```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
```
``````use crate::prelude::{Rint, Scalar};
use crate::scalar::macros::*;
use crate::*;
use std::cmp::Ordering::*;
use std::convert::TryFrom;
use std::ops::{Add, Div, Mul, Neg, Sub};

/// `Rfloat` is a wrapper for `f64` in the context of an R's integer vector.
///
/// `Rfloat` has a special `NA` value, obtained from R headers via `R_NaReal`.
///
/// `Rfloat` has the same footprint as an `f64` value allowing us to use it in zero copy slices.
#[repr(transparent)]
pub struct Rfloat(f64);

impl Scalar<f64> for Rfloat {
fn inner(&self) -> f64 {
self.0
}

fn new(val: f64) -> Self {
Rfloat(val)
}
}

impl Rfloat {
pub fn is_nan(&self) -> bool {
self.0.is_nan()
}
pub fn is_sign_positive(&self) -> bool {
self.0.is_sign_positive()
}
pub fn is_sign_negative(&self) -> bool {
self.0.is_sign_negative()
}
pub fn is_infinite(&self) -> bool {
self.0.is_infinite()
}
pub fn is_subnormal(&self) -> bool {
self.0.is_subnormal()
}
pub fn abs(&self) -> Rfloat {
self.0.abs().into()
}
pub fn sqrt(&self) -> Rfloat {
self.0.sqrt().into()
}

/// ```
/// use extendr_api::prelude::*;
/// test! {
///     assert!(Rfloat::na().min(Rfloat::default()).is_na());
///     assert!(Rfloat::default().min(Rfloat::na()).is_na());
///     assert_eq!(Rfloat::default().min(Rfloat::default()), Rfloat::default());
///     assert_eq!(Rfloat::from(1).min(Rfloat::from(2)), Rfloat::from(1));
///     assert_eq!(Rfloat::from(2).min(Rfloat::from(1)), Rfloat::from(1));
/// }
/// ```
pub fn min(&self, other: Self) -> Self {
match self.partial_cmp(&other) {
Some(Less | Equal) => *self,
Some(Greater) => other,
_ => Self::na(),
}
}

/// ```
/// use extendr_api::prelude::*;
/// test! {
///     assert!(Rfloat::na().max(Rfloat::default()).is_na());
///     assert!(Rfloat::default().max(Rfloat::na()).is_na());
///     assert_eq!(Rfloat::default().max(Rfloat::default()), Rfloat::default());
///     assert_eq!(Rfloat::from(1).max(Rfloat::from(2)), Rfloat::from(2));
///     assert_eq!(Rfloat::from(2).max(Rfloat::from(1)), Rfloat::from(2));
/// }
/// ```
pub fn max(&self, other: Self) -> Self {
match self.partial_cmp(&other) {
Some(Less) => other,
Some(Greater | Equal) => *self,
_ => Self::na(),
}
}
}

// `NA_real_` is a `NaN` with specific bit representation.
// Check that underlying `f64` is `NA_real_`.
gen_trait_impl!(Rfloat, f64, |x: &Rfloat| x.inner().is_na(), f64::na());
gen_from_primitive!(Rfloat, f64);

impl From<Rfloat> for Option<f64> {
fn from(v: Rfloat) -> Self {
if v.is_na() {
None
} else {
Some(v.0)
}
}
}

gen_sum_iter!(Rfloat);
gen_partial_ord!(Rfloat, f64);

// Generate binary ops for +, -, * and /
gen_binop!(
Rfloat,
f64,
|lhs: f64, rhs: f64| Some(lhs + rhs),
"Add two Rfloat values or an option of f64."
);
gen_binop!(
Rfloat,
f64,
Sub,
|lhs: f64, rhs: f64| Some(lhs - rhs),
"Subtract two Rfloat values or an option of f64."
);
gen_binop!(
Rfloat,
f64,
Mul,
|lhs: f64, rhs: f64| Some(lhs * rhs),
"Multiply two Rfloat values or an option of f64."
);
gen_binop!(
Rfloat,
f64,
Div,
|lhs: f64, rhs: f64| Some(lhs / rhs),
"Divide two Rfloat values or an option of f64."
);
gen_binopassign!(
Rfloat,
f64,
|lhs: f64, rhs: f64| Some(lhs + rhs),
"Add two Rfloat values or an option of f64, modifying the left-hand side in place. Overflows to NA."
);
gen_binopassign!(
Rfloat,
f64,
SubAssign,
|lhs: f64, rhs: f64| Some(lhs - rhs),
"Subtract two Rfloat values or an option of f64, modifying the left-hand side in place. Overflows to NA."
);
gen_binopassign!(
Rfloat,
f64,
MulAssign,
|lhs: f64, rhs: f64| Some(lhs * rhs),
"Multiply two Rfloat values or an option of f64, modifying the left-hand side in place. Overflows to NA."
);
gen_binopassign!(
Rfloat,
f64,
DivAssign,
|lhs: f64, rhs: f64| Some(lhs / rhs),
"Divide two Rfloat values or an option of f64, modifying the left-hand side in place. Overflows to NA."
);

// Generate unary ops for -, !
gen_unop!(Rfloat, Neg, |lhs: f64| Some(-lhs), "Negate a Rfloat value.");

impl From<i32> for Rfloat {
fn from(value: i32) -> Self {
Rfloat::from(value as f64)
}
}

impl From<Rint> for Rfloat {
fn from(value: Rint) -> Self {
if value.is_na() {
Rfloat::na()
} else {
Rfloat::from(value.inner())
}
}
}

impl TryFrom<&Robj> for Rfloat {
type Error = Error;

fn try_from(robj: &Robj) -> Result<Self> {
let f64_val: Result<f64> = robj.try_into();
match f64_val {
Ok(val) => Ok(Rfloat::from(val)),
// TODO: Currently this results in an extra protection of robj
Err(Error::MustNotBeNA(_)) => Ok(Rfloat::na()),
Err(e) => Err(e),
}
}
}

impl std::fmt::Debug for Rfloat {
/// Debug format.
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
if self.is_na() {
write!(f, "NA_REAL")
} else {
self.inner().fmt(f)
}
}
}
``````