Ok, we're going to include site data that is needed to do a complete hazard calculation to the relevant model repositories. These datasets may become large down the road, but that is not today's problem to solve. The priority is to make it as easy as possible for users to check out the code nshmp-haz and a model nshm-* and reproduce our published results. This approach will also obviate the need for AWS deployed containers to include a local web service for site data. We'll leave nshmp-ws alone for now but will soon figure out how to pull and manage site data from various models.
nshmp-lib:
CalcConfig will continue to support the SiteData.serviceOptional<URL> field. If it's empty, we'll use the local model data, which we'll expect to find in a site-data directory. If the URL is present, we'll call the service (this allows the use of new or different data sets without having to muck about putting data into a branch of a model)
I'm not sure about the full implementation in nshmp-lib just yet. In building a Site object, we'll be able to loop over the defined polygons doing hit tests (as is done now) and updating z*p* values as needed. What we don't want is an enum in nshmp-lib defining basins or other polygons that are model specific. An enum of basin identifiers could go up in nshmp-haz, however.
nshm-conus:
Copy nshmp-ws/src/main/resources/basin to nshm-conus/site-data/basin (PP)
Amendments to the above. Site data directory was added in !192 (merged), and modifications to Site, Site.Builder, and Sites in !208 (merged). Models now have their NSHM values included locally and there is no need for a service, which has been removed from CalcConfig.