sources:

  conus404-catalog:
    args:
      path: 'https://raw.githubusercontent.com/hytest-org/hytest/main/dataset_catalog/subcatalogs/conus404-catalog.yml'
    description: 'Catalog for CONUS404 datasets'
    driver: intake.catalog.local.YAMLFileCatalog
    
  conus404-drb-eval-tutorial-catalog:
    args:
      path: 'https://raw.githubusercontent.com/hytest-org/hytest/main/dataset_catalog/subcatalogs/conus404-drb-eval-tutorial-catalog.yml'
    description: 'Catalog for CONUS404 DRB tutorial intermediate datasets'
    driver: intake.catalog.local.YAMLFileCatalog

  nhm-v1.0-daymet-catalog:
    args:
      path: 'https://raw.githubusercontent.com/hytest-org/hytest/main/dataset_catalog/subcatalogs/nhm-v1.0-daymet-catalog.yml'
    description: 'Catalog for National Hydrologic Model version 1.0 model output variables with Daymet v3 forcings'
    driver: intake.catalog.local.YAMLFileCatalog

  nhm-v1.1-c404-bc-catalog:
    args:
      path: 'https://raw.githubusercontent.com/hytest-org/hytest/main/dataset_catalog/subcatalogs/nhm-v1.1-c404-bc-catalog.yml'
    description: 'Catalog for National Hydrologic Model version 1.1 model output variables with CONUS404 bias-corrected forcings'
    driver: intake.catalog.local.YAMLFileCatalog

  nhm-v1.1-gridmet-catalog:
    args:
      path: 'https://raw.githubusercontent.com/hytest-org/hytest/main/dataset_catalog/subcatalogs/nhm-v1.1-gridmet-catalog.yml'
    description: 'Catalog for National Hydrologic Model version 1.1 model output variables with Gridmet forcings'
    driver: intake.catalog.local.YAMLFileCatalog
    
  trends-and-drivers-catalog:
    args:
      path: 'https://raw.githubusercontent.com/hytest-org/hytest/main/dataset_catalog/subcatalogs/trends-and-drivers-catalog.yml'
    description: 'Catalog for USGS Trends and Drivers project data. Note: This data is stored in a credentialed bucket; please contact the HyTEST project if you need credentials to read this data.'
    driver: intake.catalog.local.YAMLFileCatalog
    
  nhm-prms-v1.1-gridmet-format-testing-catalog:
    args:
      path: 'https://raw.githubusercontent.com/hytest-org/hytest/main/dataset_catalog/subcatalogs/nhm-prms-v1.1-gridmet-format-testing-catalog.yml'
    description: 'Catalog of NHM-PRMS v1.1 with Gridmet Forcings data stored in various file formats, used for testing purposes with the USGS National Modeled Water Atlas Project. Note: This data is stored in a credentialed bucket; please contact the HyTEST project if you need credentials to read this data.'
    driver: intake.catalog.local.YAMLFileCatalog

  nwis-streamflow-usgs-gages-onprem:
    driver: zarr
    description: "Streamflow from NWIS, extracted and rechunked into time series (NWM2.1 time period)"
    args:
      urlpath: '/caldera/projects/usgs/hazards/cmgp/woodshole/rsignell/conus404/zarr/nwis_chanobs.zarr'
      consolidated: true
      
  nwis-streamflow-usgs-gages-osn:
    driver: zarr
    description: "Streamflow from NWIS, extracted and rechunked into time series (NWM2.1 time period)"
    args:
      urlpath: 's3://hytest/tutorials/evaluation/nwm/nwis_chanobs.zarr'
      consolidated: true
      storage_options:
        anon: true
        requester_pays: false
        client_kwargs:
          endpoint_url: https://usgs.osn.mghpcc.org
      
  nwm21-streamflow-usgs-gages-onprem:
    driver: zarr
    description: "Streamflow from NWM2.1, extracted and rechunked into time series"
    args:
      urlpath: '/caldera/projects/usgs/hazards/cmgp/woodshole/rsignell/conus404/zarr/chanobs.zarr'
      consolidated: true
      
