{"cells": [{"cell_type": "markdown", "metadata": {}, "source": ["# Pairwise distances with ONNX (pdist)\n", "\n", "Function [pdist](https://docs.scipy.org/doc/scipy/reference/generated/scipy.spatial.distance.pdist.html) computes pairwise distances between observations in n-dimensional space. It is not that difficult to convert that into *ONNX* when the dimension of the input is always the same. What if not?"]}, {"cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [{"data": {"text/html": ["<div id=\"my_id_menu_nb\">run previous cell, wait for 2 seconds</div>\n", "<script>\n", "function repeat_indent_string(n){\n", "    var a = \"\" ;\n", "    for ( ; n > 0 ; --n)\n", "        a += \"    \";\n", "    return a;\n", "}\n", "// look up into all sections and builds an automated menu //\n", "var update_menu_string = function(begin, lfirst, llast, sformat, send, keep_item, begin_format, end_format) {\n", "    var anchors = document.getElementsByClassName(\"section\");\n", "    if (anchors.length == 0) {\n", "        anchors = document.getElementsByClassName(\"text_cell_render rendered_html\");\n", "    }\n", "    var i,t;\n", "    var text_menu = begin;\n", "    var text_memo = \"<pre>\\nlength:\" + anchors.length + \"\\n\";\n", "    var ind = \"\";\n", "    var memo_level = 1;\n", "    var href;\n", "    var tags = [];\n", "    var main_item = 0;\n", "    var format_open = 0;\n", "    for (i = 0; i <= llast; i++)\n", "        tags.push(\"h\" + i);\n", "\n", "    for (i = 0; i < anchors.length; i++) {\n", "        text_memo += \"**\" + anchors[i].id + \"--\\n\";\n", "\n", "        var child = null;\n", "        for(t = 0; t < tags.length; t++) {\n", "            var r = anchors[i].getElementsByTagName(tags[t]);\n", "            if (r.length > 0) {\n", "child = r[0];\n", "break;\n", "            }\n", "        }\n", "        if (child == null) {\n", "            text_memo += \"null\\n\";\n", "            continue;\n", "        }\n", "        if (anchors[i].hasAttribute(\"id\")) {\n", "            // when converted in RST\n", "            href = anchors[i].id;\n", "            text_memo += \"#1-\" + href;\n", "            // passer \u00e0 child suivant (le chercher)\n", "        }\n", "        else if (child.hasAttribute(\"id\")) {\n", "            // in a notebook\n", "            href = child.id;\n", "            text_memo += \"#2-\" + href;\n", "        }\n", "        else {\n", "            text_memo += \"#3-\" + \"*\" + \"\\n\";\n", "            continue;\n", "        }\n", "        var title = child.textContent;\n", "        var level = parseInt(child.tagName.substring(1,2));\n", "\n", "        text_memo += \"--\" + level + \"?\" + lfirst + \"--\" + title + \"\\n\";\n", "\n", "        if ((level < lfirst) || (level > llast)) {\n", "            continue ;\n", "        }\n", "        if (title.endsWith('\u00b6')) {\n", "            title = title.substring(0,title.length-1).replace(\"<\", \"&lt;\")\n", "         .replace(\">\", \"&gt;\").replace(\"&\", \"&amp;\");\n", "        }\n", "        if (title.length == 0) {\n", "            continue;\n", "        }\n", "\n", "        while (level < memo_level) {\n", "            text_menu += end_format + \"</ul>\\n\";\n", "            format_open -= 1;\n", "            memo_level -= 1;\n", "        }\n", "        if (level == lfirst) {\n", "            main_item += 1;\n", "        }\n", "        if (keep_item != -1 && main_item != keep_item + 1) {\n", "            // alert(main_item + \" - \" + level + \" - \" + keep_item);\n", "            continue;\n", "        }\n", "        while (level > memo_level) {\n", "            text_menu += \"<ul>\\n\";\n", "            memo_level += 1;\n", "        }\n", "        text_menu += repeat_indent_string(level-2);\n", "        text_menu += begin_format + sformat.replace(\"__HREF__\", href).replace(\"__TITLE__\", title);\n", "        format_open += 1;\n", "    }\n", "    while (1 < memo_level) {\n", "        text_menu += end_format + \"</ul>\\n\";\n", "        memo_level -= 1;\n", "        format_open -= 1;\n", "    }\n", "    text_menu += send;\n", "    //text_menu += \"\\n\" + text_memo;\n", "\n", "    while (format_open > 0) {\n", "        text_menu += end_format;\n", "        format_open -= 1;\n", "    }\n", "    return text_menu;\n", "};\n", "var update_menu = function() {\n", "    var sbegin = \"\";\n", "    var sformat = '<a href=\"#__HREF__\">__TITLE__</a>';\n", "    var send = \"\";\n", "    var begin_format = '<li>';\n", "    var end_format = '</li>';\n", "    var keep_item = -1;\n", "    var text_menu = update_menu_string(sbegin, 2, 4, sformat, send, keep_item,\n", "       begin_format, end_format);\n", "    var menu = document.getElementById(\"my_id_menu_nb\");\n", "    menu.innerHTML=text_menu;\n", "};\n", "window.setTimeout(update_menu,2000);\n", "            </script>"], "text/plain": ["<IPython.core.display.HTML object>"]}, "execution_count": 2, "metadata": {}, "output_type": "execute_result"}], "source": ["from jyquickhelper import add_notebook_menu\n", "add_notebook_menu()"]}, {"cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": ["%load_ext mlprodict"]}, {"cell_type": "markdown", "metadata": {}, "source": ["## Function pdist\n", "\n", "The function [pdist](https://docs.scipy.org/doc/scipy/reference/generated/scipy.spatial.distance.pdist.html) distances. Let's denote a list of vectors $(X_1, ..., X_n)$, function [pdist](https://docs.scipy.org/doc/scipy/reference/generated/scipy.spatial.distance.pdist.html) returns the matrix $D=(d_{ij})$ where $d_{ij}=dist(X_i, X_j)=\\lVert X_i - X_j \\rVert^2$."]}, {"cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [{"data": {"text/plain": ["array([[0.  , 2.  , 0.02, 5.  ],\n", "       [2.  , 0.  , 1.62, 1.  ],\n", "       [0.02, 1.62, 0.  , 4.42],\n", "       [5.  , 1.  , 4.42, 0.  ]])"]}, "execution_count": 4, "metadata": {}, "output_type": "execute_result"}], "source": ["import numpy\n", "from scipy.spatial.distance import pdist, squareform\n", "\n", "M = numpy.array([[0, 1],\n", "                 [1, 2],\n", "                 [0.1, 1.1],\n", "                 [2, 2]], dtype=float)\n", "\n", "d1 = squareform(pdist(M, metric='sqeuclidean'))\n", "d1"]}, {"cell_type": "markdown", "metadata": {}, "source": ["The two following functions are implemented to reduce the number of allocations the algorithm requires."]}, {"cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [{"data": {"text/plain": ["array([[0.  , 2.  , 0.02, 5.  ],\n", "       [2.  , 0.  , 1.62, 1.  ],\n", "       [0.02, 1.62, 0.  , 4.42],\n", "       [5.  , 1.  , 4.42, 0.  ]])"]}, "execution_count": 5, "metadata": {}, "output_type": "execute_result"}], "source": ["def custom_pdist(M):\n", "    n = M.shape[0]\n", "    res = numpy.zeros((n, n))\n", "    buffer = numpy.empty(M.shape)\n", "    for i in range(n):\n", "        numpy.subtract(M, M[i], out=buffer)  # broadcasted substraction\n", "        numpy.square(buffer, out=buffer)\n", "        res[i, :] = numpy.sum(buffer, axis=1)\n", "    return res\n", "\n", "d2 = custom_pdist(M)\n", "d2"]}, {"cell_type": "markdown", "metadata": {}, "source": ["This function computes $n^2$ distances wheres only $\\frac{n(n-1)}{2}$ are necessary since the final matrix is symmetric. Let's change the implementation to reflect that."]}, {"cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [{"data": {"text/plain": ["array([[0.  , 2.  , 0.02, 5.  ],\n", "       [2.  , 0.  , 1.62, 1.  ],\n", "       [0.02, 1.62, 0.  , 4.42],\n", "       [5.  , 1.  , 4.42, 0.  ]])"]}, "execution_count": 6, "metadata": {}, "output_type": "execute_result"}], "source": ["def custom_pdist_lower(M):\n", "    n = M.shape[0]\n", "    res = numpy.zeros((n, n))\n", "    buffer = numpy.empty((M.shape[0]-1, M.shape[1]))\n", "    a = numpy.empty(M.shape[0])\n", "    for i in range(1, n):\n", "        numpy.subtract(M[:i], M[i], out=buffer[:i])  # broadcasted substraction\n", "        numpy.square(buffer[:i], out=buffer[:i])\n", "        numpy.sum(buffer[:i], axis=1, out=a[:i])\n", "        res[:i, i] = a[:i]\n", "        res[i, :i] = a[:i]\n", "    return res\n", "\n", "d3 = custom_pdist_lower(M)\n", "d3"]}, {"cell_type": "markdown", "metadata": {}, "source": ["## Loop mechanism in ONNX\n", "\n", "Operator [Loop](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Loop) seems appropriate but it is just a loop wheras [Scan](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Scan) holds accumulator. The first graph is what is repeated inside the loop."]}, {"cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [{"data": {"text/html": ["<div id=\"M45bc8edb8c134f73a10d432beb7c906d-cont\"><div id=\"M45bc8edb8c134f73a10d432beb7c906d\" style=\"width:;height:;\"></div></div>\n", "<script>\n", "\n", "require(['http://www.xavierdupre.fr/js/vizjs/viz.js'], function() { var svgGraph = Viz(\"digraph{\\n  size=7;\\n  orientation=portrait;\\n  ranksep=0.25;\\n  nodesep=0.05;\\n\\n  sum_in [shape=box color=red label=\\\"sum_in\\nfloat((0,))\\\" fontsize=10];\\n  next [shape=box color=red label=\\\"next\\nfloat((0,))\\\" fontsize=10];\\n\\n  sum_out [shape=box color=green label=\\\"sum_out\\nfloat(('?',))\\\" fontsize=10];\\n  scan_out [shape=box color=green label=\\\"scan_out\\nfloat(('?',))\\\" fontsize=10];\\n\\n\\n  Ad_Add [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Add\\n(Ad_Add)\\\" fontsize=10];\\n  sum_in -> Ad_Add;\\n  next -> Ad_Add;\\n  Ad_Add -> sum_out;\\n\\n  Id_Identity [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Identity\\n(Id_Identity)\\\" fontsize=10];\\n  sum_out -> Id_Identity;\\n  Id_Identity -> scan_out;\\n}\");\n", "document.getElementById('M45bc8edb8c134f73a10d432beb7c906d').innerHTML = svgGraph; });\n", "\n", "</script>"], "text/plain": ["<jyquickhelper.jspy.render_nb_js_dot.RenderJsDot at 0x1f9f9edee30>"]}, "execution_count": 7, "metadata": {}, "output_type": "execute_result"}], "source": ["from skl2onnx.algebra.onnx_ops import OnnxAdd, OnnxIdentity, OnnxScan\n", "from skl2onnx.common.data_types import FloatTensorType\n", "\n", "initial = numpy.array([0, 0]).astype(numpy.float32).reshape((2,))\n", "x = numpy.array([1, 2, 3, 4, 5, 6]).astype(numpy.float32).reshape((3, 2))\n", "\n", "add_node = OnnxAdd('sum_in', 'next', output_names=['sum_out'], op_version=12)\n", "id_node = OnnxIdentity(add_node, output_names=['scan_out'], op_version=12)\n", "\n", "scan_body = id_node.to_onnx(\n", "    {'sum_in': initial, 'next': initial},\n", "    outputs=[('sum_out', FloatTensorType()),\n", "             ('scan_out', FloatTensorType())])\n", "\n", "# add -l 1 if nothing shows up\n", "%onnxview scan_body"]}, {"cell_type": "markdown", "metadata": {}, "source": ["The operator [Scan](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Scan) repeats this graph a couple of times. *sum_in* is an accumulator, *next* is the iterated row from the input matrix."]}, {"cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [{"data": {"text/html": ["<div id=\"Mc1e02e56cd3d4d4181984779e3889c39-cont\"><div id=\"Mc1e02e56cd3d4d4181984779e3889c39\" style=\"width:;height:;\"></div></div>\n", "<script>\n", "\n", "require(['http://www.xavierdupre.fr/js/vizjs/viz.js'], function() { var svgGraph = Viz(\"digraph{\\n  size=7;\\n  orientation=portrait;\\n  ranksep=0.25;\\n  nodesep=0.05;\\n\\n  initial [shape=box color=red label=\\\"initial\\nfloat((0,))\\\" fontsize=10];\\n  x [shape=box color=red label=\\\"x\\nfloat((0, 2))\\\" fontsize=10];\\n\\n  y [shape=box color=green label=\\\"y\\nfloat(('?',))\\\" fontsize=10];\\n  z [shape=box color=green label=\\\"z\\nfloat(('?',))\\\" fontsize=10];\\n\\n\\n  Sc_Scan [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Scan\\n(Sc_Scan)\\nbody=node {\\n  input: 'sum_in'\\...\\nnum_scan_inputs=1\\\" fontsize=10];\\n  initial -> Sc_Scan;\\n  x -> Sc_Scan;\\n  Sc_Scan -> y;\\n  Sc_Scan -> z;\\n}\");\n", "document.getElementById('Mc1e02e56cd3d4d4181984779e3889c39').innerHTML = svgGraph; });\n", "\n", "</script>"], "text/plain": ["<jyquickhelper.jspy.render_nb_js_dot.RenderJsDot at 0x1f9f59b4190>"]}, "execution_count": 8, "metadata": {}, "output_type": "execute_result"}], "source": ["node = OnnxScan('initial', 'x', output_names=['y', 'z'],\n", "                num_scan_inputs=1, body=scan_body.graph)\n", "\n", "model_def = node.to_onnx(\n", "    {'initial': initial, 'x': x},\n", "    outputs=[('y', FloatTensorType()),\n", "             ('z', FloatTensorType())])\n", "\n", "# add -l 1 if nothing shows up\n", "%onnxview model_def"]}, {"cell_type": "markdown", "metadata": {}, "source": ["All together in the same graph."]}, {"cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [{"data": {"text/html": ["<div id=\"M2fa3d20a67de41a391631d72ce1bd040-cont\"><div id=\"M2fa3d20a67de41a391631d72ce1bd040\" style=\"width:;height:;\"></div></div>\n", "<script>\n", "\n", "require(['http://www.xavierdupre.fr/js/vizjs/viz.js'], function() { var svgGraph = Viz(\"digraph{\\n  size=7;\\n  orientation=portrait;\\n  ranksep=0.25;\\n  nodesep=0.05;\\n\\n  initial [shape=box color=red label=\\\"initial\\nfloat((0,))\\\" fontsize=10];\\n  x [shape=box color=red label=\\\"x\\nfloat((0, 2))\\\" fontsize=10];\\n\\n  y [shape=box color=green label=\\\"y\\nfloat(('?',))\\\" fontsize=10];\\n  z [shape=box color=green label=\\\"z\\nfloat(('?',))\\\" fontsize=10];\\n\\n\\n  subgraph cluster_Scan2173153644832_2172776124464 {\\n    label=\\\"Scan\\n(Sc_Scan)\\nbody=node {\\n  input: 'sum_in'\\...\\nnum_scan_inputs=1\\\";\\n    fontsize=10;\\n    color=black;\\n    B_sum_in [shape=box color=red label=\\\"sum_in\\nfloat((0,))\\\" fontsize=10];\\n    B_next [shape=box color=red label=\\\"next\\nfloat((0,))\\\" fontsize=10];\\n  \\n    B_sum_out [shape=box color=green label=\\\"sum_out\\nfloat(('?',))\\\" fontsize=10];\\n    B_scan_out [shape=box color=green label=\\\"scan_out\\nfloat(('?',))\\\" fontsize=10];\\n  \\n  \\n    B_Ad_Add [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Add\\n(Ad_Add)\\\" fontsize=10];\\n    B_sum_in -> B_Ad_Add;\\n    B_next -> B_Ad_Add;\\n    B_Ad_Add -> B_sum_out;\\n  \\n    B_Id_Identity [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Identity\\n(Id_Identity)\\\" fontsize=10];\\n    B_sum_out -> B_Id_Identity;\\n    B_Id_Identity -> B_scan_out;\\n  }\\n  initial -> B_sum_in;\\n  x -> B_next;\\n  B_sum_out -> y;\\n  B_scan_out -> z;\\n  Sc_Scan -> B_Ad_Add [lhead=cluster_Scan2173153644832_2172776124464];\\n  initial -> Sc_Scan;\\n  x -> Sc_Scan;\\n  Sc_Scan -> y;\\n  Sc_Scan -> z;\\n}\");\n", "document.getElementById('M2fa3d20a67de41a391631d72ce1bd040').innerHTML = svgGraph; });\n", "\n", "</script>"], "text/plain": ["<jyquickhelper.jspy.render_nb_js_dot.RenderJsDot at 0x1f9ef6e7820>"]}, "execution_count": 9, "metadata": {}, "output_type": "execute_result"}], "source": ["# add -l 1 if nothing shows up\n", "%onnxview model_def -r 1"]}, {"cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [{"data": {"text/plain": ["array([ 9., 12.], dtype=float32)"]}, "execution_count": 10, "metadata": {}, "output_type": "execute_result"}], "source": ["from mlprodict.onnxrt import OnnxInference\n", "oinf = OnnxInference(model_def)\n", "res = oinf.run({'initial': initial, 'x': x})\n", "res['y']"]}, {"cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [{"data": {"text/plain": ["array([[ 1.,  2.],\n", "       [ 4.,  6.],\n", "       [ 9., 12.]], dtype=float32)"]}, "execution_count": 11, "metadata": {}, "output_type": "execute_result"}], "source": ["res['z']"]}, {"cell_type": "markdown", "metadata": {}, "source": ["## Back to pdist\n", "\n", "[sklearn-onnx](https://github.com/onnx/sklearn-onnx) implements function *pdist* with *ONNX* operators. The parameter ``inputs=[('x', FloatTensorType())`` tels the method ``to_onnx`` that the dimension of the inputs is not fixed and should not be checked."]}, {"cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [{"data": {"text/html": ["<div id=\"Md6b2e28b2e55475b9077148617dfef2d-cont\"><div id=\"Md6b2e28b2e55475b9077148617dfef2d\" style=\"width:;height:;\"></div></div>\n", "<script>\n", "\n", "require(['http://www.xavierdupre.fr/js/vizjs/viz.js'], function() { var svgGraph = Viz(\"digraph{\\n  size=7;\\n  orientation=portrait;\\n  ranksep=0.25;\\n  nodesep=0.05;\\n\\n  x [shape=box color=red label=\\\"x\\nfloat(('?',))\\\" fontsize=10];\\n\\n  Y [shape=box color=green label=\\\"Y\\nfloat((0, 0))\\\" fontsize=10];\\n\\n\\n  scan0_{idself} [shape=box label=\\\"scan0_{idself}\\\" fontsize=10];\\n  scan1_{idself} [shape=box label=\\\"scan1_{idself}\\\" fontsize=10];\\n  Sc_Scan [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Scan\\n(Sc_Scan)\\nbody=node {\\n  input: 'next_in'...\\nnum_scan_inputs=1\\\" fontsize=10];\\n  x -> Sc_Scan;\\n  x -> Sc_Scan;\\n  Sc_Scan -> scan0_{idself};\\n  Sc_Scan -> scan1_{idself};\\n\\n  Id_Identity1 [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Identity\\n(Id_Identity1)\\\" fontsize=10];\\n  scan1_{idself} -> Id_Identity1;\\n  Id_Identity1 -> Y;\\n}\");\n", "document.getElementById('Md6b2e28b2e55475b9077148617dfef2d').innerHTML = svgGraph; });\n", "\n", "</script>"], "text/plain": ["<jyquickhelper.jspy.render_nb_js_dot.RenderJsDot at 0x1f9fa48c850>"]}, "execution_count": 12, "metadata": {}, "output_type": "execute_result"}], "source": ["# from skl2onnx.algebra.complex_functions import squareform_pdist_\n", "\n", "from collections import OrderedDict\n", "from skl2onnx.algebra.onnx_ops import (\n", "    OnnxSub, OnnxReduceSumSquare, OnnxSqueeze,\n", "    OnnxIdentity, OnnxScan)\n", "from skl2onnx.common.data_types import FloatTensorType\n", "\n", "\n", "def squareform_pdist(X, op_version=17, **kwargs):\n", "    \"\"\"Returns the ONNX graph which computes\n", "    ``squareform(pdist(X, metric='sqeuclidean')``.\"\"\"\n", "\n", "    # The subgraph executed at every iteration.\n", "    opv = op_version\n", "    diff = OnnxSub('next_in', 'next', output_names=['diff'], op_version=opv)\n", "    id_next = OnnxIdentity('next_in', output_names=['next_out'], op_version=opv)\n", "    norm = OnnxReduceSumSquare(diff, output_names=['norm'], axes=[1], op_version=opv)\n", "    flat = OnnxSqueeze(norm, numpy.array([1], dtype=numpy.int64),\n", "                       output_names=['scan_out'], op_version=opv)\n", "    scan_body = id_next.to_onnx(\n", "        OrderedDict([('next_in', FloatTensorType()),\n", "                     ('next', FloatTensorType())]),\n", "        # Size must be empty otherwise onnxruntime fails\n", "        # at execution time if it receives a matrix\n", "        # with a different shape. With 'None', the same ONNX graph\n", "        # can compute pairwise distance for any shape.\n", "        outputs=[('next_out', FloatTensorType([None, None])),\n", "                 ('scan_out', FloatTensorType([None]))],\n", "        other_outputs=[flat])\n", "\n", "    # The loop.\n", "    # 'scan0_{idself}' means the variable name will include\n", "    # id(OnnxScan), this is needed if squareform_pdist is used\n", "    # twice in the same graph.\n", "    node = OnnxScan(X, X, output_names=['scan0_{idself}', 'scan1_{idself}'],\n", "                    num_scan_inputs=1, body=scan_body.graph, op_version=opv,\n", "                    **kwargs)\n", "    return node[1]    \n", "\n", "opv = 17\n", "onnx_fct = OnnxIdentity(squareform_pdist('x', op_version=opv),\n", "                        output_names='Y', op_version=opv)\n", "model_def = onnx_fct.to_onnx(inputs=[('x', FloatTensorType())],\n", "                             target_opset=opv)\n", "\n", "# add -l 1 if nothing shows up\n", "%onnxview model_def"]}, {"cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [], "source": ["from collections import OrderedDict\n", "from skl2onnx.algebra.onnx_ops import (\n", "    OnnxSub, OnnxReduceSumSquare, OnnxSqueeze,\n", "    OnnxIdentity, OnnxScan)\n", "from skl2onnx.common.data_types import FloatTensorType\n", "\n", "\n", "def squareform_pdist(X, op_version=17, **kwargs):\n", "    # The subgraph executed at every iteration.\n", "    opv = op_version\n", "    diff = OnnxSub('next_in', 'next', output_names=['diff'], op_version=opv)\n", "    id_next = OnnxIdentity('next_in', output_names=['next_out'], op_version=opv)\n", "    norm = OnnxReduceSumSquare(diff, output_names=['norm'], axes=[1], op_version=opv)\n", "    flat = OnnxSqueeze(norm, numpy.array([1], dtype=numpy.int64),\n", "                       output_names=['scan_out'], op_version=opv)\n", "    scan_body = id_next.to_onnx(\n", "        OrderedDict([('next_in', FloatTensorType()),\n", "                     ('next', FloatTensorType())]),\n", "        outputs=[('next_out', FloatTensorType([None, None])),\n", "                 ('scan_out', FloatTensorType([None]))],\n", "        other_outputs=[flat])\n", "\n", "    # The loop.\n", "    node = OnnxScan(X, X, output_names=['scan0_{idself}', 'scan1_{idself}'],\n", "                    num_scan_inputs=1, body=scan_body.graph, op_version=opv,\n", "                    **kwargs)\n", "    return node[1]    \n", "\n", "opv = 17\n", "onnx_fct = OnnxIdentity(squareform_pdist('x', op_version=opv),\n", "                        output_names='Y', op_version=opv)\n", "model_def = onnx_fct.to_onnx(inputs=[('x', FloatTensorType())])"]}, {"cell_type": "markdown", "metadata": {}, "source": ["Notice the double arrow. Input _x_ is used twice, once as an permanent state involved in broacasted substract, another time to iterator rows. On the other side, the first output of operator *Scan* is a permanent state equal to the input, the second one is an aggregation of results produced at each iteration. Each of those produces a row of a final matrix."]}, {"cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [{"data": {"text/html": ["<div id=\"M1c083263fa82410f863c0065b8fdcf5a-cont\"><div id=\"M1c083263fa82410f863c0065b8fdcf5a\" style=\"width:;height:;\"></div></div>\n", "<script>\n", "\n", "require(['http://www.xavierdupre.fr/js/vizjs/viz.js'], function() { var svgGraph = Viz(\"digraph{\\n  size=7;\\n  orientation=portrait;\\n  ranksep=0.25;\\n  nodesep=0.05;\\n\\n  next_in [shape=box color=red label=\\\"next_in\\nfloat(('?',))\\\" fontsize=10];\\n  next [shape=box color=red label=\\\"next\\nfloat(('?',))\\\" fontsize=10];\\n\\n  next_out [shape=box color=green label=\\\"next_out\\nfloat((0, 0))\\\" fontsize=10];\\n  scan_out [shape=box color=green label=\\\"scan_out\\nfloat((0,))\\\" fontsize=10];\\n\\n  Sq_Squeezecst [shape=box label=\\\"Sq_Squeezecst\\nint64((1,))\\n[1]\\\" fontsize=10];\\n\\n  Id_Identity [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Identity\\n(Id_Identity)\\\" fontsize=10];\\n  next_in -> Id_Identity;\\n  Id_Identity -> next_out;\\n\\n  diff [shape=box label=\\\"diff\\\" fontsize=10];\\n  Su_Sub [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Sub\\n(Su_Sub)\\\" fontsize=10];\\n  next_in -> Su_Sub;\\n  next -> Su_Sub;\\n  Su_Sub -> diff;\\n\\n  norm [shape=box label=\\\"norm\\\" fontsize=10];\\n  Re_ReduceSumSquare [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"ReduceSumSquare\\n(Re_ReduceSumSquare)\\naxes=[1]\\\" fontsize=10];\\n  diff -> Re_ReduceSumSquare;\\n  Re_ReduceSumSquare -> norm;\\n\\n  Sq_Squeeze [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Squeeze\\n(Sq_Squeeze)\\\" fontsize=10];\\n  norm -> Sq_Squeeze;\\n  Sq_Squeezecst -> Sq_Squeeze;\\n  Sq_Squeeze -> scan_out;\\n}\");\n", "document.getElementById('M1c083263fa82410f863c0065b8fdcf5a').innerHTML = svgGraph; });\n", "\n", "</script>"], "text/plain": ["<jyquickhelper.jspy.render_nb_js_dot.RenderJsDot at 0x1f9fa460f10>"]}, "execution_count": 14, "metadata": {}, "output_type": "execute_result"}], "source": ["oinf = OnnxInference(model_def)\n", "body = oinf['Sc_Scan', 'body']\n", "\n", "# add -l 1 if nothing shows up\n", "%onnxview body.g"]}, {"cell_type": "markdown", "metadata": {}, "source": ["All together."]}, {"cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [{"data": {"text/html": ["<div id=\"Mbe44289537d945aa9c160a86db48dd6e-cont\"><div id=\"Mbe44289537d945aa9c160a86db48dd6e\" style=\"width:;height:;\"></div></div>\n", "<script>\n", "\n", "require(['http://www.xavierdupre.fr/js/vizjs/viz.js'], function() { var svgGraph = Viz(\"digraph{\\n  size=7;\\n  orientation=portrait;\\n  ranksep=0.25;\\n  nodesep=0.05;\\n\\n  x [shape=box color=red label=\\\"x\\nfloat(('?',))\\\" fontsize=10];\\n\\n  Y [shape=box color=green label=\\\"Y\\nfloat((0, 0))\\\" fontsize=10];\\n\\n\\n  scan0_{idself} [shape=box label=\\\"scan0_{idself}\\\" fontsize=10];\\n  scan1_{idself} [shape=box label=\\\"scan1_{idself}\\\" fontsize=10];\\n  subgraph cluster_Scan2173157244640_2172776124464 {\\n    label=\\\"Scan\\n(Sc_Scan)\\nbody=node {\\n  input: 'next_in'...\\nnum_scan_inputs=1\\\";\\n    fontsize=10;\\n    color=black;\\n    B_next_in [shape=box color=red label=\\\"next_in\\nfloat(('?',))\\\" fontsize=10];\\n    B_next [shape=box color=red label=\\\"next\\nfloat(('?',))\\\" fontsize=10];\\n  \\n    B_next_out [shape=box color=green label=\\\"next_out\\nfloat((0, 0))\\\" fontsize=10];\\n    B_scan_out [shape=box color=green label=\\\"scan_out\\nfloat((0,))\\\" fontsize=10];\\n  \\n    B_Sq_Squeezecst [shape=box label=\\\"Sq_Squeezecst\\nint64((1,))\\n[1]\\\" fontsize=10];\\n  \\n    B_Id_Identity [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Identity\\n(Id_Identity)\\\" fontsize=10];\\n    B_next_in -> B_Id_Identity;\\n    B_Id_Identity -> B_next_out;\\n  \\n    B_diff [shape=box label=\\\"diff\\\" fontsize=10];\\n    B_Su_Sub [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Sub\\n(Su_Sub)\\\" fontsize=10];\\n    B_next_in -> B_Su_Sub;\\n    B_next -> B_Su_Sub;\\n    B_Su_Sub -> B_diff;\\n  \\n    B_norm [shape=box label=\\\"norm\\\" fontsize=10];\\n    B_Re_ReduceSumSquare [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"ReduceSumSquare\\n(Re_ReduceSumSquare)\\naxes=[1]\\\" fontsize=10];\\n    B_diff -> B_Re_ReduceSumSquare;\\n    B_Re_ReduceSumSquare -> B_norm;\\n  \\n    B_Sq_Squeeze [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Squeeze\\n(Sq_Squeeze)\\\" fontsize=10];\\n    B_norm -> B_Sq_Squeeze;\\n    B_Sq_Squeezecst -> B_Sq_Squeeze;\\n    B_Sq_Squeeze -> B_scan_out;\\n  }\\n  x -> B_next_in;\\n  x -> B_next;\\n  B_next_out -> scan0_{idself};\\n  B_scan_out -> scan1_{idself};\\n  Sc_Scan -> B_Id_Identity [lhead=cluster_Scan2173157244640_2172776124464];\\n  x -> Sc_Scan;\\n  x -> Sc_Scan;\\n  Sc_Scan -> scan0_{idself};\\n  Sc_Scan -> scan1_{idself};\\n\\n  Id_Identity1 [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Identity\\n(Id_Identity1)\\\" fontsize=10];\\n  scan1_{idself} -> Id_Identity1;\\n  Id_Identity1 -> Y;\\n}\");\n", "document.getElementById('Mbe44289537d945aa9c160a86db48dd6e').innerHTML = svgGraph; });\n", "\n", "</script>"], "text/plain": ["<jyquickhelper.jspy.render_nb_js_dot.RenderJsDot at 0x1f9ef5f50c0>"]}, "execution_count": 15, "metadata": {}, "output_type": "execute_result"}], "source": ["# add -l 1 if nothing shows up\n", "%onnxview model_def -r 1"]}, {"cell_type": "markdown", "metadata": {}, "source": ["Let's now execute the graph and compare it with the original graph."]}, {"cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [{"data": {"text/plain": ["array([[0.  , 2.  , 0.02, 5.  ],\n", "       [2.  , 0.  , 1.62, 1.  ],\n", "       [0.02, 1.62, 0.  , 4.42],\n", "       [5.  , 1.  , 4.42, 0.  ]])"]}, "execution_count": 16, "metadata": {}, "output_type": "execute_result"}], "source": ["d1 = squareform(pdist(M, metric='sqeuclidean'))\n", "d1"]}, {"cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [{"data": {"text/plain": ["array([[0.  , 2.  , 0.02, 5.  ],\n", "       [2.  , 0.  , 1.62, 1.  ],\n", "       [0.02, 1.62, 0.  , 4.42],\n", "       [5.  , 1.  , 4.42, 0.  ]])"]}, "execution_count": 17, "metadata": {}, "output_type": "execute_result"}], "source": ["oinf.run({'x': M})['Y']"]}, {"cell_type": "code", "execution_count": 17, "metadata": {}, "outputs": [{"name": "stdout", "output_type": "stream", "text": ["9.57 \u00b5s \u00b1 166 ns per loop (mean \u00b1 std. dev. of 7 runs, 100,000 loops each)\n"]}], "source": ["%timeit squareform(pdist(M, metric='sqeuclidean'))"]}, {"cell_type": "code", "execution_count": 18, "metadata": {}, "outputs": [{"name": "stdout", "output_type": "stream", "text": ["36.7 \u00b5s \u00b1 2.28 \u00b5s per loop (mean \u00b1 std. dev. of 7 runs, 10,000 loops each)\n"]}], "source": ["%timeit custom_pdist(M)"]}, {"cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [{"name": "stdout", "output_type": "stream", "text": ["35.7 \u00b5s \u00b1 646 ns per loop (mean \u00b1 std. dev. of 7 runs, 10,000 loops each)\n"]}], "source": ["%timeit custom_pdist_lower(M)"]}, {"cell_type": "code", "execution_count": 20, "metadata": {}, "outputs": [{"name": "stdout", "output_type": "stream", "text": ["206 \u00b5s \u00b1 17.3 \u00b5s per loop (mean \u00b1 std. dev. of 7 runs, 1,000 loops each)\n"]}], "source": ["%timeit oinf.run({'x': M})['Y']"]}, {"cell_type": "code", "execution_count": 21, "metadata": {}, "outputs": [], "source": ["M32 = M.astype(numpy.float32)"]}, {"cell_type": "code", "execution_count": 22, "metadata": {}, "outputs": [], "source": ["model_def.ir_version = 8"]}, {"cell_type": "code", "execution_count": 23, "metadata": {}, "outputs": [{"data": {"text/plain": ["array([[0.        , 2.        , 0.02000001, 5.        ],\n", "       [2.        , 0.        , 1.6199999 , 1.        ],\n", "       [0.02000001, 1.6199999 , 0.        , 4.42      ],\n", "       [5.        , 1.        , 4.42      , 0.        ]], dtype=float32)"]}, "execution_count": 24, "metadata": {}, "output_type": "execute_result"}], "source": ["oinfrt = OnnxInference(model_def, runtime=\"onnxruntime1\")\n", "oinfrt.run({'x': M32})['Y']"]}, {"cell_type": "code", "execution_count": 24, "metadata": {}, "outputs": [{"name": "stdout", "output_type": "stream", "text": ["47.2 \u00b5s \u00b1 295 ns per loop (mean \u00b1 std. dev. of 7 runs, 10,000 loops each)\n"]}], "source": ["%timeit oinfrt.run({'x': M32})['Y']"]}, {"cell_type": "markdown", "metadata": {}, "source": ["## Benchmark"]}, {"cell_type": "code", "execution_count": 25, "metadata": {}, "outputs": [{"data": {"text/plain": ["{'average': 4.610000061802566e-05,\n", " 'deviation': 8.591571512763607e-06,\n", " 'min_exec': 3.507999936118722e-05,\n", " 'max_exec': 6.718999939039349e-05,\n", " 'repeat': 10,\n", " 'number': 10,\n", " 'nrows': 4,\n", " 'ncols': 2,\n", " 'name': 'scipy'}"]}, "execution_count": 26, "metadata": {}, "output_type": "execute_result"}], "source": ["from timeit import Timer\n", "\n", "\n", "def measure_time(name, stmt, context, repeat=10, number=10):\n", "    tim = Timer(stmt, globals=context)\n", "    res = numpy.array(tim.repeat(repeat=repeat, number=number))\n", "    res /= number\n", "    mean = numpy.mean(res)\n", "    dev = numpy.mean(res ** 2)\n", "    dev = (dev - mean**2) ** 0.5\n", "    return dict(average=mean, deviation=dev, min_exec=numpy.min(res),\n", "                max_exec=numpy.max(res), repeat=repeat, number=number,\n", "                nrows=context['M'].shape[0], ncols=context['M'].shape[1],\n", "                name=name)\n", "\n", "measure_time(\"scipy\", \"squareform(pdist(M, metric='sqeuclidean'))\",\n", "             context={'squareform': squareform, 'M': M,\n", "                      'pdist': pdist})"]}, {"cell_type": "code", "execution_count": 26, "metadata": {}, "outputs": [{"name": "stderr", "output_type": "stream", "text": ["feat=100 n=400: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 26/26 [01:32<00:00,  3.56s/it]\n"]}, {"data": {"text/html": ["<div>\n", "<style scoped>\n", "    .dataframe tbody tr th:only-of-type {\n", "        vertical-align: middle;\n", "    }\n", "\n", "    .dataframe tbody tr th {\n", "        vertical-align: top;\n", "    }\n", "\n", "    .dataframe thead th {\n", "        text-align: right;\n", "    }\n", "</style>\n", "<table border=\"1\" class=\"dataframe\">\n", "  <thead>\n", "    <tr style=\"text-align: right;\">\n", "      <th></th>\n", "      <th>average</th>\n", "      <th>deviation</th>\n", "      <th>min_exec</th>\n", "      <th>max_exec</th>\n", "      <th>repeat</th>\n", "      <th>number</th>\n", "      <th>nrows</th>\n", "      <th>ncols</th>\n", "      <th>name</th>\n", "      <th>dimres</th>\n", "    </tr>\n", "  </thead>\n", "  <tbody>\n", "    <tr>\n", "      <th>0</th>\n", "      <td>0.000045</td>\n", "      <td>7.066342e-06</td>\n", "      <td>0.000040</td>\n", "      <td>0.000065</td>\n", "      <td>10</td>\n", "      <td>10</td>\n", "      <td>5</td>\n", "      <td>5</td>\n", "      <td>scipy</td>\n", "      <td>5</td>\n", "    </tr>\n", "    <tr>\n", "      <th>1</th>\n", "      <td>0.000121</td>\n", "      <td>3.075137e-05</td>\n", "      <td>0.000084</td>\n", "      <td>0.000189</td>\n", "      <td>10</td>\n", "      <td>10</td>\n", "      <td>5</td>\n", "      <td>5</td>\n", "      <td>numpy</td>\n", "      <td>5</td>\n", "    </tr>\n", "    <tr>\n", "      <th>2</th>\n", "      <td>0.000046</td>\n", "      <td>9.946988e-07</td>\n", "      <td>0.000045</td>\n", "      <td>0.000049</td>\n", "      <td>10</td>\n", "      <td>10</td>\n", "      <td>5</td>\n", "      <td>5</td>\n", "      <td>numpy-lower</td>\n", "      <td>5</td>\n", "    </tr>\n", "    <tr>\n", "      <th>3</th>\n", "      <td>0.000400</td>\n", "      <td>1.665463e-04</td>\n", "      <td>0.000224</td>\n", "      <td>0.000716</td>\n", "      <td>10</td>\n", "      <td>10</td>\n", "      <td>5</td>\n", "      <td>5</td>\n", "      <td>onnx-py</td>\n", "      <td>5</td>\n", "    </tr>\n", "    <tr>\n", "      <th>4</th>\n", "      <td>0.000055</td>\n", "      <td>3.251956e-06</td>\n", "      <td>0.000051</td>\n", "      <td>0.000063</td>\n", "      <td>10</td>\n", "      <td>10</td>\n", "      <td>5</td>\n", "      <td>5</td>\n", "      <td>onnx-rt</td>\n", "      <td>5</td>\n", "    </tr>\n", "  </tbody>\n", "</table>\n", "</div>"], "text/plain": ["    average     deviation  min_exec  max_exec  repeat  number  nrows  ncols  \\\n", "0  0.000045  7.066342e-06  0.000040  0.000065      10      10      5      5   \n", "1  0.000121  3.075137e-05  0.000084  0.000189      10      10      5      5   \n", "2  0.000046  9.946988e-07  0.000045  0.000049      10      10      5      5   \n", "3  0.000400  1.665463e-04  0.000224  0.000716      10      10      5      5   \n", "4  0.000055  3.251956e-06  0.000051  0.000063      10      10      5      5   \n", "\n", "          name  dimres  \n", "0        scipy       5  \n", "1        numpy       5  \n", "2  numpy-lower       5  \n", "3      onnx-py       5  \n", "4      onnx-rt       5  "]}, "execution_count": 27, "metadata": {}, "output_type": "execute_result"}], "source": ["from tqdm import trange\n", "\n", "def generator():\n", "    for feat in [5, 10, 50, 100]:\n", "        for n in [5, 10, 20, 50, 100, 400, 1000]:\n", "            if n <= 500 or feat <= 10:\n", "                yield feat, n\n", "            \n", "all_values = list(generator())\n", "\n", "rows = []\n", "\n", "with trange(len(all_values)) as t:\n", "    for i in t:        \n", "        feat, n = all_values[i]\n", "        t.set_description(\"feat=%d n=%d\" % (feat, n))\n", "        M = numpy.random.rand(n, feat)\n", "\n", "        context = {'squareform': squareform, 'M': M, 'pdist': pdist}\n", "        res = measure_time(\"scipy\", \"squareform(pdist(M, metric='sqeuclidean'))\", context=context)\n", "        res['dimres'] = squareform(pdist(M, metric='sqeuclidean')).shape[0]\n", "        rows.append(res)\n", "\n", "        context = {'M': M, 'custom_pdist': custom_pdist}\n", "        res = measure_time(\"numpy\", \"custom_pdist(M)\", context=context)\n", "        res['dimres'] = custom_pdist(M).shape[0]\n", "        rows.append(res)\n", "\n", "        context = {'M': M, 'custom_pdist_lower': custom_pdist_lower}\n", "        res = measure_time(\"numpy-lower\", \"custom_pdist_lower(M)\", context=context)\n", "        res['dimres'] = custom_pdist_lower(M).shape[0]\n", "        rows.append(res)\n", "\n", "        context = {'oinf': oinf, 'M': M}\n", "        res = measure_time(\"onnx-py\", \"oinf.run({'x': M})['Y']\", context=context)\n", "        res['dimres'] = oinf.run({'x': M})['Y'].shape[0]\n", "        rows.append(res)\n", "\n", "        M32 = M.astype(numpy.float32)\n", "        context = {'oinfrt': oinfrt, 'M': M32}\n", "        res = measure_time(\"onnx-rt\", \"oinfrt.run({'x': M})['Y']\", context=context)\n", "        res['dimres'] = oinfrt.run({'x': M32})['Y'].shape[0]\n", "        rows.append(res)\n", "\n", "    \n", "from pandas import DataFrame\n", "df = DataFrame(rows)\n", "df.head()"]}, {"cell_type": "code", "execution_count": 27, "metadata": {}, "outputs": [{"data": {"text/html": ["<div>\n", "<style scoped>\n", "    .dataframe tbody tr th:only-of-type {\n", "        vertical-align: middle;\n", "    }\n", "\n", "    .dataframe tbody tr th {\n", "        vertical-align: top;\n", "    }\n", "\n", "    .dataframe thead th {\n", "        text-align: right;\n", "    }\n", "</style>\n", "<table border=\"1\" class=\"dataframe\">\n", "  <thead>\n", "    <tr style=\"text-align: right;\">\n", "      <th></th>\n", "      <th>nrows</th>\n", "      <th>5</th>\n", "      <th>10</th>\n", "      <th>20</th>\n", "      <th>50</th>\n", "      <th>100</th>\n", "    </tr>\n", "    <tr>\n", "      <th>ncols</th>\n", "      <th>name</th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "    </tr>\n", "  </thead>\n", "  <tbody>\n", "    <tr>\n", "      <th rowspan=\"5\" valign=\"top\">5</th>\n", "      <th>numpy</th>\n", "      <td>0.000121</td>\n", "      <td>0.000159</td>\n", "      <td>0.000248</td>\n", "      <td>0.000542</td>\n", "      <td>0.001130</td>\n", "    </tr>\n", "    <tr>\n", "      <th>numpy-lower</th>\n", "      <td>0.000046</td>\n", "      <td>0.000116</td>\n", "      <td>0.000239</td>\n", "      <td>0.000662</td>\n", "      <td>0.001294</td>\n", "    </tr>\n", "    <tr>\n", "      <th>onnx-py</th>\n", "      <td>0.000400</td>\n", "      <td>0.000501</td>\n", "      <td>0.000886</td>\n", "      <td>0.002395</td>\n", "      <td>0.004488</td>\n", "    </tr>\n", "    <tr>\n", "      <th>onnx-rt</th>\n", "      <td>0.000055</td>\n", "      <td>0.000092</td>\n", "      <td>0.000148</td>\n", "      <td>0.000382</td>\n", "      <td>0.000956</td>\n", "    </tr>\n", "    <tr>\n", "      <th>scipy</th>\n", "      <td>0.000045</td>\n", "      <td>0.000021</td>\n", "      <td>0.000027</td>\n", "      <td>0.000027</td>\n", "      <td>0.000070</td>\n", "    </tr>\n", "    <tr>\n", "      <th rowspan=\"5\" valign=\"top\">10</th>\n", "      <th>numpy</th>\n", "      <td>0.000063</td>\n", "      <td>0.000087</td>\n", "      <td>0.000211</td>\n", "      <td>0.001672</td>\n", "      <td>0.001868</td>\n", "    </tr>\n", "    <tr>\n", "      <th>numpy-lower</th>\n", "      <td>0.000061</td>\n", "      <td>0.000117</td>\n", "      <td>0.000280</td>\n", "      <td>0.001676</td>\n", "      <td>0.001954</td>\n", "    </tr>\n", "    <tr>\n", "      <th>onnx-py</th>\n", "      <td>0.000253</td>\n", "      <td>0.000530</td>\n", "      <td>0.000950</td>\n", "      <td>0.007125</td>\n", "      <td>0.004770</td>\n", "    </tr>\n", "    <tr>\n", "      <th>onnx-rt</th>\n", "      <td>0.000068</td>\n", "      <td>0.000110</td>\n", "      <td>0.000178</td>\n", "      <td>0.000951</td>\n", "      <td>0.001071</td>\n", "    </tr>\n", "    <tr>\n", "      <th>scipy</th>\n", "      <td>0.000032</td>\n", "      <td>0.000019</td>\n", "      <td>0.000018</td>\n", "      <td>0.000048</td>\n", "      <td>0.000096</td>\n", "    </tr>\n", "    <tr>\n", "      <th rowspan=\"5\" valign=\"top\">50</th>\n", "      <th>numpy</th>\n", "      <td>0.000051</td>\n", "      <td>0.000098</td>\n", "      <td>0.000220</td>\n", "      <td>0.000664</td>\n", "      <td>0.001796</td>\n", "    </tr>\n", "    <tr>\n", "      <th>numpy-lower</th>\n", "      <td>0.000048</td>\n", "      <td>0.000112</td>\n", "      <td>0.000248</td>\n", "      <td>0.000702</td>\n", "      <td>0.001657</td>\n", "    </tr>\n", "    <tr>\n", "      <th>onnx-py</th>\n", "      <td>0.000283</td>\n", "      <td>0.000507</td>\n", "      <td>0.000910</td>\n", "      <td>0.002543</td>\n", "      <td>0.005693</td>\n", "    </tr>\n", "    <tr>\n", "      <th>onnx-rt</th>\n", "      <td>0.000072</td>\n", "      <td>0.000109</td>\n", "      <td>0.000187</td>\n", "      <td>0.000573</td>\n", "      <td>0.001665</td>\n", "    </tr>\n", "    <tr>\n", "      <th>scipy</th>\n", "      <td>0.000030</td>\n", "      <td>0.000029</td>\n", "      <td>0.000034</td>\n", "      <td>0.000068</td>\n", "      <td>0.000167</td>\n", "    </tr>\n", "    <tr>\n", "      <th rowspan=\"5\" valign=\"top\">100</th>\n", "      <th>numpy</th>\n", "      <td>0.000066</td>\n", "      <td>0.000106</td>\n", "      <td>0.000234</td>\n", "      <td>0.000870</td>\n", "      <td>0.002847</td>\n", "    </tr>\n", "    <tr>\n", "      <th>numpy-lower</th>\n", "      <td>0.000068</td>\n", "      <td>0.000119</td>\n", "      <td>0.000262</td>\n", "      <td>0.000798</td>\n", "      <td>0.002007</td>\n", "    </tr>\n", "    <tr>\n", "      <th>onnx-py</th>\n", "      <td>0.000303</td>\n", "      <td>0.000465</td>\n", "      <td>0.000963</td>\n", "      <td>0.002710</td>\n", "      <td>0.007020</td>\n", "    </tr>\n", "    <tr>\n", "      <th>onnx-rt</th>\n", "      <td>0.000076</td>\n", "      <td>0.000104</td>\n", "      <td>0.000218</td>\n", "      <td>0.000786</td>\n", "      <td>0.002568</td>\n", "    </tr>\n", "    <tr>\n", "      <th>scipy</th>\n", "      <td>0.000028</td>\n", "      <td>0.000027</td>\n", "      <td>0.000034</td>\n", "      <td>0.000088</td>\n", "      <td>0.000305</td>\n", "    </tr>\n", "  </tbody>\n", "</table>\n", "</div>"], "text/plain": ["nrows                   5         10        20        50        100\n", "ncols name                                                         \n", "5     numpy        0.000121  0.000159  0.000248  0.000542  0.001130\n", "      numpy-lower  0.000046  0.000116  0.000239  0.000662  0.001294\n", "      onnx-py      0.000400  0.000501  0.000886  0.002395  0.004488\n", "      onnx-rt      0.000055  0.000092  0.000148  0.000382  0.000956\n", "      scipy        0.000045  0.000021  0.000027  0.000027  0.000070\n", "10    numpy        0.000063  0.000087  0.000211  0.001672  0.001868\n", "      numpy-lower  0.000061  0.000117  0.000280  0.001676  0.001954\n", "      onnx-py      0.000253  0.000530  0.000950  0.007125  0.004770\n", "      onnx-rt      0.000068  0.000110  0.000178  0.000951  0.001071\n", "      scipy        0.000032  0.000019  0.000018  0.000048  0.000096\n", "50    numpy        0.000051  0.000098  0.000220  0.000664  0.001796\n", "      numpy-lower  0.000048  0.000112  0.000248  0.000702  0.001657\n", "      onnx-py      0.000283  0.000507  0.000910  0.002543  0.005693\n", "      onnx-rt      0.000072  0.000109  0.000187  0.000573  0.001665\n", "      scipy        0.000030  0.000029  0.000034  0.000068  0.000167\n", "100   numpy        0.000066  0.000106  0.000234  0.000870  0.002847\n", "      numpy-lower  0.000068  0.000119  0.000262  0.000798  0.002007\n", "      onnx-py      0.000303  0.000465  0.000963  0.002710  0.007020\n", "      onnx-rt      0.000076  0.000104  0.000218  0.000786  0.002568\n", "      scipy        0.000028  0.000027  0.000034  0.000088  0.000305"]}, "execution_count": 28, "metadata": {}, "output_type": "execute_result"}], "source": ["from pandas import pivot_table\n", "piv = pivot_table(df, index=[\"nrows\"], columns= ['ncols', 'name'], values='average')\n", "piv.head().T"]}, {"cell_type": "code", "execution_count": 28, "metadata": {}, "outputs": [], "source": ["%matplotlib inline"]}, {"cell_type": "code", "execution_count": 29, "metadata": {}, "outputs": [{"data": {"image/png": "iVBORw0KGgoAAAANSUhEUgAAAzsAAADkCAYAAAC/rzpIAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjYuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8o6BhiAAAACXBIWXMAAAsTAAALEwEAmpwYAAC0VklEQVR4nOzdd1xV5R/A8c+5l703CKJguEUcoJa7UjP3yFHu1EwzNbVs/yotV+7cIq5M09yaZmpuBQcKOHCggCJ7X7jr+f2BmrkHcC943r18Bfeee873Mh7O9xnfRxJCIJPJZDKZTCaTyWSljcLQAchkMplMJpPJZDJZUZCTHZlMJpPJZDKZTFYqycmOTCaTyWQymUwmK5XkZEcmk8lkMplMJpOVSnKyI5PJZDKZTCaTyUolOdmRyWQymUwmk8lkpZKc7MiKhCRJ/5MkaaWh45DJZCWf3J7IZLLnJbcfMjnZkRkVSZIWSpJ0QZIkvSRJ/R7y/ChJkhIkScqUJClYkiRzA4Qpk8lKAEmShCRJOZIkZd/+t/ie5yRJkiZJkpRy+98kSZIkQ8Yrk8mMx4vcj0iS5CNJ0l5JknIlSTovSdKbxRq87D/kZEdmbMKBocDJ+5+QJKkVMA54AygPVAC+K9boZDJZSRMghLC5/W/gPY8PBjoCAUBNoB3wgQHik8lkxulF7kdWA6cAZ+BLYJ0kSa5FHbDs4eRkRwaAJEkxkiSNkSTpjCRJGZIkrZEkyeL2cx0kSTp9u/fisiRJb91+3FOSpM2SJKVKknRJkqRBLxqHEOIXIcTfQN5Dnu4LLBFCRAoh0oAfgH4vek2ZTFa4jKU9eYK+wM9CiDghRDzwM3J7IpMZnLG0H897PyJJUiWgDvCtEEIlhFgPnAW6vGhMsudjYugAZEalG/AWBb/Yh4B+kiSdBJYDXYG/gTKA7e3jfwMiAE+gCvCXJEmXhRB77j+xJEnpj7nuRCHExKeIrzqw6Z7PwwF3SZKchRApT/F6mUxWfIylPdkvSZICOAx8IoSIuf14dQrakDvCbz8mk8kMz1jaj0d55P3I7eeuCCGy7ntebl8MRE52ZPeaJYS4ASBJ0hagFlAbCBZC/HX7mPjbz3sDDYE2Qog84PTt+fB9gAcaFyGEQyHEZwNk3PP5nY9tATnZkcmMizG0J02Bo4AVMB7YKklSLSGEloe3JzaSJElCCPEsb1QmkxU6Y2g/Hudx9yP3P3fnea9CuK7sOcjT2GT3Srjn41wKfmG9gcsPOdYTSL2v5+IaRfvLnA3Y3fP5nY+zHnKsTCYzLIO3J0KI/UIItRAiHRgB+AJVbz/9sPYkW050ZDKjYPD24wkedz9y/3N3npfvVQxETnZkTxILvPKQx28ATpIk2d7zWDlu97Tc755qSA/798VTxhJJwWLiOwKAW/IUNpmsxDB0eyKAOxXXHtaeRD71O5HJZMXN0O3HvR53PxIJVLgvHrl9MSB5GpvsSZYAuyRJ2grs5fYcWSHEeUmSDgM/SZI0BqgEvA+897CTCCFsnuZikiSZUZCES4Dp7UWJaiGEnoK5uiGSJK2ioHH7Cgh5kTcnk8mKVbG1J5IkVQdMKVgYbEnBNLZ44NztQ5YDn0iStJ2CJGg0MPsF3ptMJitaJeJ+RAhxUZKk08C3kiR9BbSmoOKjXKDAQOSRHdljCSGOA/2B6RTMOf2HgjKLAD0BHwp+0TdQUHlk9wtechegAl4DFt7+uMntWP4EJlPQyF2nYJj62xe8nkwmKybF3J64A2uATODK7XO3FUJobj+/ANhCQTIUAWy7/ZhMJjNCJex+pAcQCKQBE4GuQoikF4xH9pwkeXqyTCaTyWQymUwmK43kkR2ZTCaTyWQymUxWKsnJjkwmk8lkMplMJiuV5GRHJpPJZDKZTCaTlUpysiOTyWQymUwmk8lKJTnZkclkMplMJpPJZKWSUe+z4+LiInx8fAwdhkwmu8eJEyeShRCuho7jWcntiUxmXOS2RCaTFZbHtSdGnez4+PgQFhZm6DBkMtk9JEm6ZugYnoUkSe2Adn5+fnJ7IpMZkZLWltwh35vIZMbnce1JsU1jkySpgiRJSyRJWldc15TJZDIhxBYhxGB7e3tDhyKTyUowSZLaSZK0MCMjw9ChyGSyZ/BUyY4kScGSJCVKkhRx3+NvSZJ0QZKkS5IkjXvcOYQQV4QQ779IsDKZTCaTyWSGIHecyGQl09NOYwsB5gDL7zwgSZIS+AVoAcQBoZIkbQaUwE/3vX6AECLxhaOVyWQymUwmk8lksqf0VMmOEGK/JEk+9z1cD7gkhLgCIEnSb0AHIcRPQNtCjVImk8me071rdu6n0WiIi4sjLy+v+AOTPcDCwoKyZctiampq6FBksgfIbUnJIrcnsjtepECBFxB7z+dxQP1HHSxJkjMwAagtSdLnt5Oihx03GBgMUK5cuRcITyaTPY0rGVc4dvMYPav0NHQoRUIIsQXYEhgYOOj+5+Li4rC1tcXHxwdJkgwQnewOIQQpKSnExcXh6+tr6HBkz+lQ/CEAGno1NHAkhU9uS0oOuT0p+TR6Db+d/423fN7C1erFijYWWzU2IUQKMOQpjlsILAQIDAwURR2XTPayismIYf6Z+ey4ugNLE0ve9n0be/OXay56Xl6efHNiJCRJwtnZmaSkJEOHInsO1zOvMyV0Cvvi9lG/TP1Smew8jtyWGBe5PSnZwhLCmHBsApfSL6EXevpW7/tC53uRZCce8L7n87K3H3thjxsqlslkL+Za5jUWhC9g29VtmCvN6VutL/1q9HvpEp075JsT4yF/L0qeXE0uC88sZHnUckwVpoyqO4peVXsZOiyDkH9+jYv8/Sh5klXJ/Bz2M1uvbMXT2pNZzWfRzLvZC5/3RZKdUKCiJEm+FCQ5PYB3XzgiHj9ULJPJnk9sZiwLzixg65WtmCpM6V21N/1r9MfZ0tnQoclkshJGCMHWK1uZcWIGiapE2r/SnpF1Rr7wdBNjJnfEymRFQ6vXsubCGuacmkO+Lp/BNQcz0H8gliaWhXL+p0p2JElaDTQDXCRJigO+FUIskSTpI2AnBRXYgoUQkYURlNygyGSFJy4rjoVnFrL58mZMFCb0rNKT9/3fx8XSxdChyWSyEigyOZKfjv9EeFI4NZxrMK35NAJcAwwdVpGTO2JlssJ3OvE0E45N4HzqeV7zfI3P632Oj71PoV7jaauxPXTlshBiO7C9UCNCblBkssIQnx3PojOL2HRpEwpJQY8qPXi/xvuluuf1YUpa50lMTAytW7emUaNGHD58GC8vLzZt2sTKlStZuHAharUaPz8/VqxYgZWVFf369cPS0pJTp06RmJhIcHAwy5cv58iRI9SvX5+QkBAAdu3axbfffkt+fj6vvPIKS5cuxcbGxrBvVlaipKhSmHVqFhuiN+Bo4cj3r31PB78OKKRi259c9ozk9kRmrFLzUpl+YjobL23E3cqdac2m8Wa5N4tk+qFRtlDyLsUy2fO7mX2T7458R9sNbdl8eTPvVH6H7Z23M67euJcu0YGSuRFgdHQ0w4YNIzIyEgcHB9avX0/nzp0JDQ0lPDycqlWrsmTJkrvHp6WlceTIEaZPn0779u0ZNWoUkZGRnD17ltOnT5OcnMz48ePZvXs3J0+eJDAwkGnTphnwHcpKEo1ew/LI5QVtyqXN9KnWh62dttKpYic50SkB5PZEZkx0eh1rzq+h7Ya2bL28lQE1BrC542ZalG9RZOusiq0a27OQR3ZksmeXkJPAojOL+OPSHwB0qdiFgf4D8bD2MHBksmfl6+tLrVq1AKhbty4xMTFERETw1VdfkZ6eTnZ2Nq1atbp7fLt27ZAkCX9/f9zd3fH39wegevXqxMTEEBcXR1RUFA0bFlTIUqvVvPrqq8X+vmQlz+H4w0wMncjVjKs09GrIZ0Gf4Wv/cpbyLWmjxHfI7YnMWEQkRzD+6HgiUyKp71GfL+p/QQWHCkV+XaNMdmQy2dO7lXOLxWcXsz56PQJBJ79ODPIfRBmbMk/1el1GBrknTmL7evMijlT2tMzNze9+rFQqUalU9OvXj40bNxIQEEBISAj79u174HiFQvGf1yoUCrRaLUqlkhYtWrB69epiew+yki02M5bJYZPZF7uPcrblmPP6HJqUbfJSV7gqqR2xcnsiM7T0vHRmnprJ+ovrcbV0ZUqTKbTyaVVs7YlRJjsltfdEJitOibmJLDm7hHUX16EXejr4dWBwzcF42ng+9TnyL10idugwNImJVPx7NybOcmU2Y5WVlUWZMmXQaDSsWrUKLy+vp35tgwYNGDZsGJcuXcLPz4+cnBzi4+OpVKlSEUYsK4lyNbksOruIZZHL/lNK2kxpZujQZIVIbk9kxUEv9GyI3sCMkzPIUmfRp1ofPqz1Idam1sUah1EmOyW190QmKw7JqmSWnF3C7xd/R6vX0sGvA4P8B1HWtuwznSdrz17ix44lWyiZ0mgwc81tkFMd4/XDDz9Qv359XF1dqV+/PllZWU/9WldXV0JCQujZsyf5+fkAjB8/Xr45kd0lhGDb1W1MD5t+t5T0iDojcLNyM3RosiIgtyeyohaVEsWEoxM4k3yGuu51+bL+l1R0rGiQWCQhhEEu/DQCAwNFWFiYocOQyYxCsiqZ4Ihg1l5Yi1avpd0r7RhcczDett5PfvE9hBCkzJ9P4qzZXHUsy48N+jG0e0PerVfuqYaUJUk6IYQIfN73YSgPa0/OnTtH1apVDRSR7GHk70nxi0yJZOKxiZxOOk115+p8Xv/zYiklLbclsqImf1+KX0Z+BrNPzWbthbU4WTgxOnA0bSu0feYpa7rsbG799BPO7w/EvMKT1wk+rj0xypEdeRqbTPavFFUKSyOWsubCGtR6NW0rtOWDmh9Qzq7cM59Ln5ND3OdfkLNrF3vL1uHPt/qztHd9KrnbFkHkMpnMmKWoUph9ajZ/RP8hl5KWyWQvRC/0bL68meknppOen867Vd9laK2h2JnZPfO5VGcjiB89Gk1cHFZ16jxVsvM4RpnsyNPYZDJIy0tjaeRSfjv/G/m6fN72fZsPan7w3JttqePiuPLBUHRXLhFcvS3Wvfvwe5tqWJgqCzdwmUxm1DR6DavPrWZ++HxUWhV9qvXhg4APsDWTOz0eR+6Ilcke7kLqBSYcm8CpxFMEuAawoMUCqjhVeebzCCFIW76cW1N/RnJ0Ylm3cXQPep3aLxifUSY7MtnLLD0vnZDIEH49/yt52jxa+7ZmSMCQFyr3mn3kCFeHjyQvX8Psph/Qa3h3WlZ/OUpSyzcoMtm/DscfZlLoJK5kXKGhV0M+DfqUCvZFX/q1NJA7YmWy/8pSZzH39FxWn1+NnZndC40Oa9PSuPn5F2Tv20dGnVcZ4d2GHGFD04y8F45TTnZkMiORkZ/BsshlrDq3CpVWxVs+bzEkYMgL1aAXQnBj6XLSp0zmho0Lm3t9zuQhb+Fhb1GIkRs3+QZFJisoJT0lbAp7Y/fibetdZKWkb0THoMlTU95fXqwuk5VWdwqa/Bz2MymqFLpV7sbw2sOxN3++zbtzQ0OJHzMWbWoqf73Ri2k2Abzm58LkrjUp62j1wvEaZbIj98TKXiYZ+Rksj1rOqnOryNHk0MqnFUNqDsHP8cV+/vVqNZFjvsBk1zaOl6lO/thvmPWWP0rFfTc3eZkQFwp+b7zQ9WQymfG5v5T0yDoj6V2td5GUkj4Q8gfm0yaQ4uZN+d0bC/38MpnM8C6lXWLCsQmE3QqjhnMN5rw+h+ou1Z/rXEKnI3n+fJJ/mUu+Wxm+eX0E0XZefP92FXrVL4/i/vuV52SUyY7cEyt7GWSqM1kZtZIVUSvI1mTTonwLhgQMoZLji/eI5t9K5FS/wdhfvcDWWq15Y+KX1PG5r7B0fjYcXwiHZ4FGBZ+cAyunF762TCYzvOIsJZ2breLPj7+i6uHtxLl44zfpx0K/hkwmM6wcTQ7zTs9j1blVWJtZ8+2r39K5YufnLmiiuXWLG2PGkhsaSlSNhnxVvjXV/cqw450AfFwKdx8eo0x2ZLLSLFudzYpzK1gRtYIsdRZvlHuDDwM+pLJT5UI5f+yh49z4eARmebnsfGckH3w5ADsL038PUOdC2BI4OANyk8GvBTT/XE50ZLJS4v5S0tOaTyuyUtKRoRHEjRpN1eTrXG70Ni1mTcDc6uWZJiuTlXZCCHbG7GRK6BQSVYl0qdiFEXVG4Gjh+NznzNq3j5uff4EmV8WCBu/xZ9m6jGlZifcbVXhw9kkhkJMdmayY3OlpnRo6lZS8FJp7N2doraHPVbHkUQ7ODsFu3lRUlvYkjp/NiI6N/52Tr8mDE0vhwDTISYQKzaH5F+Bdr9CuL5PJDKc4S0nr9YLN00PwXjoTJ4WSrC8n0LZ350K/jkwmM5wrGVf46dhPHL15lKpOVZnefDo1XWs+9/mEWk3itOmkhoSQ7F6ecXUG4Vy1ItveCaDiw7bASIgA18qgNH3wuWcgF9OXyYrB5fTLvL/rfT4/8DllrMvw69u/Muv1WYWW6KhU+Wx4fzTOv0wipkxFvH9bQ7tOtxcfa/Ph+CKYVQv+HAculaDfduizUU50jEBMTAxVq1Zl0KBBVK9enZYtW6JSqWjWrBl3Ni5MTk7Gx8cHgJCQEDp27EiLFi3w8fFhzpw5TJs2jdq1a9OgQQNSU1MBaNasGSNGjKBWrVrUqFGD48ePo9frqVixIklJSQDo9Xr8/Pzufi4rmTR6DSuiVtBuQzs2XdpEn2p92NppK50qdiqSROdWYhq/9RhC5UWTSfcoh+8ff1BPTnQMTm5LZIUlV5PLjBMz6LK5C5EpkXxZ/0tWt1n9QomO+vp1Yt59j9SQEP6q3JgPXv2QHp0b8ceHrz2Y6Og0sPdHWNAEjs59wXdjpCM7coECWWmRq8ll/pn5rIhcgZWpFV83+JouFbugVBTe3jbnL8RyfujHVIk/z6XGbWg1+0fMLcwKGovTq2D/VMiIBe8G0Hkh+DYptGuXJt9tiSTqRmahnrOapx3ftnvyws3o6GhWr17NokWL6NatG+vXr3/s8REREZw6dYq8vDz8/PyYNGkSp06dYtSoUSxfvpyRI0cCkJuby+nTp9m/fz8DBgwgIiKCXr16sWrVKkaOHMnu3bsJCAjA1dW1MN6uzACKu5T03j+Povn2cwIybpHQtjvNf/wChVnhFzswRk97byK3JbKSSgjB39f/ZlLoJBJyEujwSgdG1R2Fs6Xzk1/8GJnbt3Pj62/I08Hken1JrfMav3cLoLrnQ6q33YqCDR9Awhmo2R3q9Hmha4ORJjtygQJZSXd/g9HRryOj6o7CyaLw1sUIIVj/+z6cJ31FBVUmmSM+p92HfUCnhVMr4Z/JkH4NvAKh3Ux45XUo5DKzssLh6+tLrVq1AKhbty4xMTGPPb558+bY2tpia2uLvb097dq1A8Df358zZ87cPa5nz54ANGnShMzMTNLT0xkwYAAdOnRg5MiRBAcH079//yJ5T7KiFZsVy5TQoi8lfUduvpY1382h1qZgNGYWmP48m+ZtHlLBUacFTQ5YPF8JWmNWEu5N5LZE9ryuZV7jp+M/cSj+EJUcKzGp8STquNd5oXPqVSpu/fgj6b+v45KrLxPqvkvX1kF8/EZFzEzuG3XW6+DIHNgzHsztoNsKqNb+ha5/h1EmOzJZSRabGcuPx3/kYPxBKjlWYnKTydR2e9H9f/8rLUfNoglLeHPTfLQWVrgsDsazQR0IXwP/TITUK1CmFrw9FSq2kJOcp/A0vaZFxdzc/O7HSqUSlUqFiYkJer0egLy8vEcer1Ao7n6uUCjQarV3n7v/xleSJLy9vXF3d2fPnj0cP36cVatWFfr7kRWdXE0ui88uJiQyBBOFSZGWkr4jMvoGp0Z+RoPLYSRWrEngollYe7g/eGD8CdgyApwrwjtLiyweYye3JbKSJE+bx+KziwmOCMZMacZnQZ/Ro0oPTBQvliLkXbxI3MhPUF+9wppKr3O4cWfm96hLLW+HBw9OuQwbh0LsUajSFtrOAJvCGyWUkx2ZrJDk6/IJPhvM4rOLMVGY8GnQp/Ss0vOFG4z7HYlO4p8vf6TDmT/J8q1MreB5mCUfgrlDIfkiuNeAHr9C5bdLbZIjSVJHoA1gBywRQuwybESFz8fHhxMnTlCvXj3WrVv3XOdYs2YNzZs35+DBg9jb22NvX9DbPnDgQHr16kXv3r1RKgtvSqWs6NxfSrpdhXaMrDuySEpJ36HXC9b8ugv36T9QKzeN3N6DaDJuBNL9PzP52bB3AhybD9Zu0GRskcUke3ZyWyJ7lH2x+5h4fCLx2fG0qdCG0XVH42r1YkmGEIL033/n5vgJZCrNmfzaIOp2bsWWlpWxML3vZ0SvL6gO+9c3oDCFTguhZrdCv3eRkx2ZrBAciDvAT8d/IjYrltY+rRkTNKbQb0I0Oj2/bA3HbvoEOiREom/1NoH9X0XxRydIOgeuVaHbcqjSDhTGW3tEkqRgoC2QKISocc/jbwEzASWwWAgx8VHnEEJsBDZKkuQITAVKXbIzZswYunXrxsKFC2nTps1zncPCwoLatWuj0WgIDg6++3j79u3p37+/PO2khIhKiWLi8YmcSjxFdefq/NzsZ2q51SrSa97KULFm3BSa7VtLno09LouW4NGowYMHXtgB28ZAZjwEDoA3vy2VU9hKMrktkd0vLiuOiccn8k/cP7xi/wrBrYIJ8gh64fPqsrKI/+prcnbu5JRbJX57oz/f9G1CkM9DpvBnxMGmYXBlX8E0+/ZzwN7rhWN4KCGE0f6rW7eukMmM2c3sm2LknpGiRkgN0faPtuLIjSNFcp3rKTliwIT1YkdgUxFRtZpImPip0P/ymhDf2gkxq64QZ34XQqcrkmvfDwgTL/B7DTQB6gAR9zymBC4DFQAzIByoBvgDW+/753bP634G6jzNdR/WnkRFRRXRV8nwmjZtKkJDQx/6XGhoqGjUqFExR/R0SvP35FmlqFLEt4e+Ff4h/qLJb03EHxf/EDp90f+e7z56Qax4o6uIqlxFHO7RT2hSUh88KPOmEGt6F7RBc+oLce3oM1/nRdsSQ/2T25J/GXNbIkTp/r48jzxtnph7eq6ou6KuCFoZJJaeXSrUOnWhnDs3PFxENH1dnK1STYzp9In4+o9wkZOvefBAvV6IU6uE+LGsEOPLCBG6pOCxF/S49kQe2ZHJnoNGp2HFuRXMD5+PEIKPa39M3+p9i2Te/ObwG6z9ZS0jjizHyhR8OtlhrZoPlhUKhnz9u0IhVncrakKI/ZIk+dz3cD3gkhDiCoAkSb8BHYQQP1EwCvQfUsEE8onADiHEySIOuVSZOHEi8+bNk+fXGzGNXsNv539j3ul5qLQqelfrzZCAIdiaPWQfikKkUutYOPcPai/7mYD8bBTDP6HB0IH/Xa+h1xfs17X7O9Dmwetfw2sfg8nLUZFN9i+5LSlZ7p2B8pbPW4wOHI2HtccLn1fo9SQuCSZ5+gySLOxY0noUQ4Z2pKGfy4MHZ92CrSPhwnYo9xp0nAtOvi8cw5MYZbIjl56WGbPQhFDGHx3PlYwrNPduzmf1PsPLpvCHXnPytXy7KQL9mlV8HrUNC2cl3g3iMXP1gqa/QM0eoDTKX+Hn4QXE3vN5HFD/MccPB94E7CVJ8hNCzH/YQZIkDQYGA5QrV66QQi0Z9u3b99DHx40bx7hx44o3GNlTO3zjMJOO3y4l7dmQT+sVbSnpO6Ji09j6+UTePrEVlZMb5UMWYhdw354aiecKChDEHgOfxgVVHp1fKfLYZIYltyUl283sm0wKncTf1//Gx86HhS0W8qrnq4Vybm1qKhdHjkE6foTDZfy5PmAkc98JwtbiIZuARm6EraNAnQMtJ0CDD4uto9Yo75RECSjvKHv5JKuSmRo2lW1XtuFl48Wc1+fQ1LtpkVwrIj6DT1Yc4/2/Z1Ez9jK2ZVV4trBC8eY0CHj3pe9FFULMAmY9xXELgYUAgYGBoqjjksme1/2lpGe/PpumZZsWWSnpO/R6wYrtJzCf9D3tkqLJa/Imdab9hNLG5t+DNHlwYCocnAHmttBxHgT0LLUFUGSy0kCj07AsahkLwhcgSRIj6oygb7W+mCofkog8h4zDR7gycjSK7CxW13uHlp8PZVDVh4wU5abC9rEQsQ48a0OnBeBauVBieFpGmezIZMZEq9ey5sIa5pyaQ74un8E1BzPQfyCWJpaFfi29XhB86CrHNqxk2tG1KFN1uNYVOH/8BVLdvmBi/uSTlEzxgPc9n5e9/dgLk0eKZcbsTinpZZHLUCqUxVJK+o7EzDxmT1lF683zsdGpsfv6W6q82/2/CdbV/bBlJKReLhhNbjUBrB8yPUUmkxmNIzeO8OOxH4nJjOHNcm/yadCnlLEpUyjnFlotFybPQLc8mCQbF44O/oFxg9/GweohbdbFXbB5OOQmQ/MvodEoKKRk61nIyY5M9hjhSeFMODqBc6nneLXMq3xR/wt87H2K5FpJWfnMXbmaFmfm0eRIGkKnwHNEO2wH/gCmFkVyTSMSClSUJMmXgiSnB/BuYZxYHimWGSMhBNuvbmfaiWkk5hZPKel77T4TT9h3k+gRuZt8r3JUmjcLi0qV/j0gNxV2fQ2nV4KjD/TeUFAxqRR6GUrZy14OCTkJTA2bys6YnXjbejPvzXk08mpUaOfPu3GTU4OH43ApkoMV6uPz/Td8E/iQabZ5mbDrSzi5HNyqwXtroUxAocXxrORkRyZ7iPS8dGacnMH66PW4WboxtelUWpZvWWRTSk4c2YNq1w98fOk8CSfsUbg44b1gMeZVazz5xSWMJEmrgWaAiyRJccC3QoglkiR9BOykoDJbsBAi0oBhymRF5t5S0tWcq/Fz06IvJX2HSq1jxqp/qLhoCh1SY5DadaTW99+gsLw9Ui0EnFkLOz+HvIyCntgmn4KZVbHE96zkUvYyWUFRk5VRK5kXPg+90DOs1jD61+iPubLwZoNEb9xO5v++wVyj4c/2Q+j9zRBcbB5y/qv7YeMwyIyDhiOh+RcGn5UiJzsy2T30Qs/GSxuZfmI6Weos+lbry4e1PsTa1LpIrqeOO83VdV9RJ/UAsadcSIh2wPrVBnjNmIHSvnTuVSGE6PmIx7cD2wv7evI0tudjY2NDdna2ocMoVVLzUpl1chZ/RP+Bo4Uj37/2PR38OqCQimdfrKgbmSyeuJQee5dhoQC3yZNxbt/ungCvFiwgvrIXvAILChB4GH2HSwgwB1h+5wFJkpTAL0ALCoqdhEqStJmCxOen+14/QAiRePvjr26/TlaI5LakaIUmhDLh6AQuZ1ymWdlmfFbvM8rali2082vz8tk/5lvK7N7ELcey8ON4Rrau92DnrzoX/v4ejs0DpwrQ/08o97g6Q8VHTnZkstvOp55n/NHxhCeFU8etDl82+JJKjpWe/MLncSuK7F0/YHN5O64qa8LDamAen4rT+wNw++STB3colz03eRqbcdPpdKV+53WNXsOa82uYe3pusZaSvkOvFyzdd5HUn6cy6PJBNH6VqfLLTMzKly84QKeBI3Ng3yRQmEDrKRD0fokoaV9cpexf5sqOJcXL0JbcKyk3ialhU9l+dTteNl7Mfn02zbybFeo1roWf59LwkXgmXiOsbgtazByPh4vdgwfGhsLGIZByCeoNhjf/B2ZF00n8PIx3m3WZrJhkqbOYeHwi3bd2JzYrlgmNJhDyVkjRJDpZCYg/BiPmvQaX9rIy7S3iD/lhkZyD55TJuI8dKyc6L5mYmBiqVq3KoEGDqF69Oi1btkSlUtGsWTPCwsIASE5OxsfHB4CQkBA6duxIixYt8PHxYc6cOUybNo3atWvToEEDUlNTAWjWrBkjRoygVq1a1KhRg+PHj6PX66lYsSJJSUkA6PV6/Pz87n7+MEIIxo4dS40aNfD392fNmjUADBs2jM2bNwPQqVMnBgwYAEBwcDBffvklACtXrqRevXrUqlWLDz74AJ1OBxT09I4ePZqAgACOHDlSyF9R43L4xmHe2fwOk0InUdO1Jus7rGds0NhiS3QSM/MYOW0zbp8Po+3lg1i924saf6z9N9GJOwELm8Hu/4HfGzDsGNQfXCISncd4WCn7x+0PcKeUfVdJkoY87AAhxEIhRKAQItDV1bXwIi1EcltSutuSO7R6LSuiVtBuYzv+uvYXH9T8gI0dNhZqoiOEYMeMEFLe64FdehIxo76j18qZDyY62vyC0ZzglgVVG/tsgrenGFWiA8U8siMvApQZkzsLhKeGTSVFlUK3yt0YXns49uZFMH1Mq4Zj89Hvm4hOo2axti2ZOTVov28NJo4OlF21Cssa1Qv/urKnn8a2YxwknC3ci3v4Q+tHLhW4Kzo6mtWrV7No0SK6devG+vXrH3t8REQEp06dIi8vDz8/PyZNmsSpU6cYNWoUy5cvZ+TIkQDk5uZy+vRp9u/fz4ABA4iIiKBXr16sWrWKkSNHsnv3bgICAnjczdsff/zB6dOnCQ8PJzk5maCgIJo0aULjxo05cOAA7du3Jz4+nps3bwJw4MABevTowblz51izZg2HDh3C1NSUoUOHsmrVKvr06UNOTg7169fn559/fvqvZQkTmxXL1NCp7IndU6ylpO/197lbbJiyhAHH12JqbobX3F+we/12kYH8LPj7Bzi+EGzLQPdVUPWBQY+XwtOWspfbErktMbSTt04y/th4otOiaejVkC/qfUE5u8IdabyRkMr+4eMIOHuA614VqfrLDIKqPKQIQcJZ2DAEbkVArV7w1o9gYZzT75862ZEXAcpKkyvpV5hwbALHE45Tw7kGc16fQ3WXIko2Lv2NfvtnKFKj+Udfm0n6PozMukr5ncuwrFOHsrNmYuJiuFKuQi/Iy9FgaVs69+4pCdPYfH19qVWrFgB169YlJibmscc3b94cW1tbbG1tsbe3p127gnUX/v7+nDlz5u5xPXsWLI9q0qQJmZmZpKenM2DAADp06MDIkSMJDg6mf//+j73WwYMH6dmzJ0qlEnd3d5o2bUpoaCiNGzdmxowZREVFUa1aNdLS0rh58yZHjhxh1qxZLFu2jBMnThAUFASASqXCza2g0phSqaRLly7P86UyeveXkh5RZwR9qvUpllLSd+RpdEzccArHxbMYdj0UqWYAr8ycjmmZ26Vnz28r2Pci8wbUGwSvfw0WD5maUnIVSSl7uS2R2xJDSVYlM/3EdDZf3oyHtQczms3g9XKvF2rniRCCbRsPYP7T1/hnJnGjXU/enPA5SrP7SkXrtHBoBuybCJaO0PM3qNy60OIoCs8yshOCvAhQVsLlanJZcGYByyOXY2lqydcNvqZLxS4oi2LKRto1xM4vkM5v5Ybkwdfqsbj6vMbsQyvQHjuKQ7dueHz1JZKZYZOMIxsuczH0Ft2+CMLKrnQmPE/lKXpNi4q5+b+VapRKJSqVChMTE/R6PQB5eXmPPF6hUNz9XKFQoNVq7z53/x9CSZLw9vbG3d2dPXv2cPz4cVatWkVsbOzdm5whQ4YwZMhDZ/L8h5eXF+np6fz55580adKE1NRU1q5di42NDba2tggh6Nu3Lz/9dP+fArCwsCh1c+vvLyXdtkJbRtUdVWylpO+IupHJxLlbeW/nAryzk3D44AM8hn+EZGJSkNxsHwvnt4JbdXhnGXgHFWt8xaTIStk/FbktAeS2pDDo9DrWXlzL7JOzUelUDPQfyCD/QViZFm51xFuZKtZ+PZMmf60i38Iai5lzeaNVswcPTI4uGM2JD4PqnaDNNLByKtRYisJTJzvFtQhQJisKQgj2XN/DpNBJ3My5SYdXOjCq7iicLZ0L/2IaFRycgf7gdNR6mKXpxgGX7nxbVo3djG/QZWfj8d13OHbvVvjXfkZn9sZy6q/r+Df1wtK2+Df6Kg4ltRqbj48PJ06coF69eqxbt+65zrFmzRqaN2/OwYMHsbe3x/52hb+BAwfSq1cvevfujVKpxNvbm9OnTz/0HI0bN2bBggX07duX1NRU9u/fz5QpUwBo0KABM2bMYM+ePaSkpNC1a1e6du0KwBtvvEGHDh0YNWoUbm5upKamkpWVRfk7a0VKCSEExxKOMefUHMKTwou9lPQder0g+OAVwueFMPrMJkxsbSkfvATrV18FvR6OL4Ld34FeA298C68NN8jmfoWtOEvZy22J3JYUp3v3+WtQpgFf1P8CX3vfQr/O9iMXSfzmG96MDSeteh3qzp+Juet9s030eji+oGBtn6kldA2GGiVnRO1F1+w8bBHg4+rM3VkEaC9Jkp8QYv79B8gVT2SFLTYrlp+O/cSB+ANUdKzIssbLqONep/AvJASc34p+x+coMmPZrnuVWco+9H4riIWHN5D+9VJMKvrhFbIU84oVC//6z+jyyUQOrI2mQi1XGnWvVKxrCYpTSZh68jBjxoyhW7duLFy4kDZt2jzXOSwsLKhduzYajYbg4OC7j7dv357+/fs/cdoJFCwYPnLkCAEBAUiSxOTJk/Hw8AAKbl527dqFn58f5cuXJzU1lcaNGwNQrVo1xo8fT8uWLdHr9ZiamvLLL7+UqhuU0IRQfjn9CyduncDdyp3vXvuOjn4di62U9B2JWXl8ufIoddfNZ2h8OGYNXqX81MkF02NvRcKWERAXChWaQdvpBWVhS4niLGUvtyVyW1IcklXJzD41mz+i/8DN0o0pTafQqnyrQv8bnZqjZs4vG2iyeiaB+ZkoPhzOq8OHICnua7/SrsGmYRBzACq2gvazwNajUGMpckKIp/4H+AAR93zelYJ1Onc+7w3MeZZzPu5f3bp1hUz2vPK0eWLu6bmizvI6ot7KemJZxDKh0WmK5mKJF4R+WQchvrUT0d/WEN0/nyw+/+OMuHUuWlzp0lVEVa4ibnz7rdCpVEVz/WcUfzFNzBu2V6ybFCY0+dpnei0QJgrpd7w4/z2sPYmKinqm916SNG3aVISGhj70udDQUNGoUaNijujplITvSejNUNH/z/6iRkgN8fqa18Wv534V+dp8g8SyOypBdBmxSOyp01BEVq0mkhYsFHqdTgh1rhB//U+I75yEmOQrxOnfhNDrDRLjo5S0tgRoByz08/N74L2UhJ/b51VS2xIhStb3RaVRiUVnFon6q+qLWstqiamhU0W2OrtIrrXr7A3xTY9Pxdkq1cSJ15qIrJOnHjxIrxciLESICZ5CTPAS4sRyo2tD7vW49uRFR3aKZBFgSR0qlhmPg/EH+fHYj8RmxfKWz1uMCRyDu7V74V8oLxP2T0YcmUcO5kzV9CHSqxvfdqiJ98n9JLz7PzAxwWvWTOxatiz86z+H1Js5bJ93BltnC9oMrYmJWemd7yx70MSJE5k3bx6rVq0ydCglzqnEU/xy+heO3TyGi6UL4+qNo2ulroW6S/nTytPomLA1iuxVK/g2ajsmrq6UX7QSqzq14co+2DIS0q5CrfegxQ9gXQRTdl8yooSO7BQVuS0pHHqhZ8fVHcw8OZObOTdp7t2cT+p+go+9T6FfK0OlYfLqQ1QNmU6PxIuIpm8QMOVHlHb3FSjJvAlbPoboXeDTGDrOBYeSO9vqRZOdIlkEKDcosueVkJPA5NDJ/HXtL3zsfFjYYiGver5a+BcSAs6sRbfrK5Q5iazRNmOpRR+GdK3P1xUduPXDeG5s2oRl3bp4TZmMqadn4cfwHHLS89ky+zRKEwXthgdgYVPy5+w/ycvaebJv376HPj5u3DjGjRtXvMGUcOFJ4cw9PZfDNw7jZOHE2MCxdKvcDQsTC4PEc+5mJp8vPUDHnUuof+scVq+/TtkfJ6A01RUsHg5fXTBVrc9mqNDUIDHKSg+5LSk6pxJPMSV0CmeTz1LVqSoTGk0gyKNoiob8czGJZbPW8v6BZdjr83H99luce3T/7/Q4ISBiPWwbXbCHTuvJEDQI7p/aVsI8S+lpeRGgzGhp9BpWRq1kXvg8hBB8XPtj+lbvWzTlXm+Go982FkXcMaLEK/xPN5yghi1Z/7ofyksXiOkyEHVsLC5Dh+Iy9MOCKkhGQK3SsmVOOPk5WjqNroOdi6WhQyoWcueJ7HlFJEfwy+lfOBh/EEdzR0bXHU23yt0KvRLS09LrBUsPx7Bl+TbGhq3CUZ2D+1df4fhuT6Qza2DnF5CfCY3HQJMxBQuJZYVGvjeRFZbYrFimn5jOX9f+ws3SjfENx9PulXZFst4vO1/Lj5sjMF2xmNEX9yCVK0+F2TOxqHzfxuk5ybDtE4jaBGWDoON8cCkdP+vPUo1NXgT4EtPqtfx07CcO3zhMLbda1POoR5BHEGVtyxo6NEITQplwdAKXMy7TzLsZ4+qNw8vmcZtlP6fcVNgzHnFiKRnY8pNmEEmvdGVKuxr4uliTumwZiT9Pw8TJiXIhS7GuV6/wY3hOOq2eHQvOknYjhzYf1cS1XPHs3i6TlUSRKZHMOz2Pf+L+wcHcgZF1RtKzSk+DJTlQUIRg7JrTeG1dzYQLf2Hq7U25GUuwcDeHFR3h6j9Qth60mwnu1QwWZ2km35vIXlSmOpNFZxax6twqTBQmDA0YSt/qfYusbTlyOYWfQvbS9+8lVEuNwaZzZ7y++hKF1X3XO7+9YNqaKv12tcaPQWkcHbWFwSjfidx7Ylw0eg3j9o9j17Vd1C9Tn8M3DrP1ylYAPK09CfIIon6Z+gR5BOFhXXwVOpJVyfwc9jNbr2zFy8aL2a/Pppl3s8K/kF4HJ0LQ/f0D5GWwXNuCdXa9+aRbPd6o6o42JYXYIaPJ+Wc/Nm+8QZnxP2Di6Fj4cTwnIQR7Vpwj7nwab/StSrlq8tx9mexhzqeeZ+7pueyN3YudmR0f1/6Yd6u+i7WptUHj+vvcLX5cfoDBB5dRM+kSdu3a4vH1lyjDl8Afk0FpBm1+hroDSvx0E5msNNLoNfx+4Xfmhc8jIz+DDn4dGF57eJHtw6VS65i88zznf9/KN6fXYGUi4Tl1KvZt76vGp0qHPz+H8F/B3R96bwSPGkUS07PS6wUn/4yhakNPrO1fbF2kUSY7cu+J8cjX5TNm3xj2xe1jbOBY+lTvgxCCy+mXOZ5wnNCEUPbF7WPT5U0AlLMtR5BHEPU86lGvTD1cLF2ecIVnp9PrWHNhDXNOzSFPl8fgmoMZ6D8QS5MimLJx/Rj6bWNQ3DpDqL4aP4ovaPXG6/zR2BdzEyU5R44Q/+mn6DMycf/6KxzffdfoSjgf3XiFi8duUb99Baq8WsbQ4RQ7ufNE9iQX0y4y7/Q8dl/fja2ZLcNqDaNX1V7YmNkYNK48jY4ft58jcuNOfjq1Bls0lJkwAfsgL6SVrSExCqq2L5hXb/fy/W7LZMZOCMH+uP1MDZtKTGYM9TzqMTZoLFWcqhTZNU9cS2Pc6jDe/GcN31w9hFm1anhPn4bZ/aW6L+8tKCmdlQBNxkKTT8HEODYWV6u0/BUcSczZFEzNTQh4w/vJL3oMo0x2ZMZBpVUxcu9IDt84zFf1v6J7le5Awc7Jfo5++Dn68W7Vd9ELPdFp0RxPOM7xhOPsitnF+uj1APja+96d8hbkEYSTxYvttHsm6Qzjj46/u8nWl/W/LJKKJWQlIP76BunMGpJw5gf1cBQ1OrOgTVXK2FsiNBoSp80iZdEizHx9KbdoERZViq7xel5n98Vxcuc1qjf2pG7r0rknwZPInSeyR7mUdol54fPYdW0XNqY2fBjwIb2q9cLOzO7JLy5i525mMmpVGK/98zvjo/dhVrEiZSd9j/nVFRC8BOw8ocdqqPK2oUN9acgdJ7JncT71PFNDp3Is4Rg+dj7Mfn02Tcs2LbIO0Xytjul/RbNt2xG+Ovkr5VLjcOrbF7fRnyCZ3ZPEqHPgr28gdDG4VIL3/4KydYskpueRmaxi29wzpCXk0rRnJWo0ffHlEkaZ7MgNiuHlanL5aM9HhCWE8f1r39OpYqdHHquQFFR2qkxlp8r0rtYbnV7H+dTzd5OfLZe3sObCGgD8HPwKRn086hHoEYi9uf1TxZOel87MUzNZf3E9rpauRbbJFlo1HF+Abu9E9Jp8Fmg7sNulN190qEs934JETR0Xx43RY1CFh+PwTlfcP//8wfmvRuDK6ST2r7mIT00XmvQovZuGymTP6kr6FeaHz+fPmD+xMrVicM3B9KnW56nbo6Kk1wtCDscQvP4wn4WuxC85Boce3XHvWBPFjh4FvbD1h8DrX4K5vPauOMkdJ7KnkZibyJxTc9h4aSN25naMqzeObpW7YaoouuqnZ+MyGP37abxD9/FLxAbMrSzxnDcX2+bN/3vg9aMFFRvTYqDBMHjja6MqZHIjOp0dC84i9IJ2wwPwrvpiHeR3GGWyIzcohpWlzuLD3R8SkRzBT41/ok2FZ9txWalQUt2lOtVdqtO/Rn80eg1RKVGEJoRy/OZx/oj+g1/P/4qERGWnygVrfjzqU8e9DrZm//3jrRd6Nl7ayPQT08lSZ9G7Wm+G1hpaNHPoL+9Bt+1TlKnR/KOvzXRlf7q1bcb6euVQKgoShczt27n5zbcgSXhNn4Zd69aFH0chuHk5g11LInH3saPlwOoolPI8fmM1bdq0uzuSDxw4kI4dO9K6dWsaNWrE4cOH8fLyYtOmTVhaWtKsWTPq16/P3r17SU9PZ8mSJTRu3Jjp06dz9uxZgoODOXv2LD179uT48eNY3ZeE+/j40K1bN3bs2IGlpSW//vor7u7u1KxZk4sXL2JqakpmZiYBAQF3Py9NrmZcZX74fHZc3YGFiQXv+79P32p9cbBwMHRoQEERgjG/n0G7bw+zw3/H0kSB549fY5e/FTZOL5hT32MVeBlPL6zMuMjtieHkanJZFrWMpRFL0eg19KnWh0E1BxVpJ4pGp2fOnkss/iuSkREbaXTlOFaBgXhOnYKpxz1rqDV5sHcCHJ4NDt7Qbyv4NCqyuJ7HucM32LfqAnYulrQZWhMH98LrRDbKZEdmOBn5GXzw1wdcSLvAlKZTaFG+xQuf01RhSoBrAAGuAQz0H4hGp+Fs8tm7a37WnF/DiqgVKCQF1ZyqEVSmYM2PrZktU0KnEJ4UTh23OnzZ4EsqOVZ68gWfVdo1xM4vkc5v4QYefKcZg0dQR5a3qIyjdcHQrz43l4QffyRj3XosAwLw/PlnzMoWQcW3QpCWkMO2ueHYOJrTZlhNTOVNQ5/KpOOTOJ96vlDPWcWpCp/V++yRz584cYKlS5dy7NgxhBDUr1+fpk2bEh0dzerVq1m0aBHdunVj/fr19OrVCwCtVsvx48fZvn073333Hbt372bEiBE0a9aMDRs2MGHCBBYsWPDAjckd9vb2nD17luXLlzNy5Ei2bt1Ks2bN2LZtGx07duS3336jc+fOperG5HrmdeaHz2fb1W2YK83pV6Mf/ar3e+FptYVpz/lbfPHbCd4J/YO3Lx/CokYNvPoGYXZ2bEGRlBbfQ4OhoCw935fSyhBtCcjtiaHohZ6tV7Yy8+RMEnMTaVG+BaPqjMLb7sXWmTzJ+YRMRq8NRxV1jsVnf8Mh7RYuw4YVbHmhvOfv/o1TBaM5Seehbj9oOd6oRoX1esGRPy5xencsZas40mpQDSysC/fnxSiTHXkam2GkqFIY/NdgYjJimNl8Jk3KNimS65gqTanjXoc67nUYEjCEfF0+4Ynhd5OfFVErWBqxFAAnCyfGNxxP+1faF/40LI0KDs1Ef2Aaah3M0nTjdNn3+LJDbap7/tsTk3f+PPGfjEZ99SrOH3yA60fDkIy04c7JyGfL7HAUCol2w2thaWMciw1lD3fw4EE6deqEtXXBSGXnzp05cOAAvr6+1KpVC4C6desSExNz9zWdO3d+4HGFQkFISAg1a9bkgw8+oGHDho+8Zs+ePe/+f9SoUUBBD/DkyZPp2LEjS5cuZdGiRYX8Tg0jNiuWBeEL2HplK6YKU3pX7U3/Gv1xtjSeioR3ihD8vSuU8ad/xSslDqdu7XDzDEU6MQFeeQPaTgNHH0OH+tIz9nsTuT0pfqEJoUwJncK51HPUcK7BlCZTqONep0ivqdXpWXjgCjN2XaTz9SP0Ob0JUydHPJcuxbr+PVte6DRw4GfYPwWsXeG99VDxzSKN7VmpVVp2BUdy7WwK/s3K0ugdvyKZiWKUyY48ja34JeYmMmjXIG5k32DOG3N41fPVYru2udKcemUKqrdBwVDw6aTTxGbG8pbvW4U/BCwEnN+Gbsc4lJmxbNM1YKFFfwZ2bMzYAM+7SZUQgrRVv5I4eTJKe3vKBS/B+tXi+7o8K3Welq1zwlFla+j0SW3sXY1nHq4hPe0NypN6TYuTufm/ZTaVSiUqleqB55RKJVqt9u7j0dHR2NjYcOPGjbuPtWrVilu3bhEYGMjixYsB/tNpcOfjhg0bEhMTw759+9DpdNSoYRylR59XfHY8C88sZPOlzSgVSt6t+i4DagwokuqQL+LczUxG/HYK79B9zI/YiJmVOZ4fNMI2czFkOkKXJVCjC8jr7YzC096bGFNbAnJ7UhSuZV5jWtg09sTuwcPag58a/8Tbvm8Xyaag97qclM3oteFcunyDaZc38cqFE9g0bUqZiT/9d8uLxPOw4QO4eRr8u8Hbk8HSeLbEgKIpRPAo8kR+GTezb9L/z/4k5CQw7815xZroPIyVqRWveb5G9yrdCz/RSY5Gv6IzrHmPyxnQW/s1UQ1n8tuYrnSo5XW3sdampRE37CNujR+P1asN8N200agTHZ1Oz86FEaTE5/DWoBq4lTd8NSljIYTYIoQYbG9v+MXn92vcuDEbN24kNzeXnJwcNmzYQOPGjZ/5PBkZGXz88cfs37+flJQU1q1bB8DOnTs5ffr03RsTgDVr1tz9/6v3/Ez36dOHd999l/79+7/guzKcm9k3+e7Id7T9oy1bL2+le5Xu7Oi8g0+DPjWqREcIQfDBq3SfuYeufwUz5uRv2FUqS4W2OdhmrIVaPeGjUPDvWuISnZuXM0i6nmXoMF5KcntS9DLyM5h0fBIdN3bk6M2jfFz7Y7Z03ELbCm2LNNHR6wVLDl7l7ZkHMD13ll+Pz+GVK+G4ffYZZefP+zfR0evg0CxY0AQyYqHbcuiyyOgSnRvR6fw+MYyc9HzafRxQpIkOGOnIjqz4xGbFMnDnQLLUWSxsuZAA1wBDh1Q08rPgn8noj84jV2/GVE0f4v3e4/t2/vi6/LfYQc6x49z49FN0qam4f/E5jr17G3UlMyEE+1ac53pUKs17V6F8DeOZoiN7vDp16tCvXz/q1SsY1Rw4cCCOz7Eh7ahRoxg2bBiVKlViyZIlNG/enCZNmuDm9uCGdWlpadSsWRNzc3NWr1599/H33nuPr7766u60lJIkISeBxWcXsz56PRISXSt1ZaD/QNyt3Q0d2gOSsvIZ83s410PDmRu+Gue0BFxeL4eLy34kez9ouxV8n/0G1RhcOpHI7qVRuPva0fGT2kbdbpZGcntSdDQ6Db9d+I354fPJ1mTTuWJnhtUaViydKNdTchmzLpywK8l8lnacJoc2YOrpide8X7H09//3wNQrsHEoXD8CVdpC2xlg41rk8T2rqEM3+OfXoilE8CiSEKLIL/Ks7pl2Mig6OtrQ4ZRaVzOuMnDXQPJ1+SxssZBqztUMHVLhEwLO/o5251eY5NxijbYZq+36MaJ9Q5pX+W/DLbRakufOJXnefMzKlcNz2s9YVq9uoMCf3rHNVwjbHkO9dr4EtfEt8utJknRCCBFY5BcqZIGBgSIsLOw/j507d46qVasaKKLi5+PjQ1hYGC4uD/6BXrduHZs2bWLFihUGiOxfz/I9ic+OZ8nZJWy8tBGBoJNfJwb5D6KMjXFusLnn/C3Grg2n0bn9DD67GTMbczwbpGDtlAmNP4FGn4CphaHDfGZCCML/juXQ+kt4+Nrz9lD/p1ovKLclJVtpa08eRgjBnut7mHZiGtezrvNqmVcZEzSmaIolPeTaq45d58ft53DKy2L65Q3YRJ7C7u238fj+O5Q2NncOhLAlsOtrUJgWTFmr2d3oRoWLuhDB49oToxzZkdfsFL3otGgG7RqEQBDcKrhYfnGL3c0z6LaNRRl3lCjxCj+K4TR9823WNPLB3OS/Fco0N24QP2YsqpMnse/YEY+vv0JhXQTlrQtZxP54wrbHUK2RJ4Fv+xg6HFkJNXz4cHbs2MH27dsNHcpTuZ55nUVnF7H18lYkSaKjX0fe938fLxvjrJB4pwjB+n/O8/X5DQRcOYl1BUs8a17CpGI9aDcT3IxvU+KnodcLDq6N5uy+OF6p7cqb/athIleAfKmVtPbkUSJTIpkSOoUTt07wiv0rzH1jLo28GhXLiOWNdBWfrT/Dgehk+ihv8u6BpaDKxWP8D9h36fJvDBlxsOkjuLIXXnkd2s8Be+NrB4urEMGjGGWyIytaUSlRfPDXB5gpzFjUahEV7CsYOqTClZuK2DMBwoLJxIafNIPQ1OjJzDbVcbd7sNc0c9cubn71Neh0eE6ZjH27dgYI+tldDU9i/+oLlPd3pmlPedNQ2ZPdW4XpXrNnzy7eQJ7TlfQrLDy7kB1Xd2CqMKV7le70q94PD2uPJ7/YQM4nZPLx6lNI5yJZdvY3rDOScaudg5N/DlLLaVCnLyhK5vJZjVrHX0siuRqeTK03vXmtsx+SovS2Q8Zeja24lfT25FESchKYdXIWW65swcnCia8bfE3nip0xURT9LbMQgo2n4/lmUyRCoyVYfZwy29diVtEPr2khmFeseOdACP8NdnwGei20mQaBA4xuNAcgI0nF9nnFU4jgUeRk5yVzJukMQ/4ago2ZDUtaLinyOvBPS6PREBcXR15e3vOfRAhQ5yDy0sG5FTmtOqBS2PCulTnmJgpS46+SGn/v4QJ9RgZ6pRJp6hSUjo7cMDHhxrlzL/x+ippOqyc3W029vk5Y2Zly4eKFQr+GhYUFZcuWLdX7I8hKhgupF1h4ZiF/XfsLCxML+lTrQ9/qfY2q6MD9hBCEHI5h4vYoul85QI8zWzCzFni9nohl07bw1kSwNd4k7UlyM9Vs+yWcpOtZNO5eiZrNi/8GprjJs05Kt9isWJZHLmfDpQ0IIRhQYwAD/Qc+sNl5UUnPVfPlxgi2nbnJG446Pj27En3kWRy6dcP983EoLG9XWM1OhK2j4PxWKPcqdJwLTsbZaX0jOp0dC84i9IJ2HwfgXcUwe5vJyc5L5MStEwzdPRRnS2cWt1yMp42noUO6Ky4uDltbW3x8fJ5vhCI/G5ERh6Q1IVuU5ZbkgrudHU7WZg89nz4vD01sLHpbW0x8fTFxc0MqIb2rWo2OtIRcFE4SDh5WKItgKFgIQUpKCnFxcfj6Fv06IJnsYSKTI1lwZgF7Y/dibWrNQP+B9K7WG0cL46osdL87RQhOn73KzPO/U/5qJLZlVZR50xpl51VQqZWhQ3whaQk5bJ0TTm6GmtZD/PENML5F0DLZ04pMiSQkIoRd13ahkBS0rdCWIQFDinVa7MHoZEb/fpqUbDVTnRLw/30BCIHX9GnYtW7974FRmwoSnfzsgs1BGwwFhXFOGzVEIYJHMcpkRx4qLnxHbx7l4z0f42HtwaIWi4yuSlFeXt7zJTo6DSLzBpIqFS1Kbgo3lFaOlLezwOQhSYAQAl1aGpqbN5GUSszK+6C0tSmkd1H0dDo9GYkqJMDezbJIEh0o2C/B2dmZpKSkIjm/TPY4pxNPM//MfA7FH8LWzJahAUN5t+q7hV+KvgjsOX+Lsb+fwSf2HKtOL8ckJxv3wEwc3uuL9PqXYF5y2puHuRGdzvZ5Z1AoJTp+Ugd3X7nMvazkEUJw5OYRgiOCOXbzGDamNvSt3pf3qrxXrPdHeRodk/48z9JDMdSwg+C0nSj++AvzWrXwnDIZM+/bs29UabB9LJz9HcrUgk4LjHadn14vOPzHJcJ3x+Jd1ZGWAwu3EMHzMMpkRx4qLlz74/Yzau8oytuXZ1GLRUa1e/i9ninREXrISUZk3UQIQZKwJ9vUhTIO1lg+YnGs0GrR3LiBLjMThY0NZl5eSCVoipZeL8hIVKHXCRzcrTAxLdrenNKyBkjuPCkZhBDkanNJUaXw4Y4PcTR3ZESdEfSo3AMbM+NPEPI0On7afo4Vh68yLnY7jU/uw8xWi1cPFyzeXw2etQ0d4guLDrvF7pAo7JwtaftRgLxxsazE0eq17IrZxdLIpZxPPY+rpSuj6o7inUrvFNt0tTsib2Qw8rfTRCdm86lrJm9uXoAuJQXXER/jPGgQksntW/Tov2DzcMhJgmZfFFRuVBrnvYtapWXXkkiuRRimEMGjGGWyIys8f1/7mzH7C8okLnhzAQ4WDoYO6cXlZRZMWdPlkyUsSVK44uxgi6+l6SNv0HU5OWji4hBaLabuHihdnEvUzbwQgswkFVq1DntXS0zNjXPY2hjJnSePFhISQsuWLfH0NNyUViEEOZocklRJ5Gpy0eg1jAkcwzuV3sHK1HDTHp5FaEwqn/9xloxrsayPXIhFXBL2FdR4jBmGounHoCzZf2qFEJzadZ0jGy5Txs+etz+safCeWpnxMYb25FFyNblsuLSBFVEriM+Ox9fel+9f+542FdpgpnxymfTCpNMLFu6/wrS/LuBmLrFeHMdq0VqUvr54z5mDpX+NggPzs2Dnl3ByGbhWhZ6/gWetYo31WfynEMG7lanRxHiqwpXsFlj2WDuu7uDzA59Tw6UGc9+ci51ZCZ9uoM1HZMYj5WWgFibcxB0LG0d8bC1QPqICkBACbVIS2sREJDMzzH19UViVjBuoO4QQZKXkoc7TYutkgbmVfJMhe3E6nY6QkBBq1KhhkJsTIQTZmmyScpNQaVWYKEzwsPZAspJoXq15scfzPNJz1UzccZ7fQmMZkHmI7oc3IrQ6PNp74/D5EnAsb+gQX5hep2f/mmgi98dTMdCN1/tWLfJRZVnJY+j25FH0Qs/c03NZfX416fnp1HKtxWdBn9HUuykKqfhHHGJTcxm9NpzjMan0dsmnz95gtJcv4/juu7iNHfNvEYKYg7DxQ0iPhYYjCkZ0jHgPLmMpRPAohh9bkhWJjZc28tn+z6jlVosFLRaU7ERHr4esm4jEc4i8TBKEIwnmvpRxd8fD3vKRiY5eo0F9NQZtYiJKewfMX3nlmROdmJgYqlatyqBBg6hevTotW7ZEpVKxaNEigoKCCAgIoEuXLuTm5gLQr18/PvzwQxo0aECFChXYt28fAwYMoGrVqvTr1+/ueXft2sWrr75KnTp1eOedd8jOzn5kDLkZavJyNFjZm2NpW7w9ULKiN23aNGrUqEGNGjWYMWPGI3/mAJo1a8Znn31GvXr1qFSpEgcOHABg+vTpDBgwAICzZ89So0aNuz+T9/Lx8eGzzz6jTp06rF69mrCwMN577z1q1ap19xpFTQhBZn4mVzKucD3zOlqhpYxNGSo6VsTZsmSMuAoh2HQ6njen/cPJo0fYduFH3tmzAVMrPb4/j8Jh0l+lItFR52nZPv8skfvjqdOqHC0GVJcTHSP3srUnj6LWqbmZfZNbObeYFz6PWm61WN56OSveXkHzcs2LPdERQrDuRBytZx7gfHw6Sy3O8+6ybxGZGXgvWojHN18XJDqpV2HdAAhpA5ISBvwJLb436kQn6tANNs04hYW1KV0/CzS6RAfkkZ1Sae2Ftfxw9AdeLfMqM1+fiaVJCZ1XLQTkZaDPiEehV5MhrElRuuDqYIuHxeNHN3SZmWji4xFCYOrlhdLB4blvoqKjo1m9ejWLFi2iW7durF+/ns6dOzNoUMGsqK+++oolS5YwfPhwANLS0jhy5AibN2+mffv2HDp0iMWLFxMUFMTp06cpW7Ys48ePZ/fu3VhbWzNp0iSmTZvGN99888C1VVlqcjLysbA2xdpeTnSKUsKPP5J/7nyhntO8ahU8vvjikc+fOHGCpUuXcuzYMYQQ1K9fn6ZNmz70Z65Xr14AaLVajh8/zvbt2/nuu+/YvXs3I0aMoFmzZmzYsIEJEyawYMECrB6R2Ds7O3Py5EkAFi9ezNSpUwkMLPpN7IUQZKozSVIlka/Nx0xphqeNJ/bm9gbpYX1eMck5fLUxgshLV5iS+xvlD55Dq1Lg3KIqLj/MR+FgXMVfnldORj7bfjlDcmyW0U1JMZSnXf9niLYEXq725FFUWhXJqmQy8zORJAlLU0s2ddhEBQfDlWZOy1HzxYaz7IhIoIWjjjGn1qA/fRLbFi3w+P47TBwdIScZ9k+B0CWgMIHGYwrW5pgZ7+bm9xciaDWohtHOPJGTnVJmRdQKJodOpmnZpvzc7GfMleaGDunJ9LqCf+rcgs2xdBrQaxD52UjqLNTClJuUwcbOAV8bcxSPSVqEXo82IQFtaioKCwvMvL1RmL/Y18DX15datWoBULduXWJiYoiIiOCrr74iPT2d7OxsWrX6t5Rsu3btkCQJf39/3N3d8ff3B6B69erExMQQFxdHVFQUDRs2BECtVvPqq68+cN38XA1ZqXmYWZhg62xRInq8Zc/m4MGDdOrUCWvrgj9onTt35sCBAw/9mbujc+fODzyuUCgICQmhZs2afPDBB3d/th6me/fuRfJeHkUIQUZ+BkmqJNQ6NeZKc7xsvbA3sy9RP9NqrZ6F+y+zaE8Eg/VbmRixj6xLZijcbPGZMxnL194wdIiFJvVGDlvmnCYvR8vbQ2vi42+8+xkVJ2Nf//cytCcPc2fdX7IqmRxNDgpJgbOlM84WzlxKvGTQROefi0mM/T2ctJx8fnaIp8bahQCU+ekn7Dt2QNKoYP9UODgDNDlQuzc0+xzsyhgs5qdhrIUIHsUokx25etLzWXx2MTNPzqRF+RZMajwJU0NX69DmQ/YtyLpV8P/sBHSZCWjSb6LLvAnZtzDJScQ0PwVaroZkzX9erkNJonBGZ+lMWXtLTJ/wi6TPy0MTF4c+Lw8TZ2dM3N0LZe8c83uSJaVSiUqlol+/fmzcuJGAgABCQkLYt2/fA8crFIr/vFahUKDValEqlbRo0YLVq1c/8pqafB2ZyXmYmCmxc7UsUTeFJdWTek2L08N+5u5/TqlUotVq7z4eHR2NjY0NN27cuPtYq1atuHXrFoGBgSxevBjg7o1QUdMLPen56SSrktHoNFiYWOBt642tmW2J+3k+fjWVr/84Rd3UbexK20TmUQVZKjOc3+2Ey2ffvnCHijGJv5DG9vlnMTFV0OmT2riVL8FToA3EmNoSKB3tycPcGS1OViWTp83DRGGCu7U7juaOKA2894xKrWPijnMsO3KNmnYQkrQN6Y+9mNeti+ekSZiVcYeTy2HfT5B1Eyq/DW98a7TlpO9lzIUIHsUokx1j7z0xNkII5obPZX74fN72fZsJjSZgoijCb21+FvlpN8hJiScv7Qaa9JvosxKQsm9hkpuIeV4S1upkrHSZD7xUEhJZ2JEoHG7/q0YiDgRiwzXhhl4yQS+ZgtIEpUKJq6051uaPfy93985JSECSJMzKl0dpW7QlJLOysihTpgwajYZVq1bh5fX0v+wNGjRg2LBhXLp0CT8/P3JycoiPj6dSpUpAwaahGYm5KJQS9m6WKB6xJklW8jVu3Jh+/foxbtw4hBBs2LCBFStWsHDhwmc6T0ZGBh9//DH79+/no48+Yt26dXTt2pWdO3c+9nW2trZkZWW9yFt4gF7oSctLI0WVgkavwdLEkjJ2ZbAxtSlxSU56rpqJ28+RcXI9CxW/Y3Y6h9TL1ph5l8FnyjQsb/eWlxYXjiWwZ/k57F0taTs8ADvnEjoF+iVVGtuTh9HpdaTnp5OSl4JGpzG6KbER8RmM+O0Ul5NyGOeczuubF6BLT8d19Cc49++PdGkXzP8Oks5D2SDouhTKPzi7wxjdiE5jx/wIhBC0/ziAska4PudhjDLZkT09IQTTT0xnaeRSOvl14ttXvy2SHg2hVRO++EMqJWzFijzMgXv7MvOFCUk4kCQcSMKJDGVFss2cUZm7orF0RWvtjmTjjqm9G/bWlthbmuJgZUYVS1MaWJmSnXCNcp6ez3wzJHS6gr1zMjJQWFtjWrYsimLYO+eHH36gfv36uLq6Ur9+/Wdq4F1dXQkJCaFnz57k5+cDMH78eCpVqoT+9qahAnAowk1DZcahTp069OvXj3r16gEwcOBAHB0dn/k8o0aNYtiwYVSqVIklS5bQvHlzmjRpgpub22Nf169fP4YMGYKlpSVHjhzB0vL5b271Qk9qXiopqhS0ei1WplZ42nhibWpd4pKcggIEN9i+ZS1DtSuomBrLjROu5Gbb4DSgP64fD0dhYbwLhp+VEIITO65xbPMVvCo58NYH/nJp6RKoNLUnD6PVa0nNSyU1LxWdXoelqSUeVh5GM1qs0wvm/3OZ6X9dpIwF/KE9jOWSP1D6vUK5BfOxsM2B5e3g+mFw9oNuK6BqOzCC2J9G1KEb/PPrBexcLGkztCYO7iWnsq0khDB0DI8UGBgowsLCDB2G0dILPZOOT+LX87/SvXJ3vqj/RZH0aghVGjFzu+KbFcYh6zfJd6qM3soNyc4DE7syWDh5Ym3vioO1GQ5WZlibKZ+54Tl37hxVq1Z9ptfoc3NRx8YiNFpM3N0wcXExigbveen1gvRbuWg1ehzdLTF9wohWcXjY90WSpBNCCMOtQH1OD2tPnufnTvZfOr2uIMnJS0Gn12Ftao2rpStWplbP9fto6O9JTHIOC9duosXN+TTRnyEh0oPM8xJmPj6U+elHrGqX/M1B76XT6dn/6wWiDt2kUj13Xu9dFaVp8XSyyG2J7GmodWpSVCmk5achhMDWzBYXS5en2oeruL4vsam5jFpzmrBrafRzyePd3UvQxVzFqW8fXPt0QHHwJzi3GazdoNlnUKev0W4Mer//FCKo5kSrgdWNshDB49oTw99NyZ6LXuj5/sj3rI9eT99qfRkdOLpIbvRF6lWSFnbCS3WdDT5f0bHfGIMnFEIItMnJaG8lIpmaYF6h5O2dcz8hBJnJ924aKv9qPo4kSVWBEYAL8LcQYp6BQ3rp6PQ6UvJSSFWlohM6bMxs7iY5JZFaq+fXnQdwPDaZ8dIhMtMcuHKqEtrUbJz698N1xMelajQHCkpL71wYwfWoVOq2Lk/99hUM3r7LZHfcX1nN3tweFwsXzE2MZ42cEILfT8Tx3eZITISe5WYRuIasRHJ2ptwvP2Od9zcsaQxK84LCA69+BOY2hg77qf2nEEHzsjTqatyFCB5FvqMqgbR6Ld8c+oYtV7YwuOZgPqr1UdEkOrGh5Cx7B3NNPqsrz6BPz94G/0Oo12gKihDk5KC0t8fU0xNJWbL3fRBCkJWah1r1cmwaKklSMNAWSBRC1Ljn8beAmYASWCyEmPiocwghzgFDJElSAMsBOdkpJlq9lhRVCql5qeiFHlszW1wtXbE0LbnrO06eu8TVDd/RM387Qq8g/kZDsg9fway8E+Vnz8eqTh1Dh1joctLz2fpLOCnxOTTvVYVqjYxnI0jZy+txldUMXnTpPqk5aj7/4ww7I2/xlqOWUWGr0Z8Nx65VCzxaOqM8Ngg0KgjsD00/A5vHTwM0NhlJKrbNPUP6rZJTiOBR5GSnhNHoNYzbP45d13YxvPZwBtccXCTXEZEb0a4bRKrOnq3+c/mwa2uDJzq6rCw0cXEFe+d4eqJ0dDR4TIUhN1NNXrYGKzuzl2XT0BBgDgVJCgCSJCmBX4AWQBwQKknSZgoSn5/ue/0AIUSiJEntgQ+BFcUR9MtOo9cUTCXJS0Mv9NiZ2+Fi6VJy9/EC0tPTOLZ6Aq8mrCRAyife5E3UB5LR3LyKU9++uI4c8e+O5qVISnw2W+eEk5+rpc2wmpSv7mzokGQvOWOurPYwey8k8um6M6Tn5DPT9jqV1ywGpRLPD9tgr94EoYlQtX1BhTWXkldZuKQWIngUOdkpQdQ6NaP/Gc2+2H2MCRxD3+p9C/8iQiAOzULa/Q1n9BXZX2cmIzu8ZtCkQuj1aG/dQpuSUrB3TtmypWY6iSpbTU767U1DHYxnaL4oCSH2S5Lkc9/D9YBLQogrAJIk/QZ0EEL8RMEo0MPOsxnYLEnSNuDXIgz5pabRaUjOSyYtr2C+vL25PS6WLliYlNzfQaFVc3rLHLzDZ9KKdM7bNsYmuRI5G3dgWr4c5VeuwKpuXUOHWSRiz6fy5/yzmJor6TSmDq7eRVu5siSQp8UajrFXVrufSq3jx+3nWHH0GnXs9Cy7tRXp0H4sa7yCZ514TNMWQblXoccq8K5n6HCfS0kuRPAoxZbsyI3Ji1FpVYzaO4pDNw7xZf0v6VGlR+FfRKdBbBuDdDKELboGnKs/ibFtAgya6Ojz89HExhbsnePkjIlH4eydYwzyVVqyUuRNQ2/zAmLv+TwOqP+ogyVJagZ0pqAo4PbHHDcYGAxQrly5Qgjz5aHWqUlWJZOen44QAgdzB1ysXErGRsWPIgS3jq1Fv/s7amvjiTKpTobvF5iErCcr7k8c+/TGbdSoUjmaA3D+yE32rjiPg4cVbT8KwNap5Casd8jTYksmY6+s9jBn4tIZueY0V5Jy+NIplSYb5yOyMnFtZo+T+wEku8rw5mqo3LrEVFi7l14vOLz+EuF/G3chgufxVMmO3JgYVq4ml4/2fERYQhjfv/Y9nSp2KvyL5GUg1vZFurKXOdoOZL32GeNaVzNIoyP0etDp0GVno7l5s2DvnHLlUNqVns3tNPk6MpNU8qahz0kIsQ/Y9xTHLQQWQkEFpaKNquQTQpCtySYjP4PM/EyQKEhyLF0wU5bsKZbqS/+QuulzPLIiiRbe7POfSuVzt0j/YQam3t6UX74Mq6AgQ4dZJIQQhG2P4fiWq5St4shbH/hjbllqJnaEIE+LLTFepLKaoWh1eubtu8zMv6PxNIcNeQewCN6EqbsFnm/cwKKsCzSfBbXeA2XJ/L0qLYUIHuVpvyshyI2JQWSpsxi6eyhnk8/yY+MfaVvhoTN6Xkz6dcSqbuiTLjJOMxiXxu8zrlXlQrkB1+fno0tPR5eWVvD/+z7W3v5Y26sXeUolaHUIve7u6xVW1ph6F8/eOcVFq9GTkZSLpJCwd5U3Db0tHvC+5/Oytx97YZIktQPa+fmVvHnTd4SFhbF8+XJmzZpV6OcWQpCnyyMjP4OM/Ay0ei1KhRJHS0dcLFyMblHwM0s4S/qWr3CI34deOLHC4zNer9qEsj/+QPr16zj26oXbJ6NKfEXHR9Fp9exbdZ7zRxKo0sCDZr2qoDQpPTcx8rTYZ1eU7cmjqLQqUlQpZORnGG1ltYe5lpLDqDWnOXk9nfedc+n+53x0N27gVCUb17pZKJqOgwZDwcza0KE+tzuFCDJKQSGCR3mqZEduTAwjIz+DD/76gAupF5jSdAotyrco/IvEn0D82oM8VS7vqz+jdtMOjGn58ERH6PVobtx8bPKiS09De/fxdIRK9chLK6ytUTo4oHRwAIUChaUlktIETJRISiWSqSkKW+Md0n4e+SotWckFm4Y6uluWqpuOFxQKVJQkyZeCJKcH8G5hnFgIsQXYEhgYOKgwzmcIgYGBBAYW7nYkGp2GDHUG6fnp5GvzkSQJW1Nb7M3tsTGzMcr58s8k7Rrqv37AJGodCCt+Me1Ljbc/osWf60j7cDCmXl6UW7YM6/olc17908hXaflzwVnizqcR1NaXoDY+pao9fYxCnxZbmqbEFkV78jBCCLLUWaTmpRp9ZbX7CSFYExrL91ujMEPPr1IojkvXIlnqKPd6JtZt+0CTsWDtYuhQX0j8xTT+XFBQiKDdiFqUrfzsm9CWBC8y3ibPsS8CQgiiUqPYfGkzO67uIFuTzYzmM2jq3bTwL3ZuC2L9INIkB7qpvqF182Z80qLSQ/8Yam4lEj9iBKrTpx88jyShtLO7m7iYurljUany3c+Vjo73fOxw92OF2b/TYs6dO4eZt/eD5y5GMTExtG7dmkaNGnH48GG8vLzYtGkTrVu3ZurUqQQGBpKcnExgYCAxMTGEhISwceNGcnJyiI6OZsyYMajValasWIG5uTnbt2/HycmJZs2aUbNmTfbt/Qe1WsPsaXNp+kZDqlavwuHDh3F1dUWv11OpUiWOHDmCq6urQb8ORU2SpNVAM8BFkqQ44FshxBJJkj4CdlIwOhwshIg0YJjFIicnh27duhEXF4dOp+Prr7+mQoUKjBgxgpycHMzNzfn77785ceIEU6dOZevWrfzvf//j8uXLXLp0ieTkZD799FMGDRpEnz596Ny5Mx07dgTgvffeo1u3bnTo0OHu9XR6HVmaLDLyM8hWZwNgaWpJGesy2JnbYaIomVMw/iMnBbF/CvrQxej1Egt07cir9zH9XCHl8+GkXbuO47vv4jb6ExTWJbc39kmy0/LYOiectJu5vN6nKlVfK2PokIzW00yLLQlTYou7PXkUnV5HWn4aqXmpaHQaTBQmuFm54WThZJSV1e6Xkp3PuD/O8lfULdo7qPh472w01xKx98nFvVdTlG//D5wqGDrMFxZ16Ab/rLqAnWvpKUTwKMX2l02eY/94ibmJbL2ylS2Xt3Ap/RJmCjOal2tOn2p9qOlas3AvJgQcno346xuuW1ajc9pHvPdGIKPerPjQRCf31CniPv4YfU4ubp99hln5crcTFseC5MXOrsTvdXNHdHQ0q1evZtGiRXTr1o3169c/9viIiAhOnTpFXl4efn5+TJo0iVOnTjFq1CiWL1/OyJEjEQLSU7LYvfUAJ84cY9RnHxEREUGvXr1YtWoVI0eOZPfu3QQEBJT6RAdACNHzEY9v5zEdIc/raaexHVh7keTY7EK9tou3DY27VXrk83/++Seenp5s27YNgIyMDGrXrs2aNWsICgoiMzMTy4cslj9z5gxHjx4lJyeH2rVr06ZNG95//32mT59Ox44dycjI4PDhwyxbtuzuvhUZ+RlkqjPRCz2mSlNcrVyxN7M3+mkkT02dA0fmoj84AzS5rNU2Zbd7f8a0fRWn34JJWL4cU09PyoWEYN3gkf1ypUJSbBbb5oSjztfRdngA3lVLdtnY51Ak02KNuS2B4mlPHidPm0dqXioZ+RnohR4rUyvcrdyxNbMtESPFer1g4+l4ftx+jsxcNat123BcuRedQo9XxzLYDZ8GXiV/z63SXIjgUV4k2ZHn2L8glVbFnut72HJ5C0duHkEv9NRyrcXXDb6mlU8r7M3tC/+iOi1sHwMnlnLGrjndEvvy4ZvVGfnmwxvRtLVrSfhhPKZlylBuyRIsKj2+sS0M322JJOpGZqGes5qnHd+2q/7E43x9falVqxYAdevWJSYm5rHHN2/eHFtbW2xtbbG3t6ddu3YA+Pv7c+bMGfJVWrRqHR3bdsHO2ZLW7VrwwbBM0tPTGTBgAB06dGDkyJEEBwfTv3//F32bsocw5mls/v7+jB49ms8++4y2bdvi4OBAmTJlCLq9UN7uEUU5OnTogKWlJZaWljRv3pzjx4/TsWNHhg4dSlJSEuvXr6djp46k5BfMkdfoNSgkBXbmdjiYO2BlYlV6pjNp8+HUCsQ/k5Gyb7FbH8QcRU+6tX2TWaYpJAzpQ+q1azj07IH7mDGlejQH4HpkCn8ujMDcyoQuY+vi7FVydmsvREUyLdaY2xIo2vakS5cumJg8eMt4p6hJiiqFHE3O3fU4ThZOJWoPrjNx6fxvcyQnr6czwCaKPidWorqmxsrbhDI/fINp/a4lssLa/fJVWnYtjuR6ZAo1m5elYSkrRPAoL5LsyHPsn4MQghO3TrDlyhZ2xuwkR5NDGesyDPQfSPtX2lPernzRXTwvE37vB5f/5m/n9xgY35qRb1ZhxJsVH4xTrSZhwo+kr1mDdaNGeP08FaV9ESRfRsbc/N9ebqVSiUqlwsTEBL1eD0BeXt4jj1coFHc/lyQJVU4eGYm5SBLYuVhgYfNvz4kkSXh7e+Pu7s6ePXs4fvw4q1atKsq39tJ62s6TJ/WaFoVKlSpx8uRJtm/fzldffcXrr7/+VK+7P1G583mv3r1YsHQB69au4/uZ35OsSsbGzAZ385LTu/rUNHlwagUcnA6Z8UQoq/Ft/od4+jdl0ZsVIHgB15ctK+ioCVmKdYMGho64yEUdusG+VRdw8rSm7bAAbBxLyajdYxjjtFhDtCVQ+O1Jnz59WLlyJb/99htLly79zzGPmqrmaOFYoqbDJmfnM3XnBdaExdLC4iL/ZPxO/o5k8rRK3Pu2wnHsFCST0jHqkZGUy7ZfzpCRqCq1hQge5WlLTxdrY1IaR3Zis2LZcnkLmy9vJj47HisTK1qUb0EHvw7Uda9b9Dch6bHwa3dE0nnWeIxlXExtPmlRiY/feDDR0SQmEj9iJKpTp3AeNAjXkSOKdZra04zAFCcfHx9OnDhBvXr1WLdu3ROP12n15Gaq0ar1WNiYYmKmZN36dbzZ4k0OHjyIvb099rcTx4EDB9KrVy969+6NspRMBTQ2xtx5cuPGDZycnOjVqxcODg7MnTuXmzdvEhoaSlBQEFlZWQ+ddrJp0yY+//xzcnJy2LdvH19+9yXXM6/TsH1DerTqgZu7G6/VeQ17c3tMFaXjD/VdGhWcXF6Q5GTd5KJ5db5X9yHGLogf+vlTPzeem316oo6JwaF7d9zGjkVpU7pHc4QQHN9ylbDtMZSr5kSrQTUwKz2lpR+rOKfFGvu9SWG1JxMnFuwi0q9fP+rVq4eHhwfVqlUDSv5UtTs0Oj0rjlxj+u6L1NScZbf6D6S9ieSmm2LpV5YyU2djXsW47kVexMtSiOBRnrYaW7HOsTfmm5NnkaXOYlfMLjZf3szJxJNISNQvU59htYbxRrk3iq+u/I1TBYmOJpe5ZScxJdqTMS0r8dHrDyY6qtOnift4BLqsLLymT8OudeviidGIjRkzhm7durFw4ULatGnz2GPzVVpSVTnotQJTCxPsnAv+sFhYWFC7dm00Gg3BwcF3j2/fvj39+/eXp7C9pM6ePcvYsWNRKBSYmpoyb948hBAMHz4clUqFpaUlu3fvfuB1NWvWpGmzpiQmJzJw1EC0tlpUWhWVyleiRrUadO7UGRfLkl0l6AHqXDgRAodmQPYtoi1r8o16AFEE8GErPxbU9iB7/lyuhYRg4uFOueAlWL/2mqGjLnI6rZ49K85x8dgtqjYsQ9N3K6N8CaalGIKx35u8SHvSvHlzkpOT+frrr/H09ATA3d2dqlWr0qFDB7LUWSV+qtodB6OT+W5zBC4px1nLBmyP3yQrzhITZye8pnyBbdsOpWeaL/8WIrB3s+TtoTVxcCu9hQgeRRLC+GoA3NN7Mig6OtrQ4TwTnV7HkZtH2HxpM3ti95Cvy8fX3pf2r7SnbYW2eFh7FG9A57fB+oEIK2d+cvyehefNGduqMsOaP9gzlb5uHQnffY+Juztlf5mDReXKxRbmuXPnqFq1arFdr7AJIchJzyc3U42JqRI7VwtMTAtGapo1a3a3mtv9wsLCGDVqFAcOHCjukJ/Kw74vkiSdEEIUfd3SQhYYGCjCwsL+81hJ+7lT69R8+c2XKM2V9B7aG4WkwNbMFgdzB6xNrVGpVPj7+3Py5Mm7o4clzQPfE3UOhAXDoVmQk8hl6zp8ld6GcGUN3m/ky8DGFTC7EMmNL75EfeUKDu+8g9tnn6K0Kf1rVfJzNexYcJb4C+nUb1+Buq3Ll6ibtJLWljzu3qSktSV3/O9//8PGxoYxY8Y88FxWdhY1a9Zk3d51WFhbYKIwwcnCqURNVbv3+xKbmsuErVFknv+bT5Ub8IyII/WCDZiY4jJ4ME4DB6GwsDBwxIXn3kIE5ao50bKUFyJ4XHtilD+txt578jDRadFsubyFrVe2kqRKwt7cnk5+nejg14HqztWL/w+QEHB0Luz8EuFZhy8svmB1VD6fvVWFD5u98t9D1WpuTZxI2q+rsX7tNbym/Vyw943sqei0ejKTVWjydVjamGLjaIH0FBuFTpw4kXnz5slrdYqYsU89eRKdXkeGumDDz1xNLjmaHBwsHfCy8cLWzPZuKdfdu3fz/vvvM2rUqBKb6PxHfjaELobDsyE3mSu2QXyl/ZCw9Kq816Acs5v7YZt0g6TPPyXrzz8xKVMG78WLsWnU0NCRF4vMFBVb55whIzGXN/tXo3L9Yu5IewmVxHuT55GvzWfDjg18MvQT+nzQBycHJ5wsnErcVLU7VGod8/Zd4uyBjXzEOl6JjSMxwp4UlS327dviOnospu5uhg6zUL2shQgexShHdu54WE+sMUnNS2XH1R1surSJc6nnMJFMaFy2Me1faU+Tsk0wU5o9+SRFQaeFPz+D0MXoq7TjE+1QNkak8XnrKnzQ9L+JjjY5mbgRI1GdOIHT+wNwGzUK6SEVV4paSe0Vy8/VkJmSBwJsnS2wsC5dvSbyyI5h6IWeHE0O6fnpZKmzEEJgpjTDwdwBB3MHo9+Q70Wci4qkaspOODIHclOIcajPl6ltOKLxo2vdsox4sxJu+Zkk/zKX9PXrkczNce7XF6cBA16K0RyApOtZbJ0Tjlajp/UQ/xI7/15uS4xHaamqdi8hBKfORLJ8605656+mavI1EsJdyE8By1oBuH/5JZb+/oYOs9DdW4igcY9KL00hghI3smPMPbFqnZr9cfvZdHkTB+MOohVaqjlXY1y9cbT2bY2ThYH3M8jPgt/7w6W/0L06nBFJHdkacYsv367KoCb/3QRLdfYscR8NR5eRgefPU7F/wnoU2b8eN21NJnseQgjydHmk56eTkZ+BTq9DqVDiaO6Ig7kDFiYWJWqK0jPT6yAnCTJvwN/fcd2pIV+q3uZAgi9t/MvwV8tK+JhqSVn8C5dXrETo9Tj27InLkA8wcSll65MeI+ZsMjsXR2JhbULnkXVw9nw5EjxjYMz3Js+rtFRVu59KrSU9LQVlbiKT0yaQEOHGtRgXTDw88Px5DHZvv10q29OXvRDBoxjlT7IxDhWn56Wz+sJqVp9bTVp+Gq6WrvSu1pt2r7SjouODC/0NIiMOfu0OiefQtpnO8AsB7IhI4Ks2VRnY+L+JTvofG0j43/8wcXXFZ/WvWJTgHqni9t9pa2bYOJo/1bQ1mWE86QZFCGHQP3oaneZugpOvy0eSJGzNbLE3s8fGzKZETht5JnptQZKTnYTQa9FKpvQzmcS+G940q+zK1paVqeZoSuqKlVxavBh9djb27dvhMnw4ZmXLGjr6YhWxP579qy/g4m1Lm2E1sbYv/aWljcmT7k0M3ZY8i3xtPil5KSW+qtr9tDo9GWkpWOYn4S7ySFbruLLTs2BdzseDcO7fH8VDKtKVBlEHb/DPry93IYJHMcpkx5gk5CSwPGo56y6uQ6VV0axsM7pX6U6DMg2Mq9fjxumCREedg6bHWj467sDOyAS+bluN9xv53j1MaDTcmjSZtJUrsWrQAK/p0zBxlDP/p3XvtDU7F8tSN22tNHrcDYqFhQUpKSk4OzsX602KTq8jS51Fen46OZocACxNLSljWQY7MzvjaluKil4L2UkFiY7Qka+04XK2kqMparKda7L2vSoElbUlfd06Ls2diy4pGZvmzXEdORKLyobZx8RQhF5wdNMVTu68RvkazrQcWB0zi5fgZ6QEMVRb8iweOlXNzB4ny5I7Ve0OIQRZGamY5t7CSeSjVpuQkCHB5Vjs3m6D66hRmHqUznVter3g8LpLhO95OQoRPA+jbC2NYaj4SvoVgiOC2XZlGwBvV3ib/tX74+dohMPXF3bAugFg5Yym358M253HrqhbfNuuGv0b/pvoaFNSiB8xktywMJz69cNtzGiDrM8pif4zbc1MiZ2LPG2tNChbtixxcXEkJSUV+bWEEKj1alQaFSqdCiEESoUSKxMrLE0syVfkc+v2f6WaXg/qrIIpt0KPTmlJurAkV5dJqgoq+pbn/VZuZO34k8sfzEJz/TqWdeviNnMmVnXqGDr6YqfT6Pl7WRTRYYlUb+xJkx6VXuqFxsaqONuSZ6UXenK1ueRqctHqtSglJVamVlibWpMpZZJJpqFDfCGavBzIy8QUDVq9EqFRgkaLMiMDnzp1sB0wwNAhFhm5EMHTMco7XUNOYwtPCmfJ2SXsjd2LpYkl3at0p0+1PnjaeBZ3KE/n6Hz4cxx41kL9zq8M3XyD3edu8V376vR9zefuYaqISOKGD0eXmornlMnYt2tnuJhLGJ1WT0aSCq3aMNPWbGxsyM7OLrbrvUxMTU3x9fV98oEv4HL65buVGm/l3sLG1IZWPq1o/0p7arvVNtpe4EKXk1JQdOD4QlBnk1yuNeOz27LxhiO+Lno+aVGJd+p7oDp8iJgvhpF/7hzmlSpRdv48bJo2fXm+TvfIy9GwY/5ZbkSn82qnV6jdstxL+XUwFo/riC2OtuRZJKuSCU0I5ciNI/wZ8ycqrYrabrV5t+q7NCnXpORvNiwEqac2krPzR7zzLxKb7U7G9UqYnLmMiYcHbqNHY9elS6n+fbm3EEGz9ypTvfHLUYjgeRhlslPchBAcunGIJWeXEHYrDHtzez4M+JCeVXriaGGkU7x0Wtj5ecGNQ5W25Lefx7DfL7D7XCLfd6hOn1d97h6avnEjCd98i9LFmfK/rsKyeunZFbiovUzT1nQ6HUqlPFpVGO5UatxyeQuRKZEoJSUNvRoyJmgMzco2w8Kk9Ozl8EQ5yXB4FhxfDJpc0nzbMim3Lb9dtKWMvQWTulSkS52yqM+EE9dvHLmhoZiWLYvnlMnYtWmDpHg5eykzk1VsmR1OZoqKlu9Xp2KQu6FDeukZ43riO1JUKYTeCiX0Ziiht0K5mnEVABtTG1qUb8F7Vd+jmnM1A0dZCPR61JFbyNg5AdfsC2Sq3Tl6623sQ6MwUcTj/NFHOA/oj8KqdK9Xib+Yxo4FZwHkQgRP4aVOdrR6LbtidhEcEcyFtAu4W7nzadCndKnYBStTI/5Fyc+Cde9D9E549SPym3/Lh7+Gs+d8Ij90rEHvBuWB2+tzpkwhbfkKrOrVw2vGdEycDFwtzojFxMTQunVrGjVqxOHDh3F3K8PS+at4r39Xpk6bSoNy9UlOTiYwMJCYmBhCQkLYuHEjOTk5REdHM2bMGNRqNStWrMDc3Jzt27fj5OREs2bNCAgI4J9//kGr1RIcHExgYCCVK1fm8OHDuLq6otfrqVSpEkeOHMHV1fWh8Qkh+PTTT9mxYweSJPHVV1/RvXt3hg0bRqtWrWjfvj2dOnXC0dGR4OBggoODuXz5MhMmTGDlypXMmjULtVpN/fr1mTt3LkqlEhsbGz744AN2797NL7/8QqNGjYr5q170imtarFqn5p+4f9h8efPdSo1VnaryadCntPZtjYvly1MxDIDsxIIkJ3QJaFRk+rVnWn57Qs5Z4mRtxtdt/XivfjmkmCvc/Phjsv/+G6WzM+5ff4XjO+8gmRmodL8RuBWTybZfwtHrBB1G1MKzonwjI/uvtLw0wm6FcfzmcUITQrmccRkAKxMr6rjXoZNfJ+p51KOyU+XSsQZQr0ec