5-parameter-estimation-Likelihood-Objective.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import emcee\n",
    "from scipy.optimize import minimize\n",
    "import corner\n",
    "from pyDOE import * # native python latin hypercube, no need for R\n",
    "\n",
    "import seaborn as sns #pretty\n",
    "sns.set()\n",
    "sns.set_context('talk') #readable\n",
    "\n",
    "#Our expensive physics simulation\n",
    "import run_simple_energy_loss\n",
    "\n",
    "import ptemcee\n",
    "import subprocess\n",
    "import pandas as pd\n",
    "\n",
    "from joblib import Parallel, delayed\n",
    "\n",
    "# Gaussian Process emulation\n",
    "from sklearn.gaussian_process import GaussianProcessRegressor as GPR #for using Gaussian Processes\n",
    "from sklearn.gaussian_process import kernels #same\n",
    "from sklearn.preprocessing import StandardScaler #useful for scaling data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "#this function runs the physics simulator\n",
    "simulation=run_simple_energy_loss.run_simulation\n",
    "\n",
    "pT_range = (1, 10, 10) \n",
    "pT_bins = [3, 5, 7] # p_T bins we will use for the analysis\n",
    "n_pT_bins = len(pT_bins)\n",
    "\n",
    "#ranges/boundaries of prior\n",
    "prior_ranges= {'alpha_s':(0.05,0.5), 'exponent_inel':(-3,3), 'exponent_el':(-3,3),'scale_inel':(0.1,0.5), 'scale_el':(0.1,0.5)}\n",
    "true_values= {'alpha_s': 0.2, 'exponent_inel': -1, 'exponent_el': 1, 'scale_inel': 0.3, 'scale_el': 0.3}\n",
    "\n",
    "prior_midpoint = np.array([.25, 0., 0., 0.3, 0.3]).reshape(1,-1)\n",
    "prior_bds = ((0.05,0.5), (-3,3), (-3,3), (0.1,0.5), (0.1,0.5))\n",
    "\n",
    "# We generate observables using the simulator at these parameters \n",
    "# and consider them as experimental pseudo-data\n",
    "true_params = np.array([true_values['alpha_s'],  true_values['exponent_inel'],\n",
    "                 true_values['exponent_el'],true_values['scale_inel'],true_values['scale_el']]).reshape(1,-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "running simulator with parameters : [ 0.2 -1.   1.   0.3  0.3]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 25.2 s, sys: 106 ms, total: 25.3 s\n",
      "Wall time: 25.5 s\n"
     ]
    }
   ],
   "source": [
    "# Set seed for reproducibility.\n",
    "seed = 42\n",
    "np.random.seed(seed)\n",
    "\n",
    "# Generate some synthetic data from the model.\n",
    "obs_err = 0.02 # Amplitude of noise term\n",
    "%time obs_Y = simulation(true_params, *pT_range)[:,pT_bins].flatten() # True model\n",
    "obs_dY = obs_err * np.random.randn(n_pT_bins) # Add some random noise\n",
    "obs_Y += obs_dY"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1, 5)\n",
      "(3,)\n",
      "(3,)\n"
     ]
    }
   ],
   "source": [
    "print(true_params.shape)\n",
    "print(obs_Y.shape)\n",
    "print(obs_dY.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot pseudo-data for calibration\n",
    "plt.errorbar(pT_bins, obs_Y, yerr=obs_dY, marker='o', ls='None')\n",
    "plt.xlabel(r\"$p_T$\")\n",
    "plt.ylabel(r\"$R_{AA}$\")\n",
    "plt.title('pseudo-data')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Cell below defines some conditions we require of the gaussian process regressors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# size of possible variation of parameters\n",
    "ptp = np.array([prior_ranges['alpha_s'][1],prior_ranges['exponent_inel'][1],prior_ranges['exponent_el'][1],prior_ranges['scale_inel'][1],prior_ranges['scale_el'][1]])\\\n",
    "-np.array([prior_ranges['alpha_s'][0],prior_ranges['exponent_inel'][0],prior_ranges['exponent_el'][0],prior_ranges['scale_inel'][0],prior_ranges['scale_el'][0]])\n",
    "\n",
    "#Squared Exponential Kernel\n",
    "rbf_kern = 1. * kernels.RBF(\n",
    "                      length_scale=ptp,\n",
    "                      length_scale_bounds=np.outer(ptp, (1e-2, 1e2)),\n",
    "                      )\n",
    "#This is a white noise kernel for finite statistical accuracy\n",
    "white_kern = kernels.WhiteKernel(\n",
    "                         noise_level=.1,\n",
