Package 'libscanR' reference manual

Title:	Analysis and Visualization of Laser-Induced Breakdown Spectroscopy Data
Description:	Provides a vendor-agnostic pipeline for importing, preprocessing, calibrating, and visualizing Laser-Induced Breakdown Spectroscopy (LIBS) spectral data. Supports baseline correction, peak identification against user-supplied atomic emission line references, univariate and multivariate calibration, relative elemental amounts under an explicitly specified two-stage Saha-Boltzmann model, and calibration-based detection-limit estimates. Includes exploratory chemometric analysis (principal component analysis, partial least squares discriminant analysis, clustering), heuristic tissue classification, and spatial elemental mapping. Includes an interactive 'shiny' application for spectral exploration. Designed for biomedical tissue analysis, environmental monitoring, and materials characterization workflows.
Authors:	R. Heller [aut, cre] (ORCID: <https://orcid.org/0000-0001-8006-9742>)
Maintainer:	R. Heller <[email protected]>
License:	MIT + file LICENSE
Version:	0.1.0.9001
Built:	2026-11-03 18:13:47 UTC
Source:	https://github.com/CTTIR/libscanR

Subset a LIBS Dataset

Description

Subset a LIBS Dataset

Usage

## S3 method for class 'libs_dataset'
x[i]
## S3 method for class 'libs_dataset'
x[i]

Arguments

x

A libs_dataset object.

i

Integer or logical index of spectra to keep.

Value

A new libs_dataset.

Subset a LIBS Spectrum by Wavelength

Description

Returns a new libs_spectrum with channels whose wavelengths fall in the supplied numeric vector's range.

Usage

## S3 method for class 'libs_spectrum'
x[i]
## S3 method for class 'libs_spectrum'
x[i]

Arguments

x

A libs_spectrum object.

i

Numeric vector. Wavelengths are kept if they fall in ⁠[min(i), max(i)]⁠. Can also be a logical or integer index into channels.

Value

A new libs_spectrum object.

Combine Multiple LIBS Spectra by Stacking Shots

Description

Concatenates shots from multiple spectra. All spectra must share the same wavelength axis. Shared metadata are retained; source metadata are stored in combined_metadata. If every input has a baseline, the output stores its shot-weighted mean; otherwise the aggregate baseline is unknown.

Usage

## S3 method for class 'libs_spectrum'
c(...)
## S3 method for class 'libs_spectrum'
c(...)

Arguments

...

Two or more libs_spectrum objects.

Value

A new libs_spectrum with combined shots.

Dimensions of a Dataset (spectra x channels)

Description

Dimensions of a Dataset (spectra x channels)

Usage

## S3 method for class 'libs_dataset'
dim(x)
## S3 method for class 'libs_dataset'
dim(x)

Arguments

x

A libs_dataset object.

Value

Integer vector of length 2.

Dimensions of a LIBS Spectrum (shots x channels)

Description

Dimensions of a LIBS Spectrum (shots x channels)

Usage

## S3 method for class 'libs_spectrum'
dim(x)
## S3 method for class 'libs_spectrum'
dim(x)

Arguments

x

A libs_spectrum object.

Value

Integer vector of length 2.

Test whether an object is a libs_calibration

Description

Test whether an object is a libs_calibration

Usage

is_libs_calibration(x)
is_libs_calibration(x)

Arguments

x

Object to test.

Value

Logical scalar.

Test whether an object is a libs_dataset

Description

Test whether an object is a libs_dataset

Usage

is_libs_dataset(x)
is_libs_dataset(x)

Arguments

x

Object to test.

Value

Logical scalar.

Test whether an object is a libs_spectrum

Description

Test whether an object is a libs_spectrum

Usage

is_libs_spectrum(x)
is_libs_spectrum(x)

Arguments

x

Object to test.

Value

Logical scalar.

Number of Spectra in a Dataset

Description

Number of Spectra in a Dataset

Usage

## S3 method for class 'libs_dataset'
length(x)
## S3 method for class 'libs_dataset'
length(x)

Arguments

x

A libs_dataset object.

Value

Integer scalar.

Number of Channels in a LIBS Spectrum

Description

Number of Channels in a LIBS Spectrum

Usage

## S3 method for class 'libs_spectrum'
length(x)
## S3 method for class 'libs_spectrum'
length(x)

Arguments

x

A libs_spectrum object.

Value

Integer scalar.

Average Replicate Shots

Description

Averages shots within a ls_spectrum().

Usage

ls_average_shots(x, method = "mean", trim = 0, remove_outliers = FALSE)
ls_average_shots(x, method = "mean", trim = 0, remove_outliers = FALSE)

Arguments

x

A ls_spectrum() object.

method

Character. "mean" or "median". Default "mean".

trim

Numeric. Fraction trimmed from each end for trimmed mean. Default 0.

remove_outliers

Logical. Remove shots with deviation > 3 SD from the per-channel mean. Default FALSE.

Value

A libs_spectrum with a single averaged shot.

Examples

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(), n_shots = 10, seed = 1)
ls_average_shots(spec, remove_outliers = TRUE)

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(), n_shots = 10, seed = 1)
ls_average_shots(spec, remove_outliers = TRUE)

Baseline Correction

Description

Removes continuum/baseline drift from LIBS spectra. Supported methods: "snip" (Statistics-sensitive Non-linear Iterative Peak-clipping, Morhac/Ryan), "als" (Asymmetric Least Squares smoothing, Eilers & Boelens), "rolling_ball" (morphological), "linear" (two-endpoint interpolation), or "polynomial" (low-order polynomial fit of baseline-only regions).

Usage

ls_baseline(
  x,
  method = "snip",
  iterations = 100,
  order = 3,
  lambda = 1e+05,
  p = 0.01,
  radius = 50,
  ...
)
ls_baseline(
  x,
  method = "snip",
  iterations = 100,
  order = 3,
  lambda = 1e+05,
  p = 0.01,
  radius = 50,
  ...
)

Arguments

x

A ls_spectrum() or ls_dataset() object.

method

Character. One of "snip", "als", "rolling_ball", "linear", "polynomial". Default "snip".

iterations

Integer. Iterations for iterative methods (SNIP, ALS). Default 100.

order

Integer. Polynomial order for "polynomial". Default 3.

lambda

Numeric. Smoothness parameter for ALS. Default 1e5.

p

Numeric. Asymmetry parameter for ALS (0 < p < 1). Default 0.01.

radius

Integer. Rolling-ball radius in channels. Default 50.

...

Reserved for future use.

Value

Same class as input with baseline-corrected intensities; the estimated baseline is stored in the baseline element for inspection.

Examples

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(), seed = 1)
corr <- ls_baseline(spec, method = "snip")
is.null(corr$baseline)

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(), seed = 1)
corr <- ls_baseline(spec, method = "snip")
is.null(corr$baseline)

Build a Spatial Elemental Map

Description

Constructs a spatial map from LIBS line-scan or raster-scan data. The dataset's sample_info must include x_pos and y_pos columns. Coordinates must be finite numeric values and each pair must be unique; aggregate replicate positions explicitly before mapping them.

Usage

ls_build_map(dataset, element, line_nm, calibration = NULL, window_nm = NULL)
ls_build_map(dataset, element, line_nm, calibration = NULL, window_nm = NULL)

Arguments

dataset

A ls_dataset() object with spatial coordinates.

element

Character. Element label (for metadata).

line_nm

Numeric. Emission line wavelength (nm).

calibration

Optional ls_calibration(). If provided, intensities are converted to concentration. Default NULL.

window_nm

Optional integration window. NULL reuses the calibration's fitted window, or 1 nm for an uncalibrated map.

Value

A list with S3 class libs_map containing x, y, values, element, line_nm, unit, and a rasterized grid matrix when the points form a regular grid.

Examples

ds <- ls_example_data("spatial")
m <- ls_build_map(ds, "Ca", 393.37)
dim(m$grid)
ds <- ls_example_data("spatial")
m <- ls_build_map(ds, "Ca", 393.37)
dim(m$grid)

Build Calibration Curve

Description

Constructs a calibration model from standards with known concentrations.

Usage

ls_calibrate(
  dataset,
  element,
  line_nm,
  concentrations,
  method = "univariate",
  internal_std_nm = NULL,
  window_nm = 1,
  pls_window_nm = NULL,
  n_components = 5,
  unit = "ppm",
  verbose = TRUE
)
ls_calibrate(
  dataset,
  element,
  line_nm,
  concentrations,
  method = "univariate",
  internal_std_nm = NULL,
  window_nm = 1,
  pls_window_nm = NULL,
  n_components = 5,
  unit = "ppm",
  verbose = TRUE
)

Arguments

dataset

A ls_dataset() containing standard spectra. Spectra order must correspond to concentrations.

element

Character. Target element symbol.

line_nm

Numeric. Emission line wavelength (nm).

concentrations

Numeric vector of known concentrations, one per spectrum in dataset.

method

Character. "univariate" (default), "internal_std", "pls", or "cf_libs".

internal_std_nm

Numeric. Internal standard line wavelength, required for method = "internal_std".

window_nm

Numeric. Peak integration window. Default 1.

pls_window_nm

Numeric vector of length 2. Spectral window for PLS. Default c(line_nm - 5, line_nm + 5).

n_components

Integer. Number of PLS components. Default 5.

unit

Character. Concentration unit. Default "ppm".

verbose

Logical. Default TRUE.

