Prepare release 0.15.2 (#137)

* Lint README

* Bump 0.15.2

* Fix moe smooth derivatives test and some cleanup

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "egobox"
-version = "0.15.1"
+version = "0.15.2"
 authors = ["Rémi Lafage <[email protected]>"]
 edition = "2021"
 description = "A toolbox for efficient global optimization"
@@ -31,10 +31,10 @@ persistent-moe = ["egobox-moe/persistent"]
 blas = ["ndarray/blas", "egobox-gp/blas", "egobox-moe/blas", "egobox-ego/blas"]
 
 [dependencies]
-egobox-doe = { version = "0.15.1", path = "./doe" }
-egobox-gp = { version = "0.15.1", path = "./gp" }
-egobox-moe = { version = "0.15.1", path = "./moe", features = ["persistent"] }
-egobox-ego = { version = "0.15.1", path = "./ego", features = ["persistent"] }
+egobox-doe = { version = "0.15.2", path = "./doe" }
+egobox-gp = { version = "0.15.2", path = "./gp" }
+egobox-moe = { version = "0.15.2", path = "./moe", features = ["persistent"] }
+egobox-ego = { version = "0.15.2", path = "./ego", features = ["persistent"] }
 
 linfa = { version = "0.7", default-features = false }
 

README.md

@@ -5,22 +5,23 @@
 [![linting](https://github.com/relf/egobox/workflows/lint/badge.svg)](https://github.com/relf/egobox/actions?query=workflow%3Alint)
 [![DOI](https://joss.theoj.org/papers/10.21105/joss.04737/status.svg)](https://doi.org/10.21105/joss.04737)
 
-Rust toolbox for Efficient Global Optimization algorithms inspired from [SMT](https://github.com/SMTorg/smt). 
+Rust toolbox for Efficient Global Optimization algorithms inspired from [SMT](https://github.com/SMTorg/smt).
 
-`egobox` is twofold: 
-1. for end-users: [a Python module](#the-python-module), the Python binding of the optimizer named `Egor` and the surrogate model `Gpx`, mixture of Gaussian processes, written in Rust. 
+`egobox` is twofold:
+
+1. for end-users: [a Python module](#the-python-module), the Python binding of the optimizer named `Egor` and the surrogate model `Gpx`, mixture of Gaussian processes, written in Rust.
 2. for developers: [a set of Rust libraries](#the-rust-libraries) useful to implement bayesian optimization (EGO-like) algorithms,
 
 ## The Python module
 
-Thanks to the [PyO3 project](https://pyo3.rs), which makes Rust well suited for building Python extensions. 
+Thanks to the [PyO3 project](https://pyo3.rs), which makes Rust well suited for building Python extensions.
 You can install the Python package using:
 
 ```bash
-$ pip install egobox
+pip install egobox
 ```
 
-See the [tutorial notebooks](https://github.com/relf/egobox/tree/master/doc/README.md) for usage of the optimizer 
+See the [tutorial notebooks](https://github.com/relf/egobox/tree/master/doc/README.md) for usage of the optimizer
 and mixture of Gaussian processes surrogate model.
 
 ## The Rust libraries
@@ -40,10 +41,10 @@ Depending on the sub-packages you want to use, you have to add following declara
 
 ```text
 [dependencies]
-egobox-doe = { version = "0.15.0" }
-egobox-gp  = { version = "0.15.0" }
-egobox-moe = { version = "0.15.0" }
-egobox-ego = { version = "0.15.0" }
+egobox-doe = { version = "0.15" }
+egobox-gp  = { version = "0.15" }
+egobox-moe = { version = "0.15" }
+egobox-ego = { version = "0.15" }
 ```
 
 ### Features
@@ -53,37 +54,46 @@ The table below presents the various features available depending on the subcrat
 | Name         | doe  | gp   | moe  | ego  |
 | :----------- | :--- | :--- | :--- | :--- |
 | serializable | ✔️    | ✔️    | ✔️    |      |
-| persistent   |      |      | ✔️    |      |
+| persistent   |      |      | ✔️    |  ✔️(*)  |
 | blas         |      | ✔️    | ✔️    | ✔️    |
 | nlopt        |      | ✔️    |      | ✔️    |
 
+(*) required for mixed-variable gaussian process
+
 #### serializable
+
 When selected, the serialization with [serde crate](https://serde.rs/) is enabled.
+
 #### persistent
-When selected, the save and load as a json file with [serde_json crate](https://serde.rs/) is enabled. 
+
+When selected, the save and load as a json file with [serde_json crate](https://serde.rs/) is enabled.
+
 #### blas
+
 When selected, the usage of BLAS/LAPACK backend is possible, see [below](#blaslapack-backend-optional) for more information.
+
 #### nlopt
+
 When selected, the [nlopt crate](https://github.com/adwhit/rust-nlopt) is used to provide optimizer implementations (ie Cobyla, Slsqp)
 
