Skip to content

Progress reporting#

Pass on_progress to run_contingency_analysis_pandapower and the engine will call it as outage groups finish. It is off by default.

This hook is pandapower only. get_ac_loadflow_results and the powsybl backend do not accept it. Powsybl security analysis runs entirely in Java; a callback from Python into that run is not implemented and not planned in the pypowsybl wrapper. Progress reporting will stay pandapower-local.

What it reports#

on_progress(done, total):

  • total is the outage-group count, len(grouped_contingencies). With grouping off that is the contingency list. With cfg.apply_outage_grouping several contingencies can share one group, so total can be smaller.
  • First call is (0, total), before the first load flow.
  • Later calls are about once a second, then (total, total) if the run succeeds.
  • A failing callback does not fail the analysis. The exception is logged and ignored.

Sequential and parallel#

The same callback works on both paths. Sequential calls it between load flows. Parallel calls it on the driver process; a Python callable cannot be sent into a Ray worker.

Using it#

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
from toop_engine_contingency_analysis.pandapower.contingency_analysis_pandapower import (
    run_contingency_analysis_pandapower,
)
from toop_engine_contingency_analysis.pandapower.pandapower_helpers.schemas import (
    ContingencyAnalysisConfig,
)

def report(done: int, total: int) -> None:
    print(f"{done}/{total} outage groups")

run_contingency_analysis_pandapower(
    net,
    nminus1_definition,
    job_id,
    timestep,
    cfg=ContingencyAnalysisConfig(method="ac", polars=True),
    on_progress=report,
)