Which ways do we have to measure productivity for open source software? Most of the tooling around that still focus on lines of code and cyclomatic complexity