 ### Examples
 
 Examples (in `examples/` sub-packages folder) are run as follows:
 
 ```bash
-$ cd doe && cargo run --example samplings --release
+cd doe && cargo run --example samplings --release
 ```
 
 ``` bash
-$ cd gp && cargo run --example kriging --release
+cd gp && cargo run --example kriging --release
 ```
 
 ``` bash
-$ cd moe && cargo run --example clustering --release
+cd moe && cargo run --example clustering --release
 ```
 
 ``` bash
-$ cd ego && cargo run --example ackley --release
+cd ego && cargo run --example ackley --release
 ```
 
 ### BLAS/LAPACK backend (optional)
@@ -95,20 +105,23 @@ As for `linfa`, the linear algebra routines used in `gp`, `moe` ad `ego` are pro
 Otherwise, you can choose an external BLAS/LAPACK backend available through the [ndarray-linalg](https://github.com/rust-ndarray/ndarray-linalg) crate. In this case, you have to specify the `blas` feature and a `linfa` [BLAS/LAPACK backend feature](https://github.com/rust-ml/linfa#blaslapack-backend) (more information in [linfa features](https://github.com/rust-ml/linfa#blaslapack-backend)).
 
 Thus, for instance, to use `gp` with the Intel MKL BLAS/LAPACK backend, you could specify in your `Cargo.toml` the following features:
+
 ```text
 [dependencies]
-egobox-gp = { version = "0.15.0", features = ["blas", "linfa/intel-mkl-static"] }
+egobox-gp = { version = "0.15", features = ["blas", "linfa/intel-mkl-static"] }
 ```
+
 or you could run the `gp` example as follows:
+
 ``` bash
-$ cd gp && cargo run --example kriging --release --features blas,linfa/intel-mkl-static
+cd gp && cargo run --example kriging --release --features blas,linfa/intel-mkl-static
 ```
 
 ## Citation
 
 [![DOI](https://joss.theoj.org/papers/10.21105/joss.04737/status.svg)](https://doi.org/10.21105/joss.04737)
 
-If you find this project useful for your research, you may cite it as follows: 
+If you find this project useful for your research, you may cite it as follows:
 
 ```text
 @article{
@@ -127,4 +140,3 @@ If you find this project useful for your research, you may cite it as follows:
 ```
 
 Additionally, you may consider adding a star to the repository. This positive feedback improves the visibility of the project.
-

doe/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "egobox-doe"
-version = "0.15.1"
+version = "0.15.2"
 authors = ["Rémi Lafage <[email protected]>"]
 edition = "2021"
 description = "A library for design of experiments"

ego/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "egobox-ego"
-version = "0.15.1"
+version = "0.15.2"
 authors = ["Rémi Lafage <[email protected]>"]
 edition = "2021"
 description = "A library for efficient global optimization"
@@ -16,11 +16,11 @@ persistent = ["egobox-moe/persistent"]
 blas = ["ndarray-linalg", "linfa/ndarray-linalg", "linfa-pls/blas"]
 
 [dependencies]
-egobox-doe = { version = "0.15.1", path = "../doe", features = [
+egobox-doe = { version = "0.15.2", path = "../doe", features = [
     "serializable",
 ] }
-egobox-gp = { version = "0.15.1", path = "../gp", features = ["serializable"] }
-egobox-moe = { version = "0.15.1", path = "../moe", features = [
+egobox-gp = { version = "0.15.2", path = "../gp", features = ["serializable"] }
+egobox-moe = { version = "0.15.2", path = "../moe", features = [
     "serializable",
 ] }
 

gp/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "egobox-gp"
-version = "0.15.1"
+version = "0.15.2"
 authors = ["Rémi Lafage <[email protected]>"]
 edition = "2021"
 description = "A library for gaussian process modeling"
@@ -18,7 +18,7 @@ persistent = ["serializable", "serde_json"]
 blas = ["ndarray-linalg", "linfa/ndarray-linalg", "linfa-pls/blas"]
 
 [dependencies]
-egobox-doe = { version = "0.15.1", path = "../doe" }
+egobox-doe = { version = "0.15.2", path = "../doe" }
 
 linfa = { version = "0.7", default-features = false }
 linfa-pls = { version = "0.7", default-features = false }

moe/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "egobox-moe"
-version = "0.15.1"
+version = "0.15.2"
 authors = ["Rémi Lafage <[email protected]>"]
 edition = "2021"
 description = "A library for mixture of expert gaussian processes"
@@ -29,8 +29,8 @@ serializable = [
 blas = ["ndarray-linalg", "linfa/ndarray-linalg", "linfa-pls/blas"]
 
