Deborah.MiriamThreads.MiriamThreadsRunner

struct JobArgsMiriam

Container struct holding all job parameters required for Deborah.Miriam parallel execution.

Fields

• location::String : Base directory containing data
• multi_ensemble::String : Group name for multiple ensembles
• ensembles::Vector{String} : List of ensemble subdirectories

Training Model Inputs (TrM1–TrM4)

• TrM1_X, TrM1_Y, TrM1_model : Inputs, target, model tag for $\text{Tr} \, M^{-1}$
• TrM2_X, TrM2_Y, TrM2_model : Inputs, target, model tag for $\text{Tr} \, M^{-2}$
• TrM3_X, TrM3_Y, TrM3_model : Inputs, target, model tag for $\text{Tr} \, M^{-3}$
• TrM4_X, TrM4_Y, TrM4_model : Inputs, target, model tag for $\text{Tr} \, M^{-4}$
• TrM{1-4}_read_column_X::Vector{Int} : $1$-based column index to read the numerical value from each $\text{Tr} \, M^{-n} \; (n=1,2,3,4)$ file.
• TrM{1-4}_read_column_Y::Int : $1$-based column index to read the numerical value from each $\text{Tr} \, M^{-n} \; (n=1,2,3,4)$ file.
• TrM{1-4}_index_column::Int : $1$-based column index to read the configuration index from each $\text{Tr} \, M^{-n} \; (n=1,2,3,4)$ file.

Job Sweep Parameters

• label::String : LBP identifier
• train::String : TRP identifier
• dump_original::Bool : Whether to dump original raw $\text{Tr} \, M^{-n} \; (n=1,2,3,4)$ files for reference

Input Meta

• ns::Int : Spatial lattice size
• nt::Int : Temporal lattice size
• nf::Int : Number of quark flavors $N_{\text{f}}$
• beta::Float64 : Gauge coupling $\beta$
• csw::Float64 : Clover coefficient $c_{\text{SW}}$
• kappa_list::Vector{Float64} : $\kappa$ values used for the multi-ensemble reweighting

Solver Parameters

• maxiter::Int : Maximum solver iterations
• eps::Float64 : Solver precision

Resampling Parameters

• bin_size::Int : Jackknife bin size
• ranseed::Int : Random seed for bootstrap
• N_bs::Int : Number of bootstrap resamples
• blk_size::Int : Bootstrap block size
• method::String : Block-bootstrap scheme (case-sensitive):
  • "nonoverlapping" — Nonoverlapping Block Bootstrap (NBB; resample disjoint blocks).
  • "moving" — Moving Block Bootstrap (MBB; resample sliding windows).
  • "circular" — Circular Block Bootstrap (CBB; sliding windows with wrap-around).

Trajectory Meta

• nkappaT::Int : Number of total $\kappa$ trajectories

Deborah.DeborahCore Metadata

• analysis_header::String : Prefix for analysis directories
• IDX_shift::Int : Index shift for additional analysis
• dump_X::Bool : Dump feature matrices?

Naming

• overall_name::String : Unique name for each job instance
• abbreviation::Dict{String,String} : Abbreviation map for input features
• use_abbreviation::Bool : Use abbreviations when naming outputs?
• output_base::String : Root directory for output data

generate_toml_dict(
    args::JobArgsMiriam
) -> Dict

Convert a JobArgsMiriam struct into a TOML-compatible nested OrderedCollections.OrderedDict.

The returned dictionary mirrors the structure of the config TOML file, grouped into sections such as data, input_meta, solver, etc.

Arguments

• args::JobArgsMiriam: The full job parameter struct to be serialized.

Returns

• Dict: A TOML-ready nested dictionary (OrderedCollections.OrderedDict) representing the job configuration.

TOML Sections

• data : Ensemble paths, model inputs/targets, flags
• input_meta : Lattice geometry and coupling constants
• solver : Maximum iteration and precision
• jackknife : Bin size for jackknife resampling
• bootstrap : Random seed, number of resamples, block size
• trajectory : Number of $\kappa$-trajectories
• deborah : Deborah.DeborahCore-specific metadata
• abbreviation : Feature name abbreviation map

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

Parse a configuration TOML file and construct a list of JobArgsMiriam structs to represent all job combinations of LBP/TRP for the Deborah.Miriam analysis.

Arguments

• toml_path::String: Path to the configuration TOML file.
• jobid::Union{Nothing, String} : Optional job ID string used for logging.

Returns

• Vector{JobArgsMiriam}: A list of job argument structs, each corresponding to a distinct label/train combination.

Notes

• This function extracts and resolves $\text{Tr} \, M^{-n} \; (n=1,2,3,4)$ inputs, solver parameters, bootstrap/jackknife configs, and meta info.
• It builds the overall_name for each job using either full names or abbreviations.
• The output directory base path is constructed using analysis_header, multi_ensemble, and $\text{Tr} \, M^{-n} \; (n=1,2,3,4)$ details.

run_MiriamThreads(
    toml_path::String
) -> Nothing

Run all Deborah.Miriam jobs defined in the given TOML config file using multithreading.

Arguments

• toml_path::String: Path to the batch config file containing multiple jobs.

Behavior

• Parses all jobs via parse_config_MiriamThreads.
• Spawns tasks in batches of Base.Threads.nthreads() to run run_one_job in parallel.

Returns

• Nothing (side effects: runs all jobs and generates logs/configs).

run_MiriamThreadsCheck(
    toml_path::String
) -> Nothing

Check-only mode for Deborah.Miriam: validate that all expected multi-ensemble reweighting results exist.

Arguments

• toml_path::String: Path to the batch config file.

Behavior

• Parses all jobs using parse_config_MiriamThreads.
• Verifies result presence using Deborah.DeborahEstherMiriam.MiriamExistenceManager.ensure_multi_ensemble_exists.
• Uses threading to batch check jobs in parallel.

Returns

• Nothing (side effects: logs for each check job).

run_one_check_job(
    args::JobArgsMiriam, 
    output_dir::String, 
    log_path::String
) -> Nothing

Run a single check-only job to verify multi-ensemble reweighting result existence.

Arguments

• args::JobArgsMiriam: Job parameter set.
• output_dir::String: Directory to save the generated TOML config.
• log_path::String: Log file for result check.

Behavior

• Writes a config file for the job.
• Calls Deborah.DeborahEstherMiriam.MiriamExistenceManager.ensure_multi_ensemble_exists to verify expected results.

Returns

• Nothing (side effect: logs and possible error messages).

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

Run a single full Deborah.Miriam job: generate TOML config, run ensemble checker, and launch main computation.

Arguments

• args::JobArgsMiriam: Struct containing all parameters for a Deborah.Miriam job.
• output_dir::String: Target output directory where logs and config will be stored.
• log_path::String: Path to write job-specific logs.

Behavior

• Generates a TOML config file.
• Checks ensemble availability via Deborah.DeborahEstherMiriam.MiriamDependencyManager.ensure_ensemble_exists.
• Runs the main Deborah.Miriam.MiriamRunner.run_Miriam pipeline using the given configuration.

Returns

• Nothing (side effects: file I/O and computation).