Add or modify a SYMFLUENCE data handler — forcing, geospatial attributes, remote-sensing products, or streamflow/observation datasets — across the acquisition → preprocessing → model-ready pipeline. Covers the registry + import-list mechanism, base-class contracts, mixins, config-key dispatch, and output-path conventions.
Add or modify a SYMFLUENCE hydrological model — runner, preprocessor, postprocessor, config manager, and calibration wiring — via model_manifest() and the unified R.* registries. Covers the models/ taxonomy (process-based, conceptual, ML, LSM, framework, routing) and the settings/forcing/simulations path conventions.
Add a new optimization/search algorithm to SYMFLUENCE's calibration loop — subclass OptimizationAlgorithm, implement the callback-driven optimize(), and register it in ALGORITHM_REGISTRY. Covers the algorithm↔worker contract, normalized-parameter conventions, the evaluate/record/update callbacks, and gradient support.
Discover what SYMFLUENCE offers by introspecting the live registry and config schema — available models, forcing/observation datasets, optimizers, calibration targets, metrics, presets, templates, and config keys. Start here before choosing a model/dataset to run, or before extending the platform.
Run the SYMFLUENCE workflow locally — full pipeline or single steps — via the `symfluence workflow` CLI (run/step/steps/resume/status/validate/clean). Covers the 16-step order + aliases, the config-hash stage-marker system (why a step is skipped or auto-re-runs), authoring/validating a config, and the domain_{NAME} layout.
Diagnose SYMFLUENCE calibration/optimization that misbehaves — flat or stuck scores, workers crashing, params not reaching the model, NaN metrics, parallel process-dir cross-talk, or regionalization not varying spatially. Fault-tree for the DDS/PSO/SCE-UA/DE loop and the BaseWorker apply→run→metrics cycle.