Fix incorrect reduction of dimers/polymers in standardize_formula

[vineetbansal]

This is most likely a bug that we haven't encountered yet, simply because we're getting lucky in the order of the keys (representing species->amount mapping) in the current wrapper.

The tests that should fail in this PR is test_phreeqc.py::test_equilibrate_logC_pH_carbonate_8_3 (among others).

The issue seems to be that for a simple solution like this:

Solution({"CO2(aq)": "0.001 mol/L"}, pH=8.3, volume="1 L", engine="phreeqc")

corresponding to the string:

SOLUTION 0
  temp 25.0
  units mol/kgw
  pH 8.3
  pe 8.5
  redox pe
  water 0.9970480319717386
  C(4) 0.0010029607079434515
SAVE SOLUTION 0
END

In addition to CO2, phreeqc reports a tiny amount of (CO2)2 (whatever that represents, not sure..). The following is the output from phreeqc UI, and is consistent with what the old (and our new) wrappers return.

----------------------------Distribution of species----------------------------

                                               Log       Log       Log    mole V
   Species          Molality    Activity  Molality  Activity     Gamma   cm3/mol

   OH-             2.072e-06   2.019e-06    -5.684    -5.695    -0.011     -4.12
   H+              5.138e-09   5.012e-09    -8.289    -8.300    -0.011      0.00
   H2O             5.551e+01   1.000e+00     1.744    -0.000     0.000     18.07
C(4)          1.003e-03
   HCO3-           9.822e-04   9.575e-04    -3.008    -3.019    -0.011     24.57
   CO2             1.079e-05   1.079e-05    -4.967    -4.967     0.000     34.43
   CO3-2           9.923e-06   8.959e-06    -5.003    -5.048    -0.044     -3.97
   (CO2)2          2.137e-12   2.137e-12   -11.670   -11.670     0.000     68.87
H(0)          3.556e-37
   H2              1.778e-37   1.778e-37   -36.750   -36.750     0.000     28.61
O(0)          2.636e-19
   O2              1.318e-19   1.318e-19   -18.880   -18.880     0.000     30.40

------------------------------Saturation indices-------------------------------

The standardize_formula function standardizes (CO2)2 as CO2, and can possibly end up overwriting the CO2 amount with the (CO2)2 amount (which is tiny), in the following logic in .equilibrate():

        solution.components = FormulaDict({})
        # use the output from PHREEQC to update the Solution composition
        # the .species_moles attribute should return MOLES (not moles per ___)
        for s, mol in self.ppsol.species_moles.items():
            solution.components[s] = mol

We simply get lucky that (CO2)2 is consistenly reported before CO2 by the current wrapper, so we end up with the correct amount of CO2. In the new wrappers, the order of the keys in the dict is reversed (for whatever reason, but most probably because we're building the dictionary in the same order as the output of phreeqc UI indicates), so we do encounter this bug.

The attached snippet simply surfaces the bug without fixing it. I'm not sure of a solution without the background on what it means/entails.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 84.81%. Comparing base (4d1454c) to head (2a0e8d8).
⚠️ Report is 12 commits behind head on main.

Additional details and impacted files

@@             Coverage Diff             @@
##             main     #309       +/-   ##
===========================================
+ Coverage   74.55%   84.81%   +10.25%     
===========================================
  Files          10       10               
  Lines        1521     1521               
  Branches      265      264        -1     
===========================================
+ Hits         1134     1290      +156     
+ Misses        352      199      -153     
+ Partials       35       32        -3     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[rkingsbury]

Ooof, good catch @vineetbansal . I have no idea how to interpret (CO2)2. I suppose for the time being we should add an ad-hoc patch to standardize_formula similar to what was done here (see for example - C2O4[-2]

[rkingsbury]

I think this can be removed, now that the fix is in place

[vineetbansal]

Thanks! I just wanted to see if the tests pass before removing this. Hopefully the fix makes sense.