File IO API

back_one_dir(dir)

Given a directory as input, outputs the parent directory.

get_filename(path_to_file)

Given path to a file, outputs the filename.

function read_head_pos(head_path::String)

Reads the worm head position from the file head_path::String. Returns a dictionary mapping frame => head position of the worm at that frame.

load_registration_problems(edge_file_paths::Array{String,1})

Loads a set of registration problems into an array.

Arguments:

• edge_file_paths::Array{String, 1}: paths containing registration problems

function modify_parameter_file(param_in::String, param_out::String, substitutions::Dict; is_universal=false)

Modifies an elastix transform parameter file.

Arguments

• param_in::String: Path to parameter file to modify
• param_out::String: Path to modified parameter file output
• substitutions::Dict: Dictionary of substitutions. For each key in substitutions, if it is a key in the parameter file, replace its value with the value in substitutions.
• is_universal::Bool (optional): If set to true, instead find/replace all instances of keys in substitutions regardless of if it is a key in the parameter file

read_parameter_file(parameter_file_path::String, key::String, dtype::Type)

Reads a value from a paremeter file.

Arguments

• parameter_fiile_path::String: Path to parameter file
• key::String: Value to read
• dtype::Type: Type of the value or elements of value (if it's an array)

read_nrrd(rootpath, img_prefix, nrrd_path, frame, channel)

Reads NRRD file from $(rootpath)/$(nrrd_path)/$(img_prefix)_t$(channel).nrrd and outputs resulting image.

read_activity(input)

Reads ROI activities from input.

write_activity(activity, out)

Writes list of ROI activities activity to output file out.

write_centroids(centroids, out)

Writes centroids to transformix-compatible file out

read_centroids_transformix(input)

Reads centroids from transformix output file input

read_centroids_roi(input)

Reads centroids from write_centroids output file input

load_training_set(h5_file::String)

Loads training (or validation dataset) in 3D dataset h5_file::String. Returns raw, label, and weight fields.

load_predictions(h5_file::String; threshold=nothing)

Loads UNet predictions in 3D dataset h5_file::String. If the file has a fourth dimension, assumes predictions are stored in its second entry. Returns the portion of predictions field corresponding to foreground. Can optionally set threshold to binarize predictions.

resample_img(img, scales; dtype="raw")

Scales an image down by using linear interpolation for the raw image, and bkg-gap priority interpoalation for the labels.

Arguments

• img: image to scale
• scales: array of factors to scale down (bin) by in each dimension. Must be positive.

Optional keyword arguments

• dtype::String: type of data - either "raw", "label", or "weight". Default raw.