We are currently using the python script usdGenSchema shipped with usd to generate the final usd arnold schema file called generatedSchemas.usda from schema.usda.
When we build the arnold usd procedural, usd_proc, with the static usd version, we can't use usdGenSchema as this version doesn't have the python bindings. We don't need and want the python bindings in the procedural.
The only transformation we need to generate the final generatedSchema.usda file is a flatten operation, plus some filtering as visible in https://github.com/PixarAnimationStudios/OpenUSD/blob/10b62439e9242a55101cf8b200f2c7e02420e1b0/pxr/usd/usd/usdGenSchema.py#L1525
We want to write a tool that would perform those operations and replace usdGenSchema when we build with a version of usd which doesn't have python.
We are currently using the python script usdGenSchema shipped with usd to generate the final usd arnold schema file called generatedSchemas.usda from schema.usda.
When we build the arnold usd procedural, usd_proc, with the static usd version, we can't use usdGenSchema as this version doesn't have the python bindings. We don't need and want the python bindings in the procedural.
The only transformation we need to generate the final generatedSchema.usda file is a flatten operation, plus some filtering as visible in https://github.com/PixarAnimationStudios/OpenUSD/blob/10b62439e9242a55101cf8b200f2c7e02420e1b0/pxr/usd/usd/usdGenSchema.py#L1525
We want to write a tool that would perform those operations and replace usdGenSchema when we build with a version of usd which doesn't have python.