hey there it is me again.
Has you probably should know I am participating in SoK 2020, see this post first otherwise. We are near the end of the project, and I had like to say what more I have done after my first post. you can see all the commits in this repo. Recently I have made an Merge Request(!1) to apply all the commits that I have done.
Things that I want to point it out:
- Auto save functionality
- Fix of the function that added new classes
- Navigating through items with arrow keys
- Support to images bigger than 1280 X 960
Auto save functionality
To this one I have made an checkable action in the menu “edit”, you can select it if you want to auto save a json or xml file automatically in the current working directory. This functionality can be pretty handy when annotating a big amount of items.
Fix of the function that added new classes
A bit of context, when loading a temporary state or a json/xml file, the marked classes name were duplicate (actually nth-ed) in the combo box that contained them. I am mentioning this one because it made me think a lot, with the help of Caio it turned out to be pretty simple to solve.
Navigation through items with arrow keys
This one is useful too, I got the idea of doing this because, as an user, I wanted navigate through items more easily, so I thought that using the arrow key up and down may be the right choice for the job.
Support to images bigger than 1280 X 960
The annotation for images smaller than these proportions works fine but when bigger it turn it out to be in the wrong place. To solve this one, a bit of math was needed.
Refactoring the code can be a lot harder than writing, but it is a must if you are a developer that not only want the code to work as intended but also be maintainable and more likely to get help from another person.
I thought a lot in how to make what I have wrote more understandable and show the intent of the code, although it was improved a lot by Caio later, it was a good experience and for sure I improved as developer.
It is also worth mentioning that we now are focusing in making the API more stable to later improve image annotation and also implement support to text and others data types format as well.
I want to say thanks to Tomaz Canabrava, Sandro Andrade and, of course, my mentor in this project Caio Jordão, I have learned a lot with their help during this period and I will continue to contribute to this community and learn more as well.
That is it, see you soon (hopefully in my whoAmI post ; )