Deborah.DeborahThreads.DeborahThreadsRunner

struct JobArgsDeborah

Container struct for a single Deborah.DeborahCore job configuration.

Fields

location::String: Base path for input/output.
ensemble::String: Ensemble name (e.g., "L8T4b1.60k13570").
analysis_header::String: Header string for organizing analysis subdirectories.
X::Vector{String}: Input feature names.
Y::String: Target observable name.
model::String: ML model name (e.g., "LightGBM").
read_column_X::Vector{Int}: Vector of $1$-based column indices specifying which column to read from each input feature file in X.
read_column_Y::Int: $1$-based column index specifying which column to read from the target observable file Y.
index_column::Int : $1$-based column index to read the configuration index from files (typically 1 if configuration index is in the first column).
label::String: LBP label identifier (e.g., "75").
train::String: TRP train identifier (e.g., "40").
ranseed::Int: Random seed for bootstrap.
N_bs::Int: Number of bootstrap samples.
blk_size::Int: Bootstrap block size.
bin_size::Int: Jackknife bin size.
IDX_shift::Int: Index shift.
dump_X::Bool: Whether to dump the X matrix.
overall_name::String: Filename prefix used throughout output files.
abbreviation::Dict{String, String}: Dictionary mapping observables to abbreviations.
use_abbreviation::Bool: Whether to encode names using abbreviations.
output_base::String: Directory path for outputs (not including filename prefix).

generate_toml_dict(
    args::JobArgsDeborah
) -> Dict

Generate a TOML-compatible dictionary based on a DeborahThreadsRunner.JobArgsDeborah struct. This dictionary can be serialized to disk and used as a configuration file for Deborah.DeborahCore or Deborah.Esther.

Arguments

args::JobArgsDeborah: Struct containing all parameters for a single LBP/TRP job.

Returns

Dict: A TOML-structured dictionary, with properly nested data, bootstrap, jackknife, and abbreviation sections.

parse_config_DeborahThreads(
    toml_path::String, 
    jobid::Union{Nothing, String}=nothing
) -> Vector{JobArgsDeborah}

Parse a TOML config file and generate one DeborahThreadsRunner.JobArgsDeborah struct per label/train pair. This function is used in Deborah.DeborahCore parallel execution to schedule multiple jobs.

Arguments

toml_path::String: Path to a TOML configuration file defining multiple LBP/TRP jobs.
jobid::Union{Nothing, String} : Optional job ID string used for logging.

Returns

Vector{JobArgsDeborah}: A list of job argument structs, each ready to be passed into a Deborah.DeborahCore run.

run_DeborahThreads(
    toml_path::String
) -> Nothing

Parse TOML config and run all Deborah.DeborahCore jobs in parallel, dispatching in batches according to the number of threads.

Arguments

toml_path::String: Path to the TOML configuration file containing job lists.

Side Effects

run_one_job(
    args::JobArgsDeborah, 
    output_dir::String, 
    log_path::String
) -> Nothing

Run a single Deborah.DeborahCore job based on the given arguments. Generates a TOML config, saves it, and runs Deborah.DeborahCore.DeborahRunner.run_Deborah() while logging start/end timestamps and errors.

Arguments