Command line arguments

The table below show all arguments that are available in the reformat module.

Argument	Type	Default	Description
-h, --help	Flag	–	Print help message.
Required Input/Output
-a, --alignment <alignment>	Path (file)	–	Path to .parquet or .jsonl file produced by fishnet align.
-o, --out <out>	Path (file)	–	Output path. Extension determines format (.parquet or .tsv).
Pod5 Input
-p, --pod5 <pod5>...	Path(s) (file or directory)	–	POD5 input(s). Required if alignment file lacks raw signal. Multiple allowed.
--rna	Flag	–	Set if direct RNA POD5 file(s) are provided (reverse signal).
Data Filter (one required)
-r, --ref-regions <ref-regions>...	String(s)	–	Filter by reference region(s) (<REF>:<START>-<END>). 1-based inclusive.
-R, --bed-file <bed-file>	Path (file)	–	BED file with reference regions (0-based start, exclusive end).
-P, --positions-of-interest <positions-of-interest>...	String(s)	–	Filter by positions of interest (<REF>:<SITE>-<HALF-SIZE>).
-m, --motifs <motifs>...	String(s)	–	Filter for motif sequence(s) (A/C/G/T/U). Multiple allowed.
-M, --motifs-file <motifs-file>	Path (file)	–	File with one motif per line (A/C/G/T/U only).
Processing Strategy
-s, --strategy <strategy>	Enum (stats, interpolate)	stats	Processing strategy: compute statistics or interpolate signal.
--alignment-type <alignment-type>	Enum (reference, query)	–	Select which alignment type to parse.
--skip-signal-norm	Flag	–	Skip z-standardization of signal intensity.
--skip-dwell-norm	Flag	–	Skip z-standardization of dwell times.
Strategy 1: Statistics Settings
--stats <stats>...	Enum(s) (mean, median, std, dwell, signal-to-noise)	mean std dwell	Statistics to compute per base when using --strategy stats.
Strategy 2: Interpolation Settings
--target-size <target-size>	Integer	30	Target size for interpolated signal chunks per base.
Threading Settings
-t, --threads <threads>	Integer	8	Number of parallel threads (set to 1 for single-threaded).
--queue-size <queue-size>	Integer	1000	Queue size for worker communication (affects memory use).
Input/Output Settings
--input-chunk-size <input-chunk-size>	Integer	4000	Number of alignments read per iteration. Larger = faster, more memory.
-f, --force-overwrite	Flag	–	Overwrite existing output file if it exists.
--output-shape <output-shape>	Enum (melted, exploded, nested)	nested	Determines output structure (nested only for Parquet).
--output-batch-size <output-batch-size>	Integer	4000	Number of alignments buffered before writing output.
Logging Settings
--log-level <log-level>	Enum (off, error, warn, info, debug, trace)	off	Controls verbosity of logging.
--log-path <log-path>	Path (file)	log.txt	Log file path (used if --log-level ≠ off).