Skip to content

fastflair/maana-optimization-google-satsolver

 
 

Repository files navigation

Maana Q Knowledge Microservice Template: Python (Ariadne)

Features

Maana Q Client (i.e., peer-to-peer)

It is possible, though not generally preferred, for services to depend directly on other services, passing requests through a secure CKG endpoint. This template includes the necessary authentication code for your convenience. Simply supply environment settings and use the client from the GraphQL context:

#
# ---------------APPLICATION VARIABLES-------------------
#

MAANA_ENDPOINT_URL=


#
# ---------------AUTHENTICATION VARIABLES--------------------
#

# keycloak or auth0
AUTH_PROVIDER=

# URL for auth server (without path), i.e. https://keycloakdev.knowledge.maana.io:8443
AUTH_DOMAIN=

# Keycloak or auth0 client name/id.
AUTH_CLIENT_ID=

# Client secret for client credentials grant
AUTH_SECRET=

# Auth audience for JWT
# Set to same value as REACT_APP_PORTAL_AUTH_IDENTIFIER in Maana Q deployment ENVs)
# NOTE: For use of keycloak in this app, this value should match both the realm and audience values. 
AUTH_IDENTIFIER=

And, in your resolver:

    # A resolver can access the graphql client via the context.
    client = info.context["client"]

    # Query all maana services.
    result = client.execute('''
    {
        allServices {
            id
            name
        }
    }
    ''')

    print(result)

Build

pip install uvicorn gunicorn ariadne graphqlclient asgi-lifespan python-dotenv requests

Containerize

Then you can build your image from the directory that has your Dockerfile, e.g:

docker build -t my-service ./

Run Debug Locally

To run the GraphQL service locally with hot reload:

./start-reload.sh

and visit http://0.0.0.0:4000

For details, please refer to the official documentation.

Run Locally (via Docker)

To run the GraphQL service locally (Via Docker):

docker run -it -p 4000:80 -t my-service

and visit http://0.0.0.0:4000

Run Debug Locally (via Docker)

To run the GraphQL service via Docker with hot reload:

docker run -it -p 4000:80 -v $(pwd):/app my-service /start-reload-docker.sh

and visit http://0.0.0.0:4000

For details, please refer to the official documentation.

Editing

To update any changes made to the service you will need to re run docker build.

Note that if you require additional packages such as pandas and numpy you need to add your packages to the pip install in the Dockerfile. There are a few difficulties with installing these additional packages, this: docker-library/python#381 explains the issue and the resolution.

Deploy

To simplify deployment to your Maana Q Kubernetes cluster, use the CLI mdeploy command:

gql mdeploy

and follow the prompts.

Basic Query

{ solveLinearCPProblem( vars: [ { id: "x", lowerBound: 0, upperBound: 50 } { id: "y", lowerBound: 0, upperBound: 50 } { id: "z", lowerBound: 0, upperBound: 50 } ] constraints: [ { id: "ct" lowerBound: 0 upperBound: 50 coefficients: [{ id: "x", value: 2 }, { id: "y", value: 7 }, { id: "z", value: 3 }] } { id: "ct2" lowerBound: 0 upperBound: 45 coefficients: [{ id: "x", value: 3 }, { id: "y", value: -5 }, { id: "z", value: 7 }] } { id: "ct3" lowerBound: 0 upperBound: 37 coefficients: [{ id: "x", value: 5 }, { id: "y", value: 2 }, { id: "z", value: -6 }] } ]

objective: {
  id: "obj"
  coefficients: [{ id: "x", value: 2 }, { id: "y", value: 2 }, { id: "z", value: 3 }]
  maximize: true
}

) { id status objectiveValue varValues { id } } }

About

simple constraint programming using google or tools Sat Solver

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 73.1%
  • Shell 23.8%
  • Dockerfile 3.1%