Make post processing faster

[JoeWang1127]

I noticed that post processing a library takes longer than generating it. We should optimize it.

This issue contains two parts:

1. Identify the slow commands in post processing
2. Optimize the slow commands.

More context: I don't know which command is slow but I noticed stuck when running the post processor.

[diegomarquezp]

From internal discussion:
One nice to have that may benefit us is to stream the output of the commands in real time. Now, the python script blocks until calls to scripts such as generate_library or postprocess_library are finished and then prints the resulting output. It should be a straightforward thing to do

[diegomarquezp]

We may benefit from postprocess_monorepo instead of postprocess_library regarding the overhead of setting up each one of the commands used in postprocessing. For example there is a mvn fmt:format ran for each library, or the fact that copy-code can process more than one yaml in a single run