Week 7: Markdown-based Semantic Annotation of Workflow Scripts – Project 1

Week 7 – Update

Hello everyone, this Xiaoliang Jiang from project 1. This is the seventh week for my internship. As the project approaches the end, for this week, I revised several existing functions and added some new to improve the stability of my whole R extension package, including Integrated input style of breaks, revised manually input function and changed several function logics.

In the past few weeks, in order to insert more meaningful breaks, my mentors suggested me to try several different input styles like JSON-LD. As this is just an attempt, so after I added it into my function, I did not have time to integrate different input method. So, at the beginning of this week, users can generate RDF file only in default style. In this week, I added a new function before generating RDF file to change different input styles into one format. Then, users could generate RDF file and diagram regardless input breaks styles.

In addition, I also revised my manually input function to let it could serve as a supplement for automatically generating association. Originally, I tried to use manually input function to generate the associations between entities, however, after discuss this with my mentor, I realized that it is not very user-friendly to someone who are not an expert of RDF or semantic workflow. Therefore, I created a new function to generate associations automatically. But there are also some problems, such as the association judgement library may not contain every relationship that users used. So, I added two default relationship to represent these unknown associations, and added the manually input function again. For the revised input function, if users input association property, it will replace the default relationship by using this manually input association which could help users define their semantic workflow much better.

Finally, I changed several function logics, such as adding a filter before generating the RDF diagram to filter duplicate association. I also revised outputting function to avoid same problem in generated RDF file if using manually input to replace original default association.

Although I made a lot of changes and revised many things on my existing function, it still have some defects that I need to think about. So, in the next week, I will keep revising it to let it more stable. In addition, I should start to write a documentation for my function to improve its reusability in the future.