A lot of dilemmas

I think it is time to look for another programmer for this project -- to work on the front-end and the back-end. I can handle the back-end (maybe) but I cannot just work on the front-end. It's just too much for me and I need some front-end frameworks to make all things come together properly.

I have asked some of my friends but none of them seem to know any front-end frameworks. I do know a teammate of mine from the hackathon who is good at react but I am not sure if she's willing to work with me on this project since she is in college and she's in India. Perhaps I will ask her if other options don't work out.

Or perhaps I should build the first version of the site without using any front-end frameworks -- straight up Django? A bit of javascript, bootstrap, and HTML might do the job for the initial launch. Then I can use that as a way to get other people. That might be a good idea for now. But I have to do this soon now. The first deadline is OCTOBER when some A-level exams will go on around the world. But the corona is playing on my part right now by not letting schools be able to conduct exams.

Also need to get someone who can do better AI than me or maybe not. I mean I can handle the AI part of it since it occupies a small portion of the project, yet is one of the biggest linkages to its functionality but if anyone can improve it -- not with data but with its architecture -- that would be the main point. Maybe I should not think about this right now.

Getting to a more technical part although I will use clean English here as much as possible.

Sort of a dilemma during question splitting. I rewrote the entire library again -- it is far better now, works on every A-level question paper on existence but has some bugs or some outliers that would not cause it to crash but rather extract overlapped or skipped questions.
It is mainly due to the paper using the same font on words that do not question numbers. It becomes kinda hard to parse through all those gibberish and find out the questions, sub-questions numbers. That's where some of the outliers can pass through and return some unexpected output.

But the dilemma is whether I should split questions based on their main question number or split based on sub-question numbers as well. Like each sub-question can be another topic of its own -- if classified. So, if we do that, maybe students will get more specific topic-based questions.

But then again, would they really care about this? This is where students' feedback is necessary. Perhaps this would be more useful for teachers while making their past papers because they can really select those sub-questions that matter most to their concerned chapters. Again, need to get teachers' feedbacks.

So, probably I will classify all the questions along with their sub-questions but only use the main questions for the first version. How about that? Then I will get feedback from students and teachers and then improve it.

Yeah, that sounds about right. I guess I can build things but there's always have to have people involved in it somehow.