Details

Prediction preserves the fitted wavelength axis, ordered preprocessing, integration window and internal-standard ratio or PLS columns. Univariate predictions retain inverse concentration-on-response regression. Estimated 3-sigma LOD and 10-sigma LOQ use the residual standard deviation of a separate response-on-concentration fit divided by its absolute slope. These estimates require independent verification near the limits; they are not validated instrument limits. PLS reports training RMSE separately and leaves LOD/LOQ unavailable (NA); residual prediction error alone is not a detection limit.

Value

A ls_calibration() object.

Examples

ds <- ls_example_data("calibration")
conc <- ds$sample_info$concentration
cal <- ls_calibrate(ds, "Ca", 393.37, conc, verbose = FALSE)
print(cal)
ds <- ls_example_data("calibration")
conc <- ds$sample_info$concentration
cal <- ls_calibrate(ds, "Ca", 393.37, conc, verbose = FALSE)
print(cal)

Create a LIBS Calibration Model

Description

Constructs a libs_calibration S3 object holding a fitted calibration model plus the underlying training data and figures of merit.

Usage

ls_calibration(
  element,
  wavelength_nm,
  concentrations,
  intensities,
  model,
  method = "univariate",
  lod = NULL,
  loq = NULL,
  r_squared = NULL,
  unit = "ppm",
  feature_contract = NULL,
  response_model = NULL,
  lod_basis = NULL,
  training_rmse = NULL
)
ls_calibration(
  element,
  wavelength_nm,
  concentrations,
  intensities,
  model,
  method = "univariate",
  lod = NULL,
  loq = NULL,
  r_squared = NULL,
  unit = "ppm",
  feature_contract = NULL,
  response_model = NULL,
  lod_basis = NULL,
  training_rmse = NULL
)

Arguments

element

Character. Element symbol (e.g. "Ca", "Fe").

wavelength_nm

Numeric. Emission line wavelength used.

concentrations

Numeric vector. Known concentrations of standards.

intensities

Numeric vector. Measured peak intensities or areas (or ratios for internal_std). Same length as concentrations.

model

Fitted model object (lm with one plain numeric predictor and an intercept, or mvr for PLS). Transformed terms and offsets are rejected by univariate prediction.

method

Character. One of "univariate", "internal_std", "pls", "cf_libs". Default "univariate".

lod

Numeric. Limit of detection (3-sigma). Default NULL.

loq

Numeric. Limit of quantification (10-sigma). Default NULL.

r_squared

Numeric. Coefficient of determination. Default NULL.

unit

Character. Concentration unit, e.g. "ppm", "wt_pct", "mg_kg". Default "ppm".

feature_contract

Optional fitted wavelength/preprocessing and extraction contract, normally populated by ls_calibrate().

response_model

Optional response-on-concentration regression used to estimate univariate detection limits.

lod_basis

Character description of the limit estimate, or NULL.

training_rmse

Numeric training prediction error, or NULL. This is not a detection limit or a held-out estimate.

Value

A libs_calibration S3 object.

Examples

conc <- c(100, 500, 1000, 2500, 5000)
ints <- conc * 0.12 + stats::rnorm(5, 0, 5)
m <- stats::lm(conc ~ ints)
cal <- ls_calibration("Ca", 393.37, conc, ints, model = m,
                      r_squared = summary(m)$r.squared)
print(cal)

conc <- c(100, 500, 1000, 2500, 5000)
ints <- conc * 0.12 + stats::rnorm(5, 0, 5)
m <- stats::lm(conc ~ ints)
cal <- ls_calibration("Ca", 393.37, conc, ints, model = m,
                      r_squared = summary(m)$r.squared)
print(cal)

Classify Unknown Spectra

Description

Applies a trained classifier (from ls_plsda() or ls_train_classifier()) to new spectra.

Usage

ls_classify(model, new_data)
ls_classify(model, new_data)

Arguments

model

A libs_plsda or libs_classifier object.

new_data

A ls_dataset() of unknown spectra.

Value

A tibble::tibble() with columns sample_id, predicted_class, and score for PLS-DA (an unbounded discriminant score, not a probability), or probability for SVM/RF. The wavelength axis and ordered preprocessing must match the fitted contract. Legacy models without this contract must be refitted before prediction.

Examples

ds <- ls_example_data("tissue")

if (requireNamespace("e1071", quietly = TRUE)) {
  clf <- ls_train_classifier(ds[1:30], "tissue", method = "svm")
  ls_classify(clf, ds[31:50])
}

ds <- ls_example_data("tissue")

if (requireNamespace("e1071", quietly = TRUE)) {
  clf <- ls_train_classifier(ds[1:30], "tissue", method = "svm")
  ls_classify(clf, ds[31:50])
}

Cluster LIBS Spectra

Description

Unsupervised clustering via k-means, hierarchical, or DBSCAN (DBSCAN requires the dbscan package in Suggests).

Usage

ls_cluster(dataset, method = "kmeans", k = 3, ...)
ls_cluster(dataset, method = "kmeans", k = 3, ...)

Arguments

dataset

A ls_dataset() object.

method

Character. "kmeans", "hclust", or "dbscan". Default "kmeans".

k

Integer. Number of clusters for kmeans/hclust. Default 3.

...

Additional arguments passed to the underlying clustering function.

Value

An S3 object of class libs_clusters with elements cluster, method, k, silhouette (mean score excluding DBSCAN noise; singleton clusters contribute zero; NA when not computable), sample_info.

Examples

ds <- ls_example_data("tissue")
cl <- ls_cluster(ds, method = "kmeans", k = 5)
cl$cluster
ds <- ls_example_data("tissue")
cl <- ls_cluster(ds, method = "kmeans", k = 5)
cl$cluster

Crop Spectral Range

Description

Subsets a spectrum or dataset to a wavelength range.

Usage

ls_crop(x, min_nm = NULL, max_nm = NULL)
ls_crop(x, min_nm = NULL, max_nm = NULL)

Arguments

x

A ls_spectrum() or ls_dataset() object.

min_nm

Numeric. Minimum wavelength (nm). Default NULL = no lower bound.

max_nm

Numeric. Maximum wavelength (nm). Default NULL = no upper bound.

Value

Same class as input, cropped.

Examples

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(), seed = 1)
ls_crop(spec, 380, 450)
spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(), seed = 1)
ls_crop(spec, 380, 450)

Create a LIBS Dataset

Description

Constructs a libs_dataset: a collection of ls_spectrum() objects sharing a common wavelength axis, plus a sample-info table.

Usage

ls_dataset(spectra, sample_info = NULL)
ls_dataset(spectra, sample_info = NULL)

Arguments

spectra

List of libs_spectrum objects. All must share the same wavelength axis (exact channel values). Resample deliberately before combining measurements on different axes; no channels are relabelled.

sample_info

Optional data.frame or tibble with one row per spectrum. Must contain column sample_id matching metadata$sample_id in each spectrum. Additional columns (e.g. material, group, concentration, x_pos, y_pos) are optional. If NULL, built from spectrum metadata.

Value

A libs_dataset S3 object.

Examples

specs <- lapply(1:3, function(i) {
  ls_simulate_spectrum(line_data = ls_demo_peaks(), elements = c(Ca = 1000 * i, Na = 500), seed = i)
})
specs <- Map(function(s, id) {
  s$metadata$sample_id <- paste0("s", id)
  s
}, specs, seq_along(specs))
ds <- ls_dataset(specs)
print(ds)

specs <- lapply(1:3, function(i) {
  ls_simulate_spectrum(line_data = ls_demo_peaks(), elements = c(Ca = 1000 * i, Na = 500), seed = i)
})
specs <- Map(function(s, id) {
  s$metadata$sample_id <- paste0("s", id)
  s
}, specs, seq_along(specs))
ds <- ls_dataset(specs)
print(ds)

Fictional Peaks for Software Demonstrations

Description

Returns invented peak strengths at the package's existing teaching and heuristic integration windows. Component names are labels used by the example scenarios, not validated elemental identifications. This table is not derived from an atomic database: it contains no transition rates, energy levels or statistical weights, and must not be used as a reference for experimental spectra. It is accepted only as explicit simulation input.

Usage

ls_demo_peaks()
ls_demo_peaks()

Value

A tibble with component, wavelength_nm and relative_strength. Every strength is the arbitrary multiplier 0.001.

Examples

peaks <- ls_demo_peaks()
spec <- ls_simulate_spectrum(seed = 1, line_data = peaks)
ls_plot_spectrum(spec)
peaks <- ls_demo_peaks()
spec <- ls_simulate_spectrum(seed = 1, line_data = peaks)
ls_plot_spectrum(spec)

Read a User-Supplied Atomic Line Reference

Description

Reads a local CSV or RDS table supplied by the user. No atomic reference compilation is bundled and this function does not access the network. Configure a file with options(libscanR.element_db = "/path/lines.csv") or the environment variable LIBSCANR_ELEMENT_DB. An explicit path takes precedence over the option, which takes precedence over the variable.

Usage

ls_element_db(
  elements = NULL,
  range_nm = c(190, 900),
  min_aki = 0,
  ionization = c(1L, 2L),
  persistent_only = FALSE,
  path = NULL
)
ls_element_db(
  elements = NULL,
  range_nm = c(190, 900),
  min_aki = 0,
  ionization = c(1L, 2L),
  persistent_only = FALSE,
  path = NULL
)

Arguments

elements

Character vector of element symbols, or NULL for all.

range_nm

Two ordered finite wavelength limits, or NULL for all.

min_aki

Minimum transition probability in units of 10^8 s^-1. Missing rates are excluded only when this is positive.

ionization

Ionization states: 1 = neutral, 2 = singly ionized, or NULL for all supplied positive integer stages.

persistent_only

Keep only rows explicitly marked persistent.

path

Local CSV/RDS file. Default NULL uses the configured path.

