{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import warnings\n",
    "\n",
    "def along_axis(y, x, newx, axis, inverse=False, method='linear'):\n",
    "    \"\"\" Interpolate vertical profiles, e.g. of atmospheric variables\n",
    "    using vectorized numpy operations\n",
    "\n",
    "    - This function assumes that the x-xoordinate increases monotonically\n",
    "\n",
    "    args\n",
    "    ----\n",
    "    y : nd-array\n",
    "        The variable to be interpolated\n",
    "    x : nd-array or 1d-array\n",
    "        The coordinate associated with y, along which to interpolate.\n",
    "        If nd-array, this variable should have the same shape as y\n",
    "        If 1d, len(x) should be equal to the length of y along `axis`\n",
    "    newx : nd-array or 1d-array or float\n",
    "           The new coordinates to which y should be interpolated.\n",
    "           If nd-array, this variable should have the same shape as y\n",
    "           If 1d, len(x) should be equal to the length of y along `axis`\n",
    "    method : string\n",
    "             'linear', straightforward linear interpolation\n",
    "             'cubic', Algorithm from: http://www.paulinternet.nl/?page=bicubic\n",
    "                    f(x) = ax**3+bx**2+cx+d with\n",
    "                    a = 2f(0) - 2f(1) + f'(0) + f'(1)\n",
    "                    b = -3f(0) + 3f(1) - 2f'(0) - f'(1)\n",
    "                    c = f'(0)\n",
    "                    d = f(0)\n",
    "\n",
    "             'hermite', Algorithm from https://en.wikipedia.org/wiki/Cubic_Hermite_spline\n",
    "                    f(x) = h00(x)*f(0)+h10(x)*f'(0)+h01(x)*f(1)+h11(x)*f'(1) with\n",
    "                    h00(x) = 2x**3-3x**2+1\n",
    "                    h10(x) = x**3-2x**2+x\n",
    "                    h01(x) = -2x**3+3x**2\n",
    "                    h11(x) = x**3-x**2\n",
    "\n",
    "             'cspline', Algorithm from https://en.wikipedia.org/wiki/Spline_interpolation\n",
    "                    f(x) = (1-x)*f(0) + x*f(1) + x*(1-x)*(a*(10x)+bx) with\n",
    "                    a = f'(0)*(x_up-x_low) - (f(1)-f(0))\n",
    "                    a = -f'(1)*(x_up-x_low) + (f(1)-f(0))\n",
    "\n",
    "            'natural', Algorithm from https://en.wikiversity.org/wiki/Cubic_Spline_Interpolation#Methods\n",
    "                    ddf0*a**3/(6*c) + ddf1*b**3/(6*c) + (y0-ddf0*c**2/6)*a/c + y1-ddf1*c**2/6)*b/c\n",
    "                    a = x1 - x\n",
    "                    b = x - x0\n",
    "                    c = x1 - x0\n",
    "                    where ddf is solved for using TDMA.\n",
    "\n",
    "    Notes\n",
    "    -----\n",
    "    * Updated to work with irregularly spaced x-coordinate.\n",
    "    * Updated to work with irregularly spaced newx-coordinate\n",
    "    * Updated to easily inverse the direction of the x-coordinate\n",
    "    * Updated to fill with nans outside extrapolation range\n",
    "    * Updated to include a linear interpolation method as well\n",
    "        (it was initially written for a cubic function)\n",
    "    * Updated for https://github.com/numpy/numpy/pull/9686 (makes it ugly!)\n",
    "    * Updated to work with either 1d or nd input for x and newx.\n",
    "    * Added two new algorithms for computing a cubic spline: 'hermite' and 'cspline'\n",
    "      Theoretically, they should yield the same results, but it seems to work better than the old method 'cubic'\n",
    "    * Added option 'gradient', which let you choose between numpy and a cardinal gradient computation.\n",
