finished power transform; removed bug; removed globals

multitask/Demo.ipynb

@@ -27,27 +27,19 @@
  "metadata": {},
  "outputs": [],
  "source": [
- "def predict_property_neural(prop=None, mol_id=-1):\n",
+ "def predict_property_neural(*, clusters, A_all, X_all, E_all, prop=None, mol_id=-1):\n",
+ " kwargs = {\"clusters\": clusters, \"A_all\": A_all, \"X_all\": X_all, \"E_all\": E_all}\n",
  " if mol_id == -1:\n",
  " raise ValueError(\"ID must be between 1 and 133885\")\n",
  " if prop == 'gap':\n",
- " lumo = predict_property_neural(prop='lumo', mol_id=mol_id)\n",
- " homo = predict_property_neural(prop='homo', mol_id=mol_id)\n",
+ " lumo = predict_property_neural(prop='lumo', mol_id=mol_id, **kwargs)\n",
+ " homo = predict_property_neural(prop='homo', mol_id=mol_id, **kwargs)\n",
  " return lumo - homo\n",
  " \n",
  " if not any([prop in cluster for cluster in clusters]):\n",
  " raise ValueError(\"Property was not found in clusters list\")\n",
  " \n",
- " return predict_property(prop, mol_id, clusters, N=N, F=F, S=S)"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
- "predict_property_neural('A', 1)"
+ " return predict_property(prop=prop, mol_id=mol_id, **kwargs)"
  ]
  },
  {
@@ -267,11 +259,7 @@
  {
  "cell_type": "code",
  "execution_count": null,
- "metadata": {
- "jupyter": {
- "source_hidden": true
- }
- },
+ "metadata": {},
  "outputs": [],
  "source": [
  "def batch_process(start_num, end_num, thermochemical=False):\n",
@@ -330,7 +318,7 @@
  "metadata": {},
  "outputs": [],
  "source": [
- "def lookup_property(prop=None, mol_id=-1):\n",
+ "def lookup_property(y_all, prop=None, mol_id=-1):\n",
  " return y_all.loc[mol_id-1, prop]"
  ]
  },
@@ -340,7 +328,7 @@
  "metadata": {},
  "outputs": [],
  "source": [
- "def prompt_user_for_calculation():\n",
+ "def prompt_user_for_calculation(*, clusters, y_all, A_all, X_all, E_all):\n",
  " while True:\n",
  " num = -1\n",
  " while num < 1 or num > 133885:\n",
@@ -376,11 +364,12 @@
  " else:\n",
  " print('Beginning neural calculation:')\n",
  " start = time.time()\n",
- " print(predict_property_neural(prop=prop, mol_id=num))\n",
+ " kwargs = {\"clusters\": clusters, \"A_all\": A_all, \"X_all\": X_all, \"E_all\": E_all}\n",
+ " print(predict_property_neural(prop=prop, mol_id=num, **kwargs))\n",
  " end = time.time()\n",
  " print('Neural method took', end-start, 's')\n",
  " print('Actual data:')\n",
- " print(lookup_property(prop=prop, mol_id=num))"
+ " print(lookup_property(y_all, prop=prop, mol_id=num))"
  ]
  },
  {
@@ -389,7 +378,13 @@
  "metadata": {},
  "outputs": [],
  "source": [
- "prompt_user_for_calculation()"
+ "A_all, X_all, E_all, y_all = load_data()\n",
+ "\n",
+ "clusters = [['A', 'B', 'alpha'], \n",
+ " ['C', 'r2', 'u0'],\n",
+ " ['zpve', 'g298', 'cv'],\n",
+ " ['lumo', 'u298', 'h298'],\n",
+ " ['mu', 'homo']]"
  ]
  },
  {
@@ -398,19 +393,15 @@
  "metadata": {},
  "outputs": [],
  "source": [
- "properties = ['A', 'B', 'C', 'mu', 'homo', 'lumo', 'gap', 'zpve', 'u0', 'u298', 'h298', 'g298', 'cv']\n",
- "\n",
- "for prop in properties:\n",
- " print(prop)\n",
- " errors = list()\n",
- " for index in np.random.choice(133885, 10):\n",
- " pred = predict_property_neural(prop=prop, mol_id=index+1)\n",
- " actual = lookup_property(prop=prop, mol_id=index+1)\n",
- " err = abs((pred-actual)/actual)\n",
- " print(f\"error on molecule {index+1} is {err:.2%}\")\n",
- " errors.append(err)\n",
- " print(f'avg error of {prop} is {sum(errors)/len(errors):.2%}')\n"
+ "prompt_user_for_calculation(clusters=clusters, A_all=A_all, X_all=X_all, E_all=E_all, y_all=y_all)"
  ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": []
  }
  ],
  "metadata": {