/sapienz

Multi-objective Automated Testing for Android Applications

Primary LanguagePythonOtherNOASSERTION

Sapienz

Sapienz: Multi-objective Automated Testing for Android Applications

Sapienz is an implementation of the Sapienz approach for multi-objective automated testing for Android applications.

Publication:

@InProceedings{mao:sapienz:16,
    author = "Ke Mao and Mark Harman and Yue Jia",
    title = "Sapienz: Multi-objective Automated Testing for {Android} Applications",
    booktitle = "Proc. of ISSTA'16",
    year = "2016", 
    pages = {94--105}
} 

Demo Videos in a Real-world Setting

Available at [sapienz.tk] (http://www.sapienz.tk)

Installation

Simply clone the source code from this repository and apply the following environment configuration:

Environment Configration

  • Python: 2.7

  • Android SDK: API 19

  • Linux: sudo apt-get install libfreetype6-dev libxml2-dev libxslt1-dev python-dev

  • Mac OS: brew install coreutils for gtimeout

Install Python dependencies:

sudo pip install -r requirements.txt

Usage

python main.py <apk_path | source_folder_path>

where apk_path is path to the subject apk under test
or you can specify source_folder_path for the subject app with source code

Subject Requirement:

  • instrumented apk should be compiled and named with suffix "-debug.apk"
  • closed-source/non-instrumented apk name should end with ".apk"

Settings

  • ANDROID_HOME and WORKING_DIR in settings.py should be set before starting Sapeinz.

Output

  • for open-sourced apps, outputs are stored under the given source folder
  • for closed-sourced apps, output are stored under <apk_file_path>_output

Output content:

/coverages - Coverage reports are stored here

/crashes - Crash reports and corresponding test cases that lead to the crashes 
(and also recorded videos files when using real devices)

/intermediate - Generated test event sequences for each generation; logbook of the genetic evolution; 
and line charts showing the variation trend for each objectives.

Notes

  • This implementation has been tested with Android 4.4, running on Ubuntu 14.04 and Mac OS 10.10
  • If measure statement coverage for open-sourced apps, the subjects need to be processed to support EMMA instrumentation: (Please refer to Dynodroid https://code.google.com/archive/p/dyno-droid/)
  • This version is ready for emulators. It also supports real devices, you may need to adapt related code for your specific devices.

License

[BSD License] (https://github.com/Rhapsod/sapienz/blob/master/LICENSE)

Contact

dreamict@gmail.com