Details

The table must contain element, ionization, wavelength_nm, aki (transition probability in units of 10^8 s^-1), wavelength_medium ("air" or "vacuum") and a nonempty source attribution per row. Missing transition rates remain missing. Optional ei_ev, ek_ev and persistent columns default to missing energies and FALSE; no persistent line designation is inferred. Other columns and provenance are retained. The installed package provides scripts/import_nist.py for converting a separately downloaded NIST ASD CSV; see the README for exact settings.

Value

A tibble::tibble() retaining the reference columns. Physical fields required by ls_saha_boltzmann() must be supplied and independently checked; loading a reference does not establish its scientific validity or reuse rights.

Examples

# After downloading and importing a reference as described in the README:
# options(libscanR.element_db = "local-data/atomic-lines.csv")
# ls_element_db(elements = "Ca")
# After downloading and importing a reference as described in the README:
# options(libscanR.element_db = "local-data/atomic-lines.csv")
# ls_element_db(elements = "Ca")

Generate Example LIBS Dataset

Description

Creates a synthetic libs_dataset suitable for demonstrating package features without an atomic reference download. All scenarios use explicit fictional peaks from ls_demo_peaks(); component names and tissue labels are teaching labels, not measured concentrations or biological ground truth. Scenarios:

Usage

ls_example_data(
  scenario = c("tissue", "calibration", "spatial", "all"),
  seed = 42,
  n_channels = 1024
)
ls_example_data(
  scenario = c("tissue", "calibration", "spatial", "all"),
  seed = 42,
  n_channels = 1024
)

Arguments

scenario

Character. One of "tissue", "calibration", "spatial", "all". Default "tissue".

seed

Integer. Random seed. Default 42.

n_channels

Integer. Number of spectral channels. Default 1024 (reduced from 2048 for faster examples).

Details

"tissue" — 50 spectra from 5 tissue types (bone, liver, kidney, muscle, fat), 10 per type, with deliberately different synthetic profiles.
"calibration" — 27 spectra of Ca standards (9 concentrations x 3 replicates) with constant matrix (Na, K, Mg).
"spatial" — 400 spectra on a 20x20 grid simulating a tissue cross-section with Ca/Fe gradients and a Zn-enriched hotspot.
"all" — named list of the above.

Value

A libs_dataset object (or named list for "all").

Examples

ds <- ls_example_data("tissue")
print(ds)
ds <- ls_example_data("tissue")
print(ds)

Export a LIBS Dataset to a Directory

Description

Writes each spectrum in a dataset as a separate CSV file, plus a sample_info.csv metadata table. IDs must be unique portable filenames. Existing output is left intact if overwrite = FALSE; if any target exists, the whole export is skipped. All files are staged before replacement, with rollback on ordinary write failures. Concurrent writers and power-loss recovery are not supported.

Usage

ls_export_spectra(dataset, dir, overwrite = FALSE)
ls_export_spectra(dataset, dir, overwrite = FALSE)

Arguments

dataset

A libs_dataset object.

dir

Character. Output directory (created if missing).

overwrite

Logical. Overwrite existing files. Default FALSE.

Value

Invisibly returns dir.

Examples

ds <- ls_example_data("tissue")[1:3]
tmp <- tempfile()
ls_export_spectra(ds, tmp)
list.files(tmp)
unlink(tmp, recursive = TRUE)

ds <- ls_example_data("tissue")[1:3]
tmp <- tempfile()
ls_export_spectra(ds, tmp)
list.files(tmp)
unlink(tmp, recursive = TRUE)

Find Emission Peaks

Description

Detects peaks in a LIBS spectrum using local maxima detection with prominence and signal-to-noise filtering.

Usage

ls_find_peaks(
  x,
  snr_threshold = 3,
  min_prominence = 0.01,
  min_distance_nm = 0.5
)
ls_find_peaks(
  x,
  snr_threshold = 3,
  min_prominence = 0.01,
  min_distance_nm = 0.5
)

Arguments

x

A ls_spectrum() object.

snr_threshold

Numeric. Minimum signal-to-noise ratio. Default 3.

min_prominence

Numeric. Minimum peak prominence as a fraction of maximum intensity (0-1). Default 0.01.

min_distance_nm

Numeric. Minimum distance between peaks (nm). Default 0.5.

Value

A tibble::tibble() with columns wavelength_nm, intensity, snr, prominence, fwhm_nm, area, sorted by intensity descending.

Examples

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(),
                             elements = c(Ca = 5000, Na = 1000), seed = 1)
peaks <- ls_find_peaks(spec, snr_threshold = 5)
head(peaks)

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(),
                             elements = c(Ca = 5000, Na = 1000), seed = 1)
peaks <- ls_find_peaks(spec, snr_threshold = 5)
head(peaks)

Gate Delay Optimization

Description

Analyzes SNR and signal-to-background ratios across spectra acquired at different gate delays to recommend optimal timing for a target line.

Usage

ls_gate_optimize(spectra, element, line_nm, window_nm = 1)
ls_gate_optimize(spectra, element, line_nm, window_nm = 1)

Arguments

spectra

List of ls_spectrum() objects. Each spectrum's metadata should contain gate_delay_us.

element

Character. Target element symbol.

line_nm

Numeric. Target emission line wavelength (nm).

window_nm

Numeric. Window around the line. Default 1.

Value

A tibble::tibble() with columns gate_delay_us, snr, sbr, peak_intensity, with attribute "recommended" giving the row index of the highest SNR.

Examples

specs <- lapply(c(0.5, 1, 2, 5, 10), function(g) {
  s <- ls_simulate_spectrum(line_data = ls_demo_peaks(), seed = round(g * 10))
  s$metadata$gate_delay_us <- g
  s
})
ls_gate_optimize(specs, "Ca", 393.37)
specs <- lapply(c(0.5, 1, 2, 5, 10), function(g) {
  s <- ls_simulate_spectrum(line_data = ls_demo_peaks(), seed = round(g * 10))
  s$metadata$gate_delay_us <- g
  s
})
ls_gate_optimize(specs, "Ca", 393.37)

Identify Peaks Against the NIST Line Database

Description

Matches detected peaks to known atomic emission lines. Each peak is assigned its nearest wavelength candidate from the user-configured reference. No atomic reference is bundled; configure ls_element_db() first. This is a candidate identification, not proof of elemental presence or resolution of blended transitions. Check the measured air/vacuum convention.

Usage

ls_identify_peaks(
  peaks,
  elements = NULL,
  tolerance_nm = 0.2,
  ionization = c(1L, 2L)
)
ls_identify_peaks(
  peaks,
  elements = NULL,
  tolerance_nm = 0.2,
  ionization = c(1L, 2L)
)

Arguments

peaks

A tibble from ls_find_peaks().

elements

Character vector. Restrict search. Default NULL = all.

tolerance_nm

Numeric. Maximum wavelength deviation. Default 0.2.

ionization

Integer vector of ionization states. Default c(1L, 2L).

Value

A tibble with all peak columns plus: element, ionization, nist_wavelength_nm, nist_aki, deviation_nm, confidence (0-1), n_candidates, and wavelength_medium. The legacy confidence name denotes only 1 - abs(deviation_nm) / tolerance_nm, not a probability.

Examples


spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(),
                             elements = c(Ca = 5000, Na = 1000), seed = 1)
pk <- ls_find_peaks(spec, snr_threshold = 5)
id <- ls_identify_peaks(pk, elements = c("Ca", "Na"))
head(id)

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(),
                             elements = c(Ca = 5000, Na = 1000), seed = 1)
pk <- ls_find_peaks(spec, snr_threshold = 5)
id <- ls_identify_peaks(pk, elements = c("Ca", "Na"))
head(id)

Limit of Detection (3-sigma)

Description

Calculates LOD using the 3-sigma criterion.

Usage

ls_lod(calibration, blank = NULL, window_nm = NULL)
ls_lod(calibration, blank = NULL, window_nm = NULL)

Arguments

calibration

A ls_calibration() object.

blank

Optional ls_spectrum() of a blank sample. If NULL (the default), LOD is recomputed from the residual standard error of the response-on-concentration calibration model. A supplied blank must contain at least two replicate shots on the fitted axis with the same preprocessing; its standard deviation is computed from the same integrated response (or internal-standard ratio) for each shot, not variation across wavelengths.

window_nm

Optional window; NULL uses the fitted window. A different value is rejected.

Value

Estimated numeric LOD in the calibration's concentration unit; NA for PLS. This 3-sigma estimate requires independent confirmation near the detection limit and is not an instrument qualification.

Examples

ds <- ls_example_data("calibration")
conc <- ds$sample_info$concentration
cal <- ls_calibrate(ds, "Ca", 393.37, conc, verbose = FALSE)
ls_lod(cal)
ds <- ls_example_data("calibration")
conc <- ds$sample_info$concentration
cal <- ls_calibrate(ds, "Ca", 393.37, conc, verbose = FALSE)
ls_lod(cal)

Limit of Quantification (10-sigma)

Description

Calculates LOQ using the 10-sigma criterion.

Usage

ls_loq(calibration, blank = NULL, window_nm = NULL)
ls_loq(calibration, blank = NULL, window_nm = NULL)

Arguments

calibration

A ls_calibration() object.

blank

Optional ls_spectrum() of a blank sample.

window_nm

Optional window; NULL uses the fitted window.

Value

Numeric LOQ in the calibration's concentration unit.