  nwm21-streamflow-usgs-gages-osn:
    driver: zarr
    description: "Streamflow from NWM2.1, extracted and rechunked into time series"
    args:
      urlpath: 's3://hytest/tutorials/evaluation/nwm/chanobs.zarr'
      consolidated: true
      storage_options:
        anon: true
        requester_pays: false
        client_kwargs:
          endpoint_url: https://usgs.osn.mghpcc.org
      
  nwm21-streamflow-cloud:
    driver: zarr
    description: "National Water Model 2.1 CHRTOUT on AWS"
    args:
      urlpath: 's3://noaa-nwm-retrospective-2-1-zarr-pds/chrtout.zarr' 
      consolidated: true
      storage_options:
        anon: true

  geofabric_v1_1-zip-osn:
    driver: geoparquet # note - this is the wrong driver, and you cannot open this dataset with intake
    description: "GIS Features of the Geospatial Fabric for the National Hydrologic Model, version 1.1 (https://doi.org/10.5066/P971JAGF). This dataset cannot be opened with intake because the GDB driver doesn't exist; however it is still being catalogged here for the purpose of being able to read the urlpath into your workflow to locate the data."
    args:
      urlpath: 's3://hytest/geofabric_v1_1/GFv1.1.gdb.zip'
      storage_options:
        anon: true
        client_kwargs:
          endpoint_url: https://usgs.osn.mghpcc.org

  geofabric_v1_1_POIs_v1_1-osn:
    driver: geoparquet
    description: "POIs_v1_1 layer of the GIS Features of the Geospatial Fabric for the National Hydrologic Model, version 1.1 (https://doi.org/10.5066/P971JAGF) converted to geoparquet format"
    args:
      urlpath: 's3://hytest/geofabric_v1_1/GFv1.1_POIs_v1_1.geoparquet'
      storage_options:
        anon: true
        client_kwargs:
          endpoint_url: https://usgs.osn.mghpcc.org

  geofabric_v1_1_TBtoGFv1_POIs-osn:
    driver: geoparquet
    description: "TBtoGFv1_POIs layer of the GIS Features of the Geospatial Fabric for the National Hydrologic Model, version 1.1 (https://doi.org/10.5066/P971JAGF) converted to geoparquet format"
    args:
      urlpath: 's3://hytest/geofabric_v1_1/GFv1.1_TBtoGFv1_POIs.geoparquet'
      storage_options:
        anon: true
        client_kwargs:
          endpoint_url: https://usgs.osn.mghpcc.org
          
  geofabric_v1_1_nhru_v1_1-osn:
    driver: geoparquet
    description: "nhru_v1_1 layer of the GIS Features of the Geospatial Fabric for the National Hydrologic Model, version 1.1 (https://doi.org/10.5066/P971JAGF) converted to geoparquet format"
    args:
      urlpath: 's3://hytest/geofabric_v1_1/GFv1.1_nhru_v1_1.geoparquet'
      storage_options:
        anon: true
        client_kwargs:
          endpoint_url: https://usgs.osn.mghpcc.org

  geofabric_v1_1_nhru_v1_1_simp-osn:
    driver: geoparquet
    description: "nhru_v1_1_simp layer of the GIS Features of the Geospatial Fabric for the National Hydrologic Model, version 1.1 (https://doi.org/10.5066/P971JAGF) converted to geoparquet format"
    args:
      urlpath: 's3://hytest/geofabric_v1_1/GFv1.1_nhru_v1_1_simp.geoparquet'
      storage_options:
        anon: true
        client_kwargs:
          endpoint_url: https://usgs.osn.mghpcc.org
          
  geofabric_v1_1_nsegment_v1_1-osn:
    driver: geoparquet
    description: "nsegment_v1_1 layer of the GIS Features of the Geospatial Fabric for the National Hydrologic Model, version 1.1 (https://doi.org/10.5066/P971JAGF) converted to geoparquet format"
    args:
      urlpath: 's3://hytest/geofabric_v1_1/GFv1.1_nsegment_v1_1.geoparquet'
      storage_options:
        anon: true
        client_kwargs:
          endpoint_url: https://usgs.osn.mghpcc.org

