feat: python bindings for building encoderfiles (#295)

Right now, just loads from a config file. Example usage:

```python3
import encoderfile as ef

builder = ef.EncoderfileBuilder.from_config("encoderfile.yml")

builder.build()
```

To test it out, make sure you have maturin installed and run `make
build-py`

Author: besaleli

encoderfile-py/python/encoderfile/encoderfile.pyi

@@ -0,0 +1,15 @@
+from typing import Optional
+
+class EncoderfileBuilder:
+    @classmethod
+    def from_config(
+        cls,
+        config: str,
+        output_path: Optional[str] = None,
+        base_binary_path: Optional[str] = None,
+        platform: Optional[str] = None,
+        version: Optional[str] = None,
+        no_download: bool = False,
+        directory: Optional[str] = None,
+    ) -> "EncoderfileBuilder": ...
+    def build(self, cache_dir: Optional[str] = None): ...

encoderfile-py/python/tests/test_all.py

@@ -1,5 +1,5 @@
-import encoderfile
+# import encoderfile as ef
 
 
-def test_sum_as_string():
-    assert encoderfile.sum_as_string(1, 1) == "2"
+def test_todo():
+    assert True

encoderfile-py/src/builder.rs

@@ -0,0 +1,63 @@
+use std::path::PathBuf;
+
+use encoderfile::build_cli::cli::{BuildArgs, GlobalArguments};
+use pyo3::{
+    exceptions::{PyRuntimeError, PyValueError},
+    prelude::*,
+    types::PyType,
+};
+
+#[pyclass]
+pub struct EncoderfileBuilder(BuildArgs);
+
+#[pymethods]
+impl EncoderfileBuilder {
+    #[allow(clippy::too_many_arguments)]
+    #[classmethod]
+    #[pyo3(signature = (
+        config,
+        output_path = None,
+        base_binary_path = None,
+        platform = None,
+        version = None,
+        no_download = false,
+        directory = None))
+        ]
+    fn from_config(
+        _cls: &Bound<'_, PyType>,
+        config: PathBuf,
+        output_path: Option<PathBuf>,
+        base_binary_path: Option<PathBuf>,
+        platform: Option<String>,
+        version: Option<String>,
+        no_download: Option<bool>,
+        directory: Option<PathBuf>,
+    ) -> PyResult<EncoderfileBuilder> {
+        let platform = platform
+            .as_ref()
+            .map(|i| std::str::FromStr::from_str(i.as_str()))
+            .transpose()
+            .map_err(|_| PyValueError::new_err(format!("Invalid platform: {:?}", &platform)))?;
+
+        Ok(EncoderfileBuilder(BuildArgs {
+            config,
+            output_path,
+            base_binary_path,
+            platform,
+            version,
+            no_download: no_download.unwrap_or(false),
+            working_dir: directory,
+        }))
+    }
+
+    #[pyo3(signature = (cache_dir = None))]
+    fn build(&self, cache_dir: Option<PathBuf>) -> PyResult<()> {
+        let global = GlobalArguments { cache_dir };
+
+        self.0
+            .run(&global)
+            .map_err(|e| PyRuntimeError::new_err(format!("Error building encoderfile: {:?}", e)))?;
+
+        Ok(())
+    }
+}

encoderfile-py/src/lib.rs

@@ -1,13 +1,10 @@
 use pyo3::prelude::*;
 
+mod builder;
+
 /// A Python module implemented in Rust.
 #[pymodule]
 mod encoderfile {
-    use pyo3::prelude::*;
-
-    /// Formats the sum of two numbers as string.
-    #[pyfunction]
-    fn sum_as_string(a: usize, b: usize) -> PyResult<String> {
-        Ok((a + b).to_string())
-    }
+    #[pymodule_export]
+    use super::builder::EncoderfileBuilder;
 }

encoderfile/src/build_cli/cli/build.rs

@@ -24,44 +24,44 @@ use clap_derive::Args;
 #[derive(Debug, Args)]
 pub struct BuildArgs {
     #[arg(short = 'f', help = "Path to config file. Required.")]
-    config: PathBuf,
+    pub config: PathBuf,
     #[arg(
         short = 'o',
         long = "output-path",
         help = "Output path, e.g., `./my_model.encoderfile`. Optional"
     )]
-    output_path: Option<PathBuf>,
+    pub output_path: Option<PathBuf>,
     #[arg(
         long = "base-binary-path",
         help = "Path to base binary to use. Optional."
     )]
-    base_binary_path: Option<PathBuf>,
+    pub base_binary_path: Option<PathBuf>,
     #[arg(
         long = "platform",
         help = "Target platform to build. Follows standard rust target triple format."
     )]
-    platform: Option<TargetSpec>,
+    pub platform: Option<TargetSpec>,
     #[arg(
         long,
         help = "Encoderfile version override (defaults to current version)."
     )]
-    version: Option<String>,
+    pub version: Option<String>,
     #[arg(
         long = "no-download",
         help = "Disable downloading",
         default_value = "false"
     )]
-    no_download: bool,
+    pub no_download: bool,
     #[arg(
         long = "directory", // working-dir???
         help = "Set the working directory for the build process. Optional.",
         default_value = None
     )]
-    working_dir: Option<PathBuf>,
+    pub working_dir: Option<PathBuf>,
 }
 
 impl BuildArgs {
-    pub fn run(self, global: &GlobalArguments) -> Result<()> {
+    pub fn run(&self, global: &GlobalArguments) -> Result<()> {
         terminal::info("Loading config...");
         let mut config = crate::build_cli::config::BuildConfig::load(&self.config)?;
 

encoderfile/src/build_cli/cli/mod.rs

@@ -10,6 +10,7 @@ mod runtime;
 #[cfg(feature = "dev-utils")]
 pub use build::{test_build_args, test_build_args_working_dir};
 
+pub use build::BuildArgs;
 pub use inspect::inspect_encoderfile;
 
 #[derive(Debug, Parser)]
@@ -27,7 +28,7 @@ pub struct GlobalArguments {
         long = "cache-dir",
         help = "Cache directory. This is used for build artifacts. Optional."
     )]
-    cache_dir: Option<PathBuf>,
+    pub cache_dir: Option<PathBuf>,
 }
 
 impl GlobalArguments {

pyproject.toml

@@ -1,5 +1,5 @@
 [project]
-name = "encoderfile"
+name = "encoderfile-workspace"
 version = "0.3.0"
 description = "Add your description here"
 readme = "README.md"
@@ -11,3 +11,8 @@ dev = [ "pre-commit>=4.3.0", "ruff>=0.14.2",]
 docs = [ "mkdocs>=1.6.1", "mkdocs-include-markdown-plugin>=7.2.0", "mkdocs-material>=9.7.0",]
 setup = [ "onnxruntime>=1.23.1", "optimum[onnxruntime]>=2.0.0", "transformers>=4.55.0",]
 models = [ "torch>=2.9.0", "click", {include-group = "setup"} ]
+
+[tool.uv.workspace]
+members = [
+    "encoderfile-py"
+]

test_config.yml

@@ -1,6 +1,6 @@
 encoderfile:
   name: my-model-2
-  path: models/token_classification
+  path: models/dummy_electra_token_classifier
   model_type: token_classification
   output_path: ./my-model-2.encoderfile
   transform: |