name=multi_polymer \
+```bash
+cd $DESIGN_DIR
+
+apptainer run --nv $APPTAINER_PATH $RFDPOLY_DIR/rf_diffusion/run_inference.py --config-name=multi_polymer \
 diffuser.T=50 \
-inference.ckpt_path=$MODEL_WEIGHTS_PATH \
-inference.input_pdb=$RFDPOLY_DIR/RFDpoly/rf_diffusion/test_data/DBP035.pdb \
-inference.num_designs=3 \
+inference.ckpt_path=$RFDPOLY_CKPT_PATH \
+inference.input_pdb=$RFDPOLY_DIR/rf_diffusion/test_data/DBP035.pdb \
+inference.num_designs=1 \
 contigmap.contigs=[\'33\ 33\ 75\'] \
 contigmap.polymer_chains=[\'dna\',\'rna\',\'protein\'] \
 inference.output_prefix=$DESIGN_DIR/test_outputs/basic_uncond_test01
-
-
 ```
+
 explanation: model initialization searches for an input pdb filepath, even if you aren't performing motif scaffolding. Providing a (real) dummy filepath will fix this is the default search paths are unsuccessful.
 
-**Expected output:** a three chain .pdb file, containing a complex of DNA (chain A), RNA (chain B), and protein (chain C).
 **NOTE:** using the `--config-name=multi_polymer` specification is the best way to ensure that all settings work together as expected, and are consistent with the behavior reported in the manuscript.
 
 ## If the example (demo) command above works, proceed to exploration of the full [design tutorial](https://github.com/RosettaCommons/RFDpoly/blob/main/RFDpoly_tutorial.pdf).
 The full design tutorial contains many inference commands for the types of designs reported in the RFDpoly paper.
-The design tutorial also is intented to provide documentation, and explain arguments in the context of their use-cases.
-
+The design tutorial is intended to provide documentation, and explain arguments in the context of their use-cases.
 
+Contact **[email protected]** if you have trouble accessing files or hit any bugs.
 
 # Additional Software Policy checklist items:
 ## System requirements
@@ -134,7 +125,7 @@ The design tutorial also is intented to provide documentation, and explain argum
   - Primary workflow: Linux host with [Apptainer](https://apptainer.org/) ≥ 1.1 (or Singularity equivalent). The bundled container image runs Ubuntu 22.04.2 LTS.
   - Alternative workflow: macOS 13+/Windows 11 via Conda (CPU only unless you install CUDA-capable PyTorch wheels).
 - **Software dependencies**
-  - Apptainer runtime to execute `SE3nv.sif`.
+  - Apptainer runtime to execute `SE3nv.sif`, this can be built from `SE3nv.spec` in `rf_diffusion/environment`
   - For native/Conda installs: Python 3.10, PyTorch 1.13.1, CUDA 11.7 toolchain (if using an NVIDIA GPU), PyRosetta 2023.09+, DGL 1.0.1, e3nn 0.5.1, hydra-core 1.3.2, and the packages listed in `rf_diffusion/environment/environment.yml`.
   - **NOTE:** All dependencies are specified in `rf_diffusion/environment/`, such that users can set up an equivalent working environment on any operating system.
   - Downloaded assets: model checkpoints (`*.pt`, ~2.3 GB each) and the container image (`SE3nv.sif`, ~8 GB).

rf_diffusion/environment/environment.yml

@@ -1,4 +1,4 @@
-name: base
+name: RFDpoly_env
 channels:
   - pytorch
   - https://conda.rosettacommons.org