Examples

ds <- ls_example_data("calibration")
conc <- ds$sample_info$concentration
cal <- ls_calibrate(ds, "Ca", 393.37, conc, verbose = FALSE)
ls_loq(cal)
ds <- ls_example_data("calibration")
conc <- ds$sample_info$concentration
cal <- ls_calibrate(ds, "Ca", 393.37, conc, verbose = FALSE)
ls_loq(cal)

Build Maps for Multiple Elements

Description

Build Maps for Multiple Elements

Usage

ls_map_elements(
  dataset,
  elements,
  lines_nm,
  calibrations = NULL,
  window_nm = NULL
)
ls_map_elements(
  dataset,
  elements,
  lines_nm,
  calibrations = NULL,
  window_nm = NULL
)

Arguments

dataset

A ls_dataset() object with spatial coordinates.

elements

Character vector. Elements to map.

lines_nm

Named numeric vector: wavelength per element (e.g. c(Ca = 393.37, Fe = 371.99)).

calibrations

Named list of ls_calibration() objects. Default NULL.

window_nm

Optional integration window. NULL reuses each calibration's fitted window, or 1 nm for uncalibrated maps.

Value

A named list of libs_map objects.

Examples

ds <- ls_example_data("spatial")
ls_map_elements(ds, c("Ca", "Fe"),
                c(Ca = 393.37, Fe = 371.99))
ds <- ls_example_data("spatial")
ls_map_elements(ds, c("Ca", "Fe"),
                c(Ca = 393.37, Fe = 371.99))

Normalize Spectra

Description

Normalizes intensity values. Methods:

"total" — divide each shot by total intensity
"max" — divide by per-shot maximum
"snv" — Standard Normal Variate: (x - mean) / sd
"internal_std" — divide by intensity at a reference line
"area" — divide by integrated (trapezoidal) area

Usage

ls_normalize(x, method = "total", ref_wavelength = NULL, ref_window = 1)
ls_normalize(x, method = "total", ref_wavelength = NULL, ref_window = 1)

Arguments

x

A ls_spectrum() or ls_dataset() object.

method

Character. Normalization method. Default "total".

ref_wavelength

Numeric. Required for method = "internal_std".

ref_window

Numeric. Window width (nm) around ref_wavelength. Default 1.

Value

Same class as input with normalized intensities. A zero or non-finite divisor is an error; a constant SNV shot becomes zero. Previous removed baselines are retained in metadata$baseline_history and cleared from baseline, because per-shot scaling changes their units.

Examples

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(), seed = 1)
spec_n <- ls_normalize(spec, method = "total")
sum(spec_n$intensity[1, ])  # ~1

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(), seed = 1)
spec_n <- ls_normalize(spec, method = "total")
sum(spec_n$intensity[1, ])  # ~1

Principal Component Analysis of LIBS Spectra

Description

Performs PCA on the intensity matrix of a ls_dataset().

Usage

ls_pca(dataset, n_components = 10, scale = TRUE, center = TRUE)
ls_pca(dataset, n_components = 10, scale = TRUE, center = TRUE)

Arguments

dataset

A ls_dataset() object.

n_components

Integer. Number of PCs to retain. Default 10.

scale

Logical. Scale variables to unit variance. Default TRUE.

center

Logical. Center variables. Default TRUE.

Value

A list with S3 class libs_pca containing: scores (matrix: spectra x PCs), loadings (matrix: channels x PCs), sdev, variance_explained (fraction), cumulative_variance, wavelength, sample_info, pca_obj.

Examples

ds <- ls_example_data("tissue")
pca <- ls_pca(ds, n_components = 5)
pca$variance_explained

ds <- ls_example_data("tissue")
pca <- ls_pca(ds, n_components = 5)
pca$variance_explained

Calculate Peak Area

Description

Integrates peak area at a given emission line, with a local baseline subtracted (linear interpolation between window edges).

Usage

ls_peak_area(x, center_nm, window_nm = 1, method = "trapezoidal")
ls_peak_area(x, center_nm, window_nm = 1, method = "trapezoidal")

Arguments

x

A ls_spectrum() object.

center_nm

Numeric. Peak center wavelength (nm).

window_nm

Numeric. Integration window width. Default 1. To support coarsely sampled input, the effective window is at least five median channel spacings, with up to five nearest channels used at edges. This does not resolve overlapping lines; inspect instrument resolution and the integration region before interpreting an area as an element.

method

Character. "trapezoidal" or "gaussian_fit". Default "trapezoidal". A failed, non-converged or non-physical Gaussian fit emits a warning and returns the trapezoidal area instead.

Value

Numeric. Baseline-subtracted integrated peak area.

Examples

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(),
                             elements = c(Ca = 5000), seed = 1)
ls_peak_area(spec, 393.37)
spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(),
                             elements = c(Ca = 5000), seed = 1)
ls_peak_area(spec, 393.37)

Plot a Calibration Curve

Description

Plot a Calibration Curve

Usage

ls_plot_calibration(
  calibration,
  show_lod = TRUE,
  show_loq = TRUE,
  show_prediction_interval = TRUE
)
ls_plot_calibration(
  calibration,
  show_lod = TRUE,
  show_loq = TRUE,
  show_prediction_interval = TRUE
)

Arguments

calibration

A ls_calibration() object. PLS shows fitted training concentrations against known concentrations, not a one-feature curve.

show_lod

Logical. Show LOD as a horizontal dashed line. Default TRUE.

show_loq

Logical. Show LOQ as a horizontal dashed line. Default TRUE.

show_prediction_interval

Logical. Show 95% prediction band. Default TRUE. For univariate/internal-standard fits this is the fitted model's interval for a new observation, not a confidence band for its mean. PLS plots do not supply a prediction interval.

Value

A ggplot2 object.

Examples

ds <- ls_example_data("calibration")
conc <- ds$sample_info$concentration
cal <- ls_calibrate(ds, "Ca", 393.37, conc, verbose = FALSE)
ls_plot_calibration(cal)
ds <- ls_example_data("calibration")
conc <- ds$sample_info$concentration
cal <- ls_calibrate(ds, "Ca", 393.37, conc, verbose = FALSE)
ls_plot_calibration(cal)

Plot an Element Map (alias)

Description

Convenience alias for ls_plot_map().

Usage

ls_plot_element_map(map, ...)
ls_plot_element_map(map, ...)

Arguments

map

A libs_map object.

...

Passed to ls_plot_map().

Value

A ggplot2 object.

Examples

ds <- ls_example_data("spatial")
ls_plot_element_map(ls_build_map(ds, "Ca", 393.37))
ds <- ls_example_data("spatial")
ls_plot_element_map(ls_build_map(ds, "Ca", 393.37))

Plot PCA Loadings

Description

Plots the loadings for a selected principal component across the full wavelength axis, optionally annotating the top contributing channels.

Usage

ls_plot_loadings(pca, pc = 1, n_top = 10)
ls_plot_loadings(pca, pc = 1, n_top = 10)

Arguments

pca

A libs_pca object.

pc

Integer. Which PC's loadings to show. Default 1.

n_top

Integer. Number of top-absolute contributors to annotate. Default 10.

Value

A ggplot2 object.

Examples

ds <- ls_example_data("tissue")
pca <- ls_pca(ds, n_components = 4)
ls_plot_loadings(pca, pc = 1, n_top = 5)
ds <- ls_example_data("tissue")
pca <- ls_pca(ds, n_components = 4)
ls_plot_loadings(pca, pc = 1, n_top = 5)

Plot an Elemental Map

Description

Plot an Elemental Map

Usage

ls_plot_map(map, color_scale = "viridis", title = NULL)
ls_plot_map(map, color_scale = "viridis", title = NULL)

Arguments

map

A libs_map object from ls_build_map().

color_scale

Character. "viridis", "magma", "plasma", "inferno", "hot", "jet". Default "viridis".

title

Character. Plot title. Default: element symbol.

Value

A ggplot2 object.

Examples

ds <- ls_example_data("spatial")
m <- ls_build_map(ds, "Ca", 393.37)
ls_plot_map(m)
ds <- ls_example_data("spatial")
m <- ls_build_map(ds, "Ca", 393.37)
ls_plot_map(m)

Plot Multi-Element Map Panel

Description

Plots multiple elemental maps as a faceted panel.

Usage

ls_plot_map_panel(maps, ncol = 3)
ls_plot_map_panel(maps, ncol = 3)

Arguments

maps

Named list of libs_map objects.

ncol

Integer. Facet columns. Default 3.

Value

A ggplot2 object.

Examples

ds <- ls_example_data("spatial")
ms <- ls_map_elements(ds, c("Ca", "Fe", "Zn"),
                      c(Ca = 393.37, Fe = 371.99, Zn = 213.86))
ls_plot_map_panel(ms)
ds <- ls_example_data("spatial")
ms <- ls_map_elements(ds, c("Ca", "Fe", "Zn"),
                      c(Ca = 393.37, Fe = 371.99, Zn = 213.86))
ls_plot_map_panel(ms)

Overlay Multiple Spectra

Description

Plots multiple spectra on the same axes, optionally colored by a grouping variable.

Usage

ls_plot_overlay(dataset, color_by = NULL, normalize = FALSE, alpha = 0.5)
ls_plot_overlay(dataset, color_by = NULL, normalize = FALSE, alpha = 0.5)

Arguments

dataset

A ls_dataset() object.

color_by

Character. Column in sample_info for color mapping. Default NULL.

normalize

Logical. Divide each spectrum by its positive maximum. Default FALSE.

alpha

Numeric. Line alpha. Default 0.5.

Value

A ggplot2 object.

