Week 5: Markdown-based Semantic Annotation of Workflow Scripts – Project 1

Week 5 – Update

Hello everyone, this is Xiaoliang Jiang from project 1. This is the fifth week for my internship. For this week, I made some changes on my existing functions, including created a new function to generate associations between entities, revised the RDF structure and modified logic structure in my scripts. I also reviewed some useful materials about how to create a better UI and annotations in R codes for users.

One of the most important things I made this week is revising my existing functions. After I uploaded my prototype functions on Github, my mentors tested it at the beginning of this week. In general, the mentors satisfied my working, however, they also provided me some suggestions about how to improve it. For example, originally, I use entities name to generate the RDF file which is not formal enough. And now I use entities’ ID to generate RDF file instead of their names, and added one more line to each entity to represent their titles.

More importantly, my prototype functions need users to type in properties manually to generate associations which is not user friendly, especially for some users who are not experts of semantic workflow. Therefore, I made some changes on existing code and tried my best to generating the properties automatically, since both of my mentors and I think it is not good enough. Originally, my functions use some manual inputs, like “properties=values” to generate the association. And now, I modified my function’s logic to generate the association between entities automatically by using the classes of parent entity and child entity. In this part, the functions could understand whether this is an “one way association” or a “two ways association” and build different relationship to these different association. At last, another graph function will capture these properties to generate RDF diagram automatically. However, I only added a few logic judgements into my function for testing, I will add more in the future.

In addition, in order to create a better UI for the R extension to let users learn how to use it easily, I have contacted with the author of “strcode” package and started to review some useful materials about creating UI in R studio. I reviewed some tutorial about “shinny developer” and got a better understanding about how to revise existing UI.

I also learnt a lot from some slides provided by my mentors about JSON-LD. I think I could try to revise headers of “inserted breaks” in JSON-LD style in the next week.

Leave a Reply

Your email address will not be published.