Separate notebooks for exact answer, steps in fitting (#14)

* Add utility for making calculator

Avoid copying code between notebooks

* Separate exact, sampling and approximation

* Accept "None" as an option

jitterbug/utils.py

@@ -0,0 +1,36 @@
+"""Utility functions"""
+from typing import Optional
+
+
+from ase.calculators.calculator import Calculator
+from ase.calculators.mopac import MOPAC
+from ase.calculators.psi4 import Psi4
+
+mopac_methods = ['pm7']
+"""List of methods for which we will use MOPAC"""
+
+
+def make_calculator(method: str, basis: Optional[str], **kwargs) -> Calculator:
+    """Make an ASE calculator that implements a desired method.
+
+    This function will select the appropriate quantum chemistry code depending
+    on the method name using the following rules:
+
+    1. Use MOPAC if the method is PM7.
+    2. Use Psi4 otherwise
+
+    Any keyword arguments are passed to the calculator
+
+    Args:
+        method: Name of the quantum chemistry method
+        basis: Basis set name, if appropriate
+    Returns:
+        Calculator defined according to the user's settings
+    """
+
+    if method in mopac_methods:
+        if not (basis is None or basis == "None"):
+            raise ValueError(f'Basis must be none for method: {method}')
+        return MOPAC(method=method, command='mopac PREFIX.mop > /dev/null')
+    else:
+        return Psi4(method=method, basis=basis, **kwargs)

notebooks/proof-of-concept/1_compute-random-offsets.ipynb → notebooks/1_explore-sampling-methods/0_random-directions-same-distance.ipynb

@@ -18,8 +18,7 @@
    },
    "outputs": [],
    "source": [
-    "from ase.calculators.mopac import MOPAC\n",
-    "from ase.calculators.psi4 import Psi4\n",
+    "from jitterbug.utils import make_calculator\n",
     "from ase.io import write, read\n",
     "from ase.db import connect\n",
     "from ase import Atoms\n",
@@ -40,21 +39,17 @@
   {
    "cell_type": "code",
    "execution_count": 2,
-   "id": "c6be56c5-a460-4acd-9b89-8c3d9c812f5f",
+   "id": "52252ee2-315c-48bb-8cba-07620e6e2faa",
    "metadata": {
     "tags": [
      "parameters"
     ]
    },
    "outputs": [],
    "source": [
-    "molecule_name = 'caffeine'\n",
-    "method = 'pm7'\n",
-    "mopac_methods = ['pm7']  # Use MOPAC for these methods\n",
-    "basis = None  # Set to None for MOPAC methods\n",
+    "starting_geometry = '../data/exact/water_b3lyp_def2-svpd.xyz'\n",
     "threads = min(os.cpu_count(), 12)\n",
-    "step_size: float = 0.01 # Perturbation amount, used as maximum L2 norm\n",
-    "assert (method in mopac_methods) == (basis is None), 'Use a basis of None for MOPAC computations'"
+    "step_size: float = 0.005 # Perturbation amount, used as maximum L2 norm"
    ]
   },
   {
@@ -72,7 +67,8 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "run_name = f'{molecule_name}_{method}_{basis}'"
+    "run_name = Path(starting_geometry).name[:-4]\n",
+    "name, method, basis = run_name.split(\"_\")"
    ]
   },
   {
@@ -95,7 +91,7 @@
     {
      "data": {
       "text/plain": [
-       "Atoms(symbols='O2N4C8H10', pbc=False, forces=..., calculator=SinglePointCalculator(...))"
+       "Atoms(symbols='OH2', pbc=False, forces=..., calculator=SinglePointCalculator(...))"
       ]
      },
      "execution_count": 4,
@@ -104,7 +100,7 @@
     }
    ],
    "source": [
-    "atoms = read(Path('data') / 'exact' / f'{run_name}.xyz')\n",
+    "atoms = read(starting_geometry)\n",
     "atoms"
    ]
   },
@@ -134,8 +130,8 @@
    },
    "outputs": [],
    "source": [
-    "out_dir = Path('data') / 'approx'\n",
-    "out_dir.mkdir(exist_ok=True)"
+    "out_dir = Path('data') / 'random-dir-same-dist'\n",
+    "out_dir.mkdir(exist_ok=True, parents=True)"
    ]
   },
   {
@@ -147,7 +143,7 @@
    },
    "outputs": [],
    "source": [
-    "db_path = out_dir / f'{run_name}-random-d={step_size:.2e}.db'"
+    "db_path = out_dir / f'{run_name}_d={step_size:.2e}.db'"
    ]
   },
   {
@@ -187,12 +183,17 @@
    "metadata": {
     "tags": []
    },
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "  Threads set to 1 by Python driver.\n"
+     ]
+    }
+   ],
    "source": [
-    "if method not in mopac_methods:\n",
-    "    calc = Psi4(method=method, basis=basis, num_threads=threads, memory='4096MB')\n",
-    "else:\n",
-    "    calc = MOPAC(method=method, command='mopac PREFIX.mop > /dev/null')"
+    "calc = make_calculator(method, basis, threads=threads)"
    ]
   },
   {
@@ -215,7 +216,7 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "Need to run 2701 calculations for full accuracy.\n"
+      "Need to run 55 calculations for full accuracy.\n"
      ]
     }
    ],
@@ -237,7 +238,7 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "Already done 1. 2700 left to do.\n"
+      "Already done 1. 54 left to do.\n"
      ]
     }
    ],
@@ -257,7 +258,21 @@
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 2701/2701 [03:10<00:00, 19.27it/s]"
+      "  0%|                                                                                                                                                                                                                                                              | 0/55 [00:00<?, ?it/s]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "  Threads set to 1 by Python driver.\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 55/55 [00:40<00:00,  1.30it/s]"
      ]
     }
    ],