cdflib 0.4.1

Pure-Rust port of CDFLIB: cumulative distribution functions (CDF) associated with common probability distributions
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
239
240
241

// Regression tests for the boundary-input contract: inverse_cdf/inverse_sf at
// p ∈ {0, 1} return the support endpoints, search_* reject NaN/Inf with typed
// errors instead of panicking or hanging, and cdf/sf propagate NaN without
// panicking through beta_inc / gamma_inc.

#![cfg(not(miri))]

use cdflib::traits::{ContinuousCdf, DiscreteCdf};
use cdflib::{
    Beta, Binomial, ChiSquared, ChiSquaredNoncentral, FisherSnedecor, FisherSnedecorNoncentral,
    Gamma, NegativeBinomial, Normal, Poisson, StudentsT,
};

// ---- Continuous endpoint contract ----

#[test]
fn normal_endpoints() {
    let n = Normal::new(0.0, 1.0);
    assert_eq!(n.inverse_cdf(0.0).unwrap(), f64::NEG_INFINITY);
    assert_eq!(n.inverse_cdf(1.0).unwrap(), f64::INFINITY);
    assert_eq!(n.inverse_ccdf(0.0).unwrap(), f64::INFINITY);
    assert_eq!(n.inverse_ccdf(1.0).unwrap(), f64::NEG_INFINITY);
}

#[test]
fn gamma_endpoints() {
    let g = Gamma::new(2.0, 1.5);
    assert_eq!(g.inverse_cdf(0.0).unwrap(), 0.0);
    assert_eq!(g.inverse_cdf(1.0).unwrap(), f64::INFINITY);
    assert_eq!(g.inverse_ccdf(0.0).unwrap(), f64::INFINITY);
    assert_eq!(g.inverse_ccdf(1.0).unwrap(), 0.0);
}

#[test]
fn chi_squared_endpoints() {
    let c = ChiSquared::new(5.0);
    assert_eq!(c.inverse_cdf(0.0).unwrap(), 0.0);
    assert_eq!(c.inverse_cdf(1.0).unwrap(), f64::INFINITY);
    assert_eq!(c.inverse_ccdf(0.0).unwrap(), f64::INFINITY);
    assert_eq!(c.inverse_ccdf(1.0).unwrap(), 0.0);
}

#[test]
fn chi_squared_noncentral_endpoints() {
    let c = ChiSquaredNoncentral::new(5.0, 2.0);
    assert_eq!(c.inverse_cdf(0.0).unwrap(), 0.0);
    assert_eq!(c.inverse_cdf(1.0).unwrap(), f64::INFINITY);
}

#[test]
fn beta_endpoints() {
    let b = Beta::new(2.0, 5.0);
    assert_eq!(b.inverse_cdf(0.0).unwrap(), 0.0);
    assert_eq!(b.inverse_cdf(1.0).unwrap(), 1.0);
    assert_eq!(b.inverse_ccdf(0.0).unwrap(), 1.0);
    assert_eq!(b.inverse_ccdf(1.0).unwrap(), 0.0);
}

#[test]
fn fisher_snedecor_endpoints() {
    let f = FisherSnedecor::new(5.0, 10.0);
    assert_eq!(f.inverse_cdf(0.0).unwrap(), 0.0);
    assert_eq!(f.inverse_cdf(1.0).unwrap(), f64::INFINITY);
    assert_eq!(f.inverse_ccdf(0.0).unwrap(), f64::INFINITY);
    assert_eq!(f.inverse_ccdf(1.0).unwrap(), 0.0);
}

#[test]
fn fisher_snedecor_noncentral_endpoints() {
    let f = FisherSnedecorNoncentral::new(5.0, 10.0, 2.0);
    assert_eq!(f.inverse_cdf(0.0).unwrap(), 0.0);
    assert_eq!(f.inverse_cdf(1.0).unwrap(), f64::INFINITY);
}

#[test]
fn students_t_endpoints() {
    let t = StudentsT::new(10.0);
    assert_eq!(t.inverse_cdf(0.0).unwrap(), f64::NEG_INFINITY);
    assert_eq!(t.inverse_cdf(1.0).unwrap(), f64::INFINITY);
    assert_eq!(t.inverse_ccdf(0.0).unwrap(), f64::INFINITY);
    assert_eq!(t.inverse_ccdf(1.0).unwrap(), f64::NEG_INFINITY);
}

// ---- Discrete endpoint contract ----

#[test]
fn binomial_endpoints() {
    let b = Binomial::new(10, 0.3);
    assert_eq!(b.inverse_cdf(0.0).unwrap(), 0);
    assert_eq!(b.inverse_cdf(1.0).unwrap(), 10);
    // inverse_sf returns the real-valued F90 cdfbin which=2 quantile.
    // At q=0 (p=1) the search converges at s=n; at q=1 (p=0) it walks
    // to the lower bound and fails per F90's status=1.
    let s = b.inverse_ccdf(0.0).unwrap();
    assert!((s - 10.0).abs() < 1e-6, "got s={s}");
    assert!(matches!(
        b.inverse_ccdf(1.0),
        Err(cdflib::BinomialError::Search(_))
    ));
}

#[test]
fn poisson_endpoints() {
    let p = Poisson::new(3.0);
    assert_eq!(p.inverse_cdf(0.0).unwrap(), 0);
    assert_eq!(p.inverse_cdf(1.0).unwrap(), u64::MAX);
    // inverse_sf returns the real-valued F90 cdfpoi which=2 quantile.
    // At q=0 the search walks to a large s where sf < abs_tol (F90 dstinv
    // converges by absolute tolerance, not by sign change); at q=1 it
    // hits the lower search bound and reports F90 status=1.
    let s_zero = p.inverse_ccdf(0.0).unwrap();
    assert!(s_zero > 10.0 && s_zero.is_finite(), "got {s_zero}");
    assert!(matches!(
        p.inverse_ccdf(1.0),
        Err(cdflib::PoissonError::Search(_))
    ));
}

