ML Model Toggler

[LBlend]

Closes #17

Implements a simple toggle button that toggles between the usage of the naive bayes model and the recurrent neural network model for prediction. As of writing this it is implemented through a react hook but could probably be as easily implemented through raw JS. I decided to use hooks nonetheless. I'm a react noob, and I don't know what is good practice.

This PR has some issues and the following should be considered before merging:

• A lot of code is repeated twice. This is bad practice and should probably be revised
• Text labels showing what model is set is not center aligned (by height) with the button. This should be addressed.

This is a WIP PR

[LBlend]

Closes #17

Implements a simple toggle button that toggles between the usage of the naive bayes model and the recurrent neural network model for prediction. As of writing this it is implemented through a react hook but could probably be as easily implemented through raw JS. I decided to use hooks nonetheless. I'm a react noob, and I don't know what is good practice.

This PR has some issues and the following should be considered before merging:

• A lot of code is repeated twice. This is bad practice and should probably be revised
• Text labels showing what model is set is not center aligned (by height) with the button. This should be addressed.

This is a WIP PR

~* A lot of code is repeated twice. This is bad practice and should probably be revised~

• Text labels showing what model is set is not center aligned (by height) with the button

I fixed one of the two problems listed. The only remaining problem is barely noticeable but should be fixed before merging.

[marksverdhei]

Looks really good but after playing around with it, I realize how much work is left to do on the rnn side of things. the classifier doesnt seem that good, which raises the demand for a validation script/notebook. nevertheless I approve this PR, but i'll let you decide if you still want to push it up to the main page, or stand by until we get a better model or know that the rnn performs better than the naive bayes classifier

[LBlend]

Looks really good but after playing around with it, I realize how much work is left to do on the rnn side of things. the classifier doesnt seem that good, which raises the demand for a validation script/notebook. nevertheless I approve this PR, but i'll let you decide if you still want to push it up to the main page, or stand by until we get a better model or know that the rnn performs better than the naive bayes classifier

Nice! The website runs on the main branch so I'll merge it to dev for now and probably wait for a better model.