Fix CI failures from #110

.github/workflows/code-tests.yaml

@@ -17,32 +17,25 @@ jobs:
   build_cpu:
     name: Build CPU
     strategy:
+      fail-fast: false
       matrix:
-        platform: [ubuntu-latest, macos-latest, windows-latest]
-        python-version: [3.9, "3.10"]
-    runs-on: ${{ matrix.platform }}
-    defaults:
-      run:
-        shell: bash -el {0}
+        python-version: ["3.11"]
+    runs-on: ubuntu-latest
 
     steps:
       - name: Checkout repository
         uses: actions/checkout@v7
-      - name: Setup miniconda
-        uses: conda-incubator/setup-miniconda@v4
+      - name: Setup Python
+        uses: actions/setup-python@v7
         with:
-          auto-update-conda: true
-          miniforge-version: latest
-          channels: "conda-forge, pytorch, pyg"
           python-version: ${{ matrix.python-version }}
-          use-mamba: true
       - id: cache-dependencies
         name: Cache dependencies
         uses: actions/cache@v6
         with:
           path: ${{ github.workspace }}/.venv
-          key: ${{ matrix.platform }}-py${{ matrix.python-version }}-${{ hashFiles('**/poetry.lock') }}
-          restore-keys: ${{ matrix.platform }}-py${{ matrix.python-version }}-
+          key: ${{ runner.os }}-py${{ matrix.python-version }}-${{ hashFiles('**/poetry.lock') }}
+          restore-keys: ${{ runner.os }}-py${{ matrix.python-version }}-
       - name: Install Workshop
         if: steps.cache-dependencies.cache-hit != 'true'
         run: |
@@ -53,7 +46,6 @@ jobs:
           poetry config virtualenvs.in-project true
           poetry install --no-interaction --without docs
           poetry run install_pyg_workshop
-          poe install-torchdrug
 
       - name: Run tests
         run: |

CHANGELOG.md

@@ -32,6 +32,9 @@
 * Update ICLR paper link and citation [#82](https://github.com/a-r-j/ProteinWorkshop/pull/82)
 * Add an optional group for installing plotting and analysis specific libraries to lighten the install of the core framework [#90](https://github.com/a-r-j/ProteinWorkshop/pull/90)
 * Update CI workflow actions to latest versions and replace retired Mambaforge installer with Miniforge; fix CI `pull_request` trigger, run test matrix across all platforms and fix cache key
+* Fix CI: install with `setup-python` instead of conda, test on Linux only (macOS/Windows unsupported), and switch the dummy test dataset from the retired RCSB MMTF service to PDB format
+* Require Pillow >=13.0.0; this raises the minimum supported Python version to 3.11 (drops 3.9 and 3.10)
+* Remove the `torchdrug` dependency (no release supports Python 3.11): GearNet now uses `torch.sparse_coo_tensor` directly and drops an unused torchdrug-only helper
 
 ### 0.2.5 (28/12/2023)
 

README.md

@@ -85,7 +85,7 @@ export DATA_PATH="where/you/want/data/" # e.g., `export DATA_PATH="proteinworksh
 However, for full exploration we recommend cloning the repository and building from source.
 
 ### Building from source
-With a local virtual environment activated (e.g., one created with `conda create -n proteinworkshop python=3.10`):
+With a local virtual environment activated (e.g., one created with `conda create -n proteinworkshop python=3.11`):
 1. Clone and install the project
 
     ```bash

proteinworkshop/config/dataset/dummy.yaml

@@ -6,7 +6,7 @@ datamodule:
   pin_memory: True # Pin memory for dataloader
   num_workers: 8 # Number of workers for dataloader
   obsolete_strategy: "drop"
-  format: "mmtf.gz" # Format of the raw PDB/MMTF files
+  format: "pdb" # Format of the raw PDB/MMTF files (RCSB no longer serves MMTF)
   transforms: ${transforms}
   overwrite: True
 

proteinworkshop/models/graph_encoders/layers/gear_net.py

@@ -6,7 +6,6 @@
 import torch.nn.functional as F
 from torch_geometric.data import Batch
 from torch_scatter import scatter_add
-from torchdrug import data, utils
 
 
 class MessagePassingBase(nn.Module):
@@ -53,11 +52,6 @@ def message_and_aggregate(self, graph, input):
         message = self.message(graph, input)
         return self.aggregate(graph, message)
 
