Add intent training to frontend

[zhujessica]

Created a barebones UI to allow users to enter in intents and corresponding training data. Training data must be entered as a list of comma separated values, and entities must be entered in the normal Rasa format: the entity is in square brackets, followed by the entity in parenthesis. Example is: [San Francisco](city). User can add more intents using the Add More button, and clicking Train trains the Rasa model on the new data. Text for 'Done Training' will appear when the model is done training and ready to use in the main conversation box.

[jessvb]

@16JessicaZ Just want to make sure I'm testing this correctly— What's the format for the training data? I.e.., are there semi-colons between each utterance example (e.g., make a recipe; save a recipe; create a [cookie](food) recipe)?

[zhujessica]

@16JessicaZ Just want to make sure I'm testing this correctly— What's the format for the training data? I.e.., are there semi-colons between each utterance example (e.g., make a recipe; save a recipe; create a [cookie](food) recipe)?

The training data should be separated with commas, so make a recipe, save a recipe, create a [cookie](food) recipe

[jessvb]

The training data should be separated with commas, so make a recipe, save a recipe, create a [cookie](food) recipe

Great, thanks!

I realized the main issue was that after I trained the model, I needed to restart the rasa server before it would connect to the newly-trained model. Could you make it so that the rasa server loads in the new model automatically (so that you don't have to manually restart it)?