It is always returning {} in my case. How does this part even know that what my user typed in is the missing parameter/s that a particular intent needs.
How would one even debug this project? The loggers do not even seem to be working when running the app directly with python run.py, and on top of that the whole API breaks (returning all 404s for the angular client) when the project is run directly with python run.py from within the docker container.
It looks like an interesting project, and the classifiers are well put together and can be repurposed in other projects. But its just so hard to understand the application flow, especially in endpoint/controllers.py. What is the purpose of the complete flag? Why are NER entities being used to figure out which parameters not fulfilled and against what intent? The current one or the previous/original one that asked for the parameters?
I gave up after trying to figure Flask code for a few days. I ended up just borrowing the classifiers and writing out my own speech context processor to manage missing parameters and intent changes.
Can someone explain how this works exactly https://github.com/alfredfrancis/ai-chatbot-framework/blob/master/app/endpoint/controllers.py#L96-L97 ?
It is always returning
{}in my case. How does this part even know that what my user typed in is the missing parameter/s that a particular intent needs.How would one even debug this project? The loggers do not even seem to be working when running the app directly with
python run.py, and on top of that the whole API breaks (returning all 404s for the angular client) when the project is run directly withpython run.pyfrom within the docker container.It looks like an interesting project, and the classifiers are well put together and can be repurposed in other projects. But its just so hard to understand the application flow, especially in
endpoint/controllers.py. What is the purpose of thecompleteflag? Why are NER entities being used to figure out which parameters not fulfilled and against what intent? The current one or the previous/original one that asked for the parameters?I gave up after trying to figure Flask code for a few days. I ended up just borrowing the classifiers and writing out my own speech context processor to manage missing parameters and intent changes.