Compute Entry Wise Perturbation Expectation

Description

This function computes the expected number of sign switches from perturbing each entry individually. Run after PreprocessMatrix().

Usage

ComputeEntryWisePerturbationExpectation(
  input_folder = NULL,
  PreProsMatrix = NULL,
  prefix = NULL,
  distribution_type = "truncnorm",
  input_a = 0,
  input_b = -2,
  threads = 1
)

Arguments

Value

If an input folder is specified the objects will be saved to that folder. If the PreProsMatrix object is specified, an R list object with the following: original_matrix, matrix_size, column_names, row_names, non_zero, num_switch_functions, asymptotic_stability_start, asymptotic_stability_end, num_switch_funcs_r, distributions, expected_num_switch, distributions_object

Examples

## Not run: 
# Set input file
infile <- system.file("extdata", "Modules", "IGP.csv", 
    package = "PressPurt")
# Preprocess the matrix
PreProsMatrix <- PreprocessMatrix(input_file = infile, 
    output_folder = NULL, max_bound = 10, threads = 2)
# Run ComputeEntryWisePerturbationExpectation
Entrywise <- ComputeEntryWisePerturbationExpectation(PreProsMatrix = PreProsMatrix,
    distribution_type = "truncnorm", 
    input_a = 0, input_b = -2, threads = 1)

## End(Not run)

Compute Multi Entry Perturbation Expectation

Description

This function takes a jacobian matrix and computes the multi-entry perturbation expectation.

Usage

ComputeMultiEntryPerturbationExpectation(
  input_file,
  num_iterates = 1000,
  interval_length = 0.01,
  threads = 1
)

Arguments

Value

returns a scalar

Examples

## Not run: 
infile <- system.file("extdata", "Modules", "IGP.csv", 
    package = "PressPurt") 
ComputeMultiEntryPerturbationExpectation(input_file = infile)

## End(Not run)

Generate Entry Wise Figures

Description

This function plots the number of mis-predictions versus perturbation value, overlaid with distribution over stable perturbation values. Run after ComputeEntryWisePerturbationExpectation()

Usage

GenerateEntryWiseFigures(
  input_folder = NULL,
  EntryWise = NULL,
  prefix = NULL,
  all_numswitch_plots = FALSE,
  list_of_numswitch_to_plot = NULL
)

Arguments

Value

plot or plots

Examples

## Not run: 
# Set input file
infile <- system.file("extdata", "Modules", "IGP.csv", 
    package = "PressPurt")
# Preprocess the matrix
PreProsMatrix <- PreprocessMatrix(input_file = infile, 
    output_folder = NULL, max_bound = 10, threads = 2)

# Run ComputeEntryWisePerturbationExpectation
Entrywise <- ComputeEntryWisePerturbationExpectation(PreProsMatrix = PreProsMatrix,
    distribution_type = "truncnorm", 
    input_a = 0, input_b = -2, threads = 1)

# Plot specific entries using entrywise object
list_of_numswitch_to_plot <- list(c(1, 1), c(1, 2))
GenerateEntryWiseFigures(EntryWise=Entrywise, 
    all_numswitch_plots = FALSE, 
    list_of_numswitch_to_plot=list_of_numswitch_to_plot)
     

# Plot specific entries from folder
GenerateEntryWiseFigures(input_folder = "test_r/test3", 
    all_numswitch_plots = FALSE, 
    list_of_numswitch_to_plot=list_of_numswitch_to_plot)

# Plot all numswitch plots
GenerateEntryWiseFigures(EntryWise=Entrywise, 
    all_numswitch_plots = TRUE)

## End(Not run)

Preprocess Matrix

Description

This script pre-processes a matrix by figuring out what the intervals of asymptotic stability are, as well as finding which perturbation values lead to a sign switch.

Usage

PreprocessMatrix(
  input_file,
  output_folder = NULL,
  prefix = NULL,
  max_bound = 10,
  zero_perturb = FALSE,
  threads = 1,
  verbose = FALSE
)

Arguments

Value

A list of with the following objects: matrix_size, column_names, row_names, non_zero, num_switch_functions, asymptotic_stability_start, asymptotic_stability_end, num_switch_funcs_r

Examples

## Not run: 
infile <- system.file("extdata", "Modules", "IGP.csv", 
    package = "PressPurt") 
PreProsMatrix <- PreprocessMatrix(input_file = infile, 
    output_folder = NULL, max_bound = 10, threads = 2)

## End(Not run)

Make a new conda environment

Description

This function creates a new conda environment and initializes the new conda environment. In doing so, this function sets your python version and one may specify a specific python version. This is useful if you have multiple versions of python installed. When making a new conda environment, if the python version isn't set, then your default one will be used.

Usage

create_conda_env(condaenv, version = NULL, verbose = TRUE)

Arguments

Value

None

Examples

## Not run: 
create_conda_env( 
    condaenv = "r-reticulate",
    version = "~/anaconda3/bin/python",
    verbose = TRUE)

## End(Not run)

Make a new virtual environment

Description

This function creates a new virtual environment and initializes the new virtual environment. In doing so, this function sets your python version and one may specify a specific python version. This is useful if you have multiple versions of python installed. When making a new virtual environment, if the python version isn't set, then your default one will be used.

