Doclerk

Doclerk logo
Architecture
Feature1 - Minutes Creator
Feature2 - Topic Detective
Feature3 - Text Summarizer
Minutes Creator 1
Topic Detective 1
Topic Detective 2
Text Summerizer 1

Inspiration

We, as fresh blood, have taken numerous minutes since we entered the world of work. As we have been struggling for it, a large number of people would not have enough time for writing a well-organized note during each meeting; therefore, the minutes would be messy at first. Also, it would require novices for a further time in order to beautify them. Turning to another problem, it is sometimes really difficult for people to understand the content of technical documents because there are so much jargon. That is why we developed this application which could provides great support especially for newcomers.

What it does

There are two features with which Doclerk provides users.
1. Minutes Creator
Once a user enters agenda items and transcription in each field, Minutes Creator clusters sentences in it into the number of agenda topics. Thus, this feature enables the user to focus on just transcribe what he/she hears. Also, it helps reduce the time in order to create well-organised minutes.

2. Topic Detective
When a user comes across a technically complicated document (unfortunately it is also too long to finish reading it,) Topic Detective can display only the sentences related to a topic which the user would like to know.

3. Text Summarizer
When a user wants to know a summary of sentences or documents, once he/she enters them as an input, Text Summarizer categorizes them into some groups and displays one sentence from each group. the chosen sentences represent eash bracket most and the summary consists of them.

How we built it

Three of us have been assigned to each of three functions and the others has prepared datasets, tested application, made a power point slide, and created documents.

Challenges we ran into

It was really tough for us to read documents because only English version exists. Also, there were a few sample codes; therefore, we were uncertain about whether our code was correct or not. Moreover, a invocation limit was a challenge for us because the API we used was really powerful and it took more than 10 seconds to execute it if we enter a long sentences as an input.

Accomplishments that we're proud of

We could integrate IBM Debater with Confluence. They generated the synergy and we could reaffirm that both of them were brilliant.

What we learned

We could find that Forge provides us with a powerful platform and enable us to create a new application easily. We all are delighted to have this opportunity to try it out.

What's next for Doclerk

This time we couldn't overcome a limit above. If we had a chance, we would implement a workaround or hope that the developers who work for Atlassian lift the limit in the near future. We would like to introduce our app to more and more people because the API itself is outstanding. We have used Debater API developed by IBM. We asked the manager who was in charge of the development of this API for us using it, and he allowed us to do so. We would like to show our app to him, and then, if possible, we hope that it could be improved by some professionals and prevalent in the world.