Examples

ds <- ls_example_data("tissue")
ls_plot_overlay(ds[1:10], color_by = "tissue")
ds <- ls_example_data("tissue")
ls_plot_overlay(ds[1:10], color_by = "tissue")

Plot PCA Scores

Description

Plot PCA Scores

Usage

ls_plot_pca(pca, pc_x = 1, pc_y = 2, color_by = NULL, ellipses = TRUE)
ls_plot_pca(pca, pc_x = 1, pc_y = 2, color_by = NULL, ellipses = TRUE)

Arguments

pca

A libs_pca object from ls_pca().

pc_x

Integer. PC for x-axis. Default 1.

pc_y

Integer. PC for y-axis. Default 2.

color_by

Character. Sample-info column. Numeric values use a continuous colour scale; other values define discrete groups. Default NULL.

ellipses

Logical. Draw nominal 95% normal-theory data ellipses for discrete groups with at least four non-collinear points. They describe score dispersion, not confidence intervals for a group mean. Default TRUE.

Value

A ggplot2 object.

Examples

ds <- ls_example_data("tissue")
pca <- ls_pca(ds, n_components = 4)
ls_plot_pca(pca, color_by = "tissue")
ds <- ls_example_data("tissue")
pca <- ls_pca(ds, n_components = 4)
ls_plot_pca(pca, color_by = "tissue")

Plot PLS-DA Results

Description

Plot PLS-DA Results

Usage

ls_plot_plsda(plsda, type = "scores")
ls_plot_plsda(plsda, type = "scores")

Arguments

plsda

A libs_plsda object from ls_plsda().

type

Character. "scores" (LV1 vs LV2), "confusion" (heatmap), or "vip" (VIP scores spectrum). Default "scores".

Value

A ggplot2 object.

Examples

ds <- ls_example_data("tissue")

if (requireNamespace("pls", quietly = TRUE)) {
  plsda <- ls_plsda(ds, "tissue", n_components = 3, validation = "none")
  ls_plot_plsda(plsda, type = "scores")
}

ds <- ls_example_data("tissue")

if (requireNamespace("pls", quietly = TRUE)) {
  plsda <- ls_plsda(ds, "tissue", n_components = 3, validation = "none")
  ls_plot_plsda(plsda, type = "scores")
}

Plot a Spectral Region in Detail

Description

Zooms into a specified wavelength range and annotates peaks when a user reference is configured. With default elements = NULL and no reference, the spectrum remains visible and a subtitle reports unavailable annotations. Explicitly requesting element labels requires a configured reference.

Usage

ls_plot_region(x, min_nm, max_nm, elements = NULL, snr_threshold = 5)
ls_plot_region(x, min_nm, max_nm, elements = NULL, snr_threshold = 5)

Arguments

x

A ls_spectrum() or ls_dataset() object.

min_nm

Numeric. Minimum wavelength (nm).

max_nm

Numeric. Maximum wavelength (nm).

elements

Optional character vector. Element symbols to annotate. Default NULL (auto-detect top peaks).

snr_threshold

Numeric. SNR threshold for peak detection. Default 5.

Value

A ggplot2 object.

Examples

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(), seed = 1)
ls_plot_region(spec, 380, 410)
spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(), seed = 1)
ls_plot_region(spec, 380, 410)

Plot Calibration Residuals

Description

Standard regression diagnostic: residuals vs fitted.

Usage

ls_plot_residuals(calibration)
ls_plot_residuals(calibration)

Arguments

calibration

A ls_calibration() object.

Value

A ggplot2 object.

Examples

ds <- ls_example_data("calibration")
conc <- ds$sample_info$concentration
cal <- ls_calibrate(ds, "Ca", 393.37, conc, verbose = FALSE)
ls_plot_residuals(cal)
ds <- ls_example_data("calibration")
conc <- ds$sample_info$concentration
cal <- ls_calibrate(ds, "Ca", 393.37, conc, verbose = FALSE)
ls_plot_residuals(cal)

Plot PCA Variance (Scree Plot)

Description

Plot PCA Variance (Scree Plot)

Usage

ls_plot_scree(pca)
ls_plot_scree(pca)

Arguments

pca

A libs_pca object.

Value

A ggplot2 object.

Examples

ds <- ls_example_data("tissue")
pca <- ls_pca(ds, n_components = 6)
ls_plot_scree(pca)
ds <- ls_example_data("tissue")
pca <- ls_pca(ds, n_components = 6)
ls_plot_scree(pca)

Plot a LIBS Spectrum

Description

Generic spectrum plotter. Returns a ggplot2 object.

Usage

ls_plot_spectrum(
  x,
  show_peaks = FALSE,
  show_elements = NULL,
  normalize = FALSE,
  snr_threshold = 5,
  ...
)
ls_plot_spectrum(
  x,
  show_peaks = FALSE,
  show_elements = NULL,
  normalize = FALSE,
  snr_threshold = 5,
  ...
)

Arguments

x

A ls_spectrum() object.

show_peaks

Logical. Annotate detected peaks. Default FALSE.

show_elements

Character vector of element symbols to highlight as dashed vertical lines from the configured reference. All matching rows in the displayed range are used; no persistent-line status is inferred. Default NULL.

normalize

Logical. Divide intensity by its positive maximum before plotting. Default FALSE.

snr_threshold

Numeric. SNR threshold for peak annotation when show_peaks = TRUE. Default 5.

...

Additional arguments (currently unused).

Value

A ggplot2 object.

Examples

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(), seed = 1)
ls_plot_spectrum(spec)
# Element overlays require a configured reference; see ls_element_db().
spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(), seed = 1)
ls_plot_spectrum(spec)
# Element overlays require a configured reference; see ls_element_db().

PLS-DA Classification

Description

Partial Least Squares Discriminant Analysis for class prediction.

Usage

ls_plsda(dataset, grouping, n_components = 5, validation = "CV")
ls_plsda(dataset, grouping, n_components = 5, validation = "CV")

Arguments

dataset

A ls_dataset() object.

grouping

Character. Column name in sample_info with class labels.

n_components

Integer. Number of PLS latent variables. Default 5.

validation

Character. "CV" (10-fold), "LOO", or "none". Default "CV". Accuracy and predictions use held-out predictions for CV/LOO and fitted predictions only for "none". Components are capped at the centered feature rank and training-sample bound of every fold. Every training fold must retain at least two classes. These sample-level folds do not establish patient/site-level generalization.

Value

An S3 object of class libs_plsda with elements model, predictions (tibble with predicted/observed class), confusion_matrix, accuracy, class_labels, n_components, sample_info, feature_contract, and evaluation_basis ("cross_validation", "leave_one_out", or "training").

Examples

ds <- ls_example_data("tissue")

if (requireNamespace("pls", quietly = TRUE)) {
  plsda <- ls_plsda(ds, "tissue", n_components = 3, validation = "none")
  plsda$accuracy
}

ds <- ls_example_data("tissue")

if (requireNamespace("pls", quietly = TRUE)) {
  plsda <- ls_plsda(ds, "tissue", n_components = 3, validation = "none")
  plsda$accuracy
}

Quantify Element Concentration

Description

Applies a calibration model using its saved wavelength axis, preprocessing and extraction contract. Legacy models without this contract must be refitted. Unavailable detection-limit flags are NA, not FALSE.

Usage

ls_quantify(calibration, x, window_nm = NULL)
ls_quantify(calibration, x, window_nm = NULL)

Arguments

calibration

A ls_calibration() object.

x

A ls_spectrum() or ls_dataset() object.

window_nm

Optional numeric integration window. NULL reuses the fitted window; a different supplied value is rejected.

Value

A tibble::tibble() with columns sample_id, element, concentration, unit, below_lod, below_loq.

Examples

ds <- ls_example_data("calibration")
conc <- ds$sample_info$concentration
cal <- ls_calibrate(ds, "Ca", 393.37, conc, verbose = FALSE)
ls_quantify(cal, ds)
ds <- ls_example_data("calibration")
conc <- ds$sample_info$concentration
cal <- ls_calibrate(ds, "Ca", 393.37, conc, verbose = FALSE)
ls_quantify(cal, ds)

Read Applied Spectra / Aurora Data

Description

Parses output from Applied Spectra J200 and Aurora LIBS systems. Applied Spectra files commonly have wavelength in the first column and shots in subsequent columns, optionally with a few header lines.

Usage

ls_read_aurora(path, verbose = TRUE)
ls_read_aurora(path, verbose = TRUE)

Arguments

path

Character. Path to export file (.csv or .asc).

verbose

Logical. Default TRUE.

Value

A libs_spectrum object.

Examples

tmp <- tempfile(fileext = ".csv")
writeLines(c(
  "Applied Spectra J200",
  "Model,J200",
  "wl,shot1,shot2",
  "200.0,50,52",
  "200.5,55,58",
  "201.0,52,50"
), tmp)
spec <- ls_read_aurora(tmp, verbose = FALSE)
unlink(tmp)
tmp <- tempfile(fileext = ".csv")
writeLines(c(
  "Applied Spectra J200",
  "Model,J200",
  "wl,shot1,shot2",
  "200.0,50,52",
  "200.5,55,58",
  "201.0,52,50"
), tmp)
spec <- ls_read_aurora(tmp, verbose = FALSE)
unlink(tmp)

Auto-Detect and Read LIBS Data

Description

Attempts to identify the file format (generic CSV, SciAps, Applied Spectra) and dispatches to the correct reader. When path is a directory, delegates to ls_read_dir().

Usage

ls_read_auto(path, verbose = TRUE)
ls_read_auto(path, verbose = TRUE)

Arguments

path

Character. Path to spectrum file or directory.

verbose

Logical. Default TRUE.

Value

A libs_spectrum or libs_dataset object.

