Package 'PhysioPipe'

Stack per-branch data frames from a cohort run into one table

Description

Stack per-branch data frames from a cohort run into one table

Usage

pp_bind_cohort(branches, id = NULL)

Arguments

branches	A list of per-record data frames (dynamic-branch output).
id	Optional vector of record ids, added as a record column.

Value

A single row-bound data frame.

Examples

branches <- list(
  data.frame(sdnn = 50),
  data.frame(sdnn = 61)
)
pp_bind_cohort(branches, id = c("rec100", "rec101"))

---

File paths of the bundled demo ECG record (MIT-BIH 100 excerpt)

Description

File paths of the bundled demo ECG record (MIT-BIH 100 excerpt)

Usage

pp_ecg_demo_files()

Value

Character vector of the record's companion files (.dat, .hea).

---

Read a WFDB record (by companion files or base) into a PhysioExperiment

Description

Read a WFDB record (by companion files or base) into a PhysioExperiment

Usage

pp_ecg_read(record)

Arguments

record

Either the companion files (c(".dat", ".hea")) or a single record base path (no extension).

Value

A PhysioExperiment.

---

Join time- and frequency-domain HRV into one row per channel

Description

Join time- and frequency-domain HRV into one row per channel

Usage

pp_hrv_combine(hrv_time, hrv_freq)

Arguments

hrv_time	Output of PhysioECG::ecgHRVtime().
hrv_freq	Output of PhysioECG::ecgHRVfreq().

Value

A data frame merged on channel.

Examples

hrv_time <- data.frame(channel = "MLII", sdnn = 50, rmssd = 42)
hrv_freq <- data.frame(channel = "MLII", lf = 400, hf = 330)
pp_hrv_combine(hrv_time, hrv_freq)

---

Write a simple RR tachogram PNG (headless) and return its path

Description

Write a simple RR tachogram PNG (headless) and return its path

Usage

pp_plot_tachogram(rr, path = "figures/tachogram.png")

Arguments

rr	Output of PhysioECG::ecgRRintervals().
path	Output PNG path.

Value

path (so the target can be format = "file").

---

Simulate a synthetic PhysioExperiment for demos and tests

Description

Deterministic (seeded) synthetic signals per modality. Clearly synthetic — for exercising a pipeline's mechanics, not for making claims.

Usage

pp_simulate(
  modality = c("eeg", "emg", "eda"),
  seed = 1L,
  n_channels = NULL,
  duration = NULL,
  sampling_rate = NULL
)

Arguments

modality	One of "eeg", "emg", "eda".
seed	Integer RNG seed (fixed so pipelines stay reproducible).
n_channels, duration, sampling_rate	Optional geometry overrides.

Value

A PhysioExperiment (time x channels) with a raw assay.

Examples

if (requireNamespace("PhysioCore", quietly = TRUE)) {
  pe <- pp_simulate("eeg")
}

---

Companion files of a WFDB record base

Description

Companion files of a WFDB record base

Usage

pp_wfdb_files(base)

Arguments

base	Record base path (no extension), e.g. "data/100".

Value

Character vector of existing base.* signal/header files.

Examples

# Create a stand-in record base and list its WFDB companion files.
base <- file.path(tempdir(), "100")
file.create(paste0(base, c(".dat", ".hea", ".atr")))
basename(pp_wfdb_files(base))

---

Write a data frame to Parquet (language-neutral interchange)

Description

Intermediate/terminal tables written as Parquet can be read by Python (pandas/pyarrow), DuckDB, and Nextflow processes — unlike the default .rds target store. This is the seam that keeps pipelines orchestrator-agnostic.

Usage

pp_write_parquet(df, path)

Arguments

df	A data frame.
path	Output .parquet path.

Value

path (so the target can be format = "file").

