pangyplot preprocess

Preprocess a graph with vg / odgi into the sorted GFA plus layout TSV inputs expected by pangyplot add.

The command has two modes:

  • Interactive (default): generate a shell (or SLURM) script you run yourself.

  • Execute (--run): run the pipeline directly via subprocess.

SYNOPSIS

pangyplot preprocess
pangyplot preprocess --run --input FILE [OPTION]…

DESCRIPTION

Interactive mode

Runs an interactive prompt that asks for the input file, reference path names, thread count, GPU availability, and (optionally) SLURM job parameters, then writes a self-contained bash script that:

  1. Converts the input GFA to ODGI .og format.

  2. Optionally sorts the graph with a prioritized list of reference paths.

  3. Runs odgi layout (with optional GPU acceleration) to produce the 2D coordinates TSV.

  4. Writes the sorted GFA back out.

The interactive mode does not run odgi itself; it just produces the script, so the preprocessing can run on a workstation or compute cluster independent of where PangyPlot is installed. This is the default because a bare install cannot assume vg/odgi are present or know where they live.

Execute mode (--run)

With --run, PangyPlot executes the same pipeline directly instead of writing a script. This requires vg/odgi on PATH and is intended for the Docker container, which ships the whole toolchain. Given an input graph (.vg, .gfa, .gfa.gz, or .og) it runs:

  1. vg convert --no-wline (only for .vg input, which odgi cannot read directly).

  2. odgi build — GFA → .og.

  3. odgi sort — optional 1D sort, prioritizing the paths given by --paths/--ref.

  4. odgi layout — 2D coordinates (with --gpu to use odgi_gpu).

  5. odgi view — export the sorted GFA.

On success it prints the resulting <prefix>.sorted.gfa and <prefix>.lay.tsv — the two files pangyplot add expects — along with the pangyplot add command to run next. If a required tool is missing, it exits with an error pointing you back to the interactive script generator.

OPTIONS

Interactive mode takes no command-line options; all parameters are collected via prompts. The options below apply to --run.

--run
Execute the pipeline directly instead of generating a script.
--input FILE
Input graph: .vg, .gfa, .gfa.gz, or .og. Required with --run.
--out-dir DIR
Directory to write outputs into (default: current directory).
--prefix NAME
Basename for output files (default: derived from --input).
--ref STRING
Primary reference path name to prioritize during the sort.
--paths A,B,…
Comma-separated path names to prioritize during the sort, in order. Overrides --ref.
--threads N
Number of threads for odgi (default: 4).
--gpu
Use odgi_gpu for the layout step.
--no-sort
Skip the 1D sort step.