logC-SFT.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Run this if you have not installed XGBoost already\n",
    "\n",
    "#pip install xgboost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import pickle\n",
    "import xgboost as xgb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Maximum Surface Excess Concentration Prediction\n",
    "def predict_Gamma_max(MaxSEC_filepath):     #MaxSEC_filepath is the file path to the AlvaDesc descriptors used to \n",
    "                                            #predict the maximum surface excess concentration (in the order shown in the SI)\n",
    "    X_pred_Gamma_max = pd.read_excel(MaxSEC_filepath)\n",
    "    estimator_MaxSEC = xgb.Booster()\n",
    "    estimator_MaxSEC.load_model(\"MaxSEC-model.json\")\n",
    "    D_pred_Gamma_max = xgb.DMatrix(X_pred_Gamma_max)\n",
    "    y_pred_Gamma_max = estimator_MaxSEC.predict(D_pred_Gamma_max)\n",
    "    y_pred_Gamma_max *= 10**-6\n",
    "    return(y_pred_Gamma_max)\n",
    "\n",
    "\n",
    "#logKL Prediction\n",
    "def predict_KL(KL_filepath):          #KL_filepath is the file path to the SIRMS descriptors used to \n",
    "                                      #predict KL (in the order shown in the SI)\n",
    "    X_pred_logKL = pd.read_excel(KL_filepath)\n",
    "    estimator_logKL = xgb.Booster()\n",
    "    estimator_logKL.load_model(\"logKL-model.json\")\n",
    "    D_pred_logKL = xgb.DMatrix(X_pred_logKL)\n",
    "    y_pred_logKL = estimator_logKL.predict(D_pred_logKL)\n",
    "    KL = 10**y_pred_logKL\n",
    "    return(KL)\n",
    "\n",
    "\n",
    "#logCMC Prediction\n",
    "def predict_CMC(CMC_filepath):        #CMC_filepath is the file path to the CDK descriptors used to \n",
    "                                      #predict CMC (in the order shown in the SI)\n",
    "    X_pred_logCMC = pd.read_excel(CMC_filepath)\n",
    "    estimator_logCMC = xgb.Booster()\n",
    "    estimator_logCMC.load_model(\"logCMC-model.json\")\n",
    "    D_pred_logCMC = xgb.DMatrix(X_pred_logCMC)\n",
    "    y_pred_logCMC = estimator_logCMC.predict(D_pred_logCMC)\n",
    "    CMC = 10**y_pred_logCMC\n",
    "    return(CMC)\n",
    "\n",
    "\n",
    "#Prediction for All\n",
    "def predict_all(MaxSEC_filepath, KL_filepath, CMC_filepath):\n",
    "    MaxSEC_list = predict_Gamma_max(MaxSEC_filepath)\n",
    "    KL_list = predict_KL(KL_filepath)\n",
    "    CMC_list = predict_CMC(CMC_filepath)\n",
    "    \n",
    "    for i, _ in enumerate(MaxSEC_list):\n",
    "        print('Surfactant ', i, ' MaxSEC, KL and CMC: ', MaxSEC_list[i], KL_list[i], CMC_list[i])\n",
    "        \n",
    "        \n",
    "        \n",
    "        \n",
    "\n",
    "#Plots the SFT-log(c) isotherm(s) for surfactants. MaxSEC_filepath is the file path to the AlvaDesc descriptors used to \n",
    "#predict the maximum surface excess concentration, logKL_filepath is the file path to the SIRMS descriptors used to \n",
    "#predict KL, logCMC_filepath is the file path to the CDK descriptors used to predict CMC, savedir is the figure directory\n",
    "#you wish to save in, alltogether is used to plot the surfactant isotherms all on one graph, type 'alltogether=True' if\n",
    "#you want them plotted together, otherwise ignore. These isotherms use the Szyszkowski equation.\n",
    "def predict_isotherm(MaxSEC_filepath, KL_filepath, CMC_filepath, savedir, alltogether=False):\n",
    "    MaxSEC_list = predict_Gamma_max(MaxSEC_filepath)\n",
    "    KL_list = predict_KL(KL_filepath)\n",
    "    CMC_list = predict_CMC(CMC_filepath)\n",
    "    colors = plt.cm.Blues(np.linspace(0.4,1,len(MaxSEC_list)))\n",
    "    \n",
    "    xfine = np.geomspace(1e-10, 10, 100000)\n",
    "    logxfine = np.log10(xfine)\n",
    "    \n",
    "    for i, _ in enumerate(MaxSEC_list):\n",
    "        print('Surfactant ', i, ' MaxSEC, KL and CMC: ', MaxSEC_list[i], KL_list[i], CMC_list[i])\n",
    "        MaxSEC = MaxSEC_list[i]\n",
    "        KL = KL_list[i]\n",
    "        CMC = CMC_list[i]\n",
    "        \n",
    "        yfine = []\n",
    "\n",
    "        for j, _ in enumerate(xfine):\n",
    "            if xfine[j] < CMC:\n",
    "                yfinei = 72-(8314 * 298.15 * MaxSEC * np.log(1 + (KL*xfine[j]*1000)))\n",
    "                yfine.append(yfinei)\n",
    "            else:\n",
    "                yfinei = 72-(8314 * 298.15 * MaxSEC * np.log(1 + (KL*CMC*1000)))\n",
    "                yfine.append(yfinei)\n",
    "                \n",
    "        if alltogether==True:\n",
    "            plt.plot(logxfine, yfine, color=colors[i], label='Surfactant '+str(i+1))\n",
    "            plt.title('All Surfactants')\n",
    "            plt.xlabel('log(c)')\n",
    "            plt.ylabel('SFT (mN/m)')\n",
    "            plt.legend()\n",
    "            \n",
    "            plt.savefig(savedir+'/AllSurfactants.jpg', dpi=1500, bbox_inches='tight')\n",
    "        else:\n",
    "            plt.plot(logxfine, yfine, 'b')\n",
    "            plt.title('Surfactant '+str(i+1))\n",
    "            plt.xlabel('log(c)')\n",
    "            plt.ylabel('SFT (mN/m)')\n",
    "            \n",
    "            \n",
    "            plt.savefig(savedir+'/Surfactant '+str(i+1)+'.jpg', dpi=1500, bbox_inches='tight')\n",
    "            plt.clf()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [],
   "source": [
    "# To start predicting, simply type 'predict_isotherm(MaxSEC_filepath, KL_filepath, CMC_filepath, savedir)'\n",
    "# with your descriptor filepaths and the folder you wish to save the figures within."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}