 [dependencies]
-egobox-doe = { version = "0.15.1", path = "../doe" }
-egobox-gp = { version = "0.15.1", path = "../gp" }
+egobox-doe = { version = "0.15.2", path = "../doe" }
+egobox-gp = { version = "0.15.2", path = "../gp" }
 
 linfa = { version = "0.7", default-features = false }
 linfa-clustering = { version = "0.7", default-features = false }

moe/src/clustering.rs

@@ -348,7 +348,6 @@ mod tests {
     use crate::gp_algorithm::GpMixture;
     use approx::assert_abs_diff_eq;
     use egobox_doe::{FullFactorial, Lhs, SamplingMethod};
-    use env_logger::{Builder, Env};
     #[cfg(not(feature = "blas"))]
     use linfa_linalg::norm::*;
     use ndarray::{array, Array1, Array2, Axis, Zip};
@@ -413,10 +412,6 @@ mod tests {
 
     #[test]
     fn test_find_best_cluster_nb_2d() {
-        let env = Env::new().filter_or("EGOBOX_LOG", "info");
-        let mut builder = Builder::from_env(env);
-        let builder = builder.target(env_logger::Target::Stdout);
-        builder.try_init().ok();
         let doe = egobox_doe::FullFactorial::new(&array![[-1., 1.], [-1., 1.]]);
         let xtrain = doe.sample(100);
         let ytrain = l1norm(&xtrain);

moe/src/gp_algorithm.rs

@@ -75,7 +75,6 @@ impl<R: Rng + SeedableRng + Clone> GpMixValidParams<f64, R> {
         yt: &ArrayBase<impl Data<Elem = f64>, Ix2>,
     ) -> Result<GpMixture> {
         trace!("Moe training...");
-        let _opt = env_logger::try_init().ok();
         let nx = xt.ncols();
         let data = concatenate(Axis(1), &[xt.view(), yt.view()]).unwrap();
 
@@ -1085,24 +1084,23 @@ mod tests {
         let x = Array1::linspace(0., 1., 50).insert_axis(Axis(1));
         let preds = moe.predict_values(&x).expect("MOE prediction");
         let dpreds = moe.predict_derivatives(&x).expect("MOE drv prediction");
-        println!("dpred = {dpreds}");
 
         let test_dir = "target/tests";
         std::fs::create_dir_all(test_dir).ok();
-        write_npy(format!("{test_dir}/x_hard.npy"), &x).expect("x saved");
-        write_npy(format!("{test_dir}/preds_hard.npy"), &preds).expect("preds saved");
-        write_npy(format!("{test_dir}/dpreds_hard.npy"), &dpreds).expect("dpreds saved");
+        write_npy(format!("{test_dir}/x_moe_smooth.npy"), &x).expect("x saved");
+        write_npy(format!("{test_dir}/preds_moe_smooth.npy"), &preds).expect("preds saved");
+        write_npy(format!("{test_dir}/dpreds_moe_smooth.npy"), &dpreds).expect("dpreds saved");
 
         let mut rng = Xoshiro256Plus::seed_from_u64(42);
         for _ in 0..20 {
             let x1: f64 = rng.gen_range(0.0..1.0);
 
-            let h = 1e-4;
+            let h = 1e-8;
             let xtest = array![[x1]];
 
             let x = array![[x1], [x1 + h], [x1 - h]];
-            let preds = moe.predict_derivatives(&x).unwrap();
-            let fdiff = preds[[1, 0]] - preds[[1, 0]] / 2. * h;
+            let preds = moe.predict_values(&x).unwrap();
+            let fdiff = (preds[[1, 0]] - preds[[2, 0]]) / (2. * h);
 
             let drv = moe.predict_derivatives(&xtest).unwrap();
             let df = df_test_1d(&xtest);
@@ -1115,6 +1113,7 @@ mod tests {
             println!(
                 "Test predicted derivatives at {xtest}: drv {drv}, true df {df}, fdiff {fdiff}"
             );
+            println!("preds(x, x+h, x-h)={}", preds);
             assert_abs_diff_eq!(err, 0.0, epsilon = 2.5e-1);
         }
     }

moe/src/sgp_algorithm.rs

@@ -76,7 +76,6 @@ impl<R: Rng + SeedableRng + Clone> SparseGpMixtureValidParams<f64, R> {
         yt: &ArrayBase<impl Data<Elem = f64>, Ix2>,
     ) -> Result<SparseGpMixture> {
         trace!("Sgp training...");
-        let _opt = env_logger::try_init().ok();
         let nx = xt.ncols();
         let data = concatenate(Axis(1), &[xt.view(), yt.view()]).unwrap();
 

pyproject.toml

@@ -24,7 +24,7 @@ python-source = "python"
 
 [tool.poetry]
 name = "egobox"
-version = "0.15.1"
+version = "0.15.2"
 description = "Python binding for egobox EGO optimizer written in Rust"
 authors = ["Rémi Lafage <[email protected]>"]
 packages = [{ include = "egobox", from = "python" }]