Skip to content

Latest commit

 

History

History
executable file
·
50 lines (38 loc) · 1.69 KB

README.md

File metadata and controls

executable file
·
50 lines (38 loc) · 1.69 KB
Raw

PCSE

Repository for the Python Crop Simulation Environment

Dependencies

If you git clone https://github.com/feliksik/ggcmi in ~/basedir/, the directory ~/basedir/ggcmi is created. The Docker build file assumes that the following extra files/dirs exist (currently you have to put them there manually).

  • basedir/ggcmi/deps/
  • basedir/ggcmi/deps/futil.so
  • basedir/ggcmi/deps/netcdf-4.3.2.zip
  • basedir/ggcmi/deps/netcdf4-python-master.zip
  • basedir/data/
  • basedir/data/Dump20141114.sql
  • basedir/data/OtherInputs/...
  • basedir/data/AgMERRA/...
  • basedir/data/geodata/...
  • basedir/output/

Building

You can run this repository in a docker container as follows:

  • install docker; on ubuntu you can do
    • curl -sSL https://get.docker.com/ubuntu/ | sudo sh
  • build the docker image
    • docker build -t ggcmi .
    • this will make it visible in the listing of docker images

RUNNING

  • edit config.sh to your liking, and run:
    • ./start-worker.sh
    • it will run a worker with the configuration of config.sh
    • if your config.sh specifies ggcmi-mysql as the DB_HOST, this means you wish to run a local Docker-based mysql database. It will be started, if it does not already exist.

Congratulations, you are now running the Global Gridded Crop Model Intercomparison (GGCMI) in a test environment!

If you want to run with multiple worker nodes, you go as follows:

  • set up a MySQL database with the worker data on a central server
  • on every worker node:
    • install:
      • the docker image
      • the required data directory
      • the config.sh configured to use the remote MySQL host (it is identical on each worker node)
    • ./start-worker.sh
    • relax, and see results coming in