Skip to content

A framework to define and operate Dialogflow Agents with a simple, code-first, approach 🔧

License

Notifications You must be signed in to change notification settings

dariowho/intents

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Intents ⛺

Documentation Status codecov HEAD version PyPI version

Intents is a Python framework to define and operate Conversational Agents with a simple, code-first approach. Intents comes with built-in support for Dialogflow ES and experimental Alexa and Snips connectors. Its main benefits are:

  • Agents are Python projects. You will develop with autocomplete, static type checking and everything you are already used to.
  • Versioning and CI. Agents can be versioned on Git, and programmatically deployed just like software.
  • Human-friendly Connectors. Intents are classes, predictions are their instances. Support can be extended beyond Dialogflow by implementing custom connectors.

A detailed view of the available features can be found in STATUS.md. Also, check out the Projects page to keep track of recent developments.

Install

pip install intents

Usage

Intents are defined like standard Python dataclasses:

@dataclass
class HelloIntent(Intent):
    """A little docstring for my Intent class"""
    user_name: Sys.Person = "Guido"
MyAgent.register(HelloIntent)

Their language resources are stored in separate YAML files:

utterances:
  - Hi! My name is $user_name{Guido}
  - Hello there, I'm $user_name{Mario}

responses:
  default:
    - text:
      - Hi $user_name
      - Hello $user_name, this is Bot!

Agents can be uploaded as Dialogflow ES projects directly from code:

df = DialogflowEsConnector('/path/to/service-account.json', MyAgent)
df.upload()  # You will find it in your Dialogflow Console

Intents will act transparently as a prediction client:

predicted = df.predict("Hi there, my name is Mario")
predicted.intent            # HelloIntent(user_name="Mario")
predicted.intent.user_name  # "Mario"
predicted.fulfillment_text  # "Hello Mario, this is Bot!"

For a complete working example, check out the included Example Agent. Also, Intents documentation is published at https://intents.readthedocs.io/ 📚

Disclaimer

This project is not affiliated, associated, authorized, endorsed by, or in any way officially connected with Dialogflow. The names Dialogflow, Google, as well as related names, marks, emblems and images are registered trademarks of their respective owners.

About

A framework to define and operate Dialogflow Agents with a simple, code-first, approach 🔧

Topics

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages