numrs2 0.4.0

A Rust implementation inspired by NumPy for numerical computing (NumRS2)
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238

//! Optimization operations for Python bindings

use crate::python::array::PyArray;
use pyo3::exceptions::PyValueError;
use pyo3::prelude::*;
use pyo3::types::{PyAnyMethods, PyDict, PyTuple};
use std::cell::RefCell;

/// Minimize a scalar function of one or more variables.
///
/// Bridges a Python callable to the Rust optimizer.
/// Supported methods: `"nelder-mead"` (default) and `"bfgs"` (uses numerical gradient).
/// Returns a Python dict with keys `x`, `fun`, `success`, `nit`, `nfev`, `message`.
#[pyfunction]
fn minimize(
    py: Python<'_>,
    fun: Py<PyAny>,
    x0: &PyArray,
    method: Option<String>,
    tol: Option<f64>,
) -> PyResult<Py<PyAny>> {
    let x0_vec = x0.tolist();
    if x0_vec.is_empty() {
        return Err(PyValueError::new_err("x0 must not be empty"));
    }

    let method_str = method.as_deref().unwrap_or("nelder-mead").to_lowercase();

    // Capture Python exceptions that occur inside the closure.
    let captured_err: RefCell<Option<PyErr>> = RefCell::new(None);

    let closure_result = match method_str.as_str() {
        "nelder-mead" | "nm" | "neldermead" => {
            // Build closure that calls the Python function
            let closure = |x: &[f64]| -> f64 {
                Python::attach(|inner_py| {
                    let py_list: Vec<f64> = x.to_vec();
                    let args = match PyTuple::new(inner_py, [py_list.into_pyobject(inner_py).ok()])
                    {
                        Ok(t) => t,
                        Err(e) => {
                            *captured_err.borrow_mut() = Some(e);
                            return f64::NAN;
                        }
                    };
                    match fun.call(inner_py, args, None) {
                        Ok(result) => match result.extract::<f64>(inner_py) {
                            Ok(v) => v,
                            Err(e) => {
                                *captured_err.borrow_mut() = Some(e);
                                f64::NAN
                            }
                        },
                        Err(e) => {
                            *captured_err.borrow_mut() = Some(e);
                            f64::NAN
                        }
                    }
                })
            };

            let mut config = crate::optimize::OptimizeConfig::<f64>::default();
            if let Some(t) = tol {
                config.ftol = t;
                config.xtol = t;
            }
            crate::optimize::nelder_mead(closure, &x0_vec, Some(config))
        }
        "bfgs" => {
            // BFGS with numerical gradient (central differences)
            let eps = 1e-7_f64;

            let closure = |x: &[f64]| -> f64 {
                Python::attach(|inner_py| {
                    let py_list: Vec<f64> = x.to_vec();
                    let args = match PyTuple::new(inner_py, [py_list.into_pyobject(inner_py).ok()])
                    {
                        Ok(t) => t,
                        Err(e) => {
                            *captured_err.borrow_mut() = Some(e);
                            return f64::NAN;
                        }
                    };
                    match fun.call(inner_py, args, None) {
                        Ok(result) => match result.extract::<f64>(inner_py) {
                            Ok(v) => v,
                            Err(e) => {
                                *captured_err.borrow_mut() = Some(e);
                                f64::NAN
                            }
                        },
                        Err(e) => {
                            *captured_err.borrow_mut() = Some(e);
                            f64::NAN
                        }
                    }
                })
            };

            let grad_closure = |x: &[f64]| -> Vec<f64> {
                Python::attach(|inner_py| {
                    let n = x.len();
                    let mut g = vec![0.0_f64; n];
                    for i in 0..n {
                        let mut xph = x.to_vec();
                        let mut xmh = x.to_vec();
                        xph[i] += eps;
                        xmh[i] -= eps;