30L2zgnYZlwgS+fO0bQuVDwVjX3KaezatcPtk1GYlilj6EiLnFyI4NkZ5W9AUd+c5Gnz2HRpE0sjlxKfHU8F+wr80PAH2vi2Mf79KzLib1dci4I208ir1Y8PV55g74UkJnSqwXv1CxIdbWoq8aM+IffYMRz79MZ97FgkUyN/b0YgOjqalctXMen7GfQf1Itde7djYqbExOThIx4RERGcOnWKvLw8/Pz8mDRpEqdOnWLUqFEsX76ckSNHApCbm8vp06fZv38/AwYMICIigl69erFq1SpGjhzJ7t27CQgIeGSiA/DHH39w+vRpwsPDSU5OJigoiCZNmtC4cWMOHDhA+/btiY+P5+bNmwAcOHCAHj16cO7cOdasWcOhQ4cwNTVl6NChrFq1ij59+pCTk0P9+vX5+eefC/1raSyKsjdWCEF4UjhbLm9hR8wOstRZdys1tn2lLZUcX66F9ABk3YJDMyEsGHT55FTqxGxtBxZEmmBjZsInLSowoJEv5sm3SPr6KzI2bUJhbY3ryBE49e6Nwtra0O/AoK6GJ7FrSSSWtma0Gx6Ao8fL/fWQFcjIzyAsIYzQW6EcTzhOdFo0AJYmltRxq0PbV9pSz6Me1ZyrlY7k5g69Hs5tJv/viZinniNRX4bVWb14NToev0tHMK9ZE49fZmNZq5ahIy1yep2ew+svy4UInoNR/kYU1c1JpjqTtRfWsiJqBal5qdR0qcmnQZ/SzLtZySi1eDO8INHJz4Z315Ln05whK0+w70ISP3by59365QDIi4oi9qOP0CWnUGbiTzh07GjYuJ/VjnGQcLZwz+nhD60nPvEwHx9fynlURKfRE1Q/iFvJ8Y89vnnz5tja2mJra4u9vT3tbq+F8vf358yZM3eP69mzJwBNmjQhMzOT9PR0BgwYQIcOHRg5ciTBwcH079//sdc6ePAgPXv2RKlU4u7uTtOmTQkNDaVx48bMmDGDqKgoqlWrRlpaGjdv3uTIkSPMmjWLZcuWceLECYKCggBQqVS4uRXsFq1UKunSpcsTvy6y/4rLimPLlS1svbyV61nXsTSx5I1yb9CuQjvql6mPUvESTgfMvAmHZsCJENBpUFXtykLRkTlnQCFJDG7sw5Cmr2Cbn03yz5OJW/0bSBJO/fvjPGigXBUSOLsvjgNrLuJazpY2wwKwsnt5R7eMUXEWT8rIz+DkrZMcTygYubmYdhGBwEJpQS23WrSu3ZogjyCqu1Qv+etvHkavh6iN6P6ZjDLpHHHCk19z36dZQg6NQ/dg4u6O2+RJ2LVt+1JMdS0oRBDB9chUar5eloZd5EIEz8Iok53ClpibyMqolay9uJYcTQ4NvRryfo33CXQPLDmL1y78WVBxzdIBBvxJnnNVBq84wf6LSUzs7E+PegWJTsaWrdz8+muUDg6UX7UKS/8aho27hBBCkJ2ej4nSFKWJAjsXCywszcjOzsbExAS9Xg9AXl7ef15nbv7vPhcKheLu5wqFAq1We/e5+3/OJEnC29sbd3d39uzZw/Hjx1m1ahWxsbF3E6YhQ4YwZMiQJ8bu5eVFeno6f/75J02aNCE1NZW1a9diY2ODra0tQgj69u3LTz/99MBrLSws5HU6TylLncWumF1svryZk4knkZCo51GPwTUH82b5N7E2fUl74DPibyc5y0CvJb9Gd0KUnZl+QotWJ+ge5M3w1yviqtSRunQRl4OD0eflYd+5E67Dhr0U006eROgFh/+4xOndsfjUdKHl+9UxNZd/L41NUY4SZ6mzOHnrJKEJBSM351PPIxCYKcyo5VaLobWGUs+jHjVcamCmLMVJsF4HURsR/0xGSjrPdbyYo/qAoAwTuh7ZjgQ4Dx2K88D3S/26nDvkQgQvrlQnOzEZMYREhrD58mZ0Qkcrn1YMqDGAKk5VDB3aszm2oKDimoc/9FxDnqUbg5aHcfBSMpO71KRbkDdCqyVx6s+khoRgFRiI18wZmDg7Gzry5/MUIzCFSafRk5GsIi9bjUIh4ehu9Z9qaz4+Ppw4cYJ69eqxbt2657rGmjVraN68OQcPHsTe3h57e3sABg4cSK9evejduzdKpRJvb29Onz790HM0btyYBQsW0LdvX1JTU9m/fz9TpkwBoEGDBsyYMYM9e/aQkpJC165d6dq1KwBvvPEGHTp0YNSoUbi5uZGamkpWVhbly5d/rvfyMtHqtRy+cZgtl7ewN3Yv+bp8fO19GVFnBG1821DG5iW+UU+PhYPT4dQKEHo0/j351awrU0Pzyc5X0yHAk5FvVqKcnSnpq1dzef4CdGlp2LZsievIEZhXqPDka7wEtBodu5ee4/LJRPybetGoeyUU8ibFpV6OJuc/yc251HPohR5ThSkBrgEMCRhCkEcQNV1rYq58CTaP1esgcgP8MxmSLxCrLMeU/GG4ZTvw/sktSCnJ2LVpg9voTzD1NNLquEVALkRQOEplshOZHMmSiCXsvrYbU4UpnSt2pm/1vnjbehs6tGej18HOL+DYfKj8NnRZjAoLBi0L49DlZCZ1qUm3QG+0aWnEf/IJuUeO4vjee7iP+0xen/OU7lZbA2ydLVAopQfKSo8ZM4Zu3bqxcOFC2rRp81zXsbCwoHbt2mg0GoKDg+8+3r59e/r37//EKWwAnTp14siRIwQEBCBJEpMnT8bj9iZpjRs3ZteuXfj5+VG+fHlSU1Np3LgxANWqVWP8+PG0bNkSvV6Pqakpv/zyi5zsPIIQggtpF9h8eTPbr2wnJS8FB3MHOlfsTPtX2lPduXrJGREuCmnX4OA0OLUKAF2t91hv1Y3JR3NJzs7hzarujGlVicqu1mRs3sLl2bPQ3riJVYMGuI3+BEt/fwO/AeOhylazfe5ZEq5k8FoXP2q96f1y/2yVYrmaXE4lniI0IZTQhFAiUyLRCR0mChNqutRkkP8g6nnUo6ZrzZerWqNeBxHrYf8USL5IgrkvP6g/JjWnLGMv/Ynl5fNY+PvjPnsWVnVqGzraYhV5IJ79qy/KhQgKgSSEMHQMD7hnXuyg6Ojop3qNEIKjN48SHBHM0ZtHsTW1pUeVHrxb9d2SWQXp5pmCRCfmADQYBi1/QKWFgctDOXw5hSldA+hatyx5588TN+wjtImJePzvfzh06WzoyJ/LuXPnqFq1arFdT+gLpq2psu5sEmqJiWnRzH9t1qwZU6dOJTAw8IHnwsLCGDVqFAcOHCiSa7+oh31fJEk6IYR48M0YqWdpT9ZeWMtvF34jOi0aE4UJzco2o/0r7Wnk1cj4i5cUtdSrBUnO6V9BUqCv3Zvt9j346VA28ekqXq3gzNi3KlPb24HsPXtInD4d9aXLWFSvjtvoT7B+7TVDvwOjkpGUy5bZ4WSn5vNm/2r41XUzdEjFrqS1JXcEBgaKsLCwxx6j1WvvJjahCaFEJEegFVpMJBNquNQgyCOIII8garnVwtLEspgiNyI6LUSsK0hyUi6Rau3HD9ntOZb5Cv+7tR/vkwcwcXXFdfQn2Ldv/1Ksy7lDp9Nz5N5CBINqYG5ZKscmCtXj2hOj/Oo9y7xYnV7H39f/ZknEEqJSonCxdOGTup/wTqV3sDGzKYZoC1l6LOwZD2fWFKzPaT8b6vQhV63l/ZAwjl5N4ed3AuhcpywZ27Zx88uvUNrbU37VSixr1jR09CWCVqMnM/n2JqG2Ztg4FO8moXdMnDiRefPmsWrVqmK/9svkWdqTc6nnsDKx4usGX9PKpxX25vbFEKGRS70C+3+G8NWgMEHU7c9el/eYcDCDy0lJBJS1Z1KXmjT0cyY3NJRrY6ehCg/HzMcHrxkzsG3VUh6tuE/ClQy2zT0DAjqMrEUZPwdDhyQrZALBiL0jUOvUVHeuTt/qfQnyCKK2W23j3tqiqOm0cPb3giQn9TLZDlWYYvEZa25VZkxaGIPDpiIBTh8OwWXgwJeuOmNmsoq/giNJuJIpFyIoREY5snPH0/SehESE8POJnylvV55+1fvR7pV2JXN+qyq9oNf06PyCzxt8CI1GgaUDuWotA0JCOX41lZ+7BdCxZhkSp00jdUkwlnXrUnbGdEweU7K4JCiukZ28HA1ZKXkggZ2zhVy28QlKw8jOHU/bG1uqyra+iJTLBTckZ9aC0hRRtx9Hy/RmwoE0IuIzqehmw+iWlWlV3Z38c+dInD6DnAMHMHF3x+WjYTh06oRkIn8t73flVBK7giOxtjej3fBaOLi/vDe+Ja0tedZZJ2eTzuJr71syO14Lm05b0Il7YCqkXkHtUp1Fim78fM2XLunn6RuxDWVKEnZvt8Zt9GhMvV6+RfiXTiSyd+V5EIJmvapQMVDeSPhZlLiRnWfR3q89ZWz+3959h0dZposf/z4zaZPeE0gBQggkQOhNBAIoyFpAVBBhWRTbsXvcXd3ib9Vz9GxxddVVV6Qt7q66oogrCEqT3kILPYEQEgKppLcpz++PCRAEIYEkM5Pcn+vKlZl33pk8874zd977qR24KfYm15zu1VIHO+fZB+VVn4U+98Lo30CgfXxRVZ2F+xfsYMeJYt6Y0pfbu/iQ/dDDVG7eTOC0e4n81a/a9eJ7jfXDbmsBoSaMLdRtTbguSXSAwnR7kpP2GRg9Yeh/sTfGnuRs/z6L6CATf76nD5P6RWHNPknucz+nbPlyDAEBhP/i5wRNn47Bqx2NOWiCvauz2bg4nYjO/tz6WDImP4ndrqSps7H1DpPxaVjqYN8nsOHPcPYE1ohkvur+R57fH0NSSTYfH1+Af+YRvHr2JOLtN/EeMMDRJW515jorGz9L5+CGXCK6+DNudk/8Q9th18YW5PL/2YO9ghnfebyji9F0WttnHln9Mpw9AXEpcPMr0KHP+V0qay3cv3AHO08U8+bUvow3VZB5z/1Yzpwh8n9eIeieexxWfFdiMVspK6y50G0tyFO61QjxQwVHYf0f7YOF3bxg2OMcibufP2wsZs3aY4T6evLKxJ7cOygWVVxI/v+8Qslni1Hu7oQ88gghsx/A6O/v6HfhlGw2zabF6exbk0Nc3zBufiAJNw8XrJwTorFqy+1rbm15D8pz0R36sH3IX3lqVyTWw/m8fnoJ3fZtxBgWSvhrrxEwaWK7GpdzTtGpClbOPcDZ05X0Hx/L4DviMEq3tWbn8smOS8raDN/+Fk6lQnhPmPE5dB0LDS7AN6QX8L9fHyKjoIK37u3HqNy9nPj1bzD6+tLpo0XtYrXg5tCw21pAmEm6rQnxQ/mH65OcL8DdG254kqzus3l9UzH/WXMYfy83fnlLd2bd0BnPmiqK3n6L4kWL0BYLQVPuIeTRR3EPb3+D6xvLXGdl1fyDHN9TQJ8xMdxwd7xMLS3arooC+wyyOz6EmlLoPILM4X/gl7tD2Lc6n8fyV3Hz3m8xaBvBjzxCyEMPYfRtX+NywD6p1oH1p9i4OAMPkxt3PNWXmKRgRxerzZJkpzUVHIVVL8GRZeDXESa+C32mQYPud+l55by2/BBrjxQQE2xi7vR+9FzxL059OBdT375Evf2WXFg0grZpKs7WUl0h3daEuKz8Q/busweW2JOc4U+Tkzibv2w5yxfv78fTzcgTo+N5aGQcflg4u2gh2R/OxVZaiv9ttxH21JN4xMY6+l04teryOpa9t4+8E2XceE83+ox1seUPhGissydg81/t625ZatGJt5HW+X5eP+DH+iX53Fq4gX8fWo5HcQF+t9xC+M9/jkd0+xuXA/ZK2LX/OMzx3QXEJgUzdlYS3v7SpbUlOWWy02AQoKOL0jzK8+D739tXGHf3hjEvwtDHwOPCwNTCilre/O4on+zIxtvDyK9/0oMZPYMofP55ijZuJHDKFCJ++xsMMj7nqhp2W/P288BHuq0JcUHeQfj+D3BwKXj4wI3Pkpv0AG9vOcvi9/ZjNCgeGN6FR0Z1JdRkpOTzLzj27rtY8vPxGTWS8GeewasVp4l3VSV5Vfznr3upLKllwsO9ievn2pPICHFZZ/bDpr/YW4aVAVvyVDaF38fruzT7U4sZX7STT3I2EZCVjmdSIpFv/RnvQYMcXWqHyc0o4bt5B6gqrbOvrTU2xiGzwbY3TpnsNHUQoNOqq7TXdGx6C6y1MGg2jHoefC6s+1NjtjJvYybvrztGjdnKg70CmUk2tk/e4OTmzWirlciXXiLo3qkOfCOuQ7qtiR9qc5Un1yrvQIMkxw9GPMeZng/w163FfPpuGgrFjKGdeCylK2G+HpR/+y3H//IWdSdO2FuV//x6u75IaYrTGSUse38fSikmPduPyDiZwly0IVrbu+Nv+gukfwsevliGPMoy7zt5a3sl+etzmZa/i1eOb8SjuBD3TrGEvvq/BEyahDK2z7FqNpsm9ZsT7Pg6E79QE5N/OYCIzjLGsbU4ZbLj8qwW2PMPWPsaVORB4h0w9ncQeuFiy2bTfLU3lz+tPALZWTxhO0lK4SH0kn2UaY1bhw4E3nUXgXdNxispyYFvxjVc0m0tzITRrend1t544w3mz58PwIMPPsikSZOYMGECN954I5s3byYqKoqlS5diMplISUlhyJAhrF27lpKSEubNm8eIESN48803SUtLY/78+aSlpTFt2jS2b9+Ot/fFU8x27tyZKVOm8M0332AymfjXv/5FREQEycnJHD16FHd3d8rKyujTp8/5+6Lp2kzlybU6k2ZPcg79Bzz9YeQvKOg5m3e3FfOvd/ah0UwdFMNjKV0JOpVJ+Zy/cuybFZhPncKzWzzR772L7+jR0jraSBmp+axacBDfYE9uf7IPAWHtd2pp0cbYbHD0G9j4F8jZDt6h1I78Nf+0jeNv24rwyN3P/We2MTR9K4a6WryHDSV45iv4jhrVLicfOKfibC2rFhzg1NESug2KIOW+7njIIqGtSo52c9Iajq6EVb+DgsMQPRimfASxQy7abXtGAR8tWEbYvm28VnCI8JI8ADySEvF7/HH8xozGMzFRLi4a6aJua/4e+AReW7e11NRUFixYwLZt29BaM2TIEEaNGkV6ejoff/wxH374IVOmTOHzzz9nxowZ9r9tsbB9+3aWL1/Oyy+/zKpVq3j66adJSUlhyZIlvPrqq3zwwQeXJDrnBAQEkJaWxqJFi3jmmWf4+uuvSUlJYdmyZUyaNIlPPvmEyZMnS6Ijmu70XvuYnMNf25OcUc9T1Hs2728t4qN39mCxae7pH8WjMRrvTasoe+/nlJ48CW5u+AwbRtizz+I/4ZZ2WxPbVFpr9qzKZvPnGUTGBfCTx3pj8pVux21Ju20lttTZp6Lf9BYUHoHAWMrH/p455cNYuO4M3XLW8uLprcRnpqE8PPC/43aCfzoTr+4Jji65w2XuK2TN3w9hsdgY+7NEug+NlGs7B5Bkp7mc2gXf/T84sQGCu9qTnMTbz8+wZquu5vjKtez+eClRh3fxWG0F2mjEZ8gQ/MY8hN+Y0bh37OjgN+Ec/rD9DxwuPtyofW1WjdViA8DobvjRWY56BPfg+cHPX/G1Nm7cyJ133olP/YrNkydPZsOGDXTp0oW+9bPfDRgwgBMnTpx/zuTJky/ZbjAYWLhwIcnJyTzyyCMMHz78R//mtGnTzv9+9tlnAXuL0h//+EcmTZrEggUL+PDDD698EIRoKHePPck5sgw8A2DUC5xNns3fthex6K3d1JotzI6GKeWHMcx9h+oTJ6iuj0WhDz+E79ixuAUFOfpduJSaSjNbvsjg4KbTdO0fzk2zEmVq6Tao3bUS11bArkWw5a9QdgoiepE/7l3eOt2TpctPcmPm57yfs4WQwlMYw0IJeupJgu69F7dgmVXMarax+YsM9q3NITTGl3GzexIU2f5mnXMWkuxcr7NZsPoV2L8YvEPgJ6/DgFlgdMdSVETFuu8p/m4VlZs24Wauo4u7F+XJgwidcjvBY1Iw+vk5+h24LKvZhs2qUQYwuhtpqcoST0/P87eNRiPV1dWXPGY0GrFYLOe3p6en4+vrS25u7vlt48ePJy8vj4EDBzJ37lyAi2p4zt0ePnw4J06cYN26dVitVnr16tUyb0y0Lbm7Yd0f7N1MvAIg5deUJj/AhzuKWfDWLkKKcvm19TiDMnehv8zEYjDgPXgwwfffj9/NN8kFyjWwWmzs//4UO5ZlUldtof/4WIZO7CoDjoVrqyyC7R/Atg+gpgQ63UjmsFf507EYtv/7KHdkzuej7G14VZXjmZRIyC+exG/CBJlAqd7ZM5V8O+8AhdkVJI+J5oY742U2WAeTZOdaVRXbVwTePgeUAUY8B8OfofZ0ERULF1G+eg3Vu3eD1hR6B7IlZhBeo1KY+tBEIoIlwbmSq7XAWMxWygpqsJivr9vaD40YMYJZs2bxwgsvoLVmyZIlfPTRR8yZM6dJr1NaWspTTz3F+vXreeKJJ1i8eDF33303K1euvGTfTz/9lBdeeIFPP/2UYcOGnd8+c+ZM7rvvPl588cXrfl+ijTu1yz4m5+gK8AqE0b+hrM9s5u8sYvn/rqB/5i7+VnSA0PxsUArTwIH4z/opfjffjFto6FVfXlxKa03m3kI2f55BaUE1MYlBDL+7GyFRvo4umhDXruSkfVKlXYvAUo3ucSt7O83iTwf8KVi4m7szP+fxnD0YtA2/m8YS/LOfYRowQLpl1dNac3jLGdZ/ehQ3NwO3PpZM52SJsc5Akp2mstTaE5z1r0NNKTp5GtVhd1KxbR/lr99LXWYmALVd4vmu7wRWBHQnalAffn1rIj0iZeaN69WSs63179+fWbNmMXjwYMDenSzoGrrzPPvsszz++OMkJCQwb948Ro8ezciRIwm/zPpIZ8+eJTk5GU9PTz7++OPz26dPn85vf/vb893chLhETqp9Svv0b8EUBGN+S0Xf2Xz6zX6y3v0/Bmft5q1Se8uiacAA/B+aid/4cbJO13UqOFnOxs/SyU0vISjSm9ue6ENsz2C54BOuK++gfWa1tMWgFLbeU/g+9D7eSLUSsGwzU09spFtBJsrHh6CfziBoxnQ8YmTNqIbqqi2s+9cR0nfkEZUQyE3398Q3yPPqTxStQmmtW++PKeUDfA+8pLX++mr7Dxw4UO/cubPlC9YYFQX2xbJ2zsdWlEOlGkR5ZTcqtu3DWlRkH9Q7eBCl/YbxVnUkq4oNJET48uufJJLSXS4urubQoUMkXmHtDm3TlJ+toabCjLunEf/Qa5ttzZl07tyZnTt3EnqZ2vXFixezdOlSPvroIweU7ILLnRelVKrWeqCDinTNnCqeXI/sHfYkJ2MVmILhhicoC5/Ahr//B9atJq44GwCd1JvIibfiN3487pGRDi6066s4W8PWpcc5su0MJl93Bt/WhaQbO2IwunYcciSJJQ6WtcWe5BxdAe4+WPrNZKnpTuZsLKTXnnVMztpMcEUxbtHRhMycScDkOzH6SuvlD+WdKOPbeQcoL6xm8O1d6H9L5x8dPyxazpXiSaNadpRS84HbgHytda8G228B3gKMwFyt9e+v8lLPA/9uVKmdgdaQvQ12zMWS+hUVOQbKz0ZReTIWXZuDwbcE35Ej8R07hrLkgby+KZele3IJ9fXgtTu7M2VgNG7yj/C62butVWMx25q125qzevLJJ/nmm29Yvny5o4sinEn2dlj3ezi2GkzBmPs+R/GZULJeW4lf5lzigNwOcVgefpIe906SCU+aSV2Nhd3fnWTPtyexaU3/cbH0v6UznjJ1rHBFNpu9NXjjm5C9FbxDqLnxBf5hG8eXqw4zIm0Of8hJxdNci2nwYEJ+NhPflBSZlfEytM0+A+PWL4/hHejBnc/1p0N8oKOLJS6jsdF6IfBXYNG5DUopI/AucDOQA+xQSn2FPfH5vx88/wGgD3AQ8Lq+IreC2nL03k+oWzmX8rQcKk77UF0YAhrcOgQQePcYfMeMxmfQICpsivfWHWPeB7tQwBOj43k0pSu+nvKPsDnUVJgpL67vthbu3aYuMBrO6tbQO++807oFEc7t5FZ7knN8LWYdQjl3UZJaQe0Ce7fH04HRbB85hRtm38PYITKZRXOx2TSHt5xm21fHqSqtI35gOMMmdcU/1OToognRdFYz7P/cvkZOwSEIiKVs9Gu8XzKU3Z9t4pbDr/KnvENoN3cCb7+V4Jkz8bpCb4v2rqqsjtULD3LyYDFx/cIYPaMHXj6yRISzatSVo9Z6vVKq8w82DwYytNbHAZRSnwATtdb/h70V6CJKqRTAB0gCqpVSy7XWtmsvevPTp/ZR9dmfqfh+A+XZRswVboA/Xj2TCJ02Fr8xY/Ds3h2lFBarjX/syOYv3x2lqLKOO/tF8Yvx3ekYKP8Im4PNaqOipLZNdVsTokmytsD3v8d8YD3leaGUFfWnOv0MsIWs4GjWJE2gfMhIZt0zgolxIY4ubZuSc7iYjYszKMqpIKKLPxMe6U1kXICjiyVE09VVwq6P7NNHl2ZDeBJnbnqbt7MSKPlgBbdn/Jw7ys6gg4IIfeIJgu6dKhOXXEX2wWK+W3iQumoLo+7rTs8RHdt0b5O24HqqyaOA7Ab3c4AhP7IvWuvfACilZgGFP5boKKUeBh4GiI2NvY7iNY61pIjKT9+mfOUyKjLKsdUZUG5eePdPJuQnk/AdPRr3iIiG74M1h/N4bflhMvIrGNwlmAW3JpIcHdjiZW0PLHVWqsrrqKm0gNbtotuauDZNHQPoMk5uxfL1y5Rv2kPZKX+q8iJBQ3WMiRX9bmNZcBKRvbrz3M0JDOsaIt+NZnT2TCWbvzjGiX2F+AV7Me7BnsQPCJdj3A60qXhiroaM1XBwqX08Tm0ZxN5AxuBXeHePH75/+oo7Mt8hoK4SQ7cEIl54Cv/bbpWpo6/CarWx/avj7Fp5kqAOPkx8uq/MwOgiWr1PkNZ64VUen6OUOg3c7uHhMaAlymDOzaV82edULPuCyqOnwaYweoHfwER8J/8M39HjMPhcWPzpbGUde3NK2JtdysaMAnacOEuXUB8++OkAxiVFyD/C66S1xmK2UpJXRV2NBZTCy8cNbz8PWZivDWq3YwCvwFZbS/WaL6j64n0qD+VQXeQBOhCPuC7k3zmC93QntukA+sQE8trNCYzsFipxpxlVV9Sx4+sTHFh/CqOHgWF3diV5TDRu7hJ/nJ3Ek3p1VZDxHRz4Eo6uBHMlmILQSRPZFXIbi9aUEP/xFzxwag9GbcNz5CgiH7gf78GDJJY0QllhNd/OO0BeZhlJIzpy4z3dcJfrE5dxPcnOKaDh3IPR9duuW3OvUqy1pubAQSpWr6J8xX+ozbQX08PPQvDQjvhNvA/TTx5AubtTVWdh56ky9uXksSe7hH05pZwsrgJAKYgP8+V3tycxfUgnPKRb1XWx1Fk5su0Me1dnE3+TJ5ZgKz6Bnph83WWGo7ZtIe1pDOBlaIuFmgMHqNy6jcoNa6jesw9t0aA0Xl1iCJ58B7u6DeFPh80cL6qiV5Q/829OYHR3aWVoTlazjX3rcti5/ATmGgs9R0Qx6LYuePtLDbcLWUh7jSe1FZC+0t6Ck/4dmKvAOxRLr3s44DeCdce8Ob18H/33v82jRZlYPU0ETptG+M9+ikcr9JxpK9J35LHun4dBKcY/1Iv4ATLDrqu5nmRnB9BNKdUFe5JzL3Bfs5SqGdhqa6nato3yNWuoWL0aS0EhKDCF1hI+2A3fCZMwjH+cI9WBrMopYd/SQ+zNKeFoXjm2+tm4owJN9IkJYPqQWJKjA+kdHSATDzSDypJa0tblcGBDLjWVZkJjfPHydSeko2+7Xnl84cKFjBs3jo5tfBat9jIGsCFts1Gbnk7V1q1Ubt1G1Y4d2CoqAPAMNBMUb8U7ZRzGqb/hm1wDf/v+GOmbSukR6SctyC1Aa82xXQVsWZJBWWENsT1DuOGuroR0lC4prqbdxZOaMnvXtINL7dPPW2qwuEdQ5DGWo6XhnNpfjs+Sg8SUrmGczQJAXVgkob/8JcH33I3RTxY1byxzrZUNnx7l0ObTRMb5c/MDPWWCEhfV2KmnPwZSgFClVA7wO631PKXUE8BK7LUl87XWB5qjUEqp24Hb4+Pjm/Q8S3ExFd+vp2LNGio2bURXVaM8DPhGVOM7pAq3/r052HUqH1kGsju3kgNvHKTWYo9nQd7uJEcHMq5nJH1jAkiODiTUVxaEak75WWXsXZ1Nxs58bFrTJTmUvjfF0CE+kMOHD7frRMdqtbJw4UJ69erV5pOdH9EmxgA2KB/mrCx7y822rVRt2461uBgAj9gY/PuE4WM4g3dYDW43zuJE0n8xZ38Ni+ccprTaTLdwX969rz8TekXKeg3NLC+zjE2L0zl9rJTgjj7c/lQfYpNkgoc2ptnjiaNiCQDVJXDkG/SBL6nbs47aIqiuCqKkIoGq/Dq8ysqAVEIBTy8/qmK6YBk1nKhBffDrlYRnfLxMHd1EhTnlfDv3AGfzqhhwSycG3d4Fo/Q4cVmNnY3tssu4a62XA82+GEhTurHV5ZyifOUKytespXr3brDZcAvyxS/OSkBIEcYObmzwvYn3Kkax53QknAaT+xl6RwXw06Gd6BMTSJ/oQGKCTVJz2gJsNk3m3gL2rs7mdEYp7p5GeqVEkTw6hoAw56sheeONN5g/fz4ADz74IJMmTWLChAnceOONbN68maioKJYuXYrJZCIlJYUhQ4awdu1aSkpKmDdvHiNGjODNN98kLS2N+fPnk5aWxrRp09i+fTve3t4X/a3OnTszdepUvvvuO/77v/+bnTt3Mn36dEwmE1u2bMFkcr7j42waMwYQmAP2hQBbsizmM2eo3LqVqi1bqdy2DcuZMwC4RUTgO2IE3gP74uN2EPfDC6C2HGvvqXzf8UHm7reyef0h3AyK8b0imT4klmFxMvFAcysvrmHLkmOk78jD5O9ByvTuJA7vKMmkOO9K8aQ1YwmAtSCb2lUfUbNtFbUZx6k5a6S21ANtCbY/rgxk+xrJDu6GYVA3OvTrTXLKQHp0jZHYcR1sVhv7159i8+fH8PRx446n+xLTI9jRxRLXySn7ZDWlZady8yby//Q6tk6xMDiKqICDBATmclB34h3rT1lWN5xYjzCSkwOZVt9i0y3cVxb7bGF11RYObspl39ocyotq8AvxYvjd8SQO7+i0a+WkpqayYMECtm3bhtaaIUOGMGrUKNLT0/n444/58MMPmTJlCp9//jkzZswAwGKxsH37dpYvX87LL7/MqlWrePrpp0lJSWHJkiW8+uqrfPDBB5ckOueEhISwa9cuAObOncvrr7/OwIEut6B4c2ixMYDX2lJ8NZbiYqq2b6dyy1aqtm6lLisLAGNQEN5DhuAzdAg+Q4fiHhWJSl0A638DVYVUxd3CJ34/470DHhTuyCc6yMQvxndnysAYwvykNbm51dVY2LUiiz2r7RX9AyZ0ov/4Tnh4OWccEs2iReJJS8QSrTXmnBxqDh+mNm03tbs2UZORibnEfH4fm6cfJwMi2BPbhczAjrh1607CkGRG9OzIzdGBcj3TDGoqzBzclEva9zlUFNfSqVcIY3+WiMlPxu+1BU4Z7ZvSsrPTaCbgVl8G+22lVrvzvftwDkTdQ0D8DdwTG8RLHf3xkhl1Wk1pQTX71mZzaPNpzDVWOsQHMPzueLr0CWt0DeqZ116j9tDhZi2XZ2IPIn/96yvus3HjRu6880586mfimzx5Mhs2bKBLly707dsXgAEDBly0GOjkyZMv2W4wGFi4cCHJyck88sgjDB8+/Ef/5tSpU6/9TbUtLTYGsLkmPLFWVFC1Y8f5cTe1R44AYPDxwXvQIAKn3YvPsGF4duuGMhjAaoF9n8CS30NpNsXhQ3nX//8x/1AoCs2YHoFMHxrLyG5hGKV1odnZrDYObbYvClpdbiZhcARDJ3XFL9i1xqCLa9Ii8eR6Y4mtqora9HRqDh+h9sjh879tlVXn/gIeflYMwW7kdo7lG6/erPPvi1tEJCO7hzEyIYz