Next Release Post-June-2022

[robertbartel]

Ticket to collect issues that are part of the next release (like to be represented by a new Github Project).

Docker and Infrastructure

• Update ngen docker container build to have netcdf cxx interfaces #162
• Update dependency documentation with Docker details #93
• Update and clean up design for ngen worker image #168
• Further work on Docker images for ngen workers
  • Determine how to design image configurations best to add routing support
  • Ensure routing support is optionally available in ngen-based worker image(s)
  • Add/improve mechanism(s) for controlling specific ngen repo commit
  • Design build, backend, and/or UI mechanisms to support ngen workflows using externally provided/referenced BMI module code
  • Investigate possible support of custom provided Docker image or Dockerfile
• Separate Docker functionality to isolated package
• Review and update Python package build process and dist_package.sh script
• Use Docker Secrets for managing service SSL certificates

Job and Resource Management

• Fully incorporate monitor-service into running main stack
• Enhance job startup monitoring to auto restart hung workers that don't start on their own
• Create (i.e., recreate, but with new names) allocation paradigms that are multiple MPI ranks per container
• Add client package support for requesting a job and designating a subset of the entire hydrofabric
• Dynamically deactivate/reactivate resources on a Swarm node if the node's status switch to/from “pause”

GUI Development Tasks

• Integrate current GUI code with dmod.client package
• Finish GUI views for model configuration
  • Finalize domain selection GUI view
  • Component for configuring "regular" BMI formulations
  • Component for configuring multi-BMI formulations
  • Support the uploading of all required catchment-specific model formulation BMI initialization files

Data Management

• Complete client package/CLI support for dataset upload/download functionality
• Support in data-service for deriving AORC_CSV format datasets that are a subset of the time series of the source dataset
• Support in data-service for deriving AORC_CSV format datasets that are a subset of catchments of the source dataset
• Support in data-service for deriving datasets to fulfill otherwise unfulfillable requirements for jobs (along with updating data availability check functionality)
• Support in data-service for automated routines to clean up temporary datasets
• Make sure the object-store-dataset-based Docker volume for output datasets that gets created as part of a job is cleaned up upon job completion
• Implement mechanism (e.g., different type, different format, new index within format) for distinguishing hydrofabric datasets that have been subdivided (keep in mind that a subdivided hydrofabric depends on both the original hydrofabric and the partitioning configuration)
• Support in data-service to retrieve and process raw forcing data from remote services or locations
• Complete support for datasets providing forcing data in NetCDF format #193
• Cleanup/merge of modeldata.datarequest with modeldata.data

Evaluation, Visualization, and Calibration

• Create DMOD Evaluation service Python package #140
• Build EvaluationClient class within dmod.client Python package for making requests and getting results from evaluation service
• Complete calibration execution workflow
• Build and incorporate DMOD Python package for generating hydrographs
  • Develop new Python library with primary functionality of building hydrographs from output
  • Incorporate Python library for generating hydrographs into job output GUI view
  • Investigate incorporating hydrograph generation Python package directly into Nextgen framework
  • Add CLI functionality to hydrograph generation Python package for easy command line use

Bug Fixes

• Debug issues with subset-service startup
• Invalid use of DOCKER_SECRET_REDIS_PASS #38
• Fix issues with resource release upon job completion

Other Required Items

• Change MaaSRequest to treat version as string #61
• Enhance subset-service to work with multiple hydrofabrics obtained via datasets from data-service
• Catchment grouping constructs for shared configs and catchment aggregations

Optional Items

• Data service support for detecting when a derived, regridded forcing dataset has become "stale" due to hydrofabric changes
• Investigate large use of Redis and Redis stream for job state management and service queuing
• Data service support for retrieving data from remote services or locations (possibly with opendap.catalog)
• Add tool for tracking and managing DMOD-specific node labeling for Docker Swarm service deployment constraints
• Implement SSL for object store
• Port stack control scripts to Python
• Port object store user access init script to Python