MorganCThomas / MolScore

An automated scoring function to facilitate and standardize the evaluation of goal-directed generative models for de novo molecular design
MIT License
148 stars 24 forks source link

how to specify rdock together with dask? #53

Closed wenchangzhou-qtx closed 5 months ago

wenchangzhou-qtx commented 5 months ago

Hey @MorganCThomas,

I'm using the GUI to setup my molscore scoring function, and like to include rdock with dask, do you have an example for that? I have a screenshot here but no idea what to provide, what to provide here under "Address to Dask scheduler"? Thanks!

Screenshot 2024-06-06 at 10 28 13 AM
wenchangzhou-qtx commented 5 months ago

One more thing regarding rdock, I guess I need install rDock under acegen environment right?

MorganCThomas commented 5 months ago

Hi, there is a tutorial here but I will try to make it clearer.

If input type is str, then an address is expected such as tcp://0.0.0.0:8000, you will see this after running the command 'dask scheduler` to setup your cluster. However, this is only necessary over more than one compute node. If you are only running it locally on one compute node, you can change the input type to int and simply provide the number of cores/workers to use.

Yes rDock needs to be installed following the (instructions)[https://github.com/CBDD/rDock], it should be independent of your python environment i.e., once installed, available whichever python environment you have. I think this is better than using conda install as I try to restrict additional dependencies in MolScore now.

MorganCThomas commented 5 months ago

I have tried to improve documentation with 33168987780d701c4d5faa4f40ad1aa5ed0b8b32 and e9808b401b13ddda58f14308e7866bdf3a9f7751.

I will leave this open for now to see if you have any further issues.