Don't use Python

[eschnett]

I checked BioJulia/YAML.jl, and it seems feasible to use it to parse YAML files. One needs to extend its constructor table default_yaml_constructors to know about ASDF tags, and then ASDF objects can be read in either as dictionary, or checked and converted to proper ASDF types (see #2).

Writing YAML files would need to be implemented directly. That would mostly require quoting strings properly and handling indentation; otherwise, creating correct YAML files seems straightforward.

What remains is reading and writing binary blocks. This is already implemented in C++ in asdf-cxx, and could be ported to Julia. Interfaces to compression libraries (zlib, bzip2) already exist in Julia.