-    def _message_and_aggregate(self, *tensors):
-        graph = data.Graph.from_tensors(tensors[:-1])
-        input = tensors[-1]
-        return self.message_and_aggregate(graph, input)
-
     def combine(self, input, update):
         """
         Combine node input and node update.
@@ -161,7 +155,7 @@ def message_and_aggregate(self, graph, input):
             dim_size=graph.num_nodes * graph.num_relation,
         )
         edge_weight = graph.edge_weight / degree_out[node_out]
-        adjacency = utils.sparse_coo_tensor(
+        adjacency = torch.sparse_coo_tensor(
             torch.stack([node_in, node_out]),
             edge_weight,
             (graph.num_nodes, graph.num_nodes * graph.num_relation),
@@ -251,7 +245,7 @@ def message_and_aggregate(self, graph, input):
         node_in, node_out, relation = graph.edge_index
         node_out = node_out * self.num_relation + relation
 
-        adjacency = utils.sparse_coo_tensor(
+        adjacency = torch.sparse_coo_tensor(
             torch.stack([node_in, node_out]),
             graph.edge_weight,
             (graph.num_nodes, graph.num_nodes * graph.num_relation),

pyproject.toml

@@ -14,8 +14,6 @@ classifiers = [
     "Operating System :: OS Independent",
     "Programming Language :: Python",
     "Programming Language :: Python :: 3 :: Only",
-    "Programming Language :: Python :: 3.9",
-    "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
     "Topic :: Scientific/Engineering"
 ]
@@ -28,7 +26,7 @@ packages = [
 "Documentation" = "https://proteins.sh"
 
 [tool.poetry.dependencies]
-python = ">= 3.9, < 3.11"
+python = ">= 3.11, < 3.12"
 graphein = "^1.7.5"
 loguru = "^0.7.0"
 fair-esm = "^2.0.0"
@@ -56,6 +54,7 @@ pyrootutils = "^1.0.4"
 torch-geometric = "^2.4.0"
 hydra-colorlog = "^1.2.0"
 foldcomp = "^0.0.5"
+pillow = ">=13.0.0"
 captum = {version="^0.6.0", optional=true}
 umap-learn = {version="^0.5.4", optional=true}
 matplotlib = {version="^3.7.2", optional=true}
@@ -107,7 +106,6 @@ install-torch-sparse = "pip install torch_sparse -f https://data.pyg.org/whl/tor
 install-torch-scatter = "pip install torch_scatter -f https://data.pyg.org/whl/torch-2.0.0+cpu.html"
 install-torch-spline-conv = "pip install torch_spline_conv -f https://data.pyg.org/whl/torch-2.0.0+cpu.html"
 install-openfold = "pip install git+https://github.com/NVIDIA/dllogger.git@0540a43971f4a8a16693a9de9de73c1072020769 && pip install git+https://github.com/aqlaboratory/openfold.git@4b41059694619831a7db195b7e0988fc4ff3a307"
-install-torchdrug = "pip install torchdrug"
 install-flash = "pip install lightning-flash --no-dependencies"
 
 [tool.poetry.scripts]
@@ -126,8 +124,8 @@ build-backend = "poetry.core.masonry.api"
 select = ["E", "F", "I"] # <-- Note: I suggest including the "I" for additional checks!
 ignore = ["F722"] # Ignore Jaxtyping forward references: https://github.com/google/jaxtyping/blob/v0.2.12/FAQ.md
 
-# Assuming you're developing for Python 3.10
-target-version = "py310"
+# Assuming you're developing for Python 3.11
+target-version = "py311"
 
 [tool.black]
 exclude = '''
@@ -148,7 +146,7 @@ exclude = '''
 )/
 '''
 line-length = 79
-target-version = ['py310']
+target-version = ['py311']
 
 [tool.isort]
 skip = [".env", "env", ".venv", "venv", "data"]