GEOAP

About GEOAP Platform

GeoAnalytics Platform (GeoAP) provides ready-to-use components to process satellite and drone imagery with AI-powered methods. The GeoAP platform meets the needs and requirements of almost all AI or CV-based application lifecycle participants: from developers to end-users, from development to operations. For additional information please visit https://quantumobile.com/geo-analytics/geo-analytics-platform

Prerequisites and run with docker

• Docker and docker-compose are needed for running the app. Here are the instructions:

https://docs.docker.com/engine/installation/#platform-support-matrix

https://docs.docker.com/engine/installation/linux/linux-postinstall/

https://docs.docker.com/compose/install/

To run app make the next steps:

cd sip/
docker-compose up -d

Go to localhost:3000.

To reload add make the next:

docker-compose exec webapplication touch uwsgi.ini

To run tests make the next:

docker-compose exec webapplication ./manage.py test

API

Go to localhost:9000/api/docs

Swagger

Go to localhost:9000/api/swagger

Admin

Go to localhost:9000/admin/

username: admin
password: AiL9uumi

Test user

username: test
password: jooZee9i

Usage of apikey

curl -X POST "http://localhost:9000/api/login" -H "accept: application/json" -H "Content-Type: application/json" -d "{  \"username\": \"test\",  \"password\": \"jooZee9i\"}"
{"key":"507928511085e6684cfa105c74339d46fc097ad0"}
curl -H  "accept: application/json" http://localhost:9000/api/users/current?apikey=507928511085e6684cfa105c74339d46fc097ad0
{"pk":2,"username":"test","email":"","first_name":"","last_name":""}

Work with registry.quantumobile.co

Authorization on registry.quantumobile.co

Local authorization

To auth locally visit utility special endpoint to receive bash command for authorization with login and password included, like:

docker login --username ***********@quantumobile.com --password ***************************************************** https://registry.quantumobile.co

Authorizing k8s cluster to pull images from local registry

Check k8s config for more information.

Building & Pushing images

Web application

To build image for web application use command: docker build -t registry.quantumobile.co/sip-web-application:latest -f ./webapplication/prod.Dockerfile ./webapplication

To push image into registry use command: docker push registry.quantumobile.co/sip-web-application:latest

Webserver

To build image for webserver use command:

docker build -t registry.quantumobile.co/sip-web-server:latest -f ./webserver/prod.Dockerfile ./

To push image into registry use command:

docker push registry.quantumobile.co/sip-web-server:latest

Copying images into kind cluster

To copy images from host to kind cluster use command:

kind load docker-image registry.quantumobile.co/sip-web-server:latest registry.quantumobile.co/sip-web-application:latest

Working with k8s

Check k8s config for more information.

Settings&Environment

Django settings

---

SENTINEL2_GOOGLE_API_KEY

JSON file. Google Cloud Console Service account key. Required to retrieve Sentinel2 data from Sentinel-2 data storage.

SENTINEL1_AWS_CREDS

JSON file. AWS service account key. Required to retrieve Sentinel1 data from Sentinel-1 storage.

SCIHUB_CREDS

JSON file. Copernicus Open Access Hub credentials: username&password. Like:

{   
    "user": "username",
    "password": "password"
}

NOTEBOOK_EXECUTION_ENVIRONMENT

Defines how geoap's services run:

• docker - Standard. Single machine environment with notebook execution and results publishing as separate threads. Component execution performs as docker containers.
• alternative one - k8s variant with executor and publisher as a separate deployments. Component execution performs as k8s jobs entities.

NOTEBOOK_JOB_BACKOFF_LIMIT

Number of component execution job creation retries.

NOTEBOOK_VALIDATION_JOB_ACTIVE_DEADLINE

DEPRECATED Time in seconds before validation job fail

K8S_NAME_SPACE

K8s namespace to run execution jobs in k8s cluster

IMAGE_PULL_SECRETS

Name of k8s secret with docker credentials to pull images with components into execution job's pods

NOTEBOOK_EXECUTOR_MAX_JOBS

Max number of simultaneous execution jobs

NOTEBOOK_POD_DATA_VOLUME_MOUNT_PATH

Path to mount volumes with inputs and outputs inside execution job's pods

GPU_CORES_PER_NOTEBOOK

Max number of GPU cores as a nvidia.com/gpu resource for execution job's pods