When I add multiple entity types, the output is incorrect.

For example, if I ask "do you accept cigna?"

The output is like this:
{ action: 'payment', args: { service_offered: 'cigna' insurance_company: 'cigna' form_of_payment: 'cigna' lender: 'cigna' } }

I would expect this output:
{ action: 'payment', args: { insurance_company: 'cigna' } }

Hi - this project looks really interesting and I'd like to try building something based on it, but there's no licensing info in the project -- could you add a license file to the project, or a note in the README, to make clear the terms of use? Thanks!

hi Matt

Is there a way to get an explanation on how this works? Specifically, where in the code is the intent output?

Thanks very much for any lead!
Lior

can we have model which distinguishes sentence with intent from a given a input text ?

like below example :
"Turn the bathroom light off and remind me to take the trash out."
There are two independent intents here:
turn_lights, set_reminder.
similarly another example : "Hey what's up? do you guys offer free trial?"
there are two intents here : greetings , product_pricing

Thanks & Regards,
Achyuta nanda Sahoo

Hi Matt. Do you know any public corpus of sentences and intents? It would be great to compare opennlp with other services. I think that you don't have enough examples but I haven't found anything better/bigger yet.

Hello, first of all, I'd like to thank you for your work on this repo, it really helps me out as it's just what I need !
I have been trying for a week or two to make a webserver based on your repo but as I don't know Java at all, it hasn't been too easy to make it work...

I could have asked you how to do that two weeks ago but I dont like to give up. So I forked your repo and tried to build some sort of webserver.
I first tried to wrap the java application in a ruby app, that did not work, I tried to wrap it in a nodejs app, that did not work either. Actually, when wrapped in an other application, java doesnt seem to be able to find the right console file descriptor so it kind of waits forever regardless of the input in the console.

I came up with a solution that works but is not very efficient :
I setup an expressjs app, when a user sends a http request, node generates a hash and crate a files with the name [hash].source, in the mean time, the java application, watches a specific folder and awaits for a .source file to appear, as it appears, it reads the file, does whatever it has to do with the sentence, then crates a file with the same [hash].result. While java is doing this, node is waiting for a [hash].result file, when it finds it, it builds the output, send a response to the user and save the log.

Anyways, now that I have learnt java a little bit through all the changes I have made in your code, I'd like to know if there is a way with maven to easily build a webserver. Would you be able to help me with this ot is it just a huge pain in the butt ?

Thank you in advance for your response, and thank you again for your great work

PS. : my webserver works very well, here is the repo : https://github.com/bastienbot/nlp-intent-toolkit