stephenberry
commented
7 months ago Generating JSON schema from a file is complicated and requires slow processing, especially when having to map partial objects into cohesive objects, and it may be impossible to determine if objects are variants or partials. It is better to generate a schema from the C++ or the original data source and use this.
BEVE isn't human readable, so it can be difficult to know what is in the file. We can convert BEVE to JSON, but for large BEVE files (may be gigabytes in size) this is slow and a waste of time. If we can simply parse the BEVE into a JSON schema file, then we can generate documentation on the schema or a default JSON file from the schema. This will make determining what data the BEVE file contains much easier.