rc-foundry 0.1.1__tar.gz

rc_foundry-0.1.1/.git

@@ -0,0 +1 @@
+gitdir: /home/rohith/modelhub/.git/worktrees/release

rc_foundry-0.1.1/.github/workflows/lint_trunk.yaml

@@ -0,0 +1,36 @@
+name: lint_trunk
+
+on:
+  push:
+    branches: [main, trunk, wip/for-release, wip/remove-chemdata]
+  pull_request:
+    branches: [main, trunk, wip/for-release, wip/remove-chemdata]
+  pull_request_target:
+    types: [ready_for_review]
+  workflow_dispatch:
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  lint:
+    name: ruff (check code style)
+    # NOTE: We use an ubuntu runner to not be dependent on possibly limited digs infra
+    #   for this tiny linting job. This means we can have many lint jobs accross repos in parallel.
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.11"
+          cache: 'pip'
+      - name: Extract ruff version from requirements.txt
+        run: |
+          echo "RUFF_VERSION=$(grep 'ruff==' pyproject.toml | sed 's/.*ruff==\(.*\)/\1/')" >> $GITHUB_ENV
+      - name: Install ruff
+        run: pip install ruff==${{ env.RUFF_VERSION }}
+      - name: Ruff format
+        run: ruff format --diff src tests scripts notebooks
+      - name: Ruff check
+        run: ruff check src tests scripts notebooks

rc_foundry-0.1.1/.gitignore

@@ -0,0 +1,227 @@
+# For docs / outputs from example notebooks;
+examples/*.cif
+**.ckpt
+**.pt
+
+# Base .gitignore from https://github.com/github/gitignore/blob/main/Python.gitignore
+*.lock
+**.nfs**
+**.dat
+
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*/__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebooks (unless explicitly not ignored)
+.ipynb_checkpoints
+**/.ipynb
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#pdm.lock
+#   pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
+#   in version control.
+#   https://pdm.fming.dev/latest/usage/project/#working-with-version-control
+.pdm.toml
+.pdm-python
+.pdm-build/
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/
+
+# VS Code
+.vscode
+.cursor
+.history/
+
+# Ruff
+ruff.toml
+.ruff_cache
+
+# Development
+dev.py
+
+# Pytest
+*.benchmarks/
+
+# Images
+*.png
+*.pdf
+*.svg
+*.jpg
+*.jpeg
+*.gif
+*.bmp
+*.tiff
+
+# Versioning
+**/version.py
+
+# W&B
+wandb/
+
+# Hydra
+.hydra/
+
+# Logs
+logs/
+
+# Other
+*.sif
+*.out
+
+# Misc
+**/notebooks/
+**/scratch/
+**/outputs/
+**/slurm_logs/
+**/tmp/
+*.ckpt
+**/distillation/
+**/benchmarks/
+
+# AI
+**/CLAUDE.md
+**/.claude

