Examples

Packaged examples are copied with trasgu_examples. Run examples are self-contained directories: each one contains its own trasgu.yaml and local input files.

trasgu_examples --list
trasgu_examples minimal ./minimal
cd minimal

Choosing an example

Situation	Example	Shows
First local run	minimal	Basic trasgu.yaml, default remote Chimera store, default output paths.
CSV input or selected columns	csv_columns	Comma-separated data_file plus the columns option.
Local or offline Chimera store	local_chimera	trasgu_url pointing to a local chimera.zarr.
Custom fitting controls	custom_controls	controls_file with a pickled pyvinecopulib.FitControlsVinecop object.
Larger local run	parallel_debug	Larger input, larger chunk_size, max_workers, and debug.
Cluster profiles	profiles	Snakemake profile files for SLURM-style execution.

Minimal local run

trasgu_examples minimal ./minimal
cd minimal
trasgu_count_chunks
trasgu_run --dry-run
trasgu_run

Use this first. Intermediate chunks are written to .trasgu_minimal, and the default combined output is fit_minimal.csv.

CSV input with selected columns

trasgu_examples csv_columns ./csv-columns
cd csv-columns

This example uses a comma-separated input file with eight columns and selects four variables:

data_file: input8_100_gauss_low.csv
columns: [1, 3, 5, 7]
chunk_size: 1000

Local Chimera Zarr

trasgu_examples local_chimera ./local-chimera
cd local-chimera

This example uses trasgu_url to point to a local Chimera Zarr store:

trasgu_url: /scratch/user/chimera.zarr

Use this after downloading Chimera with trasgu_download_zarr /scratch/user, or when running on compute nodes without external HTTP access.

Custom controls

trasgu_examples custom_controls ./custom-controls
cd custom-controls

This example loads a pickled pyvinecopulib.FitControlsVinecop object from controls_file:

controls_file: controls.pkl

The copied directory also includes controls_management.py, the helper script used to create controls.pkl.

Larger local run

trasgu_examples parallel_debug ./parallel-debug
cd parallel-debug

This example uses a larger input and more workers:

data_file: input7_500_gumbel_high.txt
chunk_size: 20000
max_workers: 16
debug: true

Use it after validating installation and data access with the minimal run. It is a better starting point for tuning chunk_size and max_workers on a local workstation.

Snakemake profiles

Copy cluster profile examples separately:

trasgu_examples profiles ./profiles

Use them with a run directory:

trasgu_examples minimal ./minimal
cd minimal
trasgu_run --profile ../profiles/unican.yaml --dry-run