Greetings!

Sprint 5 has now ended, the team has done a lot in this sprint. Most of the goals have been done this sprint, the only task that we didn’t finish was to create a chat bot. The chat bot is suppose to use NLP to help the user with the application and run the application. We have decided that we will not focus on this requirement, we will instead focus on accomplishing the other requirements. The reason being that there isn’t much time before the demonstration and presentation.

The first big goal we completed this sprint was to implement speech to text, we implemented it using Googles speech API. The output form the speech to text is processed exactly the same as the input from the CLI. Thus there weren’t many changes that needed to be done. The second goal we completed was to implement a new module that can remove events from the calendar and the last one we completed was to implement a new module for rescheduling an event in the calendar. Both of these new modules required us to retrain our calendar NLP model to support the two new use cases. The functionality of the GUI didn’t get completed this sprint. There is still a lot of functionality that is missing. The most important functionality that is missing is the functionality to run the input through the application. But this next sprint more of the team is going to work on the GUI because is is very important to get done and there isn’t much time left.

Some of the other things that got implemented this sprint where user experience improvements such as adding animations to the CLI that show that the NLP modules are loading. Another one is that now the NLP modules can understand full names such as “John Doe” for example. It is also now possible to run multiple NLP modules by adding a full stop add the end of each NLP sentence, the modules will run sequentially in the same order as the input. There were also some other small improvements done to the user experience, but nothing big worth mentioning.

Lastly we also got done with some code improvement like having a pool of all the NLP models. This makes it possible to load all the modules in the beginning of the application and then the modules can request access to models that are not busy. We also did some improvements to the followup questions that the modules can ask the users and we did some minor refactoring.

The goal for the next sprint, sprint 6. The most important goal is to finish the GUI and test the application so there are no bugs in it. This is most important because there is not a lot of time left so we need to finish everything that we planed on having on the first release. And it has to work fairly well so finding and fixing bug will be a priority. The minor goal of this sprint will be to improve the NLP modules so that they understand a larger range of different ways to write the input. And also to add more documentation about how to use the release and continue working on the project.

/The RPA Tomorrow Team