oraluben / diab

Smart app that helps you manage your diabetes

Diab is a smart opensource application that helps you managing your diabetes by keeping track of your glucose values and insulin injections.

Using the data registered inside the app it's possible to generate a customized plugin that once applied to the app will provide smart insights for insulin dosages based on real-time context.

It's also possible to integrate the app with other fitness services to share the data.

Features

• Save records of glucose and insulin dosages
• Insulin suggestions plugin
• Export records as Excel file
• Reminder for checking again in case of hypoglycemia
• [optional] Google Fit integration

For upcoming features, see the bug tracker.

Open source

License

Released under the GNU GPLv3.

Working with the repository

Clone the repository with git

git clone https://github.com/bvlj/diab

Setup the work environment

• Install git pre-push hooks (all the tests must pass before pushing to master)
• Install ktlint (enforces code style guidelines)
• Prevent git from tracking changes to your trained plugin models

./_scripts/setup.sh

Compile

The builds can be compiled using gradle. The following build variants are available:

1. oss
2. googleFit

oss builds

Builds the app without any fitness services integration. The output is composed of 100% open source code.

./gradlew assembleOssRelease

googleFit builds

Builds the app with (optional) Google Fit integration. The output includes proprietary libraries from Google used for Fit.

./gradlew assembleGoogleFitRelease

Insulin suggestions plugin

⚠️ This feature could be dangerous, use with extreme caution. ⚠️

DO NOT blindly rely on it as while it's efficient enough for providing hints, it's NOT supposed to replace medical advices or suggestions from more-experienced humans.

The developer and contributors disclaim any responsibility over any form of injury derived from (mis)usage of this feature.

The insulin suggestion plugin allows the app to provide smart insulin dosage hints.

The plugin is created by using a machine-learning model created with data from the app which is tested against a wide set of possible scenarios.

Given the strictly sensitive and personal nature of dealt with data, no plugin is bundled with the app, nor made available for download.

Build your own suggestion plugin

To create your customized insulin suggestions plugins using ml, see the readme found in the ml module.

It's also possible to build a plugin manually, but no documentation is available for the time being.