                        let call_val = |xv: Vec<f64>| -> f64 {
                            let args =
                                match PyTuple::new(inner_py, [xv.into_pyobject(inner_py).ok()]) {
                                    Ok(t) => t,
                                    Err(e) => {
                                        *captured_err.borrow_mut() = Some(e);
                                        return f64::NAN;
                                    }
                                };
                            match fun.call(inner_py, args, None) {
                                Ok(result) => result.extract::<f64>(inner_py).unwrap_or(f64::NAN),
                                Err(e) => {
                                    *captured_err.borrow_mut() = Some(e);
                                    f64::NAN
                                }
                            }
                        };
                        g[i] = (call_val(xph) - call_val(xmh)) / (2.0 * eps);
                    }
                    g
                })
            };

            let mut config = crate::optimize::OptimizeConfig::<f64>::default();
            if let Some(t) = tol {
                config.ftol = t;
                config.gtol = t;
            }
            crate::optimize::bfgs(closure, grad_closure, &x0_vec, Some(config))
        }
        other => {
            return Err(PyValueError::new_err(format!(
                "Unknown method '{}'. Supported: 'nelder-mead', 'bfgs'",
                other
            )));
        }
    };

    // Propagate any stored Python error from inside the closure
    if let Some(err) = captured_err.into_inner() {
        return Err(err);
    }

    let opt_result = closure_result.map_err(|e| PyValueError::new_err(e.to_string()))?;

    // Return as Python dict mirroring scipy.optimize.OptimizeResult
    let dict = PyDict::new(py);
    dict.set_item("x", opt_result.x.into_pyobject(py)?.into_any().unbind())?;
    dict.set_item("fun", opt_result.fun)?;
    dict.set_item("success", opt_result.success)?;
    dict.set_item("nit", opt_result.nit)?;
    dict.set_item("nfev", opt_result.nfev)?;
    dict.set_item("message", opt_result.message)?;

    Ok(dict.into_any().unbind())
}

/// Find a root of a scalar function within a bracketing interval.
///
/// Bridges a Python callable to the Rust root-finder.
/// Supported methods: `"brentq"` (default), `"bisect"`.
/// Returns the root as f64.
#[pyfunction]
fn root_scalar(fun: Py<PyAny>, bracket: (f64, f64), method: Option<String>) -> PyResult<f64> {
    let (a, b) = bracket;

    let method_str = method.as_deref().unwrap_or("brentq").to_lowercase();

    let captured_err: RefCell<Option<PyErr>> = RefCell::new(None);

    let closure = |x: f64| -> f64 {
        Python::attach(|py| {
            let args = match PyTuple::new(py, [x]) {
                Ok(t) => t,
                Err(e) => {
                    *captured_err.borrow_mut() = Some(e);
                    return f64::NAN;
                }
            };
            match fun.call(py, args, None) {
                Ok(result) => match result.extract::<f64>(py) {
                    Ok(v) => v,
                    Err(e) => {
                        *captured_err.borrow_mut() = Some(e);
                        f64::NAN
                    }
                },
                Err(e) => {
                    *captured_err.borrow_mut() = Some(e);
                    f64::NAN
                }
            }
        })
    };

    let root_result = match method_str.as_str() {
        "brentq" | "brent" => crate::roots::brentq(closure, a, b, 1e-10, 1000)
            .map_err(|e| PyValueError::new_err(e.to_string()))?,
        "bisect" | "bisection" => crate::roots::bisect(closure, a, b, 1e-10, 1000)
            .map_err(|e| PyValueError::new_err(e.to_string()))?,
        other => {
            return Err(PyValueError::new_err(format!(
                "Unknown method '{}'. Supported: 'brentq', 'bisect'",
                other
            )));
        }
    };

    // Propagate stored Python errors
    if let Some(err) = captured_err.into_inner() {
        return Err(err);
    }

    Ok(root_result.root)
}

/// Register optimization functions
pub fn register(m: &Bound<'_, PyModule>) -> PyResult<()> {
    // Create optimize submodule
    let optimize_module = PyModule::new(m.py(), "optimize")?;

    // Add functions
    optimize_module.add_function(wrap_pyfunction!(minimize, m)?)?;
    optimize_module.add_function(wrap_pyfunction!(root_scalar, m)?)?;

    m.add_submodule(&optimize_module)?;

    Ok(())
}