export tool

usage: export [-x FORMAT] [-p PATH] [-m MODEL] [-l LABEL] [-e EVENTS]
              [-k KINDS] [--dry] source [sink]

Convert the results produced by the experiment tool to wide format
(counts matrices).

positional arguments:
  source               Set the path to the results location to operate
                       on.
  sink                 Define where to place the exported results. If
                       not provided, write to standard output.

optional arguments:
  -x, --format FORMAT  Choose the output format: one of `csv`, `tsv`
                       or `arrow`. If missing, it will be determined
                       from the SINK filename extension, falling back
                       to `csv` if SINK is also missing. (type:
                       Symbol)
  -p, --path PATH      Filter for simulation segments that have a path
                       starting with this prefix.
  -m, --model MODEL    Filter for simulation segments that have a
                       model path starting with this prefix.
  -l, --label LABEL    Filter for simulation segments that have a
                       label matching this regular expression
                       anywhere. (type: Regex)
  -e, --events EVENTS  Filter for simulation segments that have a an
                       events stream name matching this regular
                       expression anywhere. (type: Regex)
  -k, --kinds KINDS    Filter for trajectory dimensions that have
                       names matching this regular expression
                       anywhere. (type: Regex, default: r"")
  --dry                Don't export, only pring matching simulation
                       segments.

This will produce a counts matrix with trajectory components as rows,
simulation segments as columns, and values ending the respective
segments. Values that have no events in their segment will be empty
(for CSV) or null (for Arrow). The rows are named in the first column,
and the column names are set according to the segment's label with its
ending time appended. If these column names are not unique, labels
will be dropped and the segment id will be included.

The segments and trajectory components to include can (and typically
need to) be filtered, for example by selecting only the segments
corresponding to extraction models. The easiest way to achieve this is
to already prepare the simulation schedule such that it redirects the
targeted segments into a separate events stream by defining `"into"`,
and then to filter for that stream using the `--events` option.