Improve parse performance with header case

[kou]

We should improve performance as much as HoneyFormat: https://github.com/buren/honey_format#benchmark

[vbrazo]

@kou I'm planning to work on this one. Any suggestions on how to tackle this issue would be welcome.

[kou]

Great!
First, we need to profile which method is bottleneck.
To profile, we need to write a simple script that is slower than HoneyFormat.