bugfix for is.smiles

[allaway]

Hello - I recently ran into an issue with this function using the current CRAN version of webchem, but I believe this issue still exists in the master branch here. Apologies if I am supposed to file a pull request against a different branch than master.

is.smiles previously returned TRUE for any string regardless of validity. I suspect this was due to changes in rcdk::parse.smiles.

This change should correctly return TRUE for valid SMILES, and FALSE for un-parseable SMILES.

[codecov-io]

Codecov Report

❗ No coverage uploaded for pull request base (master@10a8858). Click here to learn what that means.
The diff coverage is 0%.

@@           Coverage Diff           @@
##             master   #140   +/-   ##
=======================================
  Coverage          ?     0%           
=======================================
  Files             ?     17           
  Lines             ?   1631           
  Branches          ?      0           
=======================================
  Hits              ?      0           
  Misses            ?   1631           
  Partials          ?      0

Impacted Files	Coverage Δ
R/utils.R	0% <0%> (ø)

---

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 10a8858...bffa4be. Read the comment docs.

[eduardszoecs]

Thanks for your contribution! I'll check and test ASAP (might take few days).
In fact, i had to turn off the tests for is. smiles() , as java is a bit b****y 😞

Could also use rdkit the test if it parses (could be implemented via rrdkit, or (better) via reticulate.

[allaway]

Eliminating the rJava dependency (via removing dependency on rcdk) would not be a bad thing! :)

[eduardszoecs]

Thanks for your contribution!
Incorporated you changes in master and fixed tests and continuous integration (rcdk installation failed).