bug fix pcross errors

[lukasbaldauf]

😅 I noticed the original implementation of the Pcross error estimation was wrong and in some cases underestimated the errors significantly.

🤓 I need error bars on the Pcross curve to compare multiple curves that are very similar but not identical. Therefore, I need the error bars to reflect the statistical significance of the differences. 📊

😵‍💫 I redid the implementation in a brute-force way, which now creates N_interfaces new infretis_data.txt files, which can be whammed. From those, the error in the crossing probability at the interface locations can then be extracted from wham_i/errPtot.txt.

😓 This requires whamming N_interfaces times, which is not ideal, but at least it gives the correct results 💯 👍 💪

Can be run as

inft pcross_errors -toml infretis0.toml -data infretis_data.txt

and then

inft wham -data infretis_0_interface_data.txt -toml infretis_0_interface.toml -folder wham_interface0
inft wham -data infretis_1_interface_data.txt -toml infretis_1_interface.toml -folder wham_interface1
inft wham -data infretis_2_interface_data.txt -toml infretis_2_interface.toml -folder wham_interface2