#[test]
fn negative_binomial_endpoints() {
    let nb = NegativeBinomial::new(5, 0.5);
    assert_eq!(nb.inverse_cdf(0.0).unwrap(), 0);
    assert_eq!(nb.inverse_cdf(1.0).unwrap(), u64::MAX);
    // Same F90 cdfnbn which=2 behavior: q=0 converges by abs_tol at large s;
    // q=1 hits the lower search bound.
    let s_zero = nb.inverse_ccdf(0.0).unwrap();
    assert!(s_zero > 10.0 && s_zero.is_finite(), "got {s_zero}");
    assert!(matches!(
        nb.inverse_ccdf(1.0),
        Err(cdflib::NegativeBinomialError::Search(_))
    ));
}

// ---- search_* NaN rejection (must produce typed errors, not hang or panic) ----

#[test]
fn normal_solve_rejects_nan_x() {
    use cdflib::NormalError;
    assert!(matches!(
        Normal::search_mean(0.5, 0.5, f64::NAN, 1.0),
        Err(NormalError::XNotFinite(_))
    ));
    assert!(matches!(
        Normal::search_sd(0.5, 0.5, f64::NAN, 0.0),
        Err(NormalError::XNotFinite(_))
    ));
}

#[test]
fn gamma_solve_rejects_nan_x() {
    use cdflib::GammaError;
    assert!(matches!(
        Gamma::search_shape(0.5, 0.5, f64::NAN, 2.0),
        Err(GammaError::XNotFinite(_))
    ));
    assert!(matches!(
        Gamma::search_rate(0.5, 0.5, f64::NAN, 2.0),
        Err(GammaError::XNotFinite(_))
    ));
    assert!(matches!(
        Gamma::search_shape(0.5, 0.5, 1.0, f64::NAN),
        Err(GammaError::RateNotFinite(_))
    ));
}

#[test]
fn chi_squared_solve_rejects_nan_x() {
    use cdflib::ChiSquaredError;
    assert!(matches!(
        ChiSquared::search_df(0.5, 0.5, f64::NAN),
        Err(ChiSquaredError::XNotFinite(_))
    ));
}

#[test]
fn chi_squared_noncentral_solve_rejects_nan() {
    use cdflib::ChiSquaredNoncentralError;
    assert!(matches!(
        ChiSquaredNoncentral::search_df(0.5, f64::NAN, 2.0),
        Err(ChiSquaredNoncentralError::XNotFinite(_))
    ));
    assert!(matches!(
        ChiSquaredNoncentral::search_ncp(0.5, f64::NAN, 5.0),
        Err(ChiSquaredNoncentralError::XNotFinite(_))
    ));
    assert!(matches!(
        ChiSquaredNoncentral::search_df(0.5, 1.0, f64::NAN),
        Err(ChiSquaredNoncentralError::NcpNotFinite(_))
    ));
}

#[test]
fn students_t_solve_rejects_nan_t() {
    use cdflib::StudentsTError;
    assert!(matches!(
        StudentsT::search_df(0.5, 0.5, f64::NAN),
        Err(StudentsTError::TNotFinite(_))
    ));
}

#[test]
fn fisher_snedecor_noncentral_solve_rejects_nan() {
    use cdflib::FisherSnedecorNoncentralError;
    assert!(matches!(
        FisherSnedecorNoncentral::search_dfn(0.5, f64::NAN, 5.0, 1.0),
        Err(FisherSnedecorNoncentralError::FNotFinite(_))
    ));
    assert!(matches!(
        FisherSnedecorNoncentral::search_dfd(0.5, 1.0, f64::NAN, 1.0),
        Err(FisherSnedecorNoncentralError::DfnNotFinite(_))
    ));
    assert!(matches!(
        FisherSnedecorNoncentral::search_ncp(0.5, 1.0, 5.0, f64::NAN),
        Err(FisherSnedecorNoncentralError::DfdNotFinite(_))
    ));
}

// ---- cdf/sf propagate NaN (do not panic through beta_inc / gamma_inc) ----

#[test]
fn continuous_cdf_nan_returns_nan() {
    assert!(Normal::new(0.0, 1.0).cdf(f64::NAN).is_nan());
    assert!(Gamma::new(2.0, 1.0).cdf(f64::NAN).is_nan());
    assert!(ChiSquared::new(5.0).cdf(f64::NAN).is_nan());
    assert!(ChiSquaredNoncentral::new(5.0, 2.0).cdf(f64::NAN).is_nan());
    assert!(Beta::new(2.0, 5.0).cdf(f64::NAN).is_nan());
    assert!(FisherSnedecor::new(5.0, 10.0).cdf(f64::NAN).is_nan());
    assert!(FisherSnedecorNoncentral::new(5.0, 10.0, 2.0)
        .cdf(f64::NAN)
        .is_nan());
    assert!(StudentsT::new(10.0).cdf(f64::NAN).is_nan());
}

#[test]
fn continuous_ccdf_nan_returns_nan() {
    assert!(Normal::new(0.0, 1.0).ccdf(f64::NAN).is_nan());
    assert!(Gamma::new(2.0, 1.0).ccdf(f64::NAN).is_nan());
    assert!(ChiSquared::new(5.0).ccdf(f64::NAN).is_nan());
    assert!(Beta::new(2.0, 5.0).ccdf(f64::NAN).is_nan());
    assert!(StudentsT::new(10.0).ccdf(f64::NAN).is_nan());
}