Add ProteinMPNN integration with assets and tests

- Implement ProteinMPNN native spec and path handling
- Improve working dir error messages and UTF-8 checks

Author: lyskov

.github/workflows/build.yml

@@ -90,5 +90,4 @@ jobs:
 
       - name: Run Native tests
         run: |
-          cargo install cargo-nextest --locked
           cargo nextest run -F native-tests,high-memory-tests

README.md

@@ -29,7 +29,7 @@ See [App Usage Examples](#app-usage-examples) for how to run each of these tools
 | `score` | ✓ | ✓ |  |
 | `pyrosetta` | ✓ | ✓ |  |
 | `rfdiffusion` | ✓ | ✓ | ✓ |
-| `proteinmpnn` | ✓ | ✓ |  |
+| `proteinmpnn` | ✓ | ✓ | ✓ |
 | `proteinmpnn-script` | ✓ | ✓ |  |
 | `ligandmpnn` | ✓ | ✓ | |
 | `foundry` | ✓ | ✓ | ✓ |
@@ -331,16 +331,3 @@ rc -v run rosetta score -in:file:s structure.pdb
 ```
 
 This will show detailed information including the exact command being executed and where it's being logged.
-
-## Requirements
-
-- One of the supported container engines (Docker, Singularity, or Apptainer)
-- Appropriate container images for the applications you want to run
-
-## License
-
-See LICENSE file for details.
-
-## Author
-
-Sergey Lyskov

assets/pixi/proteinmpnn.toml

@@ -0,0 +1,46 @@
+[workspace]
+name = "proteinmpnn"
+version = "0.0.1"
+description = "Pixi environment for https://github.com/dauparas/ProteinMPNN"
+platforms = ["linux-64"]
+channels = ["conda-forge"]
+
+[dependencies]
+python = "3.11.*"
+pip = "*"
+git = "*"
+wget = "*"
+ca-certificates = "*"
+
+# # Optional but nice for debugging
+# mc = "*"
+
+numpy = "*"
+biopython = "*"
+pandas = "*"
+tqdm = "*"
+
+[pypi-options]
+index-url = "https://pypi.org/simple"
+extra-index-urls = ["https://download.pytorch.org/whl/cu121"]
+
+[pypi-dependencies]
+torch = "*"
+torchvision = "*"
+torchaudio = "*"
+
+[tasks]
+#setup = { depends-on = ["install"] }
+setup = { depends-on = ["clone"] }
+
+[tasks.clone]
+cmd = "rm -rf proteinmpnn-repo-clone && git clone https://github.com/dauparas/ProteinMPNN.git proteinmpnn-repo-clone"
+
+# Run ProteinMPNN with arbitrary args:
+[tasks.execute]
+args = ["args"]
+cmd = "python protein_mpnn_run.py {{args}}"
+cwd = "proteinmpnn-repo-clone"
+
+[activation]
+env = { PYTHONWARNINGS = "ignore::FutureWarning" }

src/app.rs

@@ -191,7 +191,7 @@ impl App {
             App::Rosetta => todo!("not implemented"), // rosetta::native_spec(app_args),
             App::PyRosetta => todo!("not implemented"), // pyrosetta::native_spec(app_args),
             App::Rfdiffusion => rfdiffusion::native_spec(app_args, working_dir),
-            App::Proteinmpnn => todo!("not implemented"), // proteinmpnn::native_spec(app_args),
+            App::Proteinmpnn => proteinmpnn::native_spec(app_args, working_dir),
             App::ProteinmpnnScript => todo!("not implemented"), // proteinmpnn_script::native_spec(app_args),
             App::Ligandmpnn => todo!("not implemented"), // ligandmpnn::native_spec(app_args),
             App::Foundry => foundry::native_spec(app_args, working_dir),

src/app/proteinmpnn.rs

@@ -1,4 +1,9 @@
-use crate::app::ContainerRunSpec;
+use camino::{Utf8Path, Utf8PathBuf};
+
+use crate::{
+    app::{ContainerRunSpec, NativeRunSpec},
+    util::include_asset,
+};
 
 pub fn container_spec(app_args: Vec<String>) -> ContainerRunSpec {
     ContainerRunSpec::with_prefixed_args(
@@ -8,3 +13,47 @@ pub fn container_spec(app_args: Vec<String>) -> ContainerRunSpec {
     )
     .working_dir("/w")
 }
+
+fn map_input_and_output_options(mut app_args: Vec<String>, working_dir: &Utf8Path) -> Vec<String> {
+    const OPTIONS: [&str; 2] = ["--pdb_path", "--out_folder"];
+    const OUTPUT_OPTION: &str = OPTIONS[1];
+
+    let mut output_option_present = false;
+
+    fn make_absolute(working_dir: &Utf8Path, path_str: &str) -> Utf8PathBuf {
+        let path = Utf8PathBuf::from(path_str);
+        if path.is_absolute() {
+            path
+        } else {
+            working_dir.join(path)
+        }
+    }
+
+    for i in 1..app_args.len() {
+        for option in OPTIONS {
+            if app_args[i - 1] == option {
+                app_args[i] = make_absolute(working_dir, &app_args[i]).into();
+                if option == OUTPUT_OPTION {
+                    output_option_present = true;
+                }
+                break;
+            }
+        }
+    }
+
+    if !output_option_present {
+        app_args.extend([OUTPUT_OPTION.into(), working_dir.to_string()]);
+    }
+    app_args
+}
+
+pub fn native_spec(app_args: Vec<String>, working_dir: &Utf8Path) -> super::NativeRunSpec {
+    let app_args = map_input_and_output_options(app_args, working_dir);
+
+    let app_args = app_args
+        .into_iter()
+        .map(|arg| shell_escape::escape(arg.into()).into())
+        .collect::<Vec<_>>();
+
+    NativeRunSpec::new(include_asset!("pixi/proteinmpnn.toml"), app_args)
+}

src/main.rs

@@ -148,8 +148,11 @@ fn main() -> Result<()> {
                 .clone()
                 .unwrap_or_else(|| Utf8PathBuf::from("."))
                 .canonicalize()
-                .unwrap();
-            let working_dir = Utf8PathBuf::try_from(working_dir).unwrap();
+                .unwrap_or_else(|_| {
+                    panic!("{}", "Specified working directory does not exist".red())
+                });
+            let working_dir = Utf8PathBuf::try_from(working_dir)
+                .unwrap_or_else(|_| panic!("{}", "Working dir path contains invalid UTF-8".red()));
 
             run::run(app, app_args.clone(), container_engine, working_dir)
         }

tests/proteinmpnn.rs

@@ -3,7 +3,7 @@ use assert_fs::TempDir;
 
 mod common;
 
-common::engine_tests!(proteinmpnn);
+common::engine_tests!(proteinmpnn; engines(docker, apptainer, singularity, none));
 
 fn proteinmpnn(engine: &str) {
     use assert_fs::assert::PathAssert;