Welcome to MultirefPredict’s documentation!¶

Automated workflow to predict multireference character of molecules in quantum chemistry calculation

Program Execution¶

A simple example of MultirefPredict’s capabilities of as follows:

>>> import MultirefPredict
>>> import qcelemental
>>> mol = qcelemental.models.Molecule.from_data("""
O                 0.000000000000     0.000000000000    -0.068516245955
H                 0.000000000000    -0.790689888800     0.543701278274
H                 0.000000000000     0.790689888800     0.543701278274
""")

Here we just built a molecule object in qcelemental format. Various multireference diagnostics can be calculated for this molecule with the computeDiagnostic syntax of the diagnostic object, which is initialized with the diagnostic_factory function.

>>> b1 = MultirefPredict.diagnostic_factory("B1",molecule=mol, molname="water", record=False).computeDiagnostic()

b1 is the returned diagnostic value

>>> b1
0.006334860365228678

The I/O of the backend quantum chemistry package is totally hidden from the user. If one would like to get detailed information about the quantum chemistry calculation, it can be easily achieved by setting the record keyword to True.

>>> b1 = MultirefPredict.diagnostic_factory("B1",molecule=mol, molname="water", record=True).computeDiagnostic()

Then the jason files recording the quantum chemistry calculations associated with this diagnostic calculation are automatically dumped in the working directory

>>> ls
B1_water_b1lyp_H.json
B1_water_b1lyp_O.json
B1_water_b1lyp_whole.json
B1_water_blyp_H.json
B1_water_blyp_O.json
B1_water_blyp_whole.json

These files are in the qcelemental result format. For example:

>>> with open('B1_water_b1lyp_H.json') as f:
         data = json.load(f)
             print(json.dumps(data,indent=4))
{
    "molecule": {
            "symbols": [
                "H"
            ],
            "geometry": [
                 0.0,
                 0.0,
                 0.0
             ],
             "molecular_charge": 0.0,
             "molecular_multiplicity": 2
     },
     "driver": "energy",
     "model": {
          "method": "b1lyp",
           "basis": "6-31g"
      },
      ...
}

Backends¶

Currently available compute backends for single results are as follow:

All backends are driven by QCEngine

Installation¶

Currently MultirefPredict can only be installed from source code

Prerequisite: have Anaconda or miniconda installed on your system

Clone MultirefPredict source from gibhub

git clone https://github.com/hjkgrp/MultirefPredict.git

Go to the folder root folder for MultirefPredict, create the conda environment from yaml file
```
cd dev_tools/conda_envs
conda env create -f test_env.yaml
```
Important

It is highly recommended to install the packages needed by creating environment from this yaml file. This help make sure that the version of package installed (e.g. psi4) is the same as the one tested by the developer. Installing the required packages from other distributions may result in unexpected behavior
Now you have created an environment called test for with the prerequisite packages for running MultirefPredict. You can rename the environment as you like by cloning this environment to a new name and remove the origional one. For example, to rename it to multiref:
```
conda create --name multiref --clone test
conda remove --name test --all
```
Activate the conda environment you just created. Go back to the root directory of MultirefPredict (where the setup.py file locates). Local install with pip
```
cd Your_root_path_to_MultirefPrect
pip install -e .
```

Usage¶

This page is still under construction

Examples¶

This page is still under construction