  gages2_nndar-osn:
    driver: parquet
    description: "Estimated streamflow at GAGESII locations (https://doi.org/10.5066/P9XT4WSP) using nearest-neighbor drainage area ratio (NNDAR), converted to geoparquet format"
    args:
      urlpath: 's3://hytest/nwis_gages2_streamflow_estimates/gages2_nndar.parquet'
      storage_options:
        anon: true
        client_kwargs:
          endpoint_url: https://usgs.osn.mghpcc.org

  wbd-zip-osn:
    driver: geoparquet # note - this is the wrong driver, and you cannot open this dataset with intake
    description: "Watershed Boundary Dataset (https://doi.org/10.5066/P9BTKP3T). This dataset cannot be opened with intake because the GDB driver doesn't exist; however it is still being catalogged here for the purpose of being able to read the urlpath into your workflow to locate the data."
    args:
      urlpath: 's3://hytest/wbd/WBD_National_GDB.zip'
      storage_options:
        anon: true
        client_kwargs:
          endpoint_url: https://usgs.osn.mghpcc.org

  huc12-geoparquet-osn:
    driver: geoparquet
    description: "WBDHU12 layer of Watershed Boundary Dataset (https://doi.org/10.5066/P9BTKP3T) converted to geoparquet format"
    args:
      urlpath: 's3://hytest/wbd/huc12/huc12.geoparquet'
      storage_options:
        anon: true
        client_kwargs:
          endpoint_url: https://usgs.osn.mghpcc.org

  huc12-gpkg-osn:
    driver: geoparquet # This should be geopackage, but this is not currently supported by intake - will be supported in intake v2
    description: "WBDHU12 layer of Watershed Boundary Dataset (https://doi.org/10.5066/P9BTKP3T) converted to geopackage format. This file format is not currently supported by intake, so you cannot currently open this dataset with intake. This will be compatible with intake v2 when in is released in early 2024."
    args:
      urlpath: 's3://hytest/wbd/huc12/huc12.gpkg'
      storage_options:
        anon: true
        client_kwargs:
          endpoint_url: https://usgs.osn.mghpcc.org

  nwm21-scores:
    description: US state information from [CivilServices](https://civil.services/)
    driver: csv
    args:
      urlpath: 'https://raw.githubusercontent.com/nhm-usgs/data-pipeline-helpers/main/hytest/results/nwm_ref_gages_assessment.csv'

  lcmap-cloud:
    driver: intake_xarray.xzarr.ZarrSource
    description: 'LCMAP, all 36 years'
    args:
      urlpath: "reference://"
      consolidated: false
      storage_options:
        target_options:
          requester_pays: true
        fo: 's3://nhgf-development/lcmap/lcmap.json'
        remote_options:
          requester_pays: true
        remote_protocol: s3

  rechunking-tutorial-osn:
    driver: zarr
    description: "Sample streamflow and velocity data used in HyTEST rechunking tutorial"
    args:
      urlpath: 's3://hytest/tutorials/dataset_preprocessing/ReChunking_Tutorial_Data.zarr'
      consolidated: true
      storage_options:
        anon: true
        requester_pays: false
        client_kwargs:
          endpoint_url: https://usgs.osn.mghpcc.org

  pointsample-tutorial-sites-osn:
    driver: csv
    description: "Sample gage/point locations used in CONUS404 point sampling tutorial"
    args:
      urlpath: 's3://hytest/tutorials/data_access/filtered_temperature_sites.csv'
      storage_options:
        anon: true
        client_kwargs:
          endpoint_url: https://usgs.osn.mghpcc.org

  pointsample-tutorial-output-osn:
    driver: netcdf
    description: "Output point sampled CONUS404 data from CONUS404 point sampling tutorial"
    args:
      urlpath: 's3://hytest/tutorials/data_access/daily_skintemp_at_filtered_temperature_sites.nc'
      storage_options:
        anon: true
        client_kwargs:
          endpoint_url: https://usgs.osn.mghpcc.org