prep_data: Prepare data assimilation with templated scripts
In the-Hull/ROAR: Reproducible and Organized Assimilation Routines

Description Usage Arguments Details Value

View source: R/prep_data.R

prep_data() cycles through a meta table (created programmatically or manually) and generates a convenient database-like folder structure and processing scripts based on templates for each dataset entry in the meta table.

prep_data(
  meta_table_path,
  data_processed_path,
  verbose = TRUE,
  overwrite = FALSE,
  oid = NULL,
  template = NULL,
  data_list = NULL
)

`meta_table_path`	character, txt (tab delimited) with some pre-defined headers, including from automated additions from previous runs of this function
`data_processed_path`	character, path to `data_processed` folder (i.e. where cleaned data should live)
`verbose`	logical, display informative messages and user prompts to proceed?
`overwrite`	should existing scripts be overwritten?
`oid`	numeric, length one, row in meta table
`template`	character, path to custom template
`data_list`	named list, where names correspond to whisker tokens in template, and values refer to columns in meta table (i.e. `meta$column[oid]`)

The meta table must contain one row per dataset that will be assimilated. It must have at least the following columns, which are used to generate database-like folder structures, processing scripts from templates, and to populate these with information from the meta table:

oid: original meta table entry id
param: the measured parameter stored in the file
id_ ..: identifier column(s) prefixed with id_.
file_raw: absolute or project/working directory-relative file path to data set

These will be used to generate a folder structure of the form:

param/id_[1]/id_[2]/.../id_[n-1]

and a processing script within that folder of the form:

param_id_[1]..id_[n].R

as well as intermediate files/outputs from the populated template script with similar naming conventions.

The meta table can include as much detail and information as is desired, and entries can be accessed through token-value combinations in data_list in a custom template. The internal template(s) all have access to fixed set of these token-value combinations (see further below), which are structured following mustache syntax through whisker::whisker.render().

Tokens that are always available and included in the package-internal templates include:

param: parameter corresponding to OID in meta table currently processed,
raw_file: path to raw file corresponding to OID in meta table currently processed,
save_dir: path to (new) directory where all outputs from OID in meta table will be saved,
base_file_name: object and file names in processing script (made from template), based on OID in meta table currently processed,
current_index: OID that is currently processed,
date_time: Sys.Date(),
excel_file: logical flag, is raw file MS Excel-based?,
text_file: logical flag, is raw file text based?
meta_table_path: path to meta table

To use a custom template with only internal tokens, supply an empty list: data_list = list().

If there are multiple values or columns in the data, that need to be represented in the meta table, split the column names by forward dashes (/). For example, if a wide table needs to be split into long format, but a unique observation is defined through multiple columns, (e.g. site, plot, tree), these values can be listed as site/plot/tree in a aptly named meta table column for later use in a template.

Nothing.

the-Hull/ROAR documentation built on Jan. 19, 2021, 4:14 p.m.