    "                         noise_level_bounds=(1e-2, 1e1)\n",
    "                        )\n",
    "#kernel is the sum of the Sq. Exponential and White Noise kernels\n",
    "my_kernel = (rbf_kern + white_kern) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# a function to return a set of trained GP regressors of \n",
    "# inputs train_X onto outputs train_Y\n",
    "def get_trained_gps(train_X, train_Y):\n",
    "    nrestarts=10\n",
    "    gps = []\n",
    "    for i in range(train_Y.shape[1]):\n",
    "        my_gp = GPR(kernel=my_kernel,\n",
    "         alpha=0.01, # the nugget, to stabilize matrix inversions\n",
    "         n_restarts_optimizer=nrestarts, \n",
    "         ).fit(train_X, train_Y[:,i])\n",
    "        gps.append(my_gp)\n",
    "    return gps"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# a function to return the predictions of a set of \n",
    "# trained GP regressors at inputs X \n",
    "def trained_gps_predict(gps, X):\n",
    "    Y = []\n",
    "    dY = [] \n",
    "    for gp in gps:\n",
    "        y, dy = gp.predict(X, return_std=True, return_cov=False)\n",
    "        Y.append(y)\n",
    "        dY.append(dy)\n",
    "    return np.array(Y), np.array(dY)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define the log of likelihood function, which includes \n",
    "# the expt. pseudo-data. This will be used in the objective fct.\n",
    "def log_likelihood(theta, obs_Y, obs_dY, gps):\n",
    "    n_dim = obs_Y.shape[0]\n",
    "  \n",
    "    # Model parameters\n",
    "    theta = np.array(theta).reshape((1, 5))\n",
    "    \n",
    "    # Model predictions given parameters\n",
    "    model_Y, model_dY = trained_gps_predict(gps, theta)\n",
    "    \n",
    "    #total covaiance assuming diagonal exp. and theor. covariance matrices\n",
    "    sigma = model_dY**2. + obs_dY**2.\n",
    "    #discrepancy\n",
    "    delta_Y = model_Y - obs_Y\n",
    "    \n",
    "    ll = -0.5*np.sum(delta_Y**2./sigma) -0.5*np.log( (2.*np.pi)**n_dim * np.prod(sigma) )\n",
    "    return ll\n",
    "\n",
    "def neg_log_likelihood(theta, obs_Y, obs_dY, gps):\n",
    "    return -1.*log_likelihood(theta, obs_Y, obs_dY, gps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Root solver to maximize the objective fct. \n",
    "def find_max_likelihood(obs_Y, obs_dY, gps):\n",
    "    x0 = prior_midpoint #pick a initial guess inside prior bounds \n",
    "    res = minimize(neg_log_likelihood, x0, \n",
    "                   args=(obs_Y, obs_dY, gps), bounds=prior_bds)\n",
    "    if (res.success != True):\n",
    "        print('minimization failed : ')\n",
    "        print(res)\n",
    "    return res.x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "#the active learning algorithm\n",
    "def active_learn(n_iters, train_X, train_Y):    \n",
    "    \n",
    "    #begin the sequential learning algo\n",
    "    print('Starting active learning stage for '+str(n_iters)+ ' iterations')\n",
    "    for iter in range(n_iters):\n",
    "        print('############')\n",
    "        print('Seq. design iter ' + str(iter))\n",
    "        \n",
    "        #train the GPs on existing data\n",
    "        print('training gps')\n",
    "        gps = get_trained_gps(train_X, train_Y )\n",
    "        \n",
    "        #search for new training point, which optimizes objective fct.\n",
    "        print('optimizing likelihood')\n",
    "        theta = find_max_likelihood(obs_Y, obs_dY, gps).reshape(1, 5)\n",
    "        \n",
    "        #get the updated model prediction\n",
    "        model_Y = simulation(theta, *pT_range)[:,pT_bins]\n",
    "        model_dY = 0. #temporary\n",
    "        #append data to gp training set\n",
    "        train_X = np.append(train_X, theta, axis=0)\n",
    "        train_Y = np.append(train_Y, model_Y, axis=0)\n",
    "        \n",
    "    return train_X, train_Y, gps"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Before the active learning, first build a coarse initial training set using a sparse LHS design"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_pts_per_dim = 5\n",
