Unable to load toolkit 'OpenEye Toolkit' with openeye-toolkit installed and OE_LICENSE set up #825
Comments
|
@mattwthompson using your suggestion to move to OEChem, decided to create a separate ticket |
|
When modifying script2 in #822 to use I see: How is it possible that OE toolkit integration tests work fine for OEChem with the licence, but OpenFF raises LicenseError? What kind of error it could be? Also when using toolkit code: I get: |
|
The code currently assumes you have each of the OpenEye toolkits we use installed, and provides a not-very-helpful error message if one of them is not found (even if the one you want to use is found). Besides Apologies again for the not-so-descriptive error message, and thanks for the feedback! I believe this is an oversight that we can fix for future releases. |
|
My licence is for OEChem only. Does it mean that I can't use the toolkit at all (to read PDB as a replacement for RDKit as in script2 or implicitly replaced as in offtop examples) because I don't have licences for How to make offtop example (script1) to use OEChem instead of RDKit? |
|
Oh, this is really interesting. Thanks for writing in about this @mieczyslaw.
Unfortunately, this is an aspect of the design that we hadn't considered before, so that is currently the case. I'll go ahead and do two things ASAP:
|
|
@mieczyslaw Here's a workaround that you can put at the top of your script to have the OpenFF toolkit attempt to use the components of OpenEye that are licensed: Could you let me know if this works for you? I tested it on the following workflow, which should exercise most of the core functionality of the OpenFF toolkit. I used this toluene PDB file. |
|
Thanks @j-wags ! As commercial licences are expensive, the case is that companies may afford only a subset of OE tools. I tried to add the lines above to script1 in #822 (off top one). It seems to pick up OpenEye (no crash because of RDKit), but shows: The same output message on my side for the the workflow like you posted above (the equivalent of script2 in #822). |
Describe the bug
I installed
openeye-toolkits 2020.2.0from conda openeye channel. I set up OE_LICENSE (for OEChem) and run openeye-toolkits integration tests to see that they pass for OEChem (and skip for other parts). The licence is fine as without setting up OE_LICENSE, OpenFF shows "LICENSE: N.B. OE_LICENSE environment variable is not set".However, when trying to parametrize molecule with OpenFF using OEChem, as suggested to sort out RDKit issues (see script1 in #822) , I still see:
Warning: Unable to load toolkit 'OpenEye Toolkit'.To Reproduce
Call script1.py from #822 with openeye-toolkits installed and OE_LICENSE for OEChem set up.
Output
Warning: Unable to load toolkit 'OpenEye Toolkit'.RDKit is still in use (and crashes as in #822).Computing environment (please complete the following information):
conda listAdditional context
I assume that OEChem is automatically loaded and takes precedence over RDKit.
The text was updated successfully, but these errors were encountered: