Closed spadavec closed 8 years ago
Thanks for catching this! We must have missed it in our automated testing of examples, which I notice doesn't have a badge on the front page. I'll add the badge.
@andrrizzi : Can you see if the setup pipeline is trying to create prepare.log
before the output
store directory is created? If so, we should move up creation of the output store to earlier in the setup process.
We should also convert this example over to the new YAML format. We should probably leave one or two examples of the old CLI format, but steer people toward the new format. Can you tackle that once you have the fellowship in?
@spadavec, thanks for catching this! I can't actually see the new error. The last two lines I see in the log are
File "/Users/FFF/anaconda/lib/python2.7/site-packages/numpy/core/fromnumeric.py", line 1028, in argmin
return argmin(axis, out)
the description of the error should be after that.
Can you see if the setup pipeline is trying to create prepare.log before the output store directory is created?
If I remember correctly, with the command line interface we always had to create the output
directory manually. All the run.sh
scripts should have a mkdir output
directive before calling yank prepare
. The YAML setup pipeline creates the folder for you correctly.
Can you tackle that once you have the fellowship in?
Of course!
By the way, I think with the new alchemy, the command line interface should not be used anymore. The default value of softcore_beta
is 1.0, the default alchemical path does not annihilate electrostatic interactions before the sterics, and there is no way to change that through a CLI option.
OK! Should we fully remove the CLI for yank setup
or just add a "This option has been deprecated, please use the YAML" version for now?
Alternatively, we could add the deprecation warning and have this option generate a YAML file you can use that matches recommended parameters.
I would add the deprecation warning for now because the YAML pipeline is currently not flexible enough to completely substitute the CLI.
and have this option generate a YAML file you can use that matches recommended parameters
Great idea! I'll open a new issue to discuss this.
The error comes up because the DSL string passed is --ligand="resname BEN"
instead of --ligand="resname MOL"
(the resname must match the one in the pdb
file for the system).
We need better error handling for this as pointed out in #383 .
We can probably raise an informative exception if the DSL fails to match any atoms
Closing this as it was only an error in the command parameters. The PR #389 adds a more informative error message.
With the latest version of YANK, running the p-xylene-implicit example, via:
yank prepare binding amber --setupdir=setup --ligand="resname BEN" --store=output --iterations=1000 --restraints=harmonic --gbsa=OBC2 --temperature=300*kelvin --verbose
Results in the following error:
To get around this, if you mkdir "output" in the same directory, and re-run the above command, the new error becomes:
This is run on OSX 10.11.13 using YANK 0.9.0