    "n_dim = 5\n",
    "n_pts = n_dim * n_pts_per_dim\n",
    "\n",
    "#make the latin hypercube of prior parameters\n",
    "criterion='maximin'\n",
    "hypcub = lhs(n_dim, samples=n_pts, criterion=criterion, iterations=9)\n",
    "#rescale the axes of the cube\n",
    "\n",
    "for i in range(n_dim):\n",
    "    p_min, p_max = prior_bds[i]\n",
    "    delta_p = p_max - p_min\n",
    "\n",
    "    #rescaling and shifting \n",
    "    hypcub[:, i] = hypcub[:, i] * delta_p + p_min"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x900 with 30 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "hypcub_df = pd.DataFrame(hypcub, columns = ['alpha_s','e_inel','e_el', 'scale_inel', 'scale_el'])\n",
    "sns.pairplot(hypcub_df)\n",
    "plt.tight_layout(True)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "#if you haven't yet generated the LHS training set (cell below)\n",
    "#turn this on\n",
    "generate_new_LHS_train = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "if generate_new_LHS_train:\n",
    "    #run the model on the LHS to build an initial sparse design\n",
    "    train_X = hypcub\n",
    "    train_Y = simulation(train_X, *pT_range)[:,pT_bins]\n",
    "\n",
    "    #save to disk for reuse:\n",
    "    with open('LHS_X.npy', 'wb') as f:\n",
    "        np.save(f, train_X)\n",
    "    with open('LHS_Y.npy', 'wb') as f:\n",
    "        np.save(f, train_Y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Now we start the active learning algo."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('LHS_X.npy', 'rb') as f:\n",
    "    train_X = np.load(f)\n",
    "with open('LHS_Y.npy', 'rb') as f:\n",
    "    train_Y = np.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Starting active learning stage for 25 iterations\n",
      "############\n",
      "Seq. design iter 0\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 1.96337190e-01 -8.07290486e-01  4.15364479e-04  3.62066877e-01\n",
      "  2.87694182e-01]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 1\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 1.93074749e-01 -9.05785455e-01  6.77590675e-04  3.64162665e-01\n",
      "  2.88731070e-01]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 2\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 1.89965396e-01 -9.93999717e-01 -3.40518356e-04  3.52185538e-01\n",
      "  2.93428436e-01]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 3\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 1.88052012e-01 -1.04940770e+00 -6.17845865e-05  3.65948282e-01\n",
      "  2.92589705e-01]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 4\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 1.85921601e-01 -1.10134069e+00 -5.50812542e-04  3.08105473e-01\n",
      "  3.02466099e-01]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 5\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 1.84982850e-01 -1.15510566e+00 -9.40819032e-04  3.64556435e-01\n",
      "  2.95528144e-01]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 6\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 0.18325522 -1.19066414 -0.00133226  0.36567471  0.29601977]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 7\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 0.18172569 -1.22023214 -0.00133766  0.36697955  0.29770259]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 8\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 1.80368924e-01 -1.24462066e+00 -9.17897506e-04  3.68203467e-01\n",
      "  2.98623584e-01]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 9\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 0.17915515 -1.26493631 -0.00190994  0.36892451  0.29865949]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 10\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 1.78084809e-01 -1.28182678e+00 -6.78326023e-04  3.10533678e-01\n",
      "  3.02988339e-01]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 11\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 1.77753949e-01 -1.31258586e+00 -6.83899220e-04  3.11205973e-01\n",
