QuantuMobileSoftware / clearcut_detection

According to the WWF (World Wildlife Fund), forests cover more than 30% of the Earth’s land surface and are considered to be the lungs of the planet. Unfortunately, people are not using this resource wisely and every day an alarmingly high amount of trees are being cut. Whether it is a natural loss of trees or human-driven – deforestation has horrifying consequences.

Clearcut_detection

Project structure info

• clearcut_detection_backend - web-service for clearcut detection
• clearcut_research - investigation about model approach, model training and model evaluation of clearcut detection

Launch requirements:

To start a web-service, do next:

• cd clearcut_detection_backend/

• create db.env based on the db.env.example

• create gcp_config.ini based on gcp_config.ini.dist AREA_TILE_SET value is responsible for region that should be fetched and processedUpdate it if needed. To get tiles Ids you can use https://mappingsupport.com/p2/gissurfer.php?center=14SQH05239974&zoom=4&basemap=USA_basemap

• create key.json based on the key.json.example In order to use Google Cloud Storage you need to generate service account key and put this file inside /clercut_detection_backend folder, file name should be specified inside gcp_config file(key.json by default). for more information about creating key.json read https://flaviocopes.com/google-api-authentication/

• create django.env based on the django.env.dist

• create rabbitmq.env based on the rabbitmq.env.dist

• cd model2/ now You should be in clearcut_detection_backend/model2/

• create model.env file based on model.env.dist, RABBITMQ_DEFAULT_USER and RABBITMQ_DEFAULT_PASS copy from rabbitmq.env created earlier, POSTGRES_USER and POSTGRES_PASSWORD copy from db.env created earlier

• put same key.json generated earlier to clearcut_detection_backend/model2/key.json

• put unet_v4.pth in to clearcut_detection_backend/model2/unet_v4.pth (trained model can be obtained from maintainers)

• Run docker build -f model.Dockerfile -t clearcut_detection/model2 . in order to build model2 docker image

before start

• cd ../clearcut_detection_backend now You should be in clearcut_detection_backend/clearcut_detection_backend/
• In settings.py edit MAX_WORKERS - depends of your CPU cores number
• In prod_settings.py edit email settings
• cd .. now You should be in clearcut_detection_backend/
• In docker-compose-stage.yml edit:
  • CUDA_VISIBLE_DEVICES - set cuda device You Want to use, default -0
  • celery -A tasks worker -Q model_predict_queue --concurrency=2 - set concurrency=x where x number of celery workers
  • VIRTUAL_HOST
  • VIRTUAL_PORT
  • LETSENCRYPT_HOST
  • LETSENCRYPT_EMAIL

At the first start:

• Run docker build -f postgis.Dockerfile -t clearcut_detection/postgis . in order to build postgis docker image
• Run docker build -f django.Dockerfile -t clearcut_detection/backend . in order to build backend image
• Run docker-compose -f docker-compose-stage.yml up -d db_stage in order to run docker for data base
• Run docker-compose -f ./docker-compose-stage.yml run --rm django_stage python /code/manage.py migrate --noinput in order to create all data base tables
• Run docker-compose -f ./docker-compose-stage.yml run --rm django_stage python /code/manage.py loaddata db.json in order to import data from db.json file to data base (if You want to use our demo data base, dont run it if You dont need information about demo regions)
• Run docker-compose -f ./docker-compose-stage.yml run --rm django_stage python /code/manage.py createsuperuser in order to create django superuser

Launch project

• Run docker-compose -f docker-compose-stage.yml up for deployment.

Collect new data from scratch

• In docker-compose-stage.yml edit: START_DATE_FOR_SCAN start date for scan images
• Run docker-compose -f ./docker-compose-stage.yml up -d model_stage_2 - in order to run prediction model
• Run docker-compose -f ./docker-compose-stage.yml run django_stage python /code/update_all.py --exit-code-from django_stage --abort-on-container-exit django_stage
  • in order to run script for collecting images for all tiles from gcp_config.ini

Fetch new data

Use if You have already collected data, and now You want to update it by new information

• Run docker-compose -f ./docker-compose-stage.yml up -d model_stage_2 - in order to run or rerun prediction model
• Run docker-compose -f ./docker-compose-stage.yml run django_stage python /code/update_new.py --exit-code-from django_stage --abort-on-container-exit django_stage
  • in order to run script for collecting new images for all staged tiles from data base

Swagger:

After the app has been launched, swagger for api can be used. Go to http://localhost/api/swagger to access swagger with full description of api endpoints. You have to login as superuser to see all API description.

Whats new

https://github.com/QuantuMobileSoftware/clearcut_detection/blob/master/clearcut_detection_backend/docs/Whats_new