2 The architecture of the RETINA project

2.1 Main pieces of CODE that make up the RETINA project

  1. The RETINA app
    • primarily php code
    • collects the data from the user
    • shows sensor data to the user
    • shows model run results to the user
  2. The RETINA database
    • PSQL data catalogue dedicated to storing RETINA project specific information such
      • as user login information
      • site details
      • site management
      • records of model runs triggered through the RETINA app
    • The data in the RETINA database is supposed to be separate from the BETY database (described below.)
  3. The PECAN code
    • A bundle of R packages that handle
      • processing and preparing the data collected from the app and other sources to
        • use it as model inputs or
        • compare with model outputs
      • reading model outputs and preparing them for visualizations, comparison with collected data and further analysis
    • For each model that is used in the RETINA project, there is a specific R PEcAn package that contains (at minimum)
      • script for preparing climate data for the model
      • script(s) for preparing configuration (i.e. parameterization) files for the model
      • script to execute the model
      • script to read the output of the model and translate it from model specific format in to a standardized format so that it can be easily comparible with other model outputs
  4. The RETINA R package
    • A series of helper scripts written in R that are specific to the RETINA project and JHI and thus shouldn’t be merged back in with the pecan project
    • These scripts complement the tasks listed more generally with the PECAN code.
  5. The BETY database
    • PSQL database.
    • While BETY does house some data, it is better to think of it as a metadatabase that primarily holds information on the location and format of data
    • The BETY database is used to track the full provinance of all data used when a model run is generated by the RETINA app including:
      • records of all raw data and processed data used for drivers, parameterization and benchmarking (for every step of processing)
      • records of run settings and model specific parameterization for model runs
      • records of model run status and output files