Examples

if (requireNamespace("arrow", quietly = TRUE)) {
  path <- pp_write_parquet(data.frame(a = 1:3, b = 4:6),
                           file.path(tempdir(), "demo.parquet"))
  file.exists(path)
}

---

ECG heart-rate-variability pipeline (target factory)

Description

Builds a standard resting-ECG pipeline as a list of targets: ⁠read -> lead -> R-peaks -> RR -> {HRV time, HRV freq} -> summary⁠, plus an RR-tachogram figure. With more than one record it fans out via dynamic branching (one branch per record) and adds a combined cohort table.

Usage

tar_ecg_hrv(
  name = "ecg",
  records = NULL,
  channel = 1L,
  detector = "pan_tompkins"
)

Arguments

name	Target-name prefix (a string; targets are "<name>_pe" etc.).
records	NULL to use the bundled demo record (MIT-BIH 100 excerpt), or a character vector of WFDB record bases (paths without extension).
channel	Lead index to analyse (MIT-BIH 100: 1 = MLII, 2 = V5).
detector	ecgDetectRpeaks() method (e.g. "pan_tompkins", "hamilton").

Value

A list of target objects to splice into a ⁠_targets.R⁠ pipeline.

See Also

tar_eeg_bandpower(), tar_emg_features(), tar_eda_scr()

Examples

# The factory returns target objects to splice into a `_targets.R` list();
# building them does not run the pipeline.
targs <- tar_ecg_hrv()          # bundled demo record (single-record pipeline)
length(targs)
class(targs[[1]])
# Cohort fan-out (dynamic branching) when given several record bases:
length(tar_ecg_hrv(records = c("data/100", "data/101")))

---

EDA skin-conductance pipeline (target factory)

Description

⁠signal -> tonic/phasic decomposition -> SCR features⁠.

Usage

tar_eda_scr(name = "eda", source = NULL)

Arguments

name	Target-name prefix.
source	Expression producing a PhysioExperiment, or NULL for a synthetic EDA demo (drifting SCL + discrete SCRs).

Value

A list of targets.

Examples

# Returns target objects (synthetic EDA demo: drifting SCL + discrete SCRs).
targs <- tar_eda_scr()
length(targs)

---

EEG band-power pipeline (target factory)

Description

⁠signal -> band power⁠ per channel. Band power is computed from the Welch PSD integrated over each band, so it is inherently frequency-selective (no pre-filter needed).

Usage

tar_eeg_bandpower(
  name = "eeg",
  source = NULL,
  bands = list(alpha = c(8, 13), beta = c(13, 30)),
  method = "welch",
  relative = FALSE
)

Arguments

name	Target-name prefix.
source	Expression producing a PhysioExperiment, or NULL for a synthetic EEG demo (10 Hz alpha + noise).
bands	Named list of c(low, high) Hz bands.
method	Welch or wavelet PSD.
relative	Report band power relative to total.

Value

A list of targets.

Examples

# Returns target objects (synthetic EEG demo); splice into a `_targets.R`.
targs <- tar_eeg_bandpower()
length(targs)
# Point it at a real file instead of the synthetic demo:
# tar_eeg_bandpower(source = quote(PhysioIO::readEDF("sub01.edf")))

---

EMG amplitude / fatigue pipeline (target factory)

Description

⁠signal -> linear envelope⁠ and ⁠signal -> spectral fatigue⁠ (median/mean frequency + RMS over windows).

Usage

tar_emg_features(name = "emg", source = NULL)

Arguments

name	Target-name prefix.
source	Expression producing a PhysioExperiment, or NULL for a synthetic bursty EMG demo.

Value

A list of targets.

Examples

# Returns target objects (synthetic bursty-EMG demo).
targs <- tar_emg_features()
length(targs)

---

Quarto report pipeline (target factory)

Description

Wraps tarchetypes::tar_quarto_raw() so a rendered report becomes a tracked target that re-renders when any embedded result changes. The .qmd reads results with targets::tar_read(); targets discovers those dependencies automatically.

Usage

tar_physio_report(name = "report", path = NULL)

Arguments

name	Target name (string).
path	Path to the .qmd. Defaults to the bundled template.

Value

A single target object (requires the tarchetypes and quarto packages plus the Quarto CLI at build time).

Examples

# list(tar_ecg_hrv(), tar_physio_report())

---