conflate_model and gpkg_from_ras should take model_name parameter rather than inferring it

[ar-siddiqui]

When submitting jobs for these processes, there is no way to provide a model_name. It is inferred from the parent folder name. This is wrong because model_name could be different. For example if model_stac_id is 'ROCKY_BRANCH' and the model_name is 'ROCKY BRANCH' then the following will fail as the GPKG name will be 'ROCKY BRANCH.gpkg'.

{
    "source_model_directory": "Z:\\collections\\ble_12040103_EastForkSanJacinto\\source_models\\ROCKY_BRANCH",
    "source_network": {
        "file_name": "Z:\\reference_data\\nwm_flowlines_with_bbox.parquet",
        "version": "2.1",
        "type": "nwm_hydrofabric"
    }
}

{
    "jobID": "0b81956a-c9ff-4b57-ab14-2e52bec5ae31",
    "processID": "conflate_model",
    "result": {
        "err": "Z:\\collections\\ble_12040103_EastForkSanJacinto\\source_models\\ROCKY_BRANCH\\ROCKY_BRANCH.gpkg: No such file or directory",
        "val": null
    },
    "status": "failed",
    "submitter": "",
    "type": "process",
    "updated": "2024-09-17 20:28:56"
}

[slawler]

Let's adjsut the payload above to include a model_name parameter (for now) we can discuss prefernces as they come along:

{
    "source_model_directory": "Z:\\collections\\ble_12040103_EastForkSanJacinto\\ROCKY_BRANCH",
    "model_name": "ROCKY_BRANCH",
    "source_network": {
        "file_name": "Z:\\reference_data\\nwm_flowlines_with_bbox.parquet",
        "version": "2.1",
        "type": "nwm_hydrofabric"
    }
}

Inference will be updated to use:

f"{source_model_directory} \ {model_name}.prj" 

[mdeshotel]

When calling ras_to_gpkg we don't yet know what the model_name is. We find the model name by first grabbing all .prj files then assessing if prj file is the RAS project file.

For conflate model, the model_name has been added as a required arg. Note that it is the path to the source model geopackage that we infer with the souce_model_directory and model_name; i.e., not the .prj file.

[ar-siddiqui]

When calling ras_to_gpkg we don't yet know what the model_name is. We find the model name by first grabbing all .prj files then assessing if prj file is the RAS project file.

For conflate model, the model_name has been added as a required arg. Note that it is the path to the source model geopackage that we infer with the souce_model_directory and model_name; i.e., not the .prj file.

@mdeshotel I didn't understand the last sentence, can you describe it with an example of API/Schema changes before and after?

[mdeshotel]

When calling ras_to_gpkg we don't yet know what the model_name is. We find the model name by first grabbing all .prj files then assessing if prj file is the RAS project file.
For conflate model, the model_name has been added as a required arg. Note that it is the path to the source model geopackage that we infer with the souce_model_directory and model_name; i.e., not the .prj file.

@mdeshotel I didn't understand the last sentence, can you describe it with an example of API/Schema changes before and after?

Seth's comment above was saying we infer the .prj file.

f"{source_model_directory} \ {model_name}.prj" 

My comment was just pointing out that we don't infer the .prj file we actually only infer the .gpkg file.

f"{source_model_directory} \ {model_name}.gpkg"