You may notice there is a somewhat unusual setup with the metadata/SQL, in the models folder.
Original developments on this app began prior to Metadata V3 release.
At that time, table metadata was aggregated entirely into a single table.yaml. This can be become more difficult to read as tables grow, so here a makeshift script/tooling setup was used to be able to define models (as both SQL DDL and the Hasura table metadata) individually.
This is not a scalable pattern after initial deployments (without bringing in another tool like https://github.com/bikeshedder/tusker that can calculate diffs between two schemas and print them as migrations) and shouldn't be taken as a proper development pattern.
The correct way to do this today is to use Metadata V3 and take advantage of it enabling individual metadata files. See here for more information: https://hasura.io/docs/latest/graphql/core/migrations/upgrade-v3.html#metadata-directory