Skip to content

Latest commit

 

History

History
113 lines (65 loc) · 2.48 KB

README.md

File metadata and controls

113 lines (65 loc) · 2.48 KB
Raw

QiwuGrader

Qiwu auto test tool, supports accuracy / pressure tests for knowledge base, QA API and server backend

Chinese version

[TOC]

1. Author

Feliciano Long

Zhiyu.Zhou

Statics

2. Supports

Python2 > 2.7 and Python3 (preferred)

3. Installation

Install from PYPI

pip install QiwuGrader

Download source code

  1. For Windows users please download (Mac users could skip this step)

https://www.python.org/ftp/python/2.7.15/python-2.7.15.msi

  1. (Optional) If Environment Variables is not set correctly, add python executable and pip script to system Path. How to

  2. Install requirements

pip install -r requirements.txt

Download release

Deprecated, please install using pip

Release executable

4. Run

last two parameters are ignorable

Run from module

qiwugrader [testcase] [session number] [test duration]

or

python -m qiwugrader.app [testcase] [session number] [test duration]

Run from source

python qiwugrader/app.py [testcase] [session number] [test duration]

5. Usage

single session test

test one case:

qiwugrader ./testcases/test1.yml

test multiple cases:

qiwugrader ./testcases/test1.yml ./testcases/test2.yml

multiple session test (multi threading)

Use multiple threads to test, in this mode only one CPU is used.

start 10 sessions in 5 seconds

qiwugrader ./testcases/test3.yml 10 5

print_conversation switch is suggested to turn off in multiple session test to make report more readable

multiple session test (multi processing)

Use multiple processes to test, in this mode the program will start logical CPU number processes (each assigned session number DIV CPU number tasks)

Similar to multi threading test, this mode will be enabled when session count larget than 1000 and start interval less than 0.1s

start 1200 sessions in 60 seconds

qiwugrader ./testcases/test4.yml 1200 60

6. Configuration

Examples are under /testcases folder

Check test1 (full explanation) and test2 (minimum test case) for knowledge tests, test3 (json request / QA dialogue) and test4 (form request / knowledge backend) for api test