vm6502q/PyQrackIsing: Mix perturbed and unperturbed OTOC(2) fraction

There was a subtle bug in the vector math for OTOC(2) samples: the (rev - fwd)-based diff_x and diff_z terms are valid vector differences to sum with the starting point of the perturbed distribution that is actually fwd, but this is mixed (or superposed) with fully self-cancelling round trips back to initial temperature given by theta. The fix is to mix a fraction of fwd + diff with unperturbed initial state, assuming a perfect mirror. Also, the hadamard formula did not yet produce a normalized distribution, which is expected for the formula and not a sign of any deeper problem, but the function should simply renormalize before returning the transformed distribution. Full Changelog: https://github.com/vm6502q/PyQrackIsing/compare/v9.16.10...v9.16.11 sha1sum results: e2b301c2cf55df08a32f46677e626eda6ed356ec pyqrackising-9.16.11-py3-none-macosx_14_0_arm64.whl c44c48967fe5b83434a68cd3d521b20b5121d5bd pyqrackising-9.16.11-py3-none-macosx_15_0_arm64.whl 81fc18dd975a4f7ffe73dd8ec4bd036a3fde6549 pyqrackising-9.16.11-py3-none-manylinux_2_35_x86_64.whl 806c78b89e7b3b5fccd463d250278d36113e5ebb pyqrackising-9.16.11-py3-none-manylinux_2_39_x86_64.whl 6371fe80149638b59a7f3de3fe7ac1e5f1449a27 pyqrackising-9.16.11-py3-none-win_amd64.whl 5304088ca4b701f5308bb3aba0de0ba82bee7217 pyqrackising-9.16.11.tar.gz