line-profiler

What it is and what it does

line_profiler is a Python profiling module that measures execution time at the individual line level within decorated functions, complementing Python's standard cProfile which only times explicit function calls. This is particularly useful for identifying performance bottlenecks in scientific computing or data-heavy code where a single statement (like a NumPy array operation) can dominate runtime without appearing as a function call.

The package provides two main interfaces: the line_profiler module itself, which can be imported and used programmatically with the @line_profiler.profile decorator, and kernprof, a command-line script that automates profiling workflows. Modern versions (4.1.0+) support environment-variable-based activation via LINE_PROFILE=1, while legacy workflows use the kernprof command directly. Results are written to .lprof binary files and can be inspected via the command line or Python module.

Use it for

• Identify which lines in a NumPy-heavy function consume the most time when function-level profiling shows only a single call.
• Optimize data processing pipelines by pinpointing slow statements in tight loops that cProfile would not break down.
• Debug performance regressions in scientific computing code by comparing line-level timings before and after changes.
• Profile web request handlers or batch jobs to find unexpectedly slow operations within a single function.
• Validate algorithmic improvements by measuring per-line execution time changes in critical functions.

Worth the install?

AI-flagged interpretation of the facts on this page. Verify before relying on it.

Install

line-profiler on PyPI

Before you install

Medium install friction due to C extension compilation, but prebuilt wheels are available for Python 3.8–3.14 on major platforms (Linux, macOS, Windows). The package is actively maintained with a recent release and no known vulnerabilities.

C compiler required when building from source; prebuilt wheels avoid this for standard Python versions on common platforms.

License in practice

BSD license is permissive and places no significant restrictions on use, modification, or distribution in proprietary or open-source projects.

Quickstart

pip install line_profiler

import line_profiler

@line_profiler.profile
def my_function():
    pass

# Set LINE_PROFILE=1 environment variable and run your script

Package facts

Evidence: line_profiler-5.0.2-cp310-cp310-macosx_10_9_universal2.whl; line_profiler-5.0.2-cp310-cp310-macosx_10_9_x86_64.whl; line_profiler-5.0.2-cp310-cp310-macosx_11_0_arm64.whl; line_profiler-5.0.2-cp310-cp310-manylinux_2_24_aarch64.manylinux_2_28_aarch64.whl; line_profiler-5.0.2-cp310-cp310-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl; line_profiler-5.0.2-cp310-cp310-musllinux_1_2_aarch64.whl; line_profiler-5.0.2-cp310-cp310-musllinux_1_2_x86_64.whl; line_profiler-5.0.2-cp310-cp310-win_amd64.whl; line_profiler-5.0.2-cp311-cp311-macosx_10_9_universal2.whl; line_profiler-5.0.2-cp311-cp311-macosx_10_9_x86_64.whl; line_profiler-5.0.2-cp311-cp311-macosx_11_0_arm64.whl; line_profiler-5.0.2-cp311-cp311-manylinux_2_24_aarch64.manylinux_2_28_aarch64.whl; line_profiler-5.0.2-cp311-cp311-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl; line_profiler-5.0.2-cp311-cp311-musllinux_1_2_aarch64.whl; line_profiler-5.0.2-cp311-cp311-musllinux_1_2_x86_64.whl; line_profiler-5.0.2-cp311-cp311-win_amd64.whl; line_profiler-5.0.2-cp311-cp311-win_arm64.whl; line_profiler-5.0.2-cp312-cp312-macosx_10_13_universal2.whl; line_profiler-5.0.2-cp312-cp312-macosx_10_13_x86_64.whl; line_profiler-5.0.2-cp312-cp312-macosx_11_0_arm64.whl

Tags

More Software Development packages

See also ipython-autotime · memory-profiler · pyinstrument · yappi · scalene · pyprof2calltree · tuna · viztracer · django-cprofile-middleware · codetiming