Examples

tmp <- tempfile(fileext = ".csv")
utils::write.csv(data.frame(wavelength = seq(200, 300, 1),
                            intensity = 1:101),
                 tmp, row.names = FALSE)
spec <- ls_read_auto(tmp, verbose = FALSE)
unlink(tmp)
tmp <- tempfile(fileext = ".csv")
utils::write.csv(data.frame(wavelength = seq(200, 300, 1),
                            intensity = 1:101),
                 tmp, row.names = FALSE)
spec <- ls_read_auto(tmp, verbose = FALSE)
unlink(tmp)

Read a LIBS Spectrum from CSV (Convenience Wrapper)

Description

Read a LIBS Spectrum from CSV (Convenience Wrapper)

Usage

ls_read_csv(path, ...)
ls_read_csv(path, ...)

Arguments

path

Character. Path to CSV file.

...

Additional arguments passed to ls_read_spectrum().

Value

A libs_spectrum object.

Read All Spectra from a Directory

Description

Batch-imports spectrum files from a directory into a libs_dataset.

Usage

ls_read_dir(
  dir,
  pattern = "\\.(csv|tsv|txt)$",
  recursive = FALSE,
  sample_info = NULL,
  ...
)
ls_read_dir(
  dir,
  pattern = "\\.(csv|tsv|txt)$",
  recursive = FALSE,
  sample_info = NULL,
  ...
)

Arguments

dir

Character. Directory path.

pattern

Character. File pattern (regex). Default "\\.(csv|tsv|txt)$".

recursive

Logical. Recurse into subdirectories. Default FALSE.

sample_info

data.frame/tibble. Optional sample metadata table. Must have a sample_id column matching all of the filenames without extension. Default NULL.

...

Additional arguments passed to ls_read_spectrum().

Value

A libs_dataset object.

Examples

dir <- tempfile()
dir.create(dir)
for (i in 1:3) {
  wl <- seq(200, 900, length.out = 100)
  int <- exp(-((wl - 393)^2) / 2) + stats::rnorm(100, 0, 0.01)
  utils::write.csv(data.frame(w = wl, i = int),
                   file.path(dir, paste0("s", i, ".csv")),
                   row.names = FALSE)
}
ds <- ls_read_dir(dir, verbose = FALSE)
unlink(dir, recursive = TRUE)

dir <- tempfile()
dir.create(dir)
for (i in 1:3) {
  wl <- seq(200, 900, length.out = 100)
  int <- exp(-((wl - 393)^2) / 2) + stats::rnorm(100, 0, 0.01)
  utils::write.csv(data.frame(w = wl, i = int),
                   file.path(dir, paste0("s", i, ".csv")),
                   row.names = FALSE)
}
ds <- ls_read_dir(dir, verbose = FALSE)
unlink(dir, recursive = TRUE)

Read SciAps LIBS Data

Description

Parses output files from SciAps handheld LIBS analyzers (Z-series). The SciAps export format typically includes metadata lines preceded by ⁠#⁠ or a keyword header, followed by a wavelength/intensity table.

Usage

ls_read_sciaps(path, verbose = TRUE)
ls_read_sciaps(path, verbose = TRUE)

Arguments

path

Character. Path to SciAps export file.

verbose

Logical. Emit messages. Default TRUE.

Value

A libs_spectrum object with SciAps-specific metadata.

Examples

# See vignette("getting-started") for realistic usage; this parser
# is best exercised against actual SciAps export files.
tmp <- tempfile(fileext = ".csv")
writeLines(c(
  "# SciAps Z-903",
  "# Serial: 12345",
  "# Gate Delay (us): 2.5",
  "wavelength,intensity",
  "200.0,50",
  "200.5,55",
  "201.0,52"
), tmp)
spec <- ls_read_sciaps(tmp, verbose = FALSE)
unlink(tmp)
# See vignette("getting-started") for realistic usage; this parser
# is best exercised against actual SciAps export files.
tmp <- tempfile(fileext = ".csv")
writeLines(c(
  "# SciAps Z-903",
  "# Serial: 12345",
  "# Gate Delay (us): 2.5",
  "wavelength,intensity",
  "200.0,50",
  "200.5,55",
  "201.0,52"
), tmp)
spec <- ls_read_sciaps(tmp, verbose = FALSE)
unlink(tmp)

Read a LIBS Spectrum from File

Description

Reads a single spectrum file. Supports CSV, TSV, and TXT formats where columns represent wavelength and intensity.

Usage

ls_read_spectrum(
  path,
  format = "auto",
  wavelength_col = 1,
  intensity_col = 2,
  skip = 0,
  metadata = list(),
  verbose = TRUE
)
ls_read_spectrum(
  path,
  format = "auto",
  wavelength_col = 1,
  intensity_col = 2,
  skip = 0,
  metadata = list(),
  verbose = TRUE
)

Arguments

path

Character. Path to spectrum file.

format

Character. One of "auto" (detect from extension), "csv", "tsv", "txt". Default "auto".

wavelength_col

Column index or name for wavelength values. Default 1.

intensity_col

Column index/indices or name(s) for intensity values. Multiple columns are interpreted as multiple shots. Default 2.

skip

Integer. Header lines to skip before the data table. Default 0.

metadata

Named list of additional metadata. Default list().

verbose

Logical. Emit import messages. Default TRUE.

Value

A libs_spectrum object.

Examples

tmp <- tempfile(fileext = ".csv")
wl <- seq(200, 900, length.out = 100)
int <- exp(-((wl - 393)^2) / 2)
utils::write.csv(data.frame(wavelength = wl, intensity = int),
                 tmp, row.names = FALSE)
spec <- ls_read_spectrum(tmp, verbose = FALSE)
unlink(tmp)

tmp <- tempfile(fileext = ".csv")
wl <- seq(200, 900, length.out = 100)
int <- exp(-((wl - 393)^2) / 2)
utils::write.csv(data.frame(wavelength = wl, intensity = int),
                 tmp, row.names = FALSE)
spec <- ls_read_spectrum(tmp, verbose = FALSE)
unlink(tmp)

Launch the libscanR Shiny Application

Description

Starts an interactive Shiny application for exploring, preprocessing, and analyzing LIBS spectral data. The app provides six tabs: Import, Preprocessing, Peaks, Calibration, Chemometrics, and Export.

Usage

ls_run_app(data = NULL, port = NULL, launch.browser = TRUE)
ls_run_app(data = NULL, port = NULL, launch.browser = TRUE)

Arguments

data

Optional. A ls_spectrum(), ls_dataset(), or file path to preload. Objects and files are validated before starting the app. Default NULL (start with upload interface).

port

Integer. Port for the app. Default NULL (auto-select).

launch.browser

Logical. Open browser on launch. Default TRUE.

Details

Upload batches are admitted together: invalid files leave the previously loaded data intact. Original upload names identify their source. New data and successful preprocessing invalidate downstream analysis results. Preprocessing controls take effect only after Apply; a failed operation retains the last successful data, and Reset restores raw data.

The complete ZIP download contains one CSV per spectrum and, for a dataset, sample_info.csv. Spectrum CSV files retain all shots, baseline and supported metadata through ls_write_csv(). ZIP creation requires an external zip command on PATH (or configured through R_ZIPCMD), including on Windows. The separate wide dataset CSV contains mean intensities only. Calibration CSV files summarize figures of merit rather than storing reusable prediction models. Load a labelled ls_dataset() through data for analyses requiring concentration or class columns in sample_info.

Example datasets use explicitly fictional teaching peaks and concentrations, not an atomic reference or experimental measurements. Peak detection and raw plots work without a reference. Optional element annotations require a user-supplied file configured for ls_element_db(); missing or invalid references leave raw detections available and report the annotation error. Peak and calibration CSVs carry data_origin = "explicit_synthetic_peaks" for wholly fictional inputs; mixed calibration inputs carry "includes_explicit_synthetic_peaks". Peak exports follow the first spectrum actually used for detection.

The previous application data option is restored when the app stops or startup fails. Each browser session keeps its own working data and results.

Value

Invisible NULL. Launches a Shiny application as a side effect.

Examples


if (interactive()) {
  ls_run_app()
}

if (interactive()) {
  ls_run_app()
}

Relative Elemental Amounts under a Two-Stage LTE Model

Description

Estimates a common Boltzmann temperature and relative elemental number fractions using explicit atomic and instrumental inputs. The Saha equation accounts for the neutral/singly-ionized balance. This research calculation requires optically thin emission, local thermodynamic equilibrium (LTE), a common sampled plasma, stoichiometric ablation and corrected spectral response. It does not establish those assumptions from an observed spectrum.

