Introduction

Use profile as a context manager, or start and stop collection explicitly. Reports show measurements beside the source that ran.

Profile Python

Context manager

from linescope import profile

with profile() as session:
    run_pipeline()

The report displays when the block ends. Workload exceptions still propagate; LineScope stops collection and restores its hooks.

Explicit start and stop

from linescope import profile

session = profile.start(display="none")
run_pipeline()
profile.stop()
session.save("pipeline.html")

Use this across notebook cells or separate steps. Results are available through session.result or profile.result. profiler is an alias for profile.

Headless execution

Set display="none" to suppress automatic display. Save with session.save().

Scripts and packages

linescope --include mypackage -m mypackage.job --input data.csv

Arguments after the target go to your program. See the CLI reference for options.

Read the results

Open Overview and follow an expensive line to its source. A dash means unavailable; sampling can miss short operations. See Reports for measurements and Backends for collector differences.