Hi everyone!
I am currently working at the development of a conversational agent/chatbot, using Rasa Open Source. I will be working with quite a lot of linguistic data, so I was looking for a good annotation tool in order to label intents and entities. Rasa has got a peculiar way to represent its training data, so I would like to know: after annotating on Prodigy, can I download my annotation file in a .txt format? How is the layout of this file?
Thanks,
Andrea
Hi! Prodigy lets you export your annotations in a simple JSON format, so using it with Rasa should hopefully be very straightforward 
You can see examples of the JSON output generated by the different interfaces here: https://prodi.gy/docs/api-interfaces
For example, if you're annotating spans within a text, Prodigy lets you export the label and the start/end character offsets into the text for each span, which you can use to identify it in the original text. If you're annotating categories that apply to the whole text, the output includes a list of IDs of the selected labels.