Tutorial01

[faranak1991]

Hello,

I am attempting to understand and replicate the steps taken in your work by following the tutorials. Before running the code, I make sure to install all the necessary packages and dependencies using Conda. However, I encountered some errors as indicated below. I would greatly appreciate it if you could inform me of any additional packages that need to be installed ?

data_set = ThermoMLDataSet.from_doi(
"10.1016/j.fluid.2013.10.034",
"10.1021/je1013476",)

len(data_set)

TypeError Traceback (most recent call last)
Cell In[9], line 1
----> 1 len(data_set)
TypeError: object of type 'NoneType' has no len()

---

from openff.evaluator.properties import EnthalpyOfVaporization

AttributeError Traceback (most recent call last)
AttributeError: module 'pint.compat' has no attribute 'upcast_types'

[lilyminium]

@faranak1991 could you please paste your conda environment (conda list)? For the second error, I have a feeling you may need to downgrade pint to 0.20.1, with something like conda install -c conda-forge pint==0.20.1 or pip uninstall pint && pip install pint==0.20.1.

[faranak1991]

Thank you for your reply @lilyminium. The error was resolved by installing the specific version of Pint. Additionally, I have provided the list of packages installed in the Conda environment in attached file.
Regarding tutorial02, the "estimated_data_set.json" file does not show the properties as expected, it appears as:
{
"@type": "openff.evaluator.datasets.datasets.PhysicalPropertyDataSet",
"properties": []
}
Do you have any recommendations or advice regarding this?

conda-list.txt

[mattwthompson]

That indicates something went wrong earlier in the tutorial when estimating properties. There should be some error messages a few cells up indicating something crashed or possibly didn't start.

[faranak1991]

Thanks @mattwthompson . I see that certain properties have already been specified as follows. Is there any possibility to include transport properties such as viscosity in the existing set?
Also, does the openff-evaluator accurately estimate these new properties?
(particularly for a molecule with 44 atoms)

<class 'openff.evaluator.properties.binding.HostGuestBindingAffinity'>
<class 'openff.evaluator.properties.density.Density'>
<class 'openff.evaluator.properties.density.ExcessMolarVolume'>
<class 'openff.evaluator.properties.dielectric.DielectricConstant'>
<class 'openff.evaluator.properties.enthalpy.EnthalpyOfMixing'>
<class 'openff.evaluator.properties.enthalpy.EnthalpyOfVaporization'>
<class 'openff.evaluator.properties.solvation.SolvationFreeEnergy'>

[mattwthompson]

The supported properties are just what are available in the API right now; PRs for others would be welcome.

Evaluator has been used to accurately, we believe, compute those properties in a handful of studies. It's hard to say it'll work for every molecule of a particular size, but any system suitable for equilibrium MD will likely work fine.