rc_foundry-0.1.1/.ipd/apptainer/rf3-dev.def

@@ -0,0 +1,114 @@
+Bootstrap: localimage
+From: /net/software/containers/versions/modelhub/pytorch_25.06-py3.sif
+IncludeCmd: yes
+
+# RF3 Development Container - Dependencies only (excludes atomworks, which we expose through a submodule for development)
+
+# To build this Apptainer, from the project root, run:
+# apptainer build rf3-dev.sif .ipd/apptainer/rf3-dev.def
+
+%labels
+    Author Institute for Protein Design
+    Version 1.0-dev
+    Description RosettaFold3 - Dependencies only (for development and model training)
+
+%setup
+    # Create a directory in the container to bind the host's current working directory
+    mkdir ${APPTAINER_ROOTFS}/foundry_host
+    # ... for mounting `/projects` with --bind
+    mkdir ${APPTAINER_ROOTFS}/projects
+    # ... for mounting `/databases` with --bind
+    mkdir ${APPTAINER_ROOTFS}/net
+    # ... for mounting `/squash` with --bind
+    mkdir ${APPTAINER_ROOTFS}/squash
+
+%files
+    /etc/localtime
+    /etc/hosts
+    pyproject.toml /opt/core_pyproject.toml
+    models/rf3/pyproject.toml /opt/rf3_pyproject.toml
+    lib/atomworks/pyproject.toml /opt/atomworks_pyproject.toml
+
+%post
+    ## GENERAL SETUP
+
+    # Common symlinks (within container)
+    ln -s /net/databases /databases
+    ln -s /net/software /software
+    ln -s /home /mnt/home
+    ln -s /projects /mnt/projects
+    ln -s /net /mnt/net
+
+    # Update system and install essential packages
+    apt-get update && apt-get install -y \
+        build-essential \
+        git \
+        libxrender1 \
+        libxrender-dev \
+        libx11-6 \
+        libx11-dev \
+        libxext6 \
+        libxext-dev \
+        && rm -rf /var/lib/apt/lists/*
+
+    ## PYTHON DEPENDENCY INSTALLATION
+
+    # Upgrade pip
+    python -m pip install --upgrade pip
+
+    # Install uv for fast dependency resolution
+    pip install uv
+
+    # Auto-generate dependency lists for both core and RF3, with all extra dependencies
+    # (Core)
+    mv /opt/core_pyproject.toml /opt/pyproject.toml
+    uv pip compile /opt/pyproject.toml --output-file /opt/core_requirements.txt --all-extras
+    rm /opt/pyproject.toml
+
+    # (RF3)
+    mv /opt/rf3_pyproject.toml /opt/pyproject.toml
+    uv pip compile /opt/pyproject.toml --output-file /opt/rf3_requirements.txt --all-extras
+    rm /opt/pyproject.toml
+
+    # (AtomWorks)
+    mv /opt/atomworks_pyproject.toml /opt/pyproject.toml
+    uv pip compile /opt/pyproject.toml --output-file /opt/atomworks_requirements.txt --all-extras
+    rm /opt/pyproject.toml
+
+    # Merge and dedupe requirements, excluding packages we don't want
+    # (atomworks is mounted from host; torch/numpy/nvidia-* provided by base image)
+    # (pynvml/packaging/pandas/markdown-it-py/triton from NGC container to avoid conflicts)
+    # Deduplicate by package name (keeping first occurrence) to handle version conflicts
+    cat /opt/core_requirements.txt /opt/rf3_requirements.txt /opt/atomworks_requirements.txt | \
+        grep -vE "^(atomworks|torch(|vision|audio)|numpy|nvidia-.*|pynvml|packaging|pandas|markdown-it-py|triton)==" | \
+        awk -F'==' '!seen[$1]++' > /opt/combined_requirements.txt
+
+    # Print combined requirements for debugging
+    echo "=== Combined requirements to install ==="
+    cat /opt/combined_requirements.txt
+    echo "========================================"
+
+    # Install the cleaned, combined requirements (use uv for better resolver)
+    uv pip install --system --break-system-packages -r /opt/combined_requirements.txt
+
+    # Clean up to minimize container size
+    uv cache clean
+    pip cache purge
+    apt-get clean
+    rm -rf /var/lib/apt/lists/* 2>/dev/null || true
+    rm -rf /root/.cache 2>/dev/null || true
+    rm -rf /opt/*.txt 2>/dev/null || true
+
+%environment
+    # (Flag to increase accessible GPU memory)
+    export PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True
+
+    # (Turn off NVLink)
+    export NCCL_P2P_DISABLE=1
+
+%runscript
+   # NOTE: The %runscript is invoked when the container is run without specifying a different command.
+   exec python "$@"
+
+%help
+    RosettaFold3 (RF3) development container with dependencies only.