    "    * Added option 'c': the tension parameter of the cardinal gradient computation.\n",
    "    * Added method 'natural', but this seems to work less well. Should it be implemented differently?\n",
    "      perhaps using tridiagonal matrix, like here: https://en.wikiversity.org/wiki/Cubic_Spline_Interpolation#Methods\n",
    "    * Fixed the 'natural' spline interpolation. This method is (nearly) equivalent to the\n",
    "      scipy routine. It finds the optimal curvature at each node under the condition that\n",
    "      the curvature is 0 at the start and end points. It uses Thomas' algorithm to solve\n",
    "      a tridiagonal matrix for the curvature at the nodes. This involves two extra loops\n",
    "      over the target axis, but it's probably still faster than applied scipy1d recursively.\n",
    "    * Removed gradient option - fall back to numpy defaults. Cardinal gradient computation\n",
    "      did not improve the results substantially, but did complicate the code and call\n",
    "\n",
    "\n",
    "    Peter Kalverla\n",
    "    March 2018; last update: October 3, 2018\n",
    "\n",
    "    \"\"\"\n",
    "    # Parse input\n",
    "    # -----------\n",
    "    _y = np.atleast_1d(y)\n",
    "    _x = np.atleast_1d(x)\n",
    "    _newx = np.atleast_1d(newx)\n",
    "\n",
    "    # This should make the shapes compatible\n",
    "    _x = np.broadcast_to(_x, _y.shape)\n",
    "    newshape = list(_y.shape)\n",
    "    newshape[axis] = len(_newx) if _newx.ndim==1 else _newx.shape[axis]\n",
    "    _newx = np.broadcast_to(_newx, newshape)\n",
    "\n",
    "    # View of x and y with axis as first dimension\n",
    "    _y = np.moveaxis(_y, axis, 0)\n",
    "    _x = np.moveaxis(_x, axis, 0)\n",
    "    _newx = np.moveaxis(_newx, axis, 0)\n",
    "\n",
    "    # Possibly inverse the arrays\n",
    "    if inverse:\n",
    "        _y = _y[::-1]\n",
    "        _x = _x[::-1]\n",
    "        _newx = _newx[::-1]\n",
    "\n",
    "    # Sanity checks: valid interpolation range and monotonic increase?\n",
    "    if np.any(_newx[0] < _x[0]) or np.any(_newx[-1] > _x[-1]):\n",
    "        warnings.warn(\"Some values are outside the interpolation range. \"\n",
    "                      \"These will be filled with NaN\")\n",
    "    if np.any(np.diff(_x, axis=0) < 0):\n",
    "        raise ValueError('x should increase monotonically')\n",
    "    if np.any(np.diff(_newx, axis=0) < 0):\n",
    "        raise ValueError('newx should increase monotonically')\n",
    "\n",
    "    ##################################################################\n",
    "    # Preprocessing: different methods need various 'helper' arrays\n",
    "    if method=='linear':\n",
    "        pass\n",
    "    elif method=='cubic':\n",
    "        # This algorithm needs a scaled gradient, so don't divide by grid spacing\n",
    "        ydx = np.gradient(_y, axis=0, edge_order=2)\n",
    "    elif method in ['hermite', 'cspline']:\n",
    "        # The other cubic spline algorithms implement their own correction for affine transformation\n",
    "        ydx = np.gradient(_y, axis=0, edge_order=2) / np.gradient(_x, axis=0, edge_order=2)\n",
    "    elif method=='natural':\n",
    "        # Initialize arrays for tridiagonal matrix algorithm\n",
    "        # a*ddf(i-1) + b*ddf(i) + c*ddf(i+1) = d ; r is a shorthand that returns often later on\n",
    "        a = np.zeros_like(_x)\n",
    "        b = np.zeros_like(_x) + 2\n",
    "        # c = 1-a, don't need to contaminate memory for that\n",
