stefanhuber / ICWE-2023

On the Energy-Efficiency of Hybrid UI Components for Mobile Cross-Platform Development

Home Page:https://link.springer.com/chapter/10.1007/978-3-031-34444-2_18

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

On the Energy-Efficiency of Hybrid UI Components for Mobile Cross-Platform Development

Setup

  • Create virtual environment: python -m venv venv
  • Activate virtual environment: source venv/bin/activate (Windows: venv\Scripts\activate)
  • Install external project dependencies: pip install -r requirements.txt
  • Add local package: pip install -e .

Scripts

Several Python scripts are available to prepare the mobile devices and execute the experiment. The commands can be executed from the project's root folder, after completing the setup.

Preparation tasks

  • Generate demo data for apps inside ./data/raw/demodata directory: python bin/generate_data.py
  • Copy demo data from ./data/raw/demodata into app projects: python bin/copy_data.py
  • Build and copy apps to ./data/apks directory: python bin/prepare_apks.py
  • Prepare the mobile device, by installing all the apks from ./data/apks: python bin/prepare_device.py
  • Start Monsoon Power Monitor: python bin/start_monsoon.py
  • Connect USB connected device via Wi-Fi, after Wi-Fi connection, USB can be removed: python bin/connect_wifi.py

Experiment execution

Before starting the experiment execution all preparation steps have to be executed. The experiment execution script has additional parameters:

  • -m to set the metering approach, only monsoon is supported and set as default
  • -f define the output file, where the data should be written to, defaults to measurements.txt
  • -c define the number of executions per experimental subject, for reliable results 30 is suggested

An examplary call of the command would be as follows:

python bin/start_experiment.py -f ./data/experiment_run_1.txt -c 30

Analysis

The data exploration and statistical analysis can be found within the notebook ./notebooks/data-analysis.ipynb. To open the notebook, execute the following command from root folder, after completing the setup:

jupyter notebook

UI Components

UI Components

Apps

The source code of the apps, which contain the UI components, resided inside the ./data/apps folder.

  • hui-android-elements: contains Dialog, Drawer, Sheet implementation for Android native
  • hui-android-scrolling: contains Scrolling implementation for Android native
  • hui-capacitor-elements: contains Dialog, Drawer, Sheet implementation for Capacitor
  • hui-capacitor-scrolling: contains Scrolling implementation for Capacitor
  • hui-elements: contains Dialog, Drawer, Sheet implementation for Hybrid UI Components
  • hui-scrolling1: contains Scrolling implementation for for Hybrid UI Components, with 1 item per WebView
  • hui-scrolling4: contains Scrolling implementation for for Hybrid UI Components, with 4 items per WebView
  • hui-scrolling10: contains Scrolling implementation for for Hybrid UI Components, with 10 items per WebView
  • instrumentation: contains the scripts to interact with the UI components

Results

Results

Monsoon Power Monitor Setup

Windows

  1. The Power Tool with drivers can be installed as described in the manual.
  2. In order to use the Monsoon Power Monitor via the Python API a USB connection is required. On Windows a filter for the usb driver can be used, thus the lib-usb32 devel filter must be used. Download and install libusb-win32-devel-filter-1.2.6.0.exe. The filter can be installed for a turned on Monsoon Power Monitor according to the following screenshots:

Screenshot 1

Screenshot 2

Mac M1

libusb might has to be installed. We faced an issue with the M1 chip architecture. The following command will install libusb:

  • brew install libusb

If the following lists USB devices, everything works as expected:

import usb.core
dev = usb.core.find()
print(dev)

With Mac M1 an issue arose, which could be solved with the following solution:

  • sudo ln -s /opt/homebrew/lib/libusb-1.0.0.dylib //usr/local/lib/libusb.dylib

About

On the Energy-Efficiency of Hybrid UI Components for Mobile Cross-Platform Development

https://link.springer.com/chapter/10.1007/978-3-031-34444-2_18


Languages

Language:Jupyter Notebook 68.4%Language:Java 18.7%Language:HTML 8.4%Language:Python 4.5%