file cleanup

Author: tom-bickley

.gitignore

@@ -1 +1,3 @@
 .DS_Store
+.venv
+__pycache__

README.md

@@ -1,9 +1,10 @@
 # extending_digitaldiscovery_data
 
-This repository contains supplimentary data for the publication "Extending Quantum Computing through Subspace, Embedding and Classical Molecular Dynamics Techniques".
+This repository contains supplementary data for the publication "*Extending Quantum Computing through Subspace, Embedding and Classical Molecular Dynamics Techniques*", available on Digital Discovery and [arXiv](https://doi.org/10.48550/arXiv.2505.16796).
 
 The data is structured as follows:
 - `input/` - contains the input data and scripts required to run the QM/MM simulation presented in Section 5.
 - `quantum/` - contains the Hamiltonian data from the QM/MM simulation and the the scripts for pre- and post-processing circuit executions.
 - `plot/` - contains the data and scripts needed to plot Figure 9.
+- `requirements.txt` - required python packages to run the scripts in `quantum/` and `plot/`. Set up a new virtual environment and run `pip install -r requirements.txt`. We used python `3.11.9`. For instructions to run the scripts for the QM/MM simulation please see `input/README.md`.
 - `LICENSE` - this repository is distributed under the GNU General Public License v2 (GPLv2), because it incorporates code adapted from [LAMMPS](https://github.com/lammps/lammps). In particular, the file `input/run.py` is adapted from the LAMMPS source code. The full GPLv2 license for LAMMPS is included here.

input/README.md

@@ -2,6 +2,6 @@
 
 This folder contains the data and scripts required to run the QM/MM simulation presented in Section 5.
 
-A prerequisite is the installation of LAMMPS with the `mdi`, `molecule`, and `kspace` flags switched on. Instructions for this can be found in the [LAMMPS documentation](https://github.com/lammps/lammps/blob/0d479cae298abb73ffd8b1460197a8a74c63bc5e/examples/QUANTUM/PySCF/README).
+A prerequisite is the installation of LAMMPS with the `mdi`, `molecule`, and `kspace` flags switched on. Instructions for this can be found in the [LAMMPS documentation](https://github.com/lammps/lammps/blob/0d479cae298abb73ffd8b1460197a8a74c63bc5e/examples/QUANTUM/PySCF/README). We recommend that you set this up in a new virtual environment, with the required python version `>=3.8, <3.11`. You will also need to install Nbed version 0.0.7 after completing the LAMMPS installation via `pip install nbed==0.0.7`.
 
 Then, the script `run.sh` should be executable and can be edited to run simulations for each of the starting geometries. The LAMMPS input geometries are found in the `data.*` files. For example, `data.19_200` corresponds to the geometry for $d_{\mathrm{OO}} = 1.9$ Å and $d_{\mathrm{OH}} = 0.200$, and similarly for the other files. The output from running the script will be the Hamiltonian data in a pickle file.

input/run.py

@@ -6,7 +6,7 @@
 # This file is adapted from LAMMPS (https://github.com/lammps/lammps/blob/develop/examples/QUANTUM/PySCF/pyscf_mdi.py)
 # Copyright 2003 Sandia Corporation – Licensed under GPLv2
 # Modified by Tom Bickley, 2024
-# The changes are from line 570 onwards, which relate to the PySCF and Nbed settings
+# The changes are from line 568 onwards, which relate to the PySCF and Nbed settings
 
 import sys,time
 
@@ -24,8 +24,6 @@
 from nbed.driver import NbedDriver
 from nbed.ham_builder import HamiltonianBuilder
 from nbed.ham_converter import HamiltonianConverter
-
-import json
 import pickle
 from numbers import Number
 

plot/README.md

@@ -1,3 +1,3 @@
 # plot/
 
-This folder contains the data and scripts needed to plot Figure 9.
+This folder contains the data and script needed to plot Figure 9.

plot/hardware_results.pdf

@@ -0,0 +1,77 @@
+import numpy as np
+import matplotlib.pyplot as plt
+import json
+
+with open("energies.json", "r") as f:
+    data = json.load(f)
+
+dOHs = [200, 275, 350, 425, 500]
+    
+hams = {
+    "1.9": [f"19_{dOH}_mean.json" for dOH in dOHs],
+    "2.6": [f"26_{dOH}_mean.json" for dOH in dOHs],
+    "3.3": [f"33_{dOH}_mean.json" for dOH in dOHs],
+}
+
+energy_keys = ["HF_energy", "DMRG_energy_2", "DMRG_energy_16", "FCI_energy", "QSCI_noisy_energy"]
+
+results = {k: {ek: [] for ek in energy_keys} for k in hams}
+
+for bond_length, hams in hams.items():
+    for ham in hams:
+        for ek in energy_keys:
+            results[bond_length][ek].append(data[ham][ek])
+
+r = np.array(dOHs) / 1000
+
+plotlist = [
+    [1.9, 'solid'],
+    [2.6, 'dashed'],
+    [3.3, 'dotted'],
+]
+
+for el in plotlist:
+    dOO = el[0]
+
+    sol_hf = np.array(results[str(dOO)]['HF_energy'])
+    sol_dmrg2 = np.array(results[str(dOO)]['DMRG_energy_2'])
+    sol_fci = np.array(results[str(dOO)]['FCI_energy'])
+    sol_qsci = np.array(results[str(dOO)]['QSCI_noisy_energy'])
+
+    if dOO == 1.9:
+        plt.plot(r, sol_hf - sol_fci, label='HF', 
+                 color='#648FFF', linestyle=el[1], marker='o')
+        plt.plot(r, sol_dmrg2 - sol_fci, label='DMRG', 
+                 color='#DC267F', linestyle=el[1], marker='o')
+        plt.plot(r, sol_qsci - sol_fci, label='QSCI', 
+                 color='#FFB000', linestyle=el[1], marker='o')
+    else:
+        plt.plot(r, sol_hf - sol_fci, color='#648FFF',
+                 linestyle=el[1], marker='o')
+        plt.plot(r, sol_dmrg2 - sol_fci, color='#DC267F',
+                 linestyle=el[1], marker='o')
+        plt.plot(r, sol_qsci - sol_fci, color='#FFB000', 
+                 linestyle=el[1], marker='o')
+
+xlim = plt.xlim()
+ylim = plt.ylim()
+
+plt.fill([0,1,1,0], [-1,-1,1.6/1000,1.6/1000], color='green', alpha=0.25)
+
+plt.xlim(xlim)
+plt.ylim(ylim)
+
+plt.xlabel('proton ratio')
+plt.ylabel(r'Δ$E$ / Ha')
+
+plt.plot([], [], label=r'$d_{\mathrm{OO}} = 1.9$ Å', 
+         linestyle='solid', color='grey')
+plt.plot([], [], label=r'$d_{\mathrm{OO}} = 2.6$ Å',
+         linestyle='dashed', color='grey')
+plt.plot([], [], label=r'$d_{\mathrm{OO}} = 3.3$ Å',
+         linestyle='dotted', color='grey')
+
+plt.legend()
+plt.tight_layout()
+plt.savefig('hardware_results.pdf')
+plt.show()