RISICO comes with two flavours:
- org.lexis.wp7.RisicoTemplate performing a RISICO post-processing without observations data in input
- org.lexis.wp7.RisicoObservationsTemplate performing a RISICO post-processing with observations data in input.
The Run workflow for the RISICO template is executing the following steps:
- getting details on the input dataset (size, locations) in DDI containing static geographical data
- asking the Dynamic Allocation Module (DAM) to select the best Cloud infrastructure where to transfer these input data
- transferring the geographical static data input dataset from DDI to the selected Cloud Staging Area
- creating a Cloud Compute instance
- then, in parallel:
- waiting for the end of geographical static data input dataset transfer on cloud staging area and SSHFS-mounting the cloud staging area
- installing and starting Docker on the compute instance
- downloading Global Forecast System (GFS) data:
- from https://nomads.ncep.noaa.gov for recent dates (less than 10 days ago)
- from https://rda.ucar.edu for historical data (more than 10 days ago)
- once Docker is installed and started, a container downloading observations data provided by CIMA is executed
- once the GFS data is downloaded, docker started, and the geographica input data filesystem mounted, the WPS GFS docker container provided by CIMA is executed
- once the preprocessing is done, the size of results is computed, and the Dynamic Allocation Module (DAM) is asked to select the best HPC infrastructure where to create a WRF computation job
- a WRF HEAppE job is created on the selected HPC infrastructure
- pre-processing results are transferred to this WRF job
- the job is then submitted, and the orchestrator monitors its execution until it ends
- once the job is done, WRF results are compressed and stored in DDI
- WRF results are also transferred to the compute instance for post-processing
- then, a RISICO container provided by CIMA is executed
- results are then transferred from the cloud staging area to DDI
- if a Dewetra SFTP server was specified by the user, the results will also be uploaded to Dewetra SFTP server
- finally, the cloud staging area is cleaned and the cloud compute instance is released
The template expects the following input properties (mandatory inputs in bold):
- token: OpenID Connect access token
- project_id: LEXIS project identifier
- preprocessing_start_date: Start date of the simulation, format YYYYMMDDHH (warmup data from the 2 previous days will be also downloaded)
- preprocessing_docker_image_gfs: Pre-processing container repository path
- default:
cimafoundation/wps-da.gfs:v2.0.3
- default:
- preprocessing_docker_image_observation_data: Repository path of container downloading observation data
- default:
cimafoundation/lexis-download-docker:v1.2.5
- default:
- preprocessing_dataset_path_geographical_data_path: Dataset containing geographical data
- default:
project/proj2bdfd9ccf5a78c3ec68ee9e1d90d2c1c/055b25ea-ba60-11eb-a44e-0050568fc9b5/static_geog_data.tar.gz
- default:
- preprocessing_decrypt_dataset_geographical_data: Should the input dataset be decrypted
- default:
false
- default:
- preprocessing_uncompress_dataset_geographical_data: Should the input dataset be uncompressed
- default:
true
- default:
- computation_heappe_command_template_name: HEAppE Command Template Name
- default:
WRF Generic
- default:
- computation_email_for_HEAppE_notifications: E-mail where to receive notifications from HEAppE about the computation job (default, no email)
- default:
""
- default:
- postprocessing_image: Post-processing container repository path
- default:
laurentg/risico:1.0.1
- default:
- postprocessing_ddi_project_path: Path where to transfer the post-processing results in DDI
- default:
project/proj2bdfd9ccf5a78c3ec68ee9e1d90d2c1c
- default:
- postprocessing_encrypt_dataset_result: Encrypt the result dataset
- default:
false
- default:
- postprocessing_compress_dataset_result: Compress the result dataset
- default:
false
- default:
- postprocessing_dewetra_sftp_server_ip: IP address of a Dewetra SPTP server where to store results (default, no sftp server upload)
The following output attribute is provided:
- attribute
destination_path
of componentCloudToDDIJob
: DDI path to Risico post-processing results - attribute
destination_path
of componentHPCToDDIJob
: DDI path to Risico WRF results