Usage

create_virtual_env(virtualenv, version = NULL, verbose = TRUE)

Arguments

Value

None

Examples

## Not run: 
create_virtual_env(version = "/usr/bin/python3", 
    virtualenv = "r-reticulate", 
    verbose = TRUE)

## End(Not run)

Find Python versions, Conda, & Virtual Environments

Description

This function lists available python versions, conda environments, and virtual environments. One may show all three or just one.

Usage

find_python(python = TRUE, conda = TRUE, virtualenv = TRUE)

Arguments

Value

None

Examples

## Not run: 
find_python()

## End(Not run)

Get PDF distribution

Description

This function retrieves the PDF (Probability Distribution Function) object from the scipy method <scipy.stats._distn_infrastructure.rv_frozen>.

Usage

get_distributions_single(
  matrix_entry,
  distribution_list,
  asymp_stab,
  points = 250
)

Arguments

Value

Probability Distribution Function from scipy

Examples

## Not run: 
k <- 1
l <- 1
np <- reticulate::import("numpy")
distributions <- reticulate::py_load_object("distributions.pkl")
single_dist <- get_distributions_single(matrix_entry = c(k,l), 
    distribution_list = distributions, 
    asymp_stab = c(combined$asymptotic_stability_start[k,l], 
    combined$asymptotic_stability_end[k,l]))

## End(Not run)

Num Switch Function to step function

Description

This function transforms a Num Switch Function to a plot ready step function with x and y values. Returns a data frame of x and y values to plot.

Usage

ns_to_step(asymp_stab_start, asymp_stab_end, num_switch_func)

Arguments

Value

plot ready x and y values from the Num Switch Function

Examples

## Not run: 
# Set input file
infile <- system.file("extdata", "Modules", "IGP.csv", 
    package = "PressPurt")
# Preprocess the matrix
PreProsMatrix <- PreprocessMatrix(input_file = infile, 
    output_folder = NULL, max_bound = 10, threads = 2)

# Run ComputeEntryWisePerturbationExpectation
Entrywise <- ComputeEntryWisePerturbationExpectation(
    PreProsMatrix = PreProsMatrix,
    distribution_type = "truncnorm", 
    input_a = 0, input_b = -2, threads = 1)

ns_step <- ns_to_step(
    asymp_stab_start = Entrywise$asymptotic_stability_start[1,1],
    asymp_stab_end = Entrywise$asymptotic_stability_end[1,1],
    num_switch_func = Entrywise$num_switch_funcs_r$`(1, 1)`)

## End(Not run)

Convert data to R format if saved to disk

Description

This function will convert objects saved to disk to R friendly objects, or the same output as ComputeEntryWisePerturbationExpectation. If you used the "save to disk" option or ran via python directly, run this function to read the data into R. Files read in: asymptotic_stability.npy, column_names.txt, distributions.pkl, expected_num_switch.csv, num_non_zero.npy, num_switch_funcs.pkl, row_names.txt and size.npy. Note how most of these objects are python based objects- numpy or pickle objects.

Usage

process_data(matrix, type = "csv", folder, prefix = NULL)

Arguments

Value

object formatted in the same way the output of ComputeEntryWisePerturbationExpectation

Examples

## Not run: 
infile <- system.file("extdata", "Modules", "IGP.csv", 
    package = "PressPurt")
data <- process_data(matrix = infile, 
    type = "csv", folder = "output")

## End(Not run)

Install Python Dependencies

Description

This function installs needed python libraries into the specified conda environment OR virtual environment. Should be the same as the one specified in set_python. Required python libraries: matplotlib, numpy, pandas, pathos, scipy and sympy On CentOS 7 pandas & scipy may need to be installed with pip install from the command line. Will get the error: /lib/libstdc++.so.6: version ‘CXXABI_1.3.9’ not found See vignette for more information.

Usage

py_depend(condaenv = NULL, virtualenv = NULL)

Arguments

Value

None

Examples

## Not run: 
# Cond env
py_depend(condaenv = "r-reticulate", 
    virtualenv = NULL)
# virtualenv:
py_depend(virtualenv = "r-reticulate", 
    condaenv = NULL)

## End(Not run)

Set Python Conda environment

Description

This function sets your conda environment. Run this command before PreprocessMatrix. Install python dependencies in the same conda environment that you set here. To make a new conda environment use the create_conda_env function.

Usage

set_python_conda(condaenv, verbose = TRUE)

Arguments

Value

None

Examples

## Not run: 
set_python_conda(
    condaenv = "r-reticulate", 
    verbose = TRUE)

## End(Not run)

Set your Python Virtual environment

Description

This function sets your virtual environment. Run this command before PreprocessMatrix. Install python dependencies in the same virtual environment that you set here. To make a new virtual environment use the create_virtual_env function.

Usage

set_python_virtual(virtualenv, verbose = TRUE)

Arguments

Value

None

Examples

## Not run: 
set_python_virtual(
    virtualenv = "r-reticulate", 
    verbose = TRUE)

## End(Not run)