globalmaksimum / sadedegel-annotator

Sadedegel extraction based summarizer document annotator tool.

Sadedegel Annotator

Electronjs based labeling tool for extraction based summarization. Developed as a part of Açık Kaynak Hackathon Programı 2020, in order to provide annotated data for Sadedegel Library.

📖 For more information, please refer to sadedegel.ai

💬 Where to ask questions

The SadedeGel project is maintained by @globalmaksmum AI team members @dafajon, @askarbozcan, @mccakir and @husnusensoy.

Type	Platforms
🚨 Bug Reports	GitHub Issue Tracker
🎁 Feature Requests	GitHub Issue Tracker

Installation

Sadedegel Annotator is a cross-platform desktop application.

• Supported Operating Systems: macOS · Linux · Windows

To install Sadedegel Annotator and its dependencies:

$ git clone https://github.com/GlobalMaksimum/sadedegel-annotator.git
$ cd sadedegel-annotator
$ npm install

Usage

$ npm start

How it works

Sadedegel Annotator enables the user to summarize texts by eliminating sentences in rounds. At each round, user identifies the sentences that contain the least important information and selects them to be extracted from the text. For each sentence, the round it is eliminated in implies its level of importance.

Sample Input

Input should be json file(s) in the following format:

{"sentences":["Şimdi sizden günlük hayatınızda devamlı olarak kullandığınız bir unsuru düşünmenizi istiyorum.",
"Onsuz yapamam dediğiniz, artık varlığı hayatınızda bir alışkanlık oluşturmuş, yokluğu ise panik sebebi olan...",
"Herkesin farklı olabilir; telefon, araba veya bilgisayar...",]}

Output

Output is saved in [selectedInputDirectory]/labeled.

References

• Extraction Based Summarization