Debug library to training script

[hwpang]

Motivation or Problem

I encountered two errors while converting a library containing this Diels-Alder reaction to training reaction.

(1) Two reactants were wrongly added with the same atom label by add_atom_labels_for_reaction because the two reactants are isomorphic.

(2) The A unit caused problem in add_rules_from_training because only cm mol s units were expected.

Reaction(reactants=[Species(label="ISP", molecule=[Molecule(smiles="C=CC(=C)C")], molecular_weight=(68.117,'amu')), Species(label="ISP", molecule=[Molecule(smiles="C=CC(=C)C")], molecular_weight=(68.117,'amu'))], products=[Species(label="DISP_2", molecule=[Molecule(smiles="C=CC1(C)CCC=C(C)C1")], molecular_weight=(136.234,'amu'))])

Description of Changes

(1) In add_atom_labels_for_reaction, I modified such that the labeled_molecule is removed after it is matched to one of the reactant. This way, two reactants won't be matched to the same atom label.

(2) In kinetics_library_to_training_tools I added an unit conversion for A from arbitrary unit to cm mol s unit.

Testing

After the change, the kinetics_library_to_training.ipynb can process the library without error.

[codecov]

Codecov Report

Merging #2057 (04f618f) into master (43a1ee9) will increase coverage by 0.06%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #2057      +/-   ##
==========================================
+ Coverage   47.54%   47.61%   +0.06%     
==========================================
  Files          88       88              
  Lines       23300    23302       +2     
  Branches     6061     6061              
==========================================
+ Hits        11079    11095      +16     
+ Misses      11062    11052      -10     
+ Partials     1159     1155       -4     

Impacted Files	Coverage Δ
rmgpy/data/kinetics/family.py	48.84% <100.00%> (+<0.01%)	⬆️
rmgpy/molecule/draw.py	54.03% <0.00%> (+1.28%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 43a1ee9...04f618f. Read the comment docs.

[mjohnson541]

This looks good to me.

[mjohnson541]

Can you rebase?

[hwpang]

Thanks for the review! It's rebased!

[hwpang]

I changed '1/s' to 's^-1' for the A unit conversion part to have consistent format with those used in add_rules_from_training.

[hwpang]

I added a line to import ScalarQuantity from rmgpy.quantity for A unit conversion. I had a separate file for my own use and some changes were lost when I copied and pasted. Sorry for the multiple follow up changes, this is the last one!

[hwpang]

I made some minor changes in the jupyter notebook while preparing for RMG workshop.

• I changed the comment part from 'all' to 'default' because the 'all' option tries to load the Surface families but the relative thermo library is not provided.
• I changed families to list(database.kinetics.famillies.keys()) to allow using 'default' option.