Usage

ls_saha_boltzmann(
  x,
  elements,
  lines_nm,
  electron_density = NULL,
  window_nm = 0.5,
  verbose = TRUE,
  atomic_data = NULL,
  partition_functions = NULL,
  ionization_energies_ev = NULL,
  temperature_k = NULL,
  intensity_scale = c("energy", "photon"),
  wavelength_medium = NULL,
  tolerance_nm = 0.01
)
ls_saha_boltzmann(
  x,
  elements,
  lines_nm,
  electron_density = NULL,
  window_nm = 0.5,
  verbose = TRUE,
  atomic_data = NULL,
  partition_functions = NULL,
  ionization_energies_ev = NULL,
  temperature_k = NULL,
  intensity_scale = c("energy", "photon"),
  wavelength_medium = NULL,
  tolerance_nm = 0.01
)

Arguments

x

A ls_spectrum() with resolved, baseline-corrected lines.

elements

Unique character vector of elements to include.

lines_nm

Named list of measured line wavelengths per element.

electron_density

Independently estimated electron density in cm^-3. Required; no density is assumed.

window_nm

Integration width in nm. Default 0.5. Effective integration windows, including the coarse-channel expansion of ls_peak_area(), must be disjoint, wholly inside the measured wavelength range and contain at least three measured channels each.

verbose

Logical. Report the temperature and scope of the result.

atomic_data

A data.frame with one unambiguous record per selected transition: element, ionization (1 = neutral, 2 = singly ionized), wavelength_nm, wavelength_medium ("air" or "vacuum"), wavelength_vacuum_nm, A_s (transition rate in s^-1), g_upper (statistical weight), energy_upper_ev (relative to that ion's ground state), response (measured signal per relative emitted intensity), and source (the transition reference). These data must be supplied explicitly; the bundled historical line compilation is insufficient for this calculation. For vacuum records the two wavelength columns must agree within 1e-8 + 1e-8 * wavelength_nm nm (numerical rounding only).

partition_functions

Named list, one entry per element. Each entry is a list with neutral and ionized, each either a function of temperature in kelvin returning the partition function, or an explicitly chosen positive constant approximation. Use the same statistical-weight convention as g_upper and include the relevant energy levels.

ionization_energies_ev

Named positive numeric vector of first ionization energies in eV, one per element.

temperature_k

Optional independently estimated common temperature in kelvin. If NULL, estimate it from a shared slope with a separate intercept per element/ion stage. Estimation requires a full-rank fit and positive residual degrees of freedom. Non-negative slopes are rejected.

intensity_scale

"energy" (default) or "photon", describing the response-corrected integrated intensities. Energy intensities require the vacuum-wavelength factor; photon counts do not.

wavelength_medium

Medium of the measured axis. If NULL, read x$metadata$wavelength_medium; an unspecified medium is an error.

tolerance_nm

Maximum difference between requested and reference wavelengths. Default 0.01. Multiple possible transitions are an error.

Details

Earlier versions silently used unit statistical weights, omitted partition functions and ignored electron density. Those outputs cannot be reused as quantitative reference results. Supply verified physical inputs and refit. A low residual does not verify LTE, optical thinness, absence of self-absorption, spectral resolution, detector response or matrix-independent ablation.

Value

A tibble with element, temperature_k, concentration_rel, electron_density, n_lines, r_squared and log_population_sd. Fractions are normalized only over the supplied elements, under a model that excludes molecules and charge states above +1; they are neither mass fractions nor absolute sample concentrations. log_population_sd measures line/stage disagreement on the log population scale, not calibrated uncertainty. Attributes line_diagnostics, assumptions, and temperature_source retain the calculation's inputs and limitations.

References

NIST Atomic Spectra Database, spectral lines and Saha/LTE documentation: https://physics.nist.gov/PhysRefData/ASD/Html/lineshelp.html.

Examples

# A complete, explicitly synthetic forward-model example:
system.file("examples", "saha-boltzmann-synthetic.R", package = "libscanR")
# A complete, explicitly synthetic forward-model example:
system.file("examples", "saha-boltzmann-synthetic.R", package = "libscanR")

Simulate a LIBS Spectrum

Description

Generates a demonstration LIBS spectrum with configurable elemental composition, noise, and continuum background. Emission lines are drawn from the user-configured atomic reference (ls_element_db()) and modelled as Gaussian peaks. Peak widths are at least 2.5 channel spacings so coarse demonstrations remain visible; this is not an instrument response model or validation of quantitative elemental concentrations or tissue identity.

Usage

ls_simulate_spectrum(
  elements = c(Ca = 5000, Fe = 200, Na = 1000),
  wavelength_range = c(200, 900),
  n_channels = 2048,
  n_shots = 10,
  noise_level = 0.02,
  continuum_level = 100,
  shot_rsd = 0.1,
  fwhm_nm = 0.1,
  seed = NULL,
  line_data = NULL
)
ls_simulate_spectrum(
  elements = c(Ca = 5000, Fe = 200, Na = 1000),
  wavelength_range = c(200, 900),
  n_channels = 2048,
  n_shots = 10,
  noise_level = 0.02,
  continuum_level = 100,
  shot_rsd = 0.1,
  fwhm_nm = 0.1,
  seed = NULL,
  line_data = NULL
)

Arguments

elements

Named numeric vector. Element concentrations (arbitrary units, interpreted as relative intensity scaling). Default c(Ca = 5000, Fe = 200, Na = 1000).

wavelength_range

Numeric vector of length 2. Range in nm. Default c(200, 900).

n_channels

Integer. Number of spectral channels. Default 2048.

n_shots

Integer. Number of replicate shots. Default 10.

noise_level

Numeric. Relative noise level (fraction of max intensity). Default 0.02.

continuum_level

Numeric. Continuum background level. Default 100.

shot_rsd

Numeric. Relative standard deviation of shot-to-shot intensity variation (0-1). Default 0.1.

fwhm_nm

Numeric. Typical line FWHM in nm. Default 0.1.

seed

Optional integer random seed for reproducibility.

line_data

Optional explicit teaching peak table, with component, wavelength_nm and relative_strength columns. Strengths are arbitrary intensity multipliers, not atomic transition probabilities. Use ls_demo_peaks() for a fictional fixture. NULL uses the configured atomic reference and retains the transition-rate scaling convention.

Value

A ls_spectrum() object.

Examples

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(),
                             elements = c(Ca = 1000, Na = 500), seed = 1)
print(spec)
spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(),
                             elements = c(Ca = 1000, Na = 500), seed = 1)
print(spec)

Smooth Spectra

Description

Applies spectral smoothing. Methods: "savgol" (Savitzky-Golay, needs the signal package when available, otherwise uses local polynomial least squares with an intercept), "moving_avg", "gaussian", "median".

Usage

ls_smooth(x, method = "savgol", window = 11, poly_order = 3)
ls_smooth(x, method = "savgol", window = 11, poly_order = 3)

Arguments

x

A ls_spectrum() or ls_dataset() object.

method

Character. Default "savgol".

window

Integer. Window size in channels (odd preferred). Default 11.

poly_order

Integer. Polynomial order for Savitzky-Golay. Default 3.

Value

Same class as input with smoothed intensities.

Examples

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(), seed = 1)
sm <- ls_smooth(spec, method = "moving_avg", window = 9)

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(), seed = 1)
sm <- ls_smooth(spec, method = "moving_avg", window = 9)

Create a LIBS Spectrum Object

Description

Constructs a libs_spectrum S3 object from a wavelength axis and one or more intensity vectors (shots).

Usage

ls_spectrum(wavelength, intensity, metadata = list(), baseline = NULL)
ls_spectrum(wavelength, intensity, metadata = list(), baseline = NULL)

Arguments

wavelength

Numeric vector of wavelengths in nanometers. Must be contain at least two distinct, finite, positive values. Channels and the optional baseline are sorted together in ascending wavelength order.

intensity

Numeric vector (single shot) or matrix (multiple shots) of emission intensities. If matrix: rows = shots, columns = wavelength channels, ncol(intensity) must equal length(wavelength). All values must be finite; signed baseline-corrected intensities are supported.

metadata

Named list of metadata (e.g. sample_id, material, gate_delay_us, integration_time_us, laser_energy_mj, atmosphere, date). Default list().

baseline

Optional numeric vector of baseline values corresponding to each wavelength channel (populated by ls_baseline()). Default NULL.

Value

An object of class libs_spectrum: a named list with elements wavelength, intensity (matrix: shots x channels), metadata, baseline, n_shots, n_channels, and range_nm.

Examples

wl <- seq(200, 900, length.out = 512)
int <- exp(-((wl - 393)^2) / 2) + stats::rnorm(512, 0, 0.05)
spec <- ls_spectrum(wl, int, metadata = list(sample_id = "demo"))
print(spec)

wl <- seq(200, 900, length.out = 512)
int <- exp(-((wl - 393)^2) / 2) + stats::rnorm(512, 0, 0.05)
spec <- ls_spectrum(wl, int, metadata = list(sample_id = "demo"))
print(spec)

Tissue Classification via LIBS

Description

Classifies tissue types based on LIBS spectral signatures. Implements three methods:

"ratio" — an unvalidated heuristic based on integrated Ca, Na, K, Fe and C line responses. It is a research illustration, not a calibrated tissue probability or clinical classifier. All required lines must be covered and the response denominators must be positive; otherwise the row is unclassified.
"plsda" — fits a PLS-DA model on the supplied labelled reference dataset and predicts class for x.
"svm" — fits an SVM on reference and predicts class.

Usage

ls_tissue_classify(
  x,
  method = "ratio",
  reference = NULL,
  group_col = "tissue",
  verbose = TRUE
)
ls_tissue_classify(
  x,
  method = "ratio",
  reference = NULL,
  group_col = "tissue",
  verbose = TRUE
)

Arguments

x

A ls_spectrum() or ls_dataset() object (unknowns).

method

Character. "ratio", "plsda", or "svm". Default "ratio".

reference

A ls_dataset() with labelled tissue reference spectra (needed for "plsda" and "svm").

group_col

Character. Column in reference sample_info with tissue labels. Default "tissue".

verbose

Logical. Default TRUE.

Value

A tibble::tibble() with columns sample_id, predicted_tissue, score, score_type, and status. The former confidence column is replaced to distinguish relative heuristic scores, unrestricted PLS-DA scores and SVM probability estimates. None establishes tissue validity.

