ML_corrections.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import os, re, shutil\n",
    "import ROOT as rt\n",
    "\n",
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "from sklearn.neural_network import MLPRegressor\n",
    "from sklearn.svm import SVR\n",
    "\n",
    "from root_numpy import root2array\n",
    "from lib.histo_utilities import create_TH1D, create_TH2D\n",
    "from lib.cebefo_style import cebefo_style"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "cebefo_style()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "input_file = '../data/test_Run768.root'\n",
    "validation_file = '../data/test_Run769.root'\n",
    "ch = 14"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data = root2array(input_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "canvas = {}\n",
    "canvas['amp'] = {}\n",
    "canvas['int'] = {}\n",
    "canvas['wave'] = {}\n",
    "canvas['pos'] = {}\n",
    "canvas['w_pos'] = {}\n",
    "\n",
    "info = {ch: {}, 9:{}}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxwAAAI8CAIAAAD0vjrdAAAABmJLR0QAAAAAAAD5Q7t/AAAgAElE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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxwAAAI8CAIAAAD0vjrdAAAABmJLR0QAAAAAAAD5Q7t/AAAgAElE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\nFQAAQAIUVQAAAAn8f+zCb4kR3teAAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for k in [9, ch]:\n",
    "    \n",
    "    '''=========================== Amplitude ==========================='''\n",
    "    canvas['amp'][k] = rt.TCanvas('c_amp_'+str(k), 'c_amp_'+str(k), 800, 600)\n",
    "    \n",
    "    name = 'h_amp_'+str(k)\n",
    "    title = 'Amplitude channel '+str(k)\n",
    "    amp = (data['amp'].T)[k]\n",
    "    h = create_TH1D(amp, name, title, \n",
    "                    binning = [100, 0, 550],\n",
    "                    axis_title = ['Peak amplitude [mV]', 'Events / 5.5 mV'])\n",
    "            \n",
    "    h.GetXaxis().SetRange(int(40/5.5)+1,int(450/5.5)+1)\n",
    "    i_max = h.GetMaximumBin()\n",
    "    h.GetXaxis().SetRange(1,100)\n",
    "    peak = h.GetBinCenter(i_max)\n",
    "    info[k]['amp_range'] = [max(20,peak*0.6), min(450, peak*1.6)]\n",
    "    res = h.Fit('landau','LQSR', '', info[k]['amp_range'][0], info[k]['amp_range'][1])\n",
    "    h.SetStats(0)\n",
    "    \n",
    "    if(h.GetMaximum() - h.GetMinimum() > 1000):\n",
    "        canvas['amp'][k].SetLogy()\n",
    "    h.DrawCopy('E1')\n",
    "    canvas['amp'][k].Update()    \n",
    "    canvas['amp'][k].Draw()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def create_sample(input_file, selection):\n",
    "    data = root2array(input_file, selection=selection)\n",
    "    amp = (data['amp'].T)[ch]\n",
    "\n",
    "#     risetime = (data['LP2_90'].T)[ch] - (data['LP2_10'].T)[ch]\n",
    "\n",
    "    x_pos = (data['x_dut'].T)[0]\n",
    "    y_pos = (data['y_dut'].T)[0]\n",
    "    \n",
    "    decay_time = (data['decaytime'].T)[ch]\n",
    "    \n",
    "    integral = (data['integral'].T)[ch]\n",
    "\n",
    "\n",
    "    x_train = np.column_stack((amp, integral, decay_time, x_pos, y_pos))\n",
    "\n",
    "    time_stamp = (data['LP2_10'].T)[ch]\n",
    "    time_ref = (data['gaus_mean'].T)[9]\n",
    "    \n",
    "    y_train = time_stamp - time_ref\n",
    "    \n",
    "    return x_train, y_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(2927, 5)\n"
     ]
    }
   ],
   "source": [
    "selection = 'amp[9]>50 && amp[9]<200 && amp['+str(ch)+']>300 && amp['+str(ch)+']<460 && x_dut[0]>-99 && y_dut[0]>-99'\n",
    "selection += ' && (LP2_10['+str(ch)+'] - gaus_mean[9]) >-11 && (LP2_10['+str(ch)+'] - gaus_mean[9]) < -10'\n",
    "\n",
    "x_train, y_train = create_sample(input_file, selection)\n",
    "print x_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "gbr = GradientBoostingRegressor(loss='ls',\n",
    "                                n_estimators=1000, max_depth=3,\n",
    "                                learning_rate=.2, min_samples_leaf=20,\n",
    "                                min_samples_split=20,\n",
    "                                verbose=0\n",
    "                                )\n",
    "\n",
    "# gbr = SVR(C=2, epsilon= )\n",
    "\n",
    "# gbr = MLPRegressor(hidden_layer_sizes=(100,50,30),\n",
    "#                    activation='tanh',\n",
    "#                    learning_rate='adaptive',\n",
    "#                    solver='sgd',\n",
    "#                    verbose=2\n",
    "#                   )\n",
    "\n",
    "gbr.fit(x_train, y_train)\n",
    "y_train_pred = gbr.predict(x_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Post:  40.4303556217\n",
      "Pre:  100.713201467\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Warning in <TCanvas::Constructor>: Deleting canvas with same name: c_train\n",
      "Warning in <TROOT::Append>: Replacing existing TH1: h_train (Potential memory leak).\n",
      "Warning in <TROOT::Append>: Replacing existing TH1: h_train (Potential memory leak).\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxwAAAI8CAIAAAD0vjrdAAAABmJLR0QAAAAAAAD5Q7t/AAAgAElE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\nAMAJKKoAAABOQFEFAABwgv8PuIt3Hpl4Kq0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "canvas = rt.TCanvas('c_train', 'c_train', 800, 600)\n",
    "\n",
    "h = create_TH1D(y_train - y_train_pred, 'h_train', 'h_train', binning=[200,-1,1])\n",
    "h.SetLineColor(4)\n",
    "print 'Post: ',1000*h.GetRMS()\n",
    "h.DrawCopy('E1')\n",
    "\n",
    "h = create_TH1D(y_train+10, 'h_train', 'h_train', binning=[200,-1,1])\n",
    "h.SetLineColor(2)\n",
    "print 'Pre: ',1000*h.GetRMS()\n",
    "h.DrawCopy('E1SAMES')\n",
    "\n",
    "canvas.Draw()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(2071, 5)\n",
      "Post:  71.7703383245\n",
      "Pre:  100.56012073\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Warning in <TROOT::Append>: Replacing existing TH1: h_val (Potential memory leak).\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxwAAAI8CAIAAAD0vjrdAAAABmJLR0QAAAAAAAD5Q7t/AAAgAElE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\nfJDP/q3KAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x_val, y_val = create_sample(validation_file, selection)\n",
    "print x_val.shape\n",
    "\n",
    "y_val_pred = gbr.predict(x_val)\n",
    "\n",
    "canvas = rt.TCanvas('c_val', 'c_val', 800, 600)\n",
    "\n",
    "h = create_TH1D(y_val - y_val_pred, 'h_val', 'h_val', binning=[100,-1,1])\n",
    "h.SetLineColor(4)\n",
    "print 'Post: ',1000*h.GetRMS()\n",
    "h.DrawCopy('E1')\n",
    "\n",
    "\n",
    "h = create_TH1D(y_val+10, 'h_val', 'h_val', binning=[100,-1,1])\n",
    "h.SetLineColor(2)\n",
    "print 'Pre: ',1000*h.GetRMS()\n",
    "h.DrawCopy('E1SAMES')\n",
    "\n",
    "\n",
    "\n",
    "canvas.Draw()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}