JavaCPP Presets for SciPy

^{Build status for all platforms: Commercial support:}

Introduction

This directory contains the JavaCPP Presets module for:

• SciPy 1.9.1 https://www.scipy.org/

Please refer to the parent README.md file for more detailed information about the JavaCPP Presets.

Documentation

Java API documentation is available here:

• http://bytedeco.org/javacpp-presets/scipy/apidocs/

∗ Call Py_Initialize(cachePackages()) instead of just Py_Initialize().

Sample Usage

Here is a simple example of SciPy based on the information available here:

• https://docs.scipy.org/doc/scipy/reference/tutorial/arpack.html

We can use Maven 3 to download and install automatically all the class files as well as the native binaries. To run this sample code, after creating the pom.xml and SparseLinalg.java source files below, simply execute on the command line:

 $ mvn compile exec:java

The pom.xml build file

<project>
    <modelVersion>4.0.0</modelVersion>
    <groupId>org.bytedeco.scipy</groupId>
    <artifactId>sparselinalg</artifactId>
    <version>1.5.8-SNAPSHOT</version>
    <properties>
        <exec.mainClass>SparseLinalg</exec.mainClass>
    </properties>
    <dependencies>
        <dependency>
            <groupId>org.bytedeco</groupId>
            <artifactId>scipy-platform</artifactId>
            <version>1.9.1-1.5.8-SNAPSHOT</version>
        </dependency>

        <!-- Additional dependencies to use bundled full version of MKL -->
        <dependency>
            <groupId>org.bytedeco</groupId>
            <artifactId>mkl-platform-redist</artifactId>
            <version>2022.2-1.5.8-SNAPSHOT</version>
        </dependency>

    </dependencies>
    <build>
        <sourceDirectory>.</sourceDirectory>
    </build>
</project>

The SparseLinalg.java source file

import org.bytedeco.javacpp.*;
import org.bytedeco.cpython.*;
import org.bytedeco.numpy.*;
import static org.bytedeco.cpython.global.python.*;
import static org.bytedeco.numpy.global.numpy.*;

public class SparseLinalg {
    public static void main(String[] args) throws Exception {
        /* try to use MKL when available */
        System.setProperty("org.bytedeco.openblas.load", "mkl");

        Py_Initialize(org.bytedeco.scipy.presets.scipy.cachePackages());
        if (_import_array() < 0) {
            System.err.println("numpy.core.multiarray failed to import");
            PyErr_Print();
            System.exit(-1);
        }
        PyObject globals = PyModule_GetDict(PyImport_AddModule("__main__"));

        PyRun_StringFlags("import numpy as np\n"
                + "from scipy.linalg import eig, eigh\n"
                + "from scipy.sparse.linalg import eigs, eigsh\n"
                + "np.set_printoptions(suppress=True)\n"

                + "np.random.seed(0)\n"
                + "X = np.random.random((100,100)) - 0.5\n"
                + "X = np.dot(X, X.T) #create a symmetric matrix\n"

                + "evals_all, evecs_all = eigh(X)\n"
                + "evals_large, evecs_large = eigsh(X, 3, which='LM')\n"
                + "print(evals_all[-3:])\n"
                + "print(evals_large)\n"
                + "print(np.dot(evecs_large.T, evecs_all[:,-3:]))\n"

                + "evals_small, evecs_small = eigsh(X, 3, sigma=0, which='LM')\n"
                + "print(evals_all[:3])\n"
                + "print(evals_small)\n"
                + "print(np.dot(evecs_small.T, evecs_all[:,:3]))\n", Py_file_input, globals, globals, null);

        if (PyErr_Occurred() != null) {
            System.err.println("Python error occurred");
            PyErr_Print();
            System.exit(-1);
        }
    }
}