    "        d = np.zeros_like(_x)\n",
    "        ddf = np.zeros_like(_x)\n",
    "        r = np.zeros_like(_x)\n",
    "\n",
    "        # Type II \"natural\" BC:\n",
    "        a[0] = a[-1] = 0\n",
    "        d[0] = d[-1] = 0\n",
    "        ddf[0] = ddf[-1] = 0\n",
    "\n",
    "        r[1:] = _x[1:]-_x[:-1]\n",
    "        a[1:-1] = r[1:-1]/(r[1:-1] + r[2:])\n",
    "        d[1:-1] = 6*np.diff(np.diff(_y, axis=0)/r[1:,...], axis=0)/(_x[2:]-_x[:-2])\n",
    "\n",
    "        # Available alternatives: TMDASolve, TMDAsolver, TMDAsolver2,\n",
    "        #                         TMDAsolver3, TMDA, TMDA2 (see below)\n",
    "        ddf = TDMAsolver3(a, b, 1-a, d)\n",
    "\n",
    "    else:\n",
    "        raise ValueError(\"interpolation method not understood (got %s)\"\n",
    "                         \"(choose 'linear', 'cubic', 'hermite', 'cspline', or 'natural')\"%method)\n",
    "\n",
    "    ##################################################################\n",
    "    # Initialize indexer arrays\n",
    "\n",
    "    # This will later be concatenated with a dynamic '0th' index\n",
    "    ind = [i for i in np.indices(_y.shape[1:])]\n",
    "\n",
    "    # Allocate the output array\n",
    "    original_dims = _y.shape\n",
    "    newdims = list(original_dims)\n",
    "    newdims[0] = len(_newx)\n",
    "    newy = np.zeros(newdims)\n",
    "\n",
    "    # set initial bounds\n",
    "    i_lower = np.zeros(_x.shape[1:], dtype=int)\n",
    "    i_upper = np.ones(_x.shape[1:], dtype=int)\n",
    "    x_lower = _x[0, ...]\n",
    "    x_upper = _x[1, ...]\n",
    "\n",
    "    ##################################################################\n",
    "    # Pass trough the array along <axis> and evaluate f(x) at _newx\n",
    "    for i, xi in enumerate(_newx):\n",
    "        # Start at the 'bottom' of the array and work upwards\n",
    "        # This only works if x and newx increase monotonically\n",
    "\n",
    "        # Update bounds where necessary and possible\n",
    "        needs_update = (xi > x_upper) & (i_upper+1<len(_x))\n",
    "        # print x_upper.max(), np.any(needs_update)\n",
    "        while np.any(needs_update):\n",
    "            i_lower = np.where(needs_update, i_lower+1, i_lower)\n",
    "            i_upper = i_lower + 1\n",
    "            x_lower = _x[tuple([i_lower]+ind)]\n",
    "            x_upper = _x[tuple([i_upper]+ind)]\n",
    "\n",
    "            # Check again\n",
    "            needs_update = (xi > x_upper) & (i_upper+1<len(_x))\n",
    "\n",
    "        # Express the position of xi relative to its neighbours (i.e. affine transformation)\n",
    "        # Note: this requires that the gradient is scaled with (x_upper-x_lower).\n",
    "        xj = (xi-x_lower)/(x_upper - x_lower)\n",
    "\n",
    "        # Determine where there is a valid interpolation range\n",
    "        within_bounds = (_x[0, ...] < xi) & (xi < _x[-1, ...])\n",
    "\n",
    "        # Get the current index values of helper arrays\n",
    "        y0, y1 = _y[tuple([i_lower]+ind)], _y[tuple([i_upper]+ind)]\n",
    "        if method in ['cubic', 'hermite', 'cspline']:\n",
    "            dy0, dy1 = ydx[tuple([i_lower]+ind)], ydx[tuple([i_upper]+ind)]\n",
    "        elif method == 'natural':\n",
    "            ddf0, ddf1 = ddf[tuple([i_lower]+ind)], ddf[tuple([i_upper]+ind)]\n",
    "            ri = r[tuple([i_upper]+ind)]\n",
    "\n",
    "        # Calculate interpolated function values\n",
    "        if method == 'linear':\n",
    "            a = y1 - y0\n",
    "            b = y0\n",
