Thanks, @ischoegl, I think this is an exciting step forward in simplifying the handling of different reaction rate parameterizations.

One thing that you only barely touched on is that by handling third bodies in the common Reaction class, these can in principle be used for any rate parameterization. However, there are a couple of quirks here that I think are worth ironing out. One is that there are a few places that single out the TwoTempPlasmaRate class as not being able to handle third bodies, but I don't understand why that would be the case now. Shouldn't any reaction type be able to handle third bodies now, with the exception that we explicitly remove them from Plog and Chebyshev reactions in acknowledgment that the third-body pressure dependence is already embedded in those rate constants?

The second is that there are a few situations where the automatic detection of specific third body colliders prevents instantiation of reactions as non-third-body reactions, even though I don't think these should be seen as problematic. For instance, the following reaction definitions raise exceptions, but were (I think correctly) permitted in Cantera 2.6:

R1 = ct.Reaction(
    equation="HO2 + O2 <=> OH + O + O2",
    rate={"type": "pressure-dependent-Arrhenius",
          "rate-constants": [
               {"P": 1013.25, "A": 1.2124e+16, "b": -0.5779, "Ea": 45491376.8},
               {"P": 101325., "A": 4.9108e+31, "b": -4.8507, "Ea": 103649395.2}
          ],
    }
)

R2 = ct.Reaction(
    equation="H + OH + 2 O2 = H2O + 2 O2",
    rate={"A": 1.234e5, "b": 0, "Ea": 4321.0}
)

Thanks, @ischoegl - I think this is good to go.