Examples

ds <- ls_example_data("tissue")
ls_tissue_classify(ds[1:3])
ds <- ls_example_data("tissue")
ls_tissue_classify(ds[1:3])

Tissue Discrimination Analysis

Description

Identifies emission channels that best discriminate two tissue types via per-wavelength t-tests with Benjamini-Hochberg FDR correction, plus log2 fold-change. Element annotations are nearest wavelength candidates, not established elemental identities; candidate_count records ambiguity. Without a configured atomic reference, statistical results remain available, with missing element annotations and annotation_status = "reference_unavailable". A configured but unreadable or invalid reference raises an error.

Usage

ls_tissue_discriminate(
  dataset,
  group_col,
  group_a,
  group_b,
  method = "t_test",
  tolerance_nm = 0.2
)
ls_tissue_discriminate(
  dataset,
  group_col,
  group_a,
  group_b,
  method = "t_test",
  tolerance_nm = 0.2
)

Arguments

dataset

A ls_dataset() with tissue labels.

group_col

Character. Label column in sample_info.

group_a

Character. First group label.

group_b

Character. Second group label.

method

Character. "t_test" (default) or "fold_change".

tolerance_nm

Numeric. Matching tolerance when annotating elements. Default 0.2.

Value

A tibble::tibble() with columns wavelength_nm, mean_a, mean_b, p_value, fold_change, fdr, significant, element.

Examples

ds <- ls_example_data("tissue")
head(ls_tissue_discriminate(ds, "tissue", "bone", "muscle"))
ds <- ls_example_data("tissue")
head(ls_tissue_discriminate(ds, "tissue", "bone", "muscle"))

Train a Classifier

Description

Trains SVM or Random Forest on spectral intensity features.

Usage

ls_train_classifier(dataset, grouping, method = "svm", ...)
ls_train_classifier(dataset, grouping, method = "svm", ...)

Arguments

dataset

A ls_dataset() with labelled samples.

grouping

Character. Column name in sample_info with class labels.

method

Character. "svm" or "rf". Default "svm".

...

Additional arguments passed to the underlying trainer.

Value

An S3 object of class libs_classifier, including predictions, a square confusion matrix, the fitted feature_contract, and evaluation_basis: "training" for SVM or "out_of_bag" for RF. SVM evaluation uses the same probability-based class decision as ls_classify(). Training accuracy is not an independent estimate of generalization.

Examples

ds <- ls_example_data("tissue")

if (requireNamespace("e1071", quietly = TRUE)) {
  clf <- ls_train_classifier(ds, "tissue", method = "svm")
  clf$accuracy
}

ds <- ls_example_data("tissue")

if (requireNamespace("e1071", quietly = TRUE)) {
  clf <- ls_train_classifier(ds, "tissue", method = "svm")
  clf$accuracy
}

Write a LIBS Spectrum to CSV

Description

Exports a libs_spectrum to a CSV file with a wavelength_nm column plus one column per shot.

Usage

ls_write_csv(x, path, include_metadata = TRUE)
ls_write_csv(x, path, include_metadata = TRUE)

Arguments

x

A libs_spectrum object.

path

Character. Output file path.

include_metadata

Logical. Prepend metadata as comment lines. Default TRUE. JSON comments preserve scalar values, arrays, nested lists, and the baseline. Named vectors become JSON objects (lists on import); R-specific classes and attributes are not preserved.

Value

Invisibly returns path.

Examples

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(), seed = 1)
tmp <- tempfile(fileext = ".csv")
ls_write_csv(spec, tmp)
unlink(tmp)

spec <- ls_simulate_spectrum(line_data = ls_demo_peaks(), seed = 1)
tmp <- tempfile(fileext = ".csv")
ls_write_csv(spec, tmp)
unlink(tmp)

Plot a LIBS Dataset

Description

Plot a LIBS Dataset

Usage

## S3 method for class 'libs_dataset'
plot(x, y, ...)
## S3 method for class 'libs_dataset'
plot(x, y, ...)

Arguments

x

A libs_dataset object.

y

Ignored.

...

Passed to ls_plot_overlay().

Value

A ggplot object.

Plot a LIBS Spectrum

Description

Plot a LIBS Spectrum

Usage

## S3 method for class 'libs_spectrum'
plot(x, y, ...)
## S3 method for class 'libs_spectrum'
plot(x, y, ...)

Arguments

x

A libs_spectrum object.

y

Ignored.

...

Arguments passed to ls_plot_spectrum().

Value

A ggplot object.

Predict Concentrations from a Calibration Model

Description

Predict Concentrations from a Calibration Model

Usage

## S3 method for class 'libs_calibration'
predict(object, newdata, ...)
## S3 method for class 'libs_calibration'
predict(object, newdata, ...)

Arguments

object

A libs_calibration object.

newdata

A finite numeric vector of already integrated responses (ratios for internal-standard calibration), or a matrix with the fitted PLS columns in their trained order. This low-level method cannot verify wavelength labels; use ls_quantify() for spectra with checked feature identity.

...

Unused.

Value

Numeric vector of predicted concentrations.

Print a LIBS Calibration

Description

Print a LIBS Calibration

Usage

## S3 method for class 'libs_calibration'
print(x, ...)
## S3 method for class 'libs_calibration'
print(x, ...)

Arguments

x

A libs_calibration object.

...

Unused.

Value

Invisibly x.

Print a LIBS Dataset

Description

Print a LIBS Dataset

Usage

## S3 method for class 'libs_dataset'
print(x, ...)
## S3 method for class 'libs_dataset'
print(x, ...)

Arguments

x

A libs_dataset object.

...

Unused.

Value

Invisibly returns x.

Print a LIBS Spectrum

Description

Print a LIBS Spectrum

Usage

## S3 method for class 'libs_spectrum'
print(x, ...)
## S3 method for class 'libs_spectrum'
print(x, ...)

Arguments

x

A libs_spectrum object.

...

Unused.

Value

Invisibly returns x.

Wavelength-Based Color Scale

Description

Maps wavelength values (nm) to approximate visible-light colors for intuitive spectral visualization. Below 380 nm: UV purple; above 780 nm: deep red.

Usage

scale_color_wavelength(name = "Wavelength (nm)", ...)
scale_color_wavelength(name = "Wavelength (nm)", ...)

Arguments

name

Character. Legend title. Default "Wavelength (nm)".

...

Additional arguments passed to ggplot2::scale_color_gradientn().

Value

A ggplot2 scale object.

Examples

df <- data.frame(x = 1:5, y = 1:5, wl = c(200, 400, 550, 700, 850))
ggplot2::ggplot(df, ggplot2::aes(x, y, colour = wl)) +
  ggplot2::geom_point(size = 5) + scale_color_wavelength()
df <- data.frame(x = 1:5, y = 1:5, wl = c(200, 400, 550, 700, 850))
ggplot2::ggplot(df, ggplot2::aes(x, y, colour = wl)) +
  ggplot2::geom_point(size = 5) + scale_color_wavelength()

Summary of a LIBS Calibration

Description

Summary of a LIBS Calibration

Usage

## S3 method for class 'libs_calibration'
summary(object, ...)
## S3 method for class 'libs_calibration'
summary(object, ...)

Arguments

object

A libs_calibration object.

...

Unused.

Value

Invisibly a list of key figures of merit.

Summary of a LIBS Dataset

Description

Summary of a LIBS Dataset

Usage

## S3 method for class 'libs_dataset'
summary(object, ...)
## S3 method for class 'libs_dataset'
summary(object, ...)

Arguments

object

A libs_dataset object.

...

Unused.

Value

Invisibly returns a list with per-group statistics.

Summary of a LIBS Spectrum

Description

Summary of a LIBS Spectrum

Usage

## S3 method for class 'libs_spectrum'
summary(object, n_peaks = 5, ...)
## S3 method for class 'libs_spectrum'
summary(object, n_peaks = 5, ...)

Arguments

object

A libs_spectrum object.

n_peaks

Integer. Number of top peaks to show. Default 5.

...

Unused.

Value

Invisibly returns a list with summary statistics.

LIBS Theme for ggplot2

Description

A clean publication-ready theme for LIBS spectral plots.

Usage

theme_libs(base_size = 12)
theme_libs(base_size = 12)

Arguments

base_size

Numeric. Base font size in points. Default 12.

Value

A ggplot2 theme object.

Examples

ggplot2::ggplot(mtcars, ggplot2::aes(mpg, wt)) +
  ggplot2::geom_point() + theme_libs()
ggplot2::ggplot(mtcars, ggplot2::aes(mpg, wt)) +
  ggplot2::geom_point() + theme_libs()

Package 'libscanR'

Help Index

Subset a LIBS Dataset

Description

Usage

Arguments

Value

Subset a LIBS Spectrum by Wavelength

Description

Usage

Arguments

Value

Combine Multiple LIBS Spectra by Stacking Shots

Description

Usage

Arguments

Value

Dimensions of a Dataset (spectra x channels)

Description

Usage

Arguments

Value

Dimensions of a LIBS Spectrum (shots x channels)

Description

Usage

Arguments

Value

Test whether an object is a libs_calibration

Description

Usage

Arguments

Value

Test whether an object is a libs_dataset

Description

Usage

Arguments

Value

Test whether an object is a libs_spectrum

Description

Usage

Arguments

Value

Number of Spectra in a Dataset

Description

Usage

Arguments

Value

Number of Channels in a LIBS Spectrum

Description

Usage

Arguments

Value

Average Replicate Shots

Description

Usage

Arguments

Value

Examples

Baseline Correction

Description

Usage

Arguments

Value

Examples

Build a Spatial Elemental Map

Description

Usage

Arguments

Value

Examples

Build Calibration Curve

Description

Usage

Arguments

Details

Value

Examples

Create a LIBS Calibration Model

Description

Usage