rc_foundry-0.1.1/.ipd/apptainer/rf3-full.def

@@ -0,0 +1,81 @@
+Bootstrap: localimage
+From: .ipd/apptainer/rf3-dev.sif
+IncludeCmd: yes
+
+# RF3 Full Production Container - Includes foundry, rf3, and atomworks packages installed
+
+# To build this Apptainer, from the project root, run:
+# apptainer build rf3-full.sif .ipd/apptainer/rf3-full.def
+
+%labels
+    Author Institute for Protein Design
+    Version 1.0.0
+    Description RosettaFold3 - Full production container with all packages installed
+
+%setup
+    # Copy entire repository to container
+    echo "Copying repository..."
+
+    # Create base directory
+    mkdir -p ${APPTAINER_ROOTFS}/opt/foundry
+
+    # Copy everything except large/unnecessary files
+    rsync -av --info=progress2 \
+        --exclude='.venv' \
+        --exclude='__pycache__' \
+        --exclude='*.pyc' \
+        --exclude='.pytest_cache' \
+        --exclude='*.egg-info' \
+        --exclude='.ruff_cache' \
+        --exclude='outputs' \
+        --exclude='logs' \
+        --exclude='*.sif' \
+        --exclude='distillation' \
+        --exclude='benchmarks' \
+        --exclude='**/slurm_logs' \
+        ./ ${APPTAINER_ROOTFS}/opt/foundry/
+
+    echo "Repository copied successfully."
+
+%post
+    echo "Installing packages into container..."
+
+    # Do editable installs so everything works as in development
+    cd /opt/foundry
+
+    # Install foundry (editable, no dependencies - already in base image)
+    echo "Installing foundry..."
+    uv pip install --system --break-system-packages --no-deps -e .
+
+    # Install rf3 (editable, no dependencies - already in base image)
+    echo "Installing rf3..."
+    cd /opt/foundry/models/rf3
+    uv pip install --system --break-system-packages --no-deps -e .
+
+    # Install atomworks (editable, no dependencies - already in base image)
+    echo "Installing atomworks..."
+    cd /opt/foundry/lib/atomworks
+    uv pip install --system --break-system-packages --no-deps -e .
+
+    # Cleanup caches only
+    echo "Cleaning up caches..."
+    uv cache clean
+    pip cache purge
+    apt-get clean
+    rm -rf /var/lib/apt/lists/* 2>/dev/null || true
+    rm -rf /root/.cache 2>/dev/null || true
+
+    echo "Package installation complete."
+
+%runscript
+    # Simple runscript - just run Python with any arguments
+    exec python "$@"
+
+%help
+    RosettaFold3 (RF3) full production container.
+
+    Contains foundry, rf3, and atomworks packages installed in site-packages.
+    All dependencies are included from the base rf3-dev container.
+
+    Usage:
+        apptainer exec --nvccli rf3-full.sif rf3 fold inputs="example.json"

