Throughout my career as a software developer and working with major IT consultants I’ve always thought to minimize the hassle, and thrive in an evolving world. In our day to the day development cycle, we use Confluence, to document all of our important stuff. But the problems start whenever we revisit to look at those documents, many times we find ourselves scrolling through pages trying to find a punchline or a specific part of a document. It can be time-consuming and mind-numbing to try and find a key topic in a document. (And this doesn’t end after one or two)
Although plenty of such NLP solutions are available online and usage is extravagant, often with the limit of the data privacy we just can’t copy-paste the content online to get the relevant inference. WordCubo integrated with the confluence provides the safe option of analyzing the documents without any worry.
The transition to online classes has, contrary to expectations, increased the average time spent for classes. To improve studying efficiency, we decided to tackle a major part of our long study times and comprehension ability: Zoom lectures. We thought it’d be nice if we could somehow condense lecture material without losing comprehension and possibly even increasing our understanding.
In 2008 a study estimated that it would take 244 hours a year for the typical American internet user to read the privacy policies of all websites he or she visits – and that was before everyone carried smartphones with dozens of apps, before cloud services, and before smart home technologies.
What it does
The core of the idea is to remove those hassles of finding that “the” information in the confluence document to give users quick and most relevant content.
WordCubo for Confluence, a forge-integrated NLP tool to provide users with a seamless learning experience by giving a fully-fledged solution that aims to reduce the time and hassles we faced day-to-day and provide a timeless solution so as to make your learning efficient.
How I built it
WordCubo is built on the Atlassian Forge with the integration of the Confluence APIs for the document extractor and Expert.AI for the NLP APIs. The front end of this app was created with the inbuilt UI-Kit for a cutting-edge user experience. Forge CLI provides continuous integration and automated testing, while Bitbucket for code management.
First, extract the body from the Confluence page by using the Confluence API and this HTML body of the page is feed into the HTML to Plain text parser which filters out the unwanted content such as hyperlinks and images and returns the plain text. Plain text gets further cleaned and prepared to be used in the NLP APIs. With the help of this cleaned data, API generates the keynotes, highlight key main sentences, and the main topic. Also did the sentimental analysis of the document using the sentiment API. All this information returned to the Forge App and finally to the user.
Challenges I ran into
In the initial phases, it was really difficult to work alone on this project and understanding the documentation of the forge. This was the first time I was working with Forge and Confluence APIs.
Accomplishments that we're proud of
I manage to finish the project in such a limited time of 2 weeks and in my free time from the routine work. I am still able to submit on time while learning and developing at the same time. I am really satisfied and proud of our final product for the hackathon.
What I learned
Through this project, I learn how to leverage the amazing Confluence APIs and got to know about Atlassian Forge. I also learn the process of developing a serverless application in minutes while the platform takes care of security, compute, and storage.
What's next for WordCubo - A NLP Solution for Confluence
- Expand the WordCubo to support the inbuilt video and sound files in the confluence document.
- Provide support for languages other than English.
- Make the UI of the tool more super-friendly.
- Providing the feature of saving the summaries.