TERRA-REF uses a suite of databases and software components that are described below.
Clowder (sensor data and computation management with web user interface)
Clowder is the primary system used to organize, annotate, and process raw data generated by the phenotyping platforms as well as information about sensors. Use Clowder to explore the raw TERRA-REF data, perform exploratory analysis, and develop custom extractors. For more information, see Using Clowder.
Globus Connect (large data transfer)
Raw data is transferred to the primary TERRA-REF compute pipeline using Globus Online. Globus also provides access to TERRA REF files, but this is not a primary portal and metadata in Clowder may be required to locate and interpret these files. Use Globus Online when you want to transfer data from the TERRA-REF system for local analysis by accessing the Terraref endpoint. For more information, see Using Globus.
BETYdb (phenotype data)
BETYdb is a database and web interface to the trait / phenotype data and agronomic metadata. This is where you can find plant and plot level trait data as well as plot locations and other information associated with agronomic experimental design. Use BETYdb to access derived trait and agronomic data. For more information, see Using BETYdb.
Algorithms (a.k.a. 'extractors')
Plant CV is an imaging processing package specific for plants that is built upon open-source software platforms OpenCV, NumPy, and MatPlotLib. Plant CV is used for trait identification, the output is stored in both Clowder and BETYdb.
Each step in the pipeline is performed by an algorithm. These are maintained in the TERRA REF GitHub organization in repositories with names that begin in
extractors-* such as github.com/terraref/extractors-hyperspectral.
The NDS Workbench enables users to access the large filesystem and databases with familiar development environments. We provide a variety of environments for developing new algorithms and integrating them into the TERRA REF pipeline. These include RStudio and Jupyter Notebooks configured for specific use cases such as sensor data processing, trait analysis, database queries, and piepline development.
CoGe contains genomic information and sequence data. For more information, see Using CoGe.