Contents

visualize$execution$dashboard

visualize an execution as an interactive dashboard

Renders a self-contained HTML dashboard for one execution from its structured <exec_id>.jsonl event stream: run summary, a time-scrubber, the collapsed dynamic execution graph (type -> rule -> type), per-rule performance, and a searchable task drill-down. The dashboard is driven entirely by the JSONL stream the execution engine writes; no resource-store access is required.

Usage

visualize.execution.dashboard(
  execution = NULL,
  file = NULL,
  open = interactive(),
  live = FALSE
)

Arguments

Argument	Description
`execution`	A run result object, a character scalar, or NULL (default). Which execution to render: NULL takes the most recent stream in the workspace execution directory; a run result object (e.g. the PGSLibrary returned by generate$models()) is resolved through its provenance; a character scalar is an execution id (e.g. "20260627-204722-0001"), a path to a .jsonl stream, or a directory holding one or more streams (the most recent is used). Aborts when no stream resolves.
`file`	Character scalar, or NULL (default). Output HTML path; NULL writes <exec_id>.dashboard.html beside the source stream. An existing file is overwritten.
`open`	Logical scalar, default interactive(). Opens the rendered file in a browser; with no usable browser it reports the path instead of failing.
`live`	Logical scalar, default FALSE. TRUE makes the dashboard fetch and tail-poll the .jsonl over HTTP instead of inlining it, so an open page follows a running execution, and requires the output file to be served from the same directory as the stream (what the engine's auto-hosting does). FALSE inlines the current stream contents for a portable snapshot.

Value

The output HTML path, a length-1 character, invisibly. Not a plot object -- nothing composes with it. Called for the side effect of writing that file, and of opening a browser when open = TRUE.

Details

Static export is fully portable: copy the single HTML file anywhere and open it. Live mode trades portability for following a run; on a remote or cluster machine the served URL may require port-forwarding, so the static export is the reliable way to share a finished run.

Task identifiers in the stream are content hashes. A readable label is harvested from each task's worker log (e.g. a GWAS accession) where available, falling back to <type>:<hash>.

Examples

models <- generate$models(sources = sources, algorithms = algorithms)
visualize$execution$dashboard(models)
visualize$execution$dashboard("20260627-204722-0001", open = FALSE)