Protocol Sequences

Advanced Mode’s Build Protocol tab lists every recorded step of the current analysis. The same steps are exported by Export Sequence.py and replayed by Run Sequence and the physplot command line.

For a step-by-step example with screenshots (record a protocol, save it, reopen and edit it, then bulk-process a folder), see Sequence Walkthrough: Record, Save, Reuse and Bulk-Run. Every control on these tabs is described in Build Protocol and Run Sequence.

Step status

After Apply This Sequence, deleting a row, or a rerun, the Status column reports what happened to each row:

  • OK: the row ran. Hover to see how long it took.

  • Failed: the row raised an error. Hover to read the error.

  • Skipped: the row did not run because an earlier row failed.

A sequence always stops at the first failing row. The spreadsheet keeps the state reached just before that row, so you can inspect the data that caused the problem. The status bar names the failing row, for example Row 2 failed (TransformColumnStep): ValueError: Column 'Volts' not found.

Statuses clear for any row whose step you replace, for example after Apply Code to Table, until the sequence runs again.

Rerun from a step

Right-click a row and choose Rerun from this step to resume the sequence at that row without replaying the rows above it. PhysPlot restores the data exactly as it was before that row last ran, then runs the row and everything after it.

Typical use:

  1. Apply the sequence. Row 5 fails.

  2. Fix row 5 in the Code view and click Apply Code to Table.

  3. Right-click row 5 and choose Rerun from this step.

Rows 1 to 4 are not replayed, which matters when they load large files or run slow steps. If you edited a row above the one you rerun from, PhysPlot starts from the earliest point it can still trust, which is normally the start of the last full run. If no run has happened yet, apply the whole sequence first.

Editing steps

Every row can be changed from the table without touching Python:

  • Edit: double-click a row, or right-click and choose Edit step…. The editor lists the row’s settings; fields marked * are required. Column fields offer the current table columns. Invalid values, such as a parameter that is not a {...} dictionary, are explained in red and the dialog stays open. After OK the sequence resumes from that row.

  • Enable or disable: untick the On checkbox, or right-click and choose Disable step. A disabled row stays in the sequence, is shown in grey, and reports Skipped; the rows after it still run. Exported sequence files keep the setting as enabled=False.

  • Reorder: right-click and choose Move up or Move down. Rows that load data stay first and nothing can move above them.

  • Insert: right-click a row and choose Insert step after…, or use Protocol > Insert Step… and Protocol > Insert Protocol Module. New rows go after the selected row, or at the end when no row is selected.

Each of these changes replays the sequence from the first affected row, the same way Rerun from this step does. If the change breaks a later row, for example moving Set Roles below Generate Plot, that row shows Failed and the status bar explains why; no dialog interrupts you.

Headless use

The same behavior is available in Python:

from physplot import PhysPlot
from physplot.workflow import load_workflow

pp = PhysPlot()
pp.load("data.csv")
steps = load_workflow("sequence.py")

results = pp.run_workflow_detailed(steps, allow_column_number_fallback=True)
for result in results:
    print(result.index + 1, result.step_name, result.status, result.error or "")

# After fixing steps[4]:
pp.rerun_from(4, steps, allow_column_number_fallback=True)

pp.run_workflow(steps) keeps its original behavior: it raises the first error. It also records per-step results on pp.last_results.

Steps can be edited in code with the same checks the editor uses:

step = steps[1]
print(step.describe())            # editable fields, types, and current values
step.update(function_name="divide", params={"divisor": 2})
step.update(enabled=False)        # keep the step but skip it
pp.move_step(3, 2)                # reorder pp.workflow
pp.insert_step(1, step)           # insert before position 1