    "            newy[i, ...] = np.where(within_bounds, a*xj+b, np.nan)\n",
    "\n",
    "        elif method=='cubic':\n",
    "            # http://www.paulinternet.nl/?page=bicubic\n",
    "            a = 2*y0 - 2*y1 + dy0 + dy1\n",
    "            b = -3*y0 + 3*y1 - 2*dy0 - dy1\n",
    "            c = dy0\n",
    "            d = y0\n",
    "            newy[i, ...] = np.where(within_bounds, a*xj**3 + b*xj**2 + c*xj + d, np.nan)\n",
    "\n",
    "        elif method=='hermite':\n",
    "            # https://en.wikipedia.org/wiki/Cubic_Hermite_spline\n",
    "            h00 = 2*xj**3 - 3*xj**2 + 1\n",
    "            h10 = xj**3 - 2*xj**2 + xj\n",
    "            h01 = -2*xj**3 + 3*xj**2\n",
    "            h11 = xj**3 - xj**2\n",
    "            scale = x_upper - x_lower\n",
    "            newy[i, ...] = np.where(within_bounds, h00*y0 + h10*scale*dy0 + h01*y1 + h11*scale*dy1, np.nan)\n",
    "\n",
    "        elif method=='cspline':\n",
    "            # https://en.wikipedia.org/wiki/Cubic_Hermite_spline\n",
    "            a = dy0*(x_upper - x_lower) - (y1-y0)\n",
    "            b = -dy1*(x_upper - x_lower) + (y1-y0)\n",
    "            newy[i, ...] = np.where(within_bounds, (1-xj)*y0+xj*y1 +xj*(1-xj)*(a*(1-xj)+b*xj), np.nan)\n",
    "\n",
    "        elif method=='natural':\n",
    "            # https://en.wikiversity.org/wiki/Cubic_Spline_Interpolation#Methods\n",
    "            newy[i] = np.where(within_bounds,\n",
    "                (ddf0*(x_upper-xi)**3/(6*ri) + ddf1*(xi-x_lower)**3/(6*ri)\n",
    "                + (y0-ddf0*ri**2/6)*(x_upper-xi)/ri + (y1-ddf1*ri**2/6)*(xi-x_lower)/ri),\n",
    "                np.nan)\n",
    "\n",
    "    ##################################################################\n",
    "    # Prepare for exit\n",
    "    if inverse:\n",
    "        newy = newy[::-1, ...]\n",
    "\n",
    "    return np.moveaxis(newy, 0, axis)\n",
    "\n",
    "################# Solvers ######################\n",
    "# https://stackoverflow.com/a/23133208/6012085\n",
    "def TDMASolve(a, b, c, d):\n",
    "    n = len(d) # n is the numbers of rows, a and c has length n-1\n",
    "    a = a[1:]  # edited by Peter (I guess that's meant by the above comment about len(a))\n",
    "    c = c[:-1] # edited by Peter (I guess that's meant by the above comment about len(c))\n",
    "    for i in range(n-1):\n",
    "        d[i+1] -= 1. * d[i] * a[i] / b[i]\n",
    "        b[i+1] -= 1. * c[i] * a[i] / b[i]\n",
    "    for i in reversed(range(n-1)):\n",
    "        d[i] -= d[i+1] * c[i] / b[i+1]\n",
    "    return d/b # edited by Peter (using numpy rather than python list comprehension)\n",
    "\n",
    "# https://gist.github.com/cbellei/8ab3ab8551b8dfc8b081c518ccd9ada9\n",
    "def TDMAsolver(a, b, c, d):\n",
    "    '''\n",
    "    TDMA solver, a b c d can be NumPy array type or Python list type.\n",
    "    refer to http://en.wikipedia.org/wiki/Tridiagonal_matrix_algorithm\n",
    "    and to http://www.cfd-online.com/Wiki/Tridiagonal_matrix_algorithm_-_TDMA_(Thomas_algorithm)\n",
    "    '''\n",
    "    nf = len(d) # number of equations\n",
    "    ac, bc, cc, dc = map(np.array, (a, b, c, d)) # copy arrays\n",
    "    for it in range(1, nf):\n",
    "        mc = ac[it-1]/bc[it-1]\n",
    "        bc[it] = bc[it] - mc*cc[it-1]\n",
    "        dc[it] = dc[it] - mc*dc[it-1]\n",
    "    xc = bc\n",
    "    xc[-1] = dc[-1]/bc[-1]\n",
    "    for il in range(nf-2, -1, -1):\n",
    "        xc[il] = (dc[il]-cc[il]*xc[il+1])/bc[il]\n",
    "    return xc\n",
    "\n",
