are we removing a variable from parallel tracks in random order mechanisms? + multiple haldanes

[dczielinski]

this is part of the new handling of random mechanisms - wasn't sure how exactly it got changed. basically, if we have two tracks of binding, from the substrates to the enzyme bound complex, e.g. , we have a constraint that K1K2 = K3K4, if reactions 1 and 2 are parallel with reactions 3 and 4 (like a half haldane). This reduces the number of free variables by one, and we should be directly removing one rate constant using this constraint, the way we used to do with haldane (but we have 0 error here so we don't have to worry about keq error like we did with the haldane, which made us stop substituting in the overall Keq)

We should also end up with multiple haldanes at the end of the day, for each possible route the enzyme, which definitely isn't the case as far as printed/fit data at least, so it seems we have more features to implement there?