I am writing a tool to dump our dedicated ML model from flatbuffers binary into human readable text format, is there any helper like text_format for protobuf?
There's currently no way to do that directly in Python, you'll need to invoke the command-line flatc tool to do so, e.g. flatc [--raw-binary] --json myschema.fbs -- mybinary.bin results in mybinary.json.
If there is no file_identifier, you need file_identifier
If necessary, it would be possible to compile the C++ json generator (and parser) into something that can be called from Python (through C), but that requires knowledge on how to create such extension libraries for Python.