Week 9 leads the team halfway through the third development sprint of the linguine application. Since last week, some significant changes in ordering to the analysis process have taken place. In order to avoid erraneous states in the application that can be caused by selecting either combinations of preprocessing steps or preprocessing steps that are not applicable to a given type of analysis, the python server will throw a cryptic 500 message and a vague error dialog to the user. We would like to limit system input so that these situations can never occur. This will be accomplished through the following process:
- Reordering the new analysis process to
analysis->corpora->preprocessingto more accurately reflect the process of NLP
- Ensuring that a specific analysis is selected before corpora and preprocessing steps are selected
- Bind certain preprocessing steps to each analysis instead of making all steps accessible
- Ensure that analyses that take a single corpus cannot be passed an entire corpora from the UI
This series of steps will hopefully resolve any error prone states that the application can enter. Once this process has been implemented, the team is going to work towards adding new analysis types into the application.
This week, the team also spent time setting up a local instance on Illinois Curator so that we can build a lightweight Java API to perform analysis using a more advanced toolset. Before the end of Sprint 4, we would like to have the ability to interface with Illinois Curator. We will make sure to limit the amount of code that is written in Java per the sponsor’s request, and instead will delegate to the Python API whenver possible. This could be done through the integration of some new libraries like Spacey.
Documentation & Artifacts for the Week
Sprint 2 Burndown
This sprint the team has fortunately been able to complete all assigned work! A total number of 29 story points (used to estimate the degree of complexity for each task to be done) were completed over the period of two weeks. We would like to work towards completing this amount of work consistently for the rest of the semester and academic year, but the overall amount of work complete per sprint will vary with the team’s courseload.
For a more specific rundown of the work we have completed over the past couple weeks, please take a look at the ‘issue tracker’ link at the top of the page, or the previous timesheets and blog posts linked to on the main page of this site.
The 4-UP charts have been a nice way for the team to keep track of each task that is getting accomplished each week, as well as any risks or needs that the team must address in sponsor meetings. If you would like to see more info on these charts, feel free to reach out to the team and we would be happy to accomodate!
Total hours worked by the team this week: 32.15