baselines 0.1.1

Baseline correction algorithms for signals, spectra, and images
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

# Rust API

Start with the small default-based API. Move to the method-chain API when you
need to tune an algorithm. Both layers use the same implementations,
validation, and numerical defaults.

## Simple API

```rust
use baselines::prelude::*;

let estimated = baseline(&y)?;
let corrected = correct_with(&y, Method::Arpls)?;
# Ok::<(), baselines::BaselineError>(())
```

The simple API currently includes AsLS, arPLS, airPLS, rolling-ball, and
polynomial methods. The method-chain and family APIs expose the full algorithm
set and all tuning parameters.

## 1D signals

```rust
use baselines::prelude::*;

let fit = Baseline::new(&y)
    .asls()
    .lambda(1.0e6)
    .p(0.01)
    .max_iter(50)
    .tol(1.0e-3)
    .fit()?;

let corrected = fit.corrected(&y)?;
# Ok::<(), baselines::BaselineError>(())
```

## 1D signals with explicit x values

Use `Baseline::new_xy` when samples are not evenly spaced or when peak masks
are easier to express in x coordinates.

```rust
use baselines::prelude::*;

let fit = Baseline::new_xy(&x, &y)?
    .arpls()
    .lambda(1.0e6)
    .exclude_range(150.0, 220.0)
    .exclude_mask(&peak_mask)?
    .fit()?;

let corrected = fit.corrected(&y)?;
# Ok::<(), baselines::BaselineError>(())
```

`exclude_mask` uses `true` for ignored points. `baseline_mask` is also
available for pybaselines-style masks where `true` marks trusted baseline
points. Multiple range and boolean masks are combined before fitting.

## 2D row-major data

```rust
use baselines::prelude::*;

let fit = Baseline2D::row_major(&data, rows, cols)?
    .asls()
    .lambda(8.0e3)
    .p(0.01)
    .cg_tol(1.0e-6)
    .fit()?;

let corrected = fit.corrected(&data)?;
# Ok::<(), baselines::BaselineError>(())
```

## Full parameter structs

Use `with_params` when params are easier to construct directly.

```rust
use baselines::prelude::*;
use baselines::whittaker::{AslsParams, WhittakerParams};

let params = AslsParams {
    whittaker: WhittakerParams {
        lambda: 1.0e6,
        max_iter: 50,
        tol: 1.0e-3,
    },
    p: 0.01,
};

let fit = Baseline::new(&y).asls().with_params(params).fit()?;
# Ok::<(), baselines::BaselineError>(())
```

## Low-level API

The family modules remain public and are still useful for advanced workflows,
including reusable workspaces and exact parameter struct calls.

```rust
use baselines::whittaker::{AslsParams, WhittakerWorkspace, asls_into};

let mut baseline = vec![0.0; y.len()];
let mut workspace = WhittakerWorkspace::new(y.len());
let report = asls_into(
    &y,
    AslsParams::default(),
    &mut baseline,
    &mut workspace,
)?;
# Ok::<(), baselines::BaselineError>(())
```

Algorithm references are documented on the underlying family functions. The
builder docs intentionally point back to those functions instead of duplicating
citations.