Week 6: Markdown-based Semantic Annotation of Workflow Scripts – Project 1

Week 6 – Update

Hello everyone, this Xiaoliang Jiang from project 1. This is the sixth week for my internship. For this week, I kept trying to made some changes on my existing functions, including tried a new structure of my inserted breaks, revised user interface, and modified my function logic. At the end of this week, my mentors also joined a meeting with professor Ludäscher to discuss my project with him.

From two-weeks ago, when I generate my first prototype RDF output file, my mentor and I were thinking a new way to create the breaks which could let users understand it easily. My mentor provided me a suggestion that using JSON-LD style structure to create the breaks, which could let the inserted breaks become more meaningful. Therefore, based on this idea, I tried to change the basic level of my function to change this breaks style. When I tried to modify it, I was stuck in a structure problem that the function always keep the title and classes of breaks in a pair of parentheses, but after I tried several different ideas, finally I solved it and generate a JSON-LD breaks successfully.

After that, I decided to add this function into existing UI to let users choose the type they want to generate between default style and JSON-LD style. But due to my limited understanding about “shiny” package which is the mainly tool created the existing UI, I failed to add a checkbox into the UI as I want. So, I tried to email Lorenz who is the author of “strcode” to get some help. Lorenz is a very nice person who not only provided me several useful tutorials about “shiny”, but also taught me how to change the user interface in detail. Thanks to his assistance, finally I revised existing UI and test its function successfully.

In this morning, my mentor and I join a meeting with professor Ludäscher who is an expert on scientific workflow in our school. My mentor and I introduced our project with him and his team, and we also discussed a lot about the relationship between my project and their YesWorkflow project. I learnt a lot and got some new ideas from this meeting.

For the next week, I think I have a lot of things to do: Firstly, I need to make some change on the manually input function (properties=values, the “key-value” pairs) to let it serves as a supplement of automatically generated association. Secondly, I need to add a function to generate RDF file (and graph) with JSON-LD style breaks. As I just added this style into my function yesterday, I did not let it works well in generating outputs. Thirdly. I need to think about how to generate an RDF file in JSON-LD style by using JSON-LD breaks. As they are similar, I think it would not be too difficult. Finally, I should think about rebuild the logic structure of my functions to meet the need of further revising.


Leave a Reply

Your email address will not be published. Required fields are marked *