Open Chemistry JupyterLab - Multiple Calculations

Running multiple calculations can be useful when many results are needed. An example is given below.



In [ ]:

    
import openchemistry as oc
import matplotlib.pyplot as plt

Start by creating a list of SMILES to perform the calculations on. Note that generating 3D structure is optional, and can be skipped for batch calculations such as these.



In [ ]:

    
structure_smiles = ['CN1CNCN(C1)c1ccc(c2c1cccc2)C1NCNCN1', 'SC1CSCC(S1)(S)c1cccnc1', 'c1ccc(nc1)c1scc(c1)c1ccnc(c1)c1ccc2c(c1)cccc2']
mols = []
for smiles in structure_smiles:
    mols.append(oc.import_structure(smiles, gen3d=False))

When viewing a molecule that does not have 3D structure, a 2D SVG will be displayed.



In [ ]:

    
mols[0].structure.show()

Multiple calculations can be ran in a single docker container if they are submitted using oc.run_calculations()



In [ ]:

    
image_name = 'openchemistry/chemml:0.6.0'
input_parameters = {}



In [ ]:

    
results = oc.run_calculations(mols, image_name, input_parameters)

All of the calculations will finish simultaneously, and their status may be tracked by requesting one of their outputs:



In [ ]:

    
results[0].properties.show()

Plotting the Results

Once the calculations are finished, their results can be obtained and plotted.



In [ ]:

    
x = []
y = []
for result in results:
    x.append(result.data()['properties']['polarizability'])
    y.append(result.data()['properties']['refractiveIndex'])



In [ ]:

    
fig, ax = plt.subplots()
ax.plot(x, y, 'ro')
ax.set(xlabel='polarizability', ylabel='refractive index', title='Machine learning...')
ax.grid()

Running a 3D Calculation on a 2D Molecule

If a 3D calculation is attempted on a 2D molecule, 3D coordinate generation will automatically start, and a message will alert the user to try again soon.



In [ ]:

    
image_name = 'openchemistry/psi4:1.2.1'
input_parameters = {
    'theory': 'dft',
    'functional': 'b3lyp',
    'basis': '6-31g'
}



In [ ]:

    
result = mols[0].energy(image_name, input_parameters)



In [ ]:

    
result.orbitals.show(mo='homo', iso=0.005)

Alternatively, 3D coordinates may also be generated manually. If this is done, the forcefield options may also be passed.



In [ ]:

    
mols[1].structure.generate_3d(forcefield='uff', steps=150)

Specifying the Type for Multiple Calculations

For typed calculations, the type can be specified by adding the appropriate "task" keyword to the input parameters.



In [ ]:

    
input_parameters['task'] = 'energy'
results = oc.run_calculations(mols, image_name, input_parameters)



In [ ]:

    
results[0].orbitals.show(mo='homo', iso=0.005)