Initialize the global RMGDatabase object upon calling it

[alongd]

Initialize the global RMGDatabase object upon calling it even (and especially) if it already exists.
This solves cases when using RMG's API where the database doesn't get appropriately updated.

[codecov]

Codecov Report

Merging #1565 into master will increase coverage by 0.01%.
The diff coverage is 50%.

@@            Coverage Diff             @@
##           master    #1565      +/-   ##
==========================================
+ Coverage   41.84%   41.86%   +0.01%     
==========================================
  Files         165      165              
  Lines       28008    28007       -1     
  Branches     5713     5713              
==========================================
+ Hits        11721    11726       +5     
+ Misses      15498    15493       -5     
+ Partials      789      788       -1

Impacted Files	Coverage Δ
rmgpy/data/rmg.py	71.54% <50%> (+0.57%)	⬆️
rmgpy/data/kinetics/family.py	58.56% <0%> (+0.29%)	⬆️

---

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 c39c888...5e907df. Read the comment docs.

[mliu49]

I think this looks good. Do we want the re-initialization message to be debug or warning? Personally, I'm leaning more towards warning since I think users should be aware that it is happening.

[alongd]

done!

[mjohnson541]

Looks good to me!

[mliu49]

I think we should use logging.warning. logging.warn is just a pointer to logging.warning, and is now deprecated in Python 3.

[alongd]

good to know, done

[alongd]

I can approve that the RMG/ARC feature works as expected when cherry-picking this commit.
In fact, and strangely, the implementation of resetting the database after each call to RMG or ARC via something like

def clear_rmg_database():
    """Set the RMG database object instance to None"""
    rmg_db = getDB()
    rmg_db.thermo = None
    rmg_db.transport = None
    rmg_db.forbiddenStructures = None
    rmg_db.kinetics = None
    rmg_db.statmech = None
    rmg_db.solvation = None

was problematic, and in the second iteration RMG comlained about missing components in the database (e.g., can't check forbidden structures since database.forbiddenStructures was None; when I manually initialized it with the correct object, it complained that a different part was None). I thought I did use this function in an order that makes sense.

In any case, I'm very happy with the implementation in this PR.

[mliu49]

Thanks! I'll go ahead and merge.