Pipeline to aggregate data from Climatology Lab

Source code

Workflow

Description

This workflow downloads NetCDF datasets from University of Idaho Gridded Surface Meteorological Dataset, aggregates gridded data to daily mean values over chosen geographies and optionally ingests it into the database.

The output of the workflow are gzipped CSV files containing aggregated data.

Optionally, the aggregated data can be ingested into a database specified in the connection parameters:

• database.ini file containing connection descriptions

• connection_name a string referring to a section in the database.ini file, identifying specific connection to be used.

The workflow can be invoked either by providing command line options as in the following example:

toil-cwl-runner --retryCount 1 --cleanWorkDir never \ 
    --outdir /scratch/work/exposures/outputs \ 
    --workDir /scratch/work/exposures \
    gridmet.cwl \  
    --database /opt/local/database.ini \ 
    --connection_name dorieh \ 
    --bands rmin rmax \ 
    --strategy auto \ 
    --geography zcta \ 
    --ram 8GB

Or, by providing a YaML file (see example) with similar options:

toil-cwl-runner --retryCount 1 --cleanWorkDir never \ 
    --outdir /scratch/work/exposures/outputs \ 
    --workDir /scratch/work/exposures \
    gridmet.cwl test_gridmet_job.yml 

Inputs

Name	Type	Default	Description
proxy	string?		HTTP/HTTPS Proxy if required
shapes	Directory?		Do we even need this parameter, as we instead downloading shapes?
geography	string		Type of geography: zip codes or counties Valid values: “zip”, “zcta” or “county”
years	string[]	['1999', '2000', '2001', '2002', '2003', '2004', '2005', '2006', '2007', '2008', '2009', '2010', '2011', '2012', '2013', '2014', '2015', '2016', '2017', '2018', '2019', '2020']
bands	string[]		University of Idaho Gridded Surface Meteorological Dataset bands
strategy	string	auto	Rasterization strategy used for spatial aggregation
ram	string	2GB	Runtime memory, available to the process. When aggregation strategy is auto, this value is used to calculate the optimal downscaling factor for the available resources.
database	File		Path to database connection file, usually database.ini
connection_name	string		The name of the section in the database.ini file
dates	string?		dates restriction, for testing purposes only
domain	string	climate

Outputs

Name	Type	Description
registry	File
registry_log	File
registry_err	File
data	array
download_log	array
download_err	array
process_log	array
process_err	array
ingest_log	array
ingest_err	array
reset_log	array
reset_err	array
index_log	array
index_err	array
vacuum_log	array
vacuum_err	array

Steps

Name	Runs	Description
init_db_schema	[‘python’, ‘-m’, ‘nsaph.util.psql’]	We need to do it because of parallel creation of tables
make_registry	registry.cwl	Writes down YAML file with the database model
init_tables	sub-workflow	creates or recreates database tables, one for each band
process	gridmet_one_file.cwl	Downloads raw data and aggregates it over shapes and time