{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# White matter challenge prediction\n",
    "\n",
    "This notebook generates a prediction for the [ISBI2015 White Matter Challenge](http://cmic.cs.ucl.ac.uk/wmmchallenge/)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Import all the necessary components from other libraries, including `numpy`, a module from `scikit-learn` and functions from `dipy` for managing gradient tables:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import sklearn.linear_model as lm\n",
    "from dipy.core import gradients as grad"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The following imports are from our current implementation of the RS-SFM model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import utils\n",
    "from model import Model, BiExponentialIsotropicModel"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A utility function reads the data from the files. We read in the 'seen' data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data = utils.read_data()\n",
    "seen_bvals = data['seen']['bvals']\n",
    "seen_bvecs = data['seen']['bvecs']\n",
    "seen_delta = data['seen']['delta']\n",
    "seen_Delta = data['seen']['Delta']\n",
    "seen_te = data['seen']['TE']\n",
    "seen_g = data['seen']['g']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And create a gradient table to describe the conditions used to collect the seen data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "seen_gtab = grad.gradient_table(seen_bvals, \n",
    "                                seen_bvecs, \n",
    "                                big_delta=seen_Delta, \n",
    "                                small_delta=seen_delta)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Similarly, a gradient table is created for the unseen data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "unseen_bvals = data['unseen']['bvals']\n",
    "unseen_bvecs = data['unseen']['bvecs']\n",
    "unseen_delta = data['unseen']['delta']\n",
    "unseen_Delta = data['unseen']['Delta']\n",
    "unseen_te = data['unseen']['TE']\n",
    "unseen_g = data['unseen']['g']\n",
    "\n",
    "unseen_gtab = grad.gradient_table(unseen_bvals, \n",
    "                                  unseen_bvecs, \n",
    "                                  big_delta=unseen_Delta, \n",
    "                                  small_delta=unseen_delta)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The model object used to fit the data is initialized with the gradient table for the seen data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model = Model(seen_gtab, \n",
    "              isotropic=BiExponentialIsotropicModel, \n",
    "              solver=lm.ElasticNet(l1_ratio=0.4, \n",
    "                                   alpha=5e-7, \n",
    "                                   positive=True, \n",
    "                                   warm_start=True, \n",
    "                                   fit_intercept=True, \n",
    "                                   normalize=True))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We initialize a list to contain the predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "predict = []"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In each iteration of the following loop, the model is fit to the seen data from a different voxel, and a prediction is made for this voxel in the unseen conditions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/arokem/anaconda/lib/python3.5/site-packages/sklearn/linear_model/coordinate_descent.py:466: ConvergenceWarning: Objective did not converge. You might want to increase the number of iterations\n",
      "  ConvergenceWarning)\n"
     ]
    }
   ],
   "source": [
    "for vox_idx in range(6):\n",
    "    signal = data['seen']['signal'][:, vox_idx]\n",
    "    fit = model.fit(signal, seen_te, seen_g)\n",
    "    predict.append(fit.predict(unseen_gtab, unseen_te))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The prediction is saved in the expected format"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "np.savetxt('unseenSignal.txt', np.array(predict).T, fmt='%.4f', header='\\%-voxel1---voxel2---voxel3---voxel4---voxel5---voxel6')"
   ]
  }
 ],
 "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": "3.5.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}