      "  3.02991677e-01]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 12\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 1.77444980e-01 -1.33946809e+00 -8.81670173e-04  3.66523521e-01\n",
      "  3.01808571e-01]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 13\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 1.76591850e-01 -1.34832413e+00 -7.17387664e-04  3.11860187e-01\n",
      "  3.03061414e-01]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 14\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 0.17635607 -1.37093698 -0.00164562  0.36642105  0.30336957]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 15\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 1.75636215e-01 -1.37596618e+00 -6.08608642e-04  3.10336120e-01\n",
      "  3.02737840e-01]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 16\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 1.75472165e-01 -1.39630430e+00 -6.41003556e-04  3.11018716e-01\n",
      "  3.02869568e-01]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 17\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 0.17532293 -1.41390582 -0.00155837  0.36470038  0.30483108]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 18\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 1.74739409e-01 -1.41496696e+00 -1.11553553e-03  3.65998904e-01\n",
      "  3.05111379e-01]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 19\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 0.17421703 -1.41508018 -0.00190518  0.36584153  0.30574392]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 20\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 0.17376368 -1.4145723  -0.00148473  0.3661206   0.30565161]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 21\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 0.17336041 -1.41376002 -0.00163321  0.36627883  0.30619272]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 22\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 1.73007422e-01 -1.41277427e+00 -7.04079177e-04  3.13716977e-01\n",
      "  3.03419174e-01]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 23\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 0.17294963 -1.42669444 -0.00159166  0.36623623  0.30725963]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "############\n",
      "Seq. design iter 24\n",
      "training gps\n",
      "optimizing likelihood\n",
      "running simulator with parameters : [ 1.72638240e-01 -1.42491026e+00 -6.26745466e-04  3.12265149e-01\n",
      "  3.03005711e-01]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: step size becomes too small\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/miniconda3/lib/python3.6/site-packages/scipy/integrate/_ode.py:1177: UserWarning: dopri5: larger nsteps is needed\n",
      "  self.messages.get(istate, unexpected_istate_msg)))\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:78: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2b=scipy.integrate.quad(lambda u, p=p: p*integrand_middle(p,u), 0, delta, limit=npts, epsabs=epsabs, epsrel=epsrel)\n",
      "/Users/derek/Work/JS-STAT-INTERNAL-current-projects/parton_energy_loss_qcd_bjorken/solver_rk.py:73: IntegrationWarning: The integral is probably divergent, or slowly convergent.\n",
      "  res_quad2a=scipy.integrate.quad(vec_integrand, pmin, p, limit=npts, epsabs=epsabs, epsrel=epsrel)\n"
     ]
    }
   ],
   "source": [
    "n_iters=25\n",
    "train_X, train_Y, gps = active_learn(n_iters=n_iters, train_X=train_X, train_Y=train_Y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's visualize where the active learning procedure placed new design points"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x900 with 30 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "active_df = pd.DataFrame(train_X[25:], columns = ['alpha_s','e_inel','e_el', 'scale_inel', 'scale_el'])\n",
    "sns.pairplot(active_df)\n",
    "plt.tight_layout(True)\n",
    "plt.show()"
   ]
  },
  {
   "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
}