andrewgy8 / apprentice Goto Github PK

We should show any user how to update the cloud function via the cli

We should be able to allow the api to accept a hyperlink or text with photos to display to the google assistant. That way a user can click on the link and find more information about whatever the api received.

Ex. This would be the link information in the today in history api

It seems that travis is trying to deploy the package with every version of python declared. For instance, https://travis-ci.com/andrewgy8/apprentice/builds/93792889

The build succeeds with 3.7, but with 3.6 and 3.5, it fails. It shouldn't be deploying with all the python versions, only one.

A how-to video embedded in the projects readme would be a nice way to break the entry barrier

We should be able to support more than one language.

The approach could use gettext or some other il8n solution.

For the local server to work as plug-and-play, it would be nice to have .env file that is read as a default to the installed main.py flask app. That way a user could execute:

1. apprentice init
2. apprentice run

And have the local server running immediately.

Dialogflow supports rich text responses for users on a device. We could add a nice interface for apprentice to construct this type of response.

Be able to handle and argument such as:

apprentice init --name=foo_bar

and the output should be a direct called /foo_bar with a main.py file containing a foo_bar function.

Includes:

• how to get started
• defining the agent
• deploying the function
• testing the agent
• how to automate testing

Instead of

export FLASK_APP=example/main.py
export FLASK_DEBUG=1
flask run

a user should be able to run apprentice local to run a local web server.

We should be using serailizers to ingest and digest the api responses. That way we can easily update the values.

Provide arguments to allow the user to specify what type of response she wants.

Use case: "Tell me about a birth that happened in history"

I user should be able to specify what type of logging they want from the functions in apprentice in order to debug.

This is what I could imagine the API for Apprentice looking like:

from apprentice import Apprentice

apr = Apprentice(__name__)

ENTITIES = {
    "name": "History",
    "entries": [
        {
            "value": "history",
            "synonyms": [
                "history",
                "past event",
                "the past",
                "past"
            ]
        }, {
            "value": "birth",
            "synonyms": [
                "birth",
                "births",
                "born"
            ]
        }
    ]
}


@apr.action(entities=['name'])
def cool_fact_generator(*args, **kwargs):
    reply = _fact_response(apr.entity)
    return apr.make_repsonse(reply)

When creating a response a user should be able to configure the view of suggestions.

The resulting payload would look like:

'payload': {
                'google': {
                    'expect_user_response': False,
                    'is_ssml': True,
                    'permissions_request': None,
                    "richResponse": {
                        "items": [],
                        "suggestions": [],
                        "linkOutSuggestion": {}
                    }
                }
            }

We should allow users to define entities in the project.

Following what Assistant did in their project, we should allow the user to define intents in their project

It would be something like this https://docs.travis-ci.com/user/coveralls/