From 94a4699d6427d60fd629320371b98c7b39d987ff Mon Sep 17 00:00:00 2001
From: Frederick Muriuki Muriithi
Date: Wed, 22 Dec 2021 11:37:09 +0300
Subject: Start a profiling topic

* Start a profiling topic to help capture some gotchas and help make
  the profiling process easier for all involved.
---
 topics/profiling_python_code.gmi | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)
 create mode 100644 topics/profiling_python_code.gmi

(limited to 'topics/profiling_python_code.gmi')

diff --git a/topics/profiling_python_code.gmi b/topics/profiling_python_code.gmi
new file mode 100644
index 0000000..d7f9ea0
--- /dev/null
+++ b/topics/profiling_python_code.gmi
@@ -0,0 +1,22 @@
+# Profiling Python code
+
+As part of improving the system, there is need to identify and fix/improve the performance bottlenecks in the code. This document details examples of how one would run various profilers, for both GeneNetwork2 and GeneNetwork3
+
+## GeneNetwork3
+
+### cProfile
+
+Syntax:
+
+* env [various-env-vars] python3 -m cProfile the-script.py
+
+where
+
+* `[various-env-vars]` is a number of environment variables that might be needed for the running of the script, e.g. `SQL_URI` which is used to define how to connect to the database.
+* `the-script.py` is the name of the python script to be run under the profiler
+
+The output can be redirected, e.g.
+
+* env [various-env-vars] python3 -m cProfile the-script.py 2>error.log 1>performance.log
+
+which should help will separating errors from the performance data, for easier analysis
-- 
cgit v1.2.3