    "# https://gist.github.com/ofan666/1875903\n",
    "def TDMAsolver2(a, b, c, d):\n",
    "    '''\n",
    "    TDMA solver, a b c d can be NumPy array type or Python list type.\n",
    "    refer to http://en.wikipedia.org/wiki/Tridiagonal_matrix_algorithm\n",
    "    '''\n",
    "    nf = len(a)     # number of equations\n",
    "    ac, bc, cc, dc = map(np.array, (a, b, c, d))     # copy the array\n",
    "    for it in range(1, nf):\n",
    "        mc = ac[it]/bc[it-1]\n",
    "        bc[it] = bc[it] - mc*cc[it-1]\n",
    "        dc[it] = dc[it] - mc*dc[it-1]\n",
    "    xc = ac\n",
    "    xc[-1] = dc[-1]/bc[-1]\n",
    "    for il in range(nf-2, -1, -1):\n",
    "        xc[il] = (dc[il]-cc[il]*xc[il+1])/bc[il]\n",
    "    del bc, cc, dc  # delete variables from memory\n",
    "    return xc\n",
    "\n",
    "# https://gist.github.com/ofan666/1875903 (simplified by peter)\n",
    "def TDMAsolver3(a, b, c, d):\n",
    "    '''\n",
    "    TDMA solver, a b c d can be NumPy array type or Python list type.\n",
    "    refer to http://en.wikipedia.org/wiki/Tridiagonal_matrix_algorithm\n",
    "    '''\n",
    "    x = np.zeros_like(a) # output array\n",
    "    nf = len(a)     # number of equations\n",
    "    # Forward sweep\n",
    "    for i in range(1, nf):\n",
    "        m = a[i]/b[i-1]\n",
    "        b[i] = b[i] - m*c[i-1]\n",
    "        d[i] = d[i] - m*d[i-1]\n",
    "    x[-1] = d[-1]/b[-1]\n",
    "    # Backward fill\n",
    "    for i in range(nf-2, -1, -1):\n",
    "        x[i] = (d[i]-c[i]*x[i+1])/b[i]\n",
    "    return x\n",
    "\n",
    "\n",
    "# Own routine (combination of the above)\n",
    "def TDMA(a, b, c, d):\n",
    "    x = np.zeros_like(d)\n",
    "    n = len(d)\n",
    "    for i in range(1, n-2):\n",
    "        d[i+1] -= 1. * d[i] * a[i] / b[i]\n",
    "        b[i+1] -= 1. * c[i] * a[i] / b[i]\n",
    "    x[-2] = d[-2] / b[-2]\n",
    "    for i in reversed(range(1, n-2)):\n",
    "        x[i] = (d[i]-c[i]*x[i+1])/b[i]\n",
    "    return x\n",
    "\n",
    "# Own implementation from wikipedia (this one is not working as well as the others)\n",
    "def TDMA2(a, b, c, d):\n",
    "    # https://en.wikipedia.org/wiki/Tridiagonal_matrix_algorithm\n",
    "    x = np.zeros_like(d)\n",
    "    n = len(d)\n",
    "    c[0] = c[0] / b[0]\n",
    "    d[0] = d[0] / b[0]\n",
    "    for i in range(1, n):\n",
    "        c /= (b[i] - a[i] * c[i-1] )\n",
    "        d = (d[i] - a[i]*d[i-1]) / (b[i] - a[i] * c[i-1])\n",
    "    x[-1] = d[-1]\n",
    "    for i in reversed(range(1, n-1)):\n",
    "        x[i] = (d[i]-c[i]*x[i+1])\n",
    "    return x\n",
    "\n",
    "#########################################################\n",
    "\n",
    "\n",
    "if __name__==\"__main__\":\n",
    "    import matplotlib.pyplot as plt\n",
    "    from scipy.interpolate import interp1d as scipy1d\n",
    "\n",
    "    # Check some random profiles\n",
    "    for case in range(1,5):\n",
    "        # toy coordinates and data\n",
    "        nx, ny, nz = 25, 30, 10\n",
    "        x = np.arange(nx)\n",
    "        y = np.arange(ny)\n",
    "        testdata = np.random.randn(nx,ny,nz) # x,y,z\n",
    "\n",
    "        # case=1\n",
    "        if case==1: # z = ndarray, znew=ndarray\n",
    "            # Original z-coordinate\n",
    "            z = np.tile(np.arange(nz), (nx,ny,1)) + np.random.randn(nx, ny, nz)*.1\n",
    "            # Desired z-coordinates (must be between bounds of z)\n",
    "            znew = np.tile(np.linspace(2,nz-2,50), (nx,ny,1)) + np.random.randn(nx, ny, 50)*0.01\n",
