rxs2tria

The goal of rxs2tria is to prepare, quality check and harmonize a collection of QWA images, corresponding ROXAS output files and accompanying metadata for submission as a dataset to the QWA database.

Note: This package is currently under development and not yet stable nor officially released. Version 1.0.0 will be made available under the organizational github.

Installation

You can install the development version of rxs2tria from GitHub with:

# install.packages("pak")
pak::pak("tria-db/rxs2tria")

# or:
remotes::install_github('tria-db/rxs2tria')

Usage

Use the package to extract all available metadata from the ROXAS output files under a given path.

library(rxs2tria)

# set input and output paths and dataset name
path_in <- './path/to/ROXAS_output_data'
path_out <- './out'
dataset_name <- 'QWA_ex_dataset'

# get overview of data to be read and extract data structure from filenames
files <- get_roxas_files(path_in)
str(files)

# define the naming convention used for your files as a regex pattern
# e.g. the following means that your filenames are structured as follows:
# `{site}_{species}_{tree}_{slide}_{image}`
# since there is exactly one sample (woodpiece) per tree, no woodpiece identifier is included in the pattern
pattern <- "^(?<site>[:alnum:]+)_(?<species>[:alnum:]+)_(?<tree>[:alnum:].+)_(?<slide>[:alnum:]+)_(?<image>[:alnum:]+)$"

# extract the data structure from the file names using the specified pattern
df_structure <- extract_data_structure(files, pattern)

# read available metadata
df_meta <- collect_metadata_from_files(df_structure,
                                       roxas_version='classic')

# write extracted metadata to file
write.csv(df_meta, file.path(path_out, paste0(dataset_name, '_meta_extracted.csv')),
          row.names = FALSE)

To provide the additional metadata required (e.g. site and tree information, author details, …) for the submission of a dataset, you can use the interactive Shiny app which is part of this package.

The app is hosted on a WSL server and accessible via webapps.wsl.ch/QWAmeta.

Or, if you want to run it locally, you can use the following command:

launch_metadata_app()

To load and preprocess the cell and ring measurements:

# read raw cells/rings data
QWA_data <- collect_raw_data(df_structure)

# clean raw data
QWA_data <- validate_QWA_data(QWA_data, df_meta)

# remove outliers
# NOTE: ROXAS does some threshold based outlier checks and assigns these a
# a negative value. Here, we replace these outliers with NAs
QWA_data <- remove_outliers(QWA_data)

# complete cell measures
QWA_data <- complete_cell_measures(QWA_data)

# save the cleaned data to files
write.csv(QWA_data$cells,
          file.path(path_out, paste0(dataset_name, '_cells.csv')),
          row.names = FALSE)

write.csv(QWA_data$rings,
          file.path(path_out, paste0(dataset_name, '_rings.csv')),
          row.names = FALSE)

To check the coverage and add ring quality flags and setting years to exclude:

launch_coverage_app()

# TODO: add export functionality to coverage app