We show that web crawling based testing is application specific. Testers have to manually configure the webcrawler for each web application and then they can perform test.
web-crawler.py is a sample of a crawler program that extracts features from a web form
web-crawler-config.yaml is the configuration file for web-crawler config. We can see that this config is application specific.
All packages are installed using pip, which is a python package manager.
-
BeautifulSoup4 (version 4.6)
pip install beautifulsoup4
-
numpy (version 1.13.3)
pip install numpy
-
scipy (version 1.0.0rc2)
If using windows: install 'wheel' for scipy package
wheel (version 0.30.0)
pip install wheel
then download scipy from http://www.lfd.uci.edu/~gohlke/pythonlibs/#scipy
install scipy using wheel
pip install C:\Users\user\Downloads\scipy-1.0.0rc2-cp36-cp36m-win32.whl
If not using Windows (on Linux or Mac OS), install scipy normally from pip
pip install scipy
Note: If on Windows, numpy and scipy can also be installed using Anaconda, which is a python
data
science platform. Anaconda comes with the necessary packages for data science projects right out
of the box. Anaconda is the preferred method for installing project dependencies.
Visit: http://www.anaconda.com for more details
-
scikit-learn (0.19.1)
pip install -U scikit-learn
-
gensim
pip install --upgrade gensim
How do we enhance web crawling based techniques to assist with the labor-intensive task of testing multiple web applications?