MarkUs Blog

MarkUs Developers Blog About Their Project

Weekly Status Update: Oct. 5, 2015

without comments


1)  Looked a little more into the the results of the branch merge. There’s definitely a few things which are broken but it’s not immediately clear what has/hasn’t been implemented. Wilson and I have informally documented a couple of things which aren’t working right now
2) –
3) Schedule a meeting with Karen, David and Wilson to set some milestones for the project. It’s still not super clear to me what has already been built and where we’d like to be at the end of the term. So I think meeting up to set a few intermediate markers will be helpful (I’ll email).


**This Week**

– Added seed data for marks spreadsheets.

– One spreadsheet is visible, with marks entered, released to students, dated immediately

– Another is hidden, has no marks entered, has not been released to students, dated two months from seeding time

– Investigated rubocop pre-commit file


– Learned that the precommit script will only run on files that have been added (staged) for a commit

**Next Week**:

– TA Analytics


1) This week I worked on converting the last .rjs files in the views folder into .erb files. This is almost done now, 7 files have been deleted/converted and only 2 more remain.

2) I ran into a roadblock that came from my lack of Rails knowledge. I realized that a lot of the .rjs files I had to convert weren’t even being used anymore. Originally I thought they were just not working properly, but after investigating the current codebase and checking in the project’s git history I found that in some cases the method or action that the .rjs file was related to in the controller was removed. This resulted in me realizing I could just delete a lot of the .rjs files that still exist (especially in the grades and assignments view). After that I was back on track.

3) This week I will finish up those last two .rjs files and submit a pull request for them (hopefully by mid-week). Then I’ll be moving onto the app/assets/javascript/ folder and working on the manage_groups.js and manage_graders.js files. I also want to spend some time looking into the Gitolite issue I had with my setup that hasn’t been fixed after the sprint weekened. David sent me a helpful resource so I’ll start there


1) Implemented a different fix to fix the broken tags list. Also setup Jeremy’s merged test branch and investigated into what is working / what is not and documented it.
2) Mostly just getting used to how RoR handles migrations. Ran into a few issues with dependencies on the test-framework branch but got those worked out.
3) Jeremy and I are going to have a meeting with David and Karen this week regarding the exact scope and features wanted for the testing framework so we can figure out how to tackle it.

The one I posted before seemed a little extra so I narrowed it down.
1) Last week with the little assessment i did, my plan to tackle the full issue of groupings in one go was unsuccessful, so I decided to break it down. This week I came up with an approach to taking the repo_name in to consideration when creating groups through a csv file upload.
2) I ran in to a lot of road blocks where I didn’t know which step to take, for example, do I reuse the group when a repo name is found etc. Also I didn’t know the existence of hooks, so it took me some time to figure out the actual flow of methods/objects.
3) This week I  plan to continue with the repo names in the csv file group uploads (edited)


Last week I made the annotations rotate when the JPG/PDF gets rotated. This was trickier than rotating just the image because the annotations are generated by coordinates, so some math needs to be done. For PDF, it wasn’t too bad because the coordinates are stored as percentages local to the PDF boundaries, but for JPG, the coordinates are in pixel space relative to the entire page. Also, rotating JPGs that aren’t square can cause them to overlap with the panel boundaries, so I added an extra translation to align them with the edges of the panel.Although I got the annotations rotating to the correct positions, I found some bugs in the annotation/image viewing code(#2308, #2307, #2306). For next week though, I’d like to fix some issues with my own code first. In no particular order:

If the grader tries to make an annotation while the image is rotated, the coordinates are stored in the db as if the image isn’t rotated.
For rotated annotations, the text sometimes doesn’t show up when the mouse is hovering over it.
Large JPG files with scrolling have wonky behaviour when the img element is rotated. The scroll bars aren’t updated to swap the width/height, and rotating the annotations on the img can bring them outside of the iamge viewing panel, which is right, but I need to figure out how to hide them until the grader scrolls over and update their positions as they scroll.
I think there’s some sort of mask at the bottom of the JPG view that needs to be resized upon rotation. When an image is uploaded in landscape orientation and rotated to portait, the bottom of the image is cut off by blank white space.

Written by William Kwan

October 5th, 2015 at 3:25 pm

Leave a Reply