rc_foundry-0.1.1/.ipd/apptainer/rfd3-full.def

@@ -0,0 +1,81 @@
+Bootstrap: localimage
+From: .ipd/apptainer/rf3-dev.sif
+IncludeCmd: yes
+
+# RFD3 Full Production Container - Includes foundry, rfd3, and atomworks packages installed
+
+# To build this Apptainer, from the project root, run:
+# apptainer build rfd3-full.sif .ipd/apptainer/rfd3-full.def
+
+%labels
+    Author Institute for Protein Design
+    Version 1.0.0
+    Description RFDiffusion3 - Full production container with all packages installed
+
+%setup
+    # Copy entire repository to container
+    echo "Copying repository..."
+
+    # Create base directory
+    mkdir -p ${APPTAINER_ROOTFS}/opt/foundry
+
+    # Copy everything except large/unnecessary files
+    rsync -av --info=progress2 \
+        --exclude='.venv' \
+        --exclude='__pycache__' \
+        --exclude='*.pyc' \
+        --exclude='.pytest_cache' \
+        --exclude='*.egg-info' \
+        --exclude='.ruff_cache' \
+        --exclude='outputs' \
+        --exclude='logs' \
+        --exclude='*.sif' \
+        --exclude='distillation' \
+        --exclude='benchmarks' \
+        --exclude='**/slurm_logs' \
+        ./ ${APPTAINER_ROOTFS}/opt/foundry/
+
+    echo "Repository copied successfully."
+
+%post
+    echo "Installing packages into container..."
+
+    # Do editable installs so everything works as in development
+    cd /opt/foundry
+
+    # Install foundry (editable, no dependencies - already in base image)
+    echo "Installing foundry..."
+    uv pip install --system --break-system-packages --no-deps -e .[all]
+
+    # Install rfd3 (editable, no dependencies - already in base image)
+    echo "Installing rfd3..."
+    cd /opt/foundry/models/rfd3
+    uv pip install --system --break-system-packages --no-deps -e .
+
+    # Install atomworks (editable, no dependencies - already in base image)
+    echo "Installing atomworks..."
+    cd /opt/foundry/lib/atomworks
+    uv pip install --system --break-system-packages --no-deps -e .
+
+    # Cleanup caches only
+    echo "Cleaning up caches..."
+    uv cache clean
+    pip cache purge
+    apt-get clean
+    rm -rf /var/lib/apt/lists/* 2>/dev/null || true
+    rm -rf /root/.cache 2>/dev/null || true
+
+    echo "Package installation complete."
+
+%runscript
+    # Simple runscript - just run Python with any arguments
+    exec python "$@"
+
+%help
+    RFDiffusion3 (RFD3) full production container.
+
+    Contains foundry, rfd3, and atomworks packages installed in site-packages.
+    All dependencies are included from the base rf3-dev container.
+
+    Usage:
+        apptainer exec --nvccli rfd3-full.sif rfd3 run_inference inputs="example.json"

rc_foundry-0.1.1/.ipd/shebang/README.md

@@ -0,0 +1,6 @@
+This directory contains scripts that are not to be run directly by the user. 
+They are [SHEBANG scripts](https://en.wikipedia.org/wiki/Shebang_(Unix)) that are used to run the appropriate apptainer container.
+
+For example, the script `rf3_exec.sh` is used to run the RF3 apptainer container with the latest apptainer image stored locally or at the IPD.
+
+The shebang lines (`#!/bin/bash` ...) at the top of entry point scripts like `train.py` redirect the system to here to find the correct apptainer container.