    "        elif case==2: # z = ndarray, znew=1d array\n",
    "            z = np.tile(np.arange(nz), (nx,ny,1)) + np.random.randn(nx, ny, nz)*.1\n",
    "            znew = np.linspace(2, nz-2, 50) + np.random.randn(50)*0.01\n",
    "        elif case==3: # z = 1d-array, znew=1d-array\n",
    "            z = np.arange(nz) + np.random.randn(nz)*.1\n",
    "            znew = np.linspace(2, nz-2, 50) + np.random.randn(50)*0.01\n",
    "        elif case==4: # z = 1d-array, znew=nd-array\n",
    "            z = np.arange(nz) + np.random.randn(nz)*.1\n",
    "            znew = np.tile(np.linspace(2,nz-2,50), (nx,ny,1)) + np.random.randn(nx, ny, 50)*0.01\n",
    "        # else # case==5 --> pass single float value to interpolation function.\n",
    "\n",
    "        # Inverse the coordinates for testing (and multiply to magnify the effect of scaling)\n",
    "        z = z[..., ::-1]*10\n",
    "        znew = znew[..., ::-1]*10\n",
    "\n",
    "        # Now use own routine\n",
    "        ynew = along_axis(testdata, z, znew, axis=2, inverse=True, method='cubic')\n",
    "        ynew2 = along_axis(testdata, z, znew, axis=2, inverse=True, method='linear')\n",
    "        ynew3 = along_axis(testdata, z, znew, axis=2, inverse=True, method='hermite')\n",
    "        ynew4 = along_axis(testdata, z, znew, axis=2, inverse=True, method='cspline')\n",
    "        ynew5 = along_axis(testdata, z, znew, axis=2, inverse=True, method='natural')\n",
    "\n",
    "        randx = np.random.randint(nx)\n",
    "        randy = np.random.randint(ny)\n",
    "\n",
    "        if case in [1,2]: # z = nd\n",
    "            checkfunc = scipy1d(z[randx, randy], testdata[randx,randy], kind='cubic')\n",
    "        else:\n",
    "            checkfunc = scipy1d(z, testdata[randx,randy], kind='cubic')\n",
    "        if case in [1,4]: # znew = nd\n",
    "            checkdata = checkfunc(znew[randx, randy])\n",
    "        else:\n",
    "            checkdata = checkfunc(znew)\n",
    "\n",
    "        fig, ax = plt.subplots()\n",
    "        if case in [1,2]: # z = nd\n",
    "            ax.plot(testdata[randx, randy], z[randx, randy], 'x', label='original data')\n",
    "        else:\n",
    "            ax.plot(testdata[randx, randy], z, 'x', label='original data')\n",
    "        if case in [1,4]: # znew = nd\n",
    "            ax.plot(checkdata, znew[randx, randy], label='scipy')\n",
    "            ax.plot(ynew[randx, randy], znew[randx, randy], '--', label='Peter - Spline')\n",
    "            ax.plot(ynew2[randx, randy], znew[randx, randy], '-.', label='Peter - Linear')\n",
    "            ax.plot(ynew3[randx, randy], znew[randx, randy], '-.', label='Peter - Hermite')\n",
    "            ax.plot(ynew4[randx, randy], znew[randx, randy], '-.', label='Peter - Cspline')\n",
    "            ax.plot(ynew5[randx, randy], znew[randx, randy], '-.', label='Peter - Natural')\n",
    "        else:\n",
    "            ax.plot(checkdata, znew, label='scipy')\n",
    "            ax.plot(ynew[randx, randy], znew, '--', label='Peter - Spline')\n",
    "            ax.plot(ynew2[randx, randy], znew, '-.', label='Peter - Linear')\n",
    "            ax.plot(ynew3[randx, randy], znew, '-.', label='Peter - Hermite')\n",
    "            ax.plot(ynew4[randx, randy], znew, '-.', label='Peter - Cspline')\n",
    "            ax.plot(ynew5[randx, randy], znew, '-.', label='Peter - Natural')\n",
    "\n",
    "        ax.legend()\n",
    "        plt.show()\n"
   ]
  },
  {
   "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}