Used Cantera reaction rate coefficient units

[jAnirudh]

Proposed fix to #318 for computing the production rates for elementary reactions with FORD reactions

The reaction rate coefficients appearing as ctuc in ceptr/production.py are computed directly from Cantera, except for Three-Body reactions, which are made to work according to Cantera's legacy behaviour as discussed in #323.

These changes are reflected in the comp_qfqr, productionRate and the productionRate_Light methods of the corresponding mechanism.H files.

This fix should immediately work for FORD reactions when not using the analytic Jacobian.

[malihass]

Thanks for the PR!
So instead of doing this, #414 adjusts the value of dim which is a little bit cleaner than the string operations.
Would you please be able to try and see if #414 fixes that issue for you?

[jAnirudh]

No worries :)

I am not a fan of the string operations myself honestly. Ideally I'd have liked Cantera to simply give me the rate coefficient unit directly without me having to do any string manipulations. Maybe this is something the Cantera folks may like to consider for us.

However, I found this way to be much easier to follow and keep track of than modifying dim when adding new reactions. Here, irrespective of the kind of reaction, if Cantera supports it we're sure to get the same dim. This makes it much easier to implement new reactions plus there are fewer and more localized changes to track which is good for maintainability IMO.

Anyway, while I haven't been able to go through all the code changes in production.py and util.py, this change and #414 give identical mechanism.H files for the 3 methods I've mentioned. So we essentially have 2 ways to solve the problem :)

[malihass]

OK, I am going to close this since this results in the same mechanisms. Thanks for bringing these FORD issues to our attention!