Pyperflog

This repository intention is to provide a single python module pyperflog.py that helps to facalitate gathering of python code performance using line_profiler.

The pyperflog.py operates as a wrapper for the module defined by PROFILER_DECORATE_MODULE whose performance is under the testing.

It is doing so by extends all functions of the PROFILER_DECORATE_MODULE by decorating them with performance.

The performance decorator utilizes LineProfiler to gather performance metrics, and store them in the following directory PROFILER_LOG_ROOT_PATH/<process_name>/<function_name>.

The metrics are updated as functions gets called through this module.

Dependencies

• line_profiler

Usage

1. Place this module in the directory where PROFILER_DECORATE_MODULE is located.
2. Rename the PROFILER_DECORATE_MODULE to some PROFILER_DECORATE_MODULE_NEW (any name).
3. Rename this module to PROFILER_DECORATE_MODULE original name.
4. Update import of the PROFILER_DECORATE_MODULE in this module to PROFILER_DECORATE_MODULE_NEW.
5. Update value stored in PROFILER_DECORATE_MODULE to PROFILER_DECORATE_MODULE_NEW.
6. Reload any user-process with the name <process_name> to use this module instead of PROFILER_DECORATE_MODULE.
7. Wait until user-process calls any function using this module.
8. Collect performance metrics in the PROFILER_LOG_ROOT_PATH/<process_name>.