perfassess 0.0.4

Access the performance of a given function and create plot.

⌛️ PERFASSESS 💾

✍ Contributor: Lucas ROUAUD

📒 Description

This module permit to evaluate the performance of a given function and create Plotly bar plot. Check the documentation at https://filouplains.github.io/perfassess/.

🗄 Dependencies

To run this software, you will need python ≥ 3.11 and this next module:

• numpy ≥ 1.26.0
• plotly ≥ 5.19.0
• pyyaml ≥ 6.0.1

⚙️ Installation

🥹 EZ way

Recommended method:

$ pipx install perfassess

If you do not have pipx installed, check out the documentation at https://pipx.pypa.io/stable/installation/. Else, classical installation method:

$ pip install perfassess

😩 From source

👬 Cloning the repository

You can clone the repository using HTTPS:

$ git clone https://github.com/FilouPlains/perfassess.git

or using SSH:

$ git clone [e-mail]:FilouPlains/perfassess.git

Then go to the 📁 perfassess/ directory:

$ cd perfassess

All next commands are assuming that you are in the 📁 perfassess/ directory. This directory will be name as 📁 ./.

🐍📦 Installing with pip

Simply launch this command in the 📁 ./ directory:

$ python3 -m pip install .

You are now able to launch the program!

⌨️ Using command line

✏️ General note

By doing:

$ perfassess --help

You will get the help to use the command line interface. Here are the used legends:

• int: Integer.
• str: String.
• [type|value]: Type of the input required, follow by the default value. So if this optional arguments is not used, “value” will be chosen.

⚠️ If you use pickle, the command line interface will not work!

📄 Normal use

To use th program, launch:

$ perfassess -s script.py \\
             -f function_name \\
             -o output_directory/ \\
             -a argument.yml

You can actually directly test the program in the repository root (📁 perfassess/) using:

$ perfassess -s src/perfassess/testor.py \\
             -f testor \\
             -a data/argument.yml \\
             -o data/

📁 Package use

Let us say that you want to test a package, which should have this kind of tree structure:

package/
└── src/
    ├── __init__.py
    └── script.py

In package, you can use relative paths. But with these, the “normal use” method do not work. Instead, use the next command, if you are in 📁 package/, in order to evaluate script.py:

$ perfassess -s src/script.py \\
             -f function_name \\
             -o output_directory/ \\
             --package src/__init__.py \\
             -a argument.yml

In addition of the previous "normal use" method, you have to indicate a __init__.py file.

Note 📝

Packages are identify with __init__.py files and allow relatives import.

🗂 Subpackage use

Let us say that you want to test a subpackage, which should have this kind of tree structure:

./package/
└── src/
    ├── __init__.py
    └── subpackage/
        ├── __init__.py
        └── script.py

In subpackage, you can use relative paths. But with these, the “normal use” method do not work. But the “package use” also do not work. Instead, use the next command, if you are in 📁 package/, in order to evaluate script.py:

$ perfassess -s src/subpackage/script.py \\
             -f function_name \\
             -o output_directory/ \\
             --package src/__init__.py \\
             --subpackage src/subpackage/__init__.py \\
             -a argument.yml

In addition of the previous “normal use” and “package use” method, you have to indicate two __init__.py files. The first one is the top package __init__.py file. The second one is the __init__.py file of the subpackage to test.

Note 📝

Packages are identify with __init__.py files and allow relatives import.

🔍 Describing possible parameters

Argument	Mandatory?	Type and usage	Description
-s --script	Yes	-s script.py	The script that contain the function to test.
-o --output	Yes	-o output_directory/	The directory where the plot have to be produced.
-f --function	No	-f main	The function to test.
-a --argument	No	-a argument.yml	The argument to passe to the function to test*.
--n_field	No	-n_field 2	The number of field to keep**.
--package	No	--package package/__init__.py	The __init__.py file of the top package to test.
--subpackage	No	-o package/subpackage/__init__.py	The __init__.py file of the subpackage to test.
-h --help	No	Flag	Display the help and exit the program.
-v --version	No	Flag	Display the version and exit the program.

• * = If you want to test a function that requires arguments, you can give to the command line interface a argument.yml file that contains all required arguments. If the tested function requires an argument like toto, you can put in the YAML file:

toto: 10

• ** = In memory usage, tested functions are going to be named something like /home/user/Documents/program/package/script.py. To shorten the name, use the --n_field tag. For instance, giving 2 will let know to the program that you want to only keep the last two field, which in the end will look something like: package/script.py.

🙇‍♂️ Aknowledgement

🔍 Code reviewing: Hubert Santuz

This work is licensed under a MIT License.