rc_foundry-0.1.1/.ipd/shebang/mpnn_exec.sh

@@ -0,0 +1,129 @@
+#!/usr/bin/bash
+
+###################
+# You can add the path to this file as the shebang line in your python script. 
+# Then by default, the python script will be executed with the python interpreter
+# in the SIF_PATH container. Here, we launch the container with nvidia gpu and slurm support.
+#
+# Example shebang: #!/usr/bin/env -S /bin/sh -c '"$(dirname "$0")/.ipd/shebang/mpnn_exec.sh" "$0" "$@"'
+###################
+
+# Let the user know this script is setting things up behind the scene
+SCRIPT_PATH=$(realpath $0)
+SCRIPT_DIR=$(dirname $SCRIPT_PATH)
+echo '################## Start shebang info ##################'
+echo "The file $SCRIPT_PATH is being run as a shebang executable.
+    It will...
+
+    1. Add 'src/foundry', 'models/mpnn/src', and 'lib/atomworks/src' to your PYTHONPATH.
+    2. Run your python script from the right container, which contains all dependencies.
+    3. Launch the container with slurm and nvidia gpu support."
+
+# Extract the path to the Python script from the arguments
+PYTHON_SCRIPT=$(realpath "$1")
+shift
+
+# Find repository root by looking for .project-root file
+find_repo_root() {
+    local current_dir="$1"
+    while [ "$current_dir" != "/" ]; do
+        if [ -f "$current_dir/.project-root" ]; then
+            echo "$current_dir"
+            return 0
+        fi
+        current_dir="$(dirname "$current_dir")"
+    done
+    return 1
+}
+
+echo
+echo "Searching for repository root directory..."
+REPO_ROOT=$(find_repo_root "$(dirname "$PYTHON_SCRIPT")")
+if [ -z "$REPO_ROOT" ]; then
+    echo "Error: Could not find .project-root file in any parent directory"
+    exit 1
+else
+    echo "... found repository root at '$REPO_ROOT'"
+fi
+
+# Function to add a directory to PYTHONPATH if it's not already included
+add_to_pythonpath() {
+    local dir_path="$1"
+    if [[ ":$PYTHONPATH:" != *":$dir_path:"* ]]; then
+        export PYTHONPATH="$dir_path:$PYTHONPATH"
+        echo "Added '$dir_path' to PYTHONPATH."
+    else
+        echo "'$dir_path' is already in PYTHONPATH."
+    fi
+}
+
+# Add foundry, mpnn, and atomworks to PYTHONPATH
+echo
+echo "Adding foundry, MPNN, and atomworks to PYTHONPATH..."
+FOUNDRY_PATH="$REPO_ROOT/src"
+MPNN_PATH="$REPO_ROOT/models/mpnn/src"
+ATOMWORKS_PATH="$REPO_ROOT/lib/atomworks/src"
+add_to_pythonpath "$FOUNDRY_PATH"
+add_to_pythonpath "$MPNN_PATH"
+add_to_pythonpath "$ATOMWORKS_PATH"
+
+echo
+echo "Fetching the appropriate apptainer image..."
+
+SIF_PATH="$REPO_ROOT/.ipd/apptainer/rf3-dev.sif"
+
+echo "... looking for a local apptainer image at '$SIF_PATH'"
+if [ ! -f "$SIF_PATH" ]; then
+    echo "... apptainer not found. To build it, run: apptainer build .ipd/apptainer/rf3-dev.sif .ipd/apptainer/rf3-dev.def"
+    echo "Attempting to run $PYTHON_SCRIPT with $(which python)"
+    SIF_PATH=""
+fi
+
+# Function to print debug=mode warning
+print_debug_warning() {
+    echo
+    echo "###############################################################################"
+    echo "#                                                                             #"
+    echo "#                               ⚠️  WARNING ⚠️                                  #"
+    echo "#                     RUNNING WITH DEBUGPY ON PORT $DEBUG_PORT                       #"
+    echo "#                      DON'T FORGET TO ATTACH A DEBUGGER                      #"
+    echo "#                                                                             #"
+    echo "###############################################################################"
+    echo
+}
+
+if [ -n "$DEBUG_PORT" ]; then
+    print_debug_warning
+    python_cmd="python -m debugpy --listen $DEBUG_PORT --wait-for-client"
+else
+    python_cmd="python"
+    echo
+fi
+
+if [ ! -z $SIF_PATH ]; then
+    echo "Running $PYTHON_SCRIPT with apptainer: $SIF_PATH."
+
+    # Determine GPU binding method: prefer --nvccli, fallback to --nv, then no GPU
+    GPU_FLAG=""
+    if command -v nvidia-container-cli &> /dev/null && [ -c /dev/nvidiactl ] 2>/dev/null; then
+        GPU_FLAG="--nvccli"
+        echo "Using --nvccli for GPU support"
+    elif [ -c /dev/nvidiactl ] 2>/dev/null; then
+        GPU_FLAG="--nv"
+        echo "Using --nv for GPU support"
+    else
+        echo "No GPU detected, running without GPU support"
+    fi
+
+    echo '################## End shebang info ####################'
+    echo
+    /usr/bin/apptainer exec $GPU_FLAG --slurm \
+        --bind "$REPO_ROOT:$REPO_ROOT" \
+        --env PYTHONPATH="\$PYTHONPATH:$PYTHONPATH" \
+        $SIF_PATH $python_cmd "$PYTHON_SCRIPT" "$@"
+else
+    echo "Running $PYTHON_SCRIPT with python: $(which python)"
+    echo '################## End shebang info ####################'
+    echo
+    $python_cmd "$PYTHON_SCRIPT" "$@"
+fi