/ig8l0FsuwJtLQXY5aWtzOLojD6vZRlT3IEZMSaBLcmi77lrf1jhlstOU2pPuHTwJCKvkeOLzhA5/gHGhkYxr+SKKBrTWnM4oYc+qbDL3FWJQivhB4fQZE0N4J39HF++6eXpeqG03Go1UV1df8pjRaMRisZzfnp6ejq+vL7m5uee3jR8/nry8PAYOHMjcuXMBzidW7YmrjAG01dRQvXs3lVu2UrltKzX7D4DVivL0xNS/H2HPPIPP0CF49eqFcmsQSrW2Dx5e879QeJQ83yRedf8dX53sTrifJ0+OieXeQTGyAHELOnmwiE2LMyjOraRDfAC3Pt6NiM6uH4vEpVoznjQllmiLhYoNG6g9csSe1Bw+bG/91fYecAYvDzxDFP4divEKNFMWGswy7wF8pYdx3NiJIV1CGNktjIcSwkiI8JWuac3IarWRuaeQfWvt3evdPAz0GBpJ75RoWTenjXLKZKcptSexQyfD0MkEGKQZt7VZLTYyduaxZ3U2hdkVePq4MWB8J3qnROMTeO3dca7WAtNSRowYwaxZs3jhhRfQWrNkyRI++ugj5syZ06TXKS0t5amnnmL9+vU88cQTLF68mLvvvpuVK1de8Xl+fn6Ul5dfz1twCc48BrA6LY2KDRuo2rqN6t270WYzGI2YkpMJefghfIYMxdSvLwbPy3y+tYbja9GrX0Hl7ua0eydeMT/LN4UDGdEtjL9N7MTYxHDcpctJiynKrWDz5xmcPFCMf6gXtzzci7h+YXKh2Ia1ZjxpUsuOwcCp536OrqrCPSYGr66x+PcJx10fw9t2EHcfK8cNsXxcdzPLbYMhrAcju4XxQkIoQ7qEYJI1XJpddXkdBzbmsv/7U1SW1OIfau9e32NYB7x83B1dPNGCnDLZaRJJclpddUUdB9afIm3dKarK6giK9CZlencShkS69CJb/fv3Z9asWQwePBiwT1AQFBTU5Nd59tlnefzxx0lISGDevHmMHj2akSNHEh5+5bn5Z82axaOPPioTFDhQ0Zw5lK9ajWdiD4JmzMBn6BBMAwZi9L1KC1z2DszfvoR79kbyVBivmx9hnWEMd93YiXWDYukc2v5a8FpTVVkd27/O5OCGU7h7uTH87nh6j4rG6C7/H4RjKIOBzn/7E4bSVOqOfo1vwWcAHLB1Yrn1LjaoYUR17cOohDD+nhBGlLT0tpj8rDLS1uaQvjMfq8VGTGIQo+7rTqdeITJBSTuhtG7xSUWu2cCBA/XOnTsdXQxRr/h0JXtXneTIdnvf1tiewfQZE0NMUvB115weOnSIxMTEZiqpaC6XOy9KqVSttcvMotCg68lD6enpV9y3Ljsbg68vbo1McnXeAUq+/h1B2d9RpP15xzKJozF3M3VoPLf0isTTzXWTf2dXW2Umc18hx1LzOXmwGDT0HBXFoFs7Y/KVMQ2uoC3HkprKMoyvx+Guzey1xbHCNpjM8JtISOzDqIRQ+sjEAi3KarVxfFcB+9bmcOZ4KW6exvNd1YI7SOVTW3SleOKULTstNXuSuDb5WWWkrsji+J4C3NzsfVuTR8cQ3FEChnB+Tel64hETc7VdACg7nUH+V78j7vQyjNqLd9RUyvs+xPQbetAtQhbtaym11RZO7C0gIzWfk4eKsVk0vsGeJI+OJunGjgRFSkwSLacpscTLx58/BzxPXXgvevfqwyMysUCrqCqr48CGU+xff4qq0joCwkzceE83etzQwWlnghUtzynPfHPNniSundaa3PQSUldkkX2wGE9vNwZO6EzymGipNRXtktaag0ePUrLy/xhU9BXRGPjSezKGEc/y4MAk6WPfQs4nOLsKOHmwyJ7gBHnSOyWa+AHhRHT2lzE5wik998wvHF2EdiMvs4x967LJSM3HZtHE9gym94xoOvUMkVnVhHMmO8JxtNZkpRWRuuIEZ46XYfL3YNidXek1MgoPqRURLuh6W4oray18s/MQto1vcXvVUtywsivkNgJu+Q2TE7o3b2EFYF+nK3Nfob0F54cJTv/6BEcuYEQrk14nzsVqsZGRmk/auhzyMstw9zLSc0QUySnRBEZcfm070T7J1asAwGbTHEvNJ3VFFkWnKvAL9mLkvQkk3tABN6mxFi7sWluKD50u47PNR/DbN58H+JIAVUVmhwmETnyZIR0kyWluP5rgjGrQgiMJjnAg6XXiHCpLazmw/hT7N+RSXVZHYIQ3I6Ym0GNopFTKistyyk+F1J60HqvZxpFtZ9i1MovSgmqCIr0ZOyuRboMiMMrgSdHO1JitLE87zadbj5Fw6guedPuScFVCSfQY9K0v06VDsqOL2KY0THCyDxZjtdgkwRFCXEJrbe+qtjaHY6n52LSmU68QklOiiUkMljghrsgpkx2pPWl55lorBzfmsvu7k1SW1BIW68ctj/Qirk+YBA3Rbj38922EHP+KNz2/oKN7HubooTDuZQJjhzq6aG1GXY2FE+dacA7YExyfQE96jYyi64BwIrtIgiOEsLOabaSn5rFvTQ4FJ8vx8DLSOyWaXilRBIZLVzXROE6Z7DRFSV4VWfuLiOsXhl+wl6OL4/RqKs3s/z6HvatzqKk007FbIGNm9rDXjMgg30bZuXMnixYt4u2333Z0UUQjNKWl+C/6dYI9VqMjesPYd3GPvwnke3HdfizB6TmyI/EDIiTBES5Bep20noqztexfn8PBjblUl5sJivRm1LQEEoZE4uHl8peuopW5/Cfm5MEiNn6WzsbP0gnv5EdcvzC69guXwWk/UFVWx97VJ0n7/hTmGiude4fQ/5bOdOga4OiiuZyBAwcycKDLLA3R7jWlpTh4xINgmY5KulMWLL5OdTUWTqQVkrGzQYIT4GFPcPqHExkXIAmOcCnS66Rlaa05fayUtLU5HN9dgE1rOvcOJXl0NNE9gqRCVlwzp0x2mlJ7kjw6htikEI7vKeDY7gK2fnmcrV8eJ7ijD3F9w+jaP4yQKN92+yUpK6pmz7cnObj5NDaLjfgB4fS/pROh0bIWyA9VVlYyZcoUcnJysFqtvPjii8TFxfH0009TWVmJp6cnq1evJjU1lddff52vv/6al156iWPHjpGRkUFhYSG//OUveeihh5g5cyaTJ09m0qRJAEyfPp0pU6YwceJEx75JcWU9fuLoEri0cwnOsdQCsg4UYTXXJzgjOtJ1QDgdJMERQvyApc5K+s489q3NoTC7Ak9vN5LHRNNrVDQBYSZHF0+0AU6Z7DS19iQwwpv+4zvRf3wnyotrOL6ngOO7C0j95gQ7l5/AP8xE175hxPULazcDXotPV7J7ZRZHt+eBgh5DI+k3rpO0eF3BihUr6NixI8uWLQOgtLSUfv368emnnzJo0CDKysowmS4NvPv27WPr1q1UVlbSr18/br31VmbPns2bb77JpEmTKC0tZfPmzfz9739v7bckRIurq7GQlVZERmr++QTHO8CDnjdKgiOE+HHlxTXs//4UBzfmUlNpJrijDynTu5MwOBJ3T5kFVjQfp0x2rodfsBd9xsTQZ0wMVWV1ZO4t4PieAvauyWb3dyfxCfQkrj7x6RgfgKGNzTiWn1VG6oosju8pwM3NQO+UaPreHINvkOuMZ9rw76MUZlc062uGxvgyYkrCFffp3bs3zz33HM8//zy33XYbgYGBdOjQgUGDBgHg7+9/2edNnDgRk8mEyWRi9OjRbN++nUmTJvHYY49RUFDA559/zl133YWbW5v7uol26nyCsyufrP0XEpykG+1d1Dp0lQRHtD0yZuf6nVuwPG1tDsf3FoLWdOkTRu/R0UQlBLbbXjiiZbXpqy9vfw96joii54goaqvMnEgr4vjuAg5tyiVtXQ5evu506RNKXN8wYnoEY3R3zcTnXPBIXZFF9sFiPExuDJzQmeTR0Zj8PBxdPJeRkJDArl27WL58Ob/97W8ZM2ZMo573w+B87v7MmTP5xz/+wSeffMKCBQuavbyiceQCpXnU1VjI2l/fgnMuwfH3IGl4R+IHSIIj2j4Zs3PtzHVW0rfbu6oVnarA08eNfjfH0HNkFP4h0lVNtKw2new05OntTvchkXQfEom51srJA0Uc211ARmo+hzadxsPLSKfeoXTtF0ZszxCXaELVWpO1v4jUb7I4c7wUk587w+7sSq+RUS69sNbVWmBaSm5uLsHBwcyYMYPAwEDee+89Tp8+zY4dOxg0aBDl5eWX7ca2dOlSfvWrX1FZWcm6dev4/e9/D8CsWbMYPHgwkZGRJCUltfbbEfXkAuXamWut9WNw7AmO5aIEJ4zIroEYJMERQjRQV2OhrLCGsoJqSgurKcmr4tiufGqrLIRE+zL6pz1IGBQhC5aLVuO6V8TXwd3TSNf+4XTtH47VbCP7cDHHdxeQubeQ9B15uLkbiO0ZQly/MDr3DsHT293RRb6IzaY5tiuf1BVZFOVU4Bvsych7E0i8oYMEj+uQlpbGL37xCwwGA+7u7rz//vtorXnyySeprq7GZDKxatWqS56XnJzM6NGjKSws5MUXX6Rjx44AREREkJiYeH6SAiGclbnWSmVJLZWl9T8ldeRllpKVZk9wTP4eJN7QwT4GJ14SHCHaM601VaV1lBZWn09oygqqKS2opqywmupy80X7e3q7Ed0jiOTR0XSIl65qovW1y2SnIaO7gc69Q+ncOxSb1UZuRinHd+XbJznYU4DBqIjuEUTXfuF06RPq0G5hVouNI9vOsGtFFqUF1QRFejP2Z4l0GxyBsY2NPXKE8ePHM378+Eu2b9269aL7KSkppKSknL+fnJzMokWLLnleVVUV6enpTJs2rdnLKkRjWOqsVJbW1ScwtVSV1tmTmjJ7QlNVv72uxnrJc03+HvS4oYO9i5okOEK0K1azjbKiCwlMWUGNPampT2wsZtv5fZUC3yAv/MO86JIcin+YCf9QEwH1v718nKvCWLQ/7T7ZachgNBDdPYjo7kGMmJpA3okyju0u4PjufNb+4zDr/gkd4gMJCDfh4eWGh5cRD5MbHl5uuDe4fW67u6cRDy/jdU+CYK61cnBjLntWnaTibC1hsX7c8kgv4vqESR95J7Vq1Spmz57Ns88+S0CArGXkKqor6tA2MLopjG4GjG4Gp/yOWS02Kkvrk5f6lpjK0lqqzrfO2JOa2irLJc81uhnwDvDAJ8CT4I4+xCQG4xPoiU+AB96BnvgEeOIT6ImHl1FqYIW4DrXVFoxG5ZRxRGtNbaXlklaZc78rSmpBX9jfzcNAQJg9gYlJCiYg1IR/mImAUBN+IV4Y3aTCVTgvp0x2nGFAsTIoIuMCiIwL4IbJXSk6VcGxXQVk7S/i5P4i6mqtmC9TG3o5bh4GexJksidC7j9IlC7crn/s3G1PIycPFLN3TTY1FWY6dgtk9IwexCQFy0WIk3jppZcuu/2mm24iKyurdQsjrtu6fx7h+O6Ci7YZjBcSH6Obwuhef/vc74tuq4u3/+AxN3cjRjeFof4xt4teQ2F0M2J0V1jNuj6ZuZC4NExqairMl5TdYFD2JCbQk8Bwb6K6BTZIXjzOJzGe3m4SP4S4Bk29Nln4y43nW0AMBoXBvT5GGBvGhgvxxdAwzvzgtuGiuKIujjENYorR3XA+wTK6GzAYDVRX1F0moamhrvriyhBvfw8CwkxEJQTVJzJe+Id5ExBmwuTnLnFDuCynTHacbUCxUorQaD9Co/0Yckfc+e3apjHXWqmrsVBXXf+7wW1zzbnHLNTVXNjPXGOhrLDOvl+tfZu26R/9+516hzBgfCc6xAe2wrsVom1pygVKzxEdie4ehNViq//RWM22BvdtDe7r8/frqi0XP2a++HHbFb7fVy+//SLEJ9ATvxAvIrsG4FOf1JxLZLz9PTH5ujtd7bEQbUlTr02GTuqKxWy9EAssNmxmG1ar/kE8sd+3WWyYaywX7d8wFtmuM5YYjAr/UBP+oV50iAu4pLuZK0zMJMS1cMpkx1Uog7K3wpjcIOjaX0dr+wVVXY21PjGynL8dEGYiJMq3+QrtxLTWUnPkRLS+9n+qzqQpFyixSSHQAhPnaduFixdLfUJkq7+gsZgvTaQMRgO+gZ54B3hg8vOQ8TJCuKA+Y2Oa/TW1TWO1Xr4ixnZJkmTfx8vXnYAwEz6BnhJLRLskyY4TUErh5mHEzcOIt3/7XBfHy8uLoqIiQkJCJOFxAlprioqK8PJyncVonZkyXPiOezq6MEIIl6UMCjeDETd3QJanEaJRJNkRTiE6OpqcnBwKCgquvrNoFV5eXkRHRzu6GEIIIYQQ10ySHeEU3N3d6dKli6OLIYQQQggh2hCZK1AIIYQQQgjRJkmyI4QQQghxFUqp25VSc0pLSx1dFCFEE0iyI4QQQghxFVrr/2itH5aFooVwLcqZp5dVShUAWUAAcKWqlGt9/HLbG7MtFCi8wt9rTld7b835/Mbse6V9mnKcL7fdkcf5cn+/JZ/fmse6uT/TnbTWYY3c12k0iCdX4+jvAbT+Z/9yrvf70Byv5azfqR/bLuetac9r67EEWvf65ce2y+ey6c+T82bnSuftx+OJ1trpf4A5LfH45bY3Zhuw01nee3M+vzH7XmmfphznHzmuDjvObflYO9tn2tl/HP09cJbzcb3fh+Z4LWf9Tsl5a70yuPpPa16//Nh2+Vw2/Xly3lzzvP3Yj6t0Y/tPCz1+ue2N3dZarvdvN+X5jdn3Svs05Thfbrsjj3Nz/H1nPdbO9pl2do7+HjiL5izXtb6Ws36nGvv3HMHVzltb15rXL415PUdxtc+lnDc7Vztvl+XU3diclVJqp9Z6oKPL0dbJcW49cqydi5wP1yTnTTgj+Vy6JjlvzcdVWnaczRxHF6CdkOPceuRYOxc5H65JzptwRvK5dE1y3pqJtOwIIYQQQggh2iRp2RFCCCGEEEK0SZLsCCGEEEIIIdokSXaEEEIIIYQQbZIkO9dJKRWnlJqnlFrs6LK0dUqpSUqpD5VSnyqlxjm6PG2ZUipRKfU3pdRipdR/Obo87Z3EGdckMUs4I4knrkdiyfWRZOcylFLzlVL5Sqn9P9h+i1LqiFIqQyn1AoDW+rjWerZjSur6mnisv9RaPwQ8Ckx1RHldWROP9SGt9aPAFGC4I8rb1kmccU0Ss4QzknjieiSWtB5Jdi5vIXBLww1KKSPwLjABSAKmKaWSWr9obc5Cmn6sf1v/uGiahTThWCul7gCWActbt5jtxkIkzriihUjMEs5nIRJPXM1CJJa0Ckl2LkNrvR4o/sHmwUBGfY1IHfAJMLHVC9fGNOVYK7s/AN9orXe1dlldXVM/11rrr7TWE4DprVvS9kHijGuSmCWckcQT1yOxpPVIstN4UUB2g/s5QJRSKkQp9Tegn1LqV44pWptz2WMNPAncBNytlHrUEQVrg37sc52ilHpbKfUB0rLTmiTOuCaJWcIZSTxxPRJLWoCbowvg6rTWRdj7UIoWprV+G3jb0eVoD7TW64B1Di6GqCdxxjVJzBLOSOKJ65FYcn2kZafxTgExDe5H128TzU+OdeuRY+1c5Hy4JjlvwhnJ59L1yDlrAZLsNN4OoJtSqotSygO4F/jKwWVqq+RYtx451s5FzodrkvMmnJF8Ll2PnLMWIMnOZSilPga2AN2VUjlKqdlaawvwBLASOAT8W2t9wJHlbAvkWLceOdbORc6Ha5LzJpyRfC5dj5yz1qO01o4ugxBCCCGEEEI0O2nZEUIIIYQQQrRJkuwIIYQQQggh2iRJdoQQQgghhBBtkiQ7QgghhBBCiDZJkh0hhBBCCCFEmyTJjhBCCCGEEKJNkmRHCCGEEEII0SZJsiOanVLKzdFlEEK0DRJPhBDNQWJJ+yXJjmgSpVRnpdQhpdSHSqkDSqlvlVImpdQ6pdRflFI7gaeVUmOVUruVUmlKqflKKU+l1CCl1Bf1rzNRKVWtlPJQSnkppY7Xb39KKXVQKbVPKfWJQ9+sEKJFSTwRQjQHiSXiSiTLFdeiGzBNa/2QUurfwF312z201gOVUl5AOjBWa31UKbUI+C/gr0Df+n1HAPuBQdg/h9vqt78AdNFa1yqlAlvl3QghHEniiRCiOUgsEZclLTviWmRqrffU304FOtff/rT+d/f6fY7W3/87MFJrbQGOKaUSgcHAG8BI7MFlQ/2++4B/KqVmAJaWfBNCCKcg8UQI0RwklojLkmRHXIvaBretXGghrGzEc9cDEwAzsAq4sf7nXEC5FXgX6A/skD62QrR5Ek+EEM1BYom4LEl2REs4AnRWSsXX3/8p8H397Q3AM8AWrXUBEIK9tmW/UsoAxGit1wLPAwGAb2sWXAjhdCSeCCGag8SSdkoyU9HstNY1Sqn7gc/qaz92AH+rf3gbEIG9FgXsTcORWmtdv+8/lFIBgALe1lqXtG7phRDOROKJEKI5SCxpv5TW2tFlEEIIIYQQQohmJ93YhBBCCCGEEG2SJDtCCCGEEEKINkmSHSGEEEIIIUSbJMmOEEIIIYQQok2SZEcIIYQQQgjRJkmyI4QQQgghhGiTJNkRQgghhBBCtEmS7AghhBBCCCHapP8PWZPh4w5zEJ0AAAAASUVORK5CYII=\n", "text/plain": ["<Figure size 1008x216 with 3 Axes>"]}, "metadata": {"needs_background": "light"}, "output_type": "display_data"}], "source": ["import matplotlib.pyplot as plt\n", "fig, ax = plt.subplots(1, 3, figsize=(14, 3))\n", "for i, ncol in enumerate([10, 50, 100]):\n", "    piv = df[df.ncols==ncol].pivot(index=\"nrows\", columns=\"name\", values=\"average\")\n", "    piv.plot(ax=ax[i], logy=True, logx=True)\n", "    ax[i].set_title(\"ncol=%d\" % ncol)\n", "ax;"]}, {"cell_type": "markdown", "metadata": {}, "source": ["Curves are not linear and rather difficult to interpret. The algorithm *numpy-lower* and *scipy* should be close as the cost of both algorithm are similar. However, *scipy* reduces the number of trips between C and python. The C implementation of the distance is here:\n", "[sqeuclidean_distance_double](https://github.com/scipy/scipy/blob/master/scipy/spatial/src/distance_impl.h#L50). The final cost is a combination of computation, multithreading, allocations..."]}, {"cell_type": "code", "execution_count": 30, "metadata": {}, "outputs": [{"name": "stderr", "output_type": "stream", "text": ["C:\\xavierdupre\\__home_\\GitHub\\pyquickhelper\\src\\pyquickhelper\\pycode\\profiling.py:541: FutureWarning: The default value of numeric_only in DataFrameGroupBy.sum is deprecated. In a future version, numeric_only will default to False. Either specify numeric_only or select only columns which should be valid for the function.\n", "  df = df.groupby(['namefct', 'file'], as_index=False).sum().sort_values(\n", "C:\\xavierdupre\\__home_\\GitHub\\pyquickhelper\\src\\pyquickhelper\\pycode\\profiling.py:541: FutureWarning: The default value of numeric_only in DataFrameGroupBy.sum is deprecated. In a future version, numeric_only will default to False. Either specify numeric_only or select only columns which should be valid for the function.\n", "  df = df.groupby(['namefct', 'file'], as_index=False).sum().sort_values(\n"]}], "source": ["from pyquickhelper.pycode.profiling import profile\n", "M = numpy.random.rand(100, 10)\n", "\n", "pr1, df1 = profile(lambda: [squareform(pdist(M, metric='sqeuclidean'))\n", "                            for i in range(0, 1000)],\n", "                   as_df=True)\n", "pr2, df2 = profile(lambda: [custom_pdist_lower(M)\n", "                            for i in range(0, 1000)],\n", "                   as_df=True)"]}, {"cell_type": "code", "execution_count": 31, "metadata": {}, "outputs": [{"data": {"image/png": "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\n", "text/plain": ["<Figure size 576x216 with 1 Axes>"]}, "metadata": {"needs_background": "light"}, "output_type": "display_data"}], "source": ["ax = df1[['namefct', 'cum_tall']].head(n=15).set_index('namefct').plot(\n", "    kind='bar', figsize=(8, 3), rot=30)\n", "ax.set_title(\"scipy\")\n", "for la in ax.get_xticklabels():\n", "    la.set_horizontalalignment('right')"]}, {"cell_type": "code", "execution_count": 32, "metadata": {}, "outputs": [{"data": {"image/png": "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\n", "text/plain": ["<Figure size 576x216 with 1 Axes>"]}, "metadata": {"needs_background": "light"}, "output_type": "display_data"}], "source": ["ax = df2[['namefct', 'cum_tall']].head(n=15).set_index('namefct').plot(\n", "    kind='bar', figsize=(8, 3), rot=30)\n", "ax.set_title(\"numpy-lower\")\n", "for la in ax.get_xticklabels():\n", "    la.set_horizontalalignment('right');    "]}, {"cell_type": "markdown", "metadata": {}, "source": ["[Universal function](https://docs.scipy.org/doc/numpy/reference/ufuncs.html) do not seem to be very efficient in our case. The last graph shows time ratio between implementations of *pdist* and the baseline *scipy*."]}, {"cell_type": "code", "execution_count": 33, "metadata": {"scrolled": true}, "outputs": [{"data": {"image/png": "iVBORw0KGgoAAAANSUhEUgAAAzUAAADkCAYAAAChZgr7AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjYuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8o6BhiAAAACXBIWXMAAAsTAAALEwEAmpwYAACjCElEQVR4nOydd1xV5R/H3+de9t4IgoJbUcGRe+Ys9ypNM7VS04b9smHDli3LsqGVmZqaZaG5MkeWew9wgBNUQNkbLnDH8/vjIGmaE7hceN6v1xXuOeee8714z/c+n+f5DkUIgUQikUgkEolEIpFYKhpzGyCRSCQSiUQikUgk94IUNRKJRCKRSCQSicSikaJGIpFIJBKJRCKRWDRS1EgkEolEIpFIJBKLRooaiUQikUgkEolEYtFIUSORSCQSiUQikUgsGilqJPeEoihvKYqy1Nx2SCQSy0f6E4lEcrdI/yGRokZiFhRFmacoyilFUUyKooy5wf7nFUVJVBQlW1GUBYqi2JrBTIlEYgEoiiIURclTFCW3+DH/qn2KoigfKYqSVvz4SFEUxZz2SiSSisO9jEcURQlSFOVvRVHyFUU5qShK93I1XnINUtRIzEUkMAk4/O8diqL0Al4BugE1gVrA2+VqnUQisTRChRBOxY8nrto+HhgIhAJNgX7ABDPYJ5FIKib3Mh75CTgCeAKvAeGKoniXtcGSGyNFTRVDUZTziqJMVRTlqKIoWYqiLFcUxa543wBFUSKKZyPOKYrSu3i7v6IoaxRFSVcU5ayiKE/eqx1CiDlCiC1AwQ12PwZ8L4Q4IYTIAN4FxtzrNSUSSelSUfzJLXgMmCWEiBdCJACzkP5EIjE7FcV/3O14RFGUekBz4E0hhE4IsQI4Bgy5V5skd4eVuQ2QmIWHgN6oN/AuYIyiKIeBxcBQYAvgBzgXH/8zcBzwBxoAmxVFOSeE+OvfJ1YUJfMm1/1QCPHhbdgXAqy+6nkk4KsoiqcQIu02Xi+RSMqPiuJPtiuKogF2A/8TQpwv3h6C6kOuEFm8TSKRmJ+K4j/+i/8cjxTvixFC5Pxrv/QvZkKKmqrJF0KISwCKoqwFwoBmwAIhxObiYxKK9wcC7YE+QogCIKI4Xn00cJ0TEUK4lYJ9TkDWVc+v/O4MSFEjkVQsKoI/6QzsBRyAGcA6RVHChBAGbuxPnBRFUYQQ4k7eqEQiKXUqgv+4GTcbj/x735X91UvhupK7QIafVU0Sr/o9H/XGDATO3eBYfyD9XzMRFyjbmzYXcLnq+ZXfc25wrEQiMS9m9ydCiO1CiCIhRCbwHBAMNCzefSN/kisFjURSITC7/7gFNxuP/Hvflf1yrGImpKiRXCEOqH2D7ZcAD0VRnK/aVoPimZN/c1X1oRs9Xr1NW06gJvVeIRRIkqFnEonFYG5/IoArFc5u5E9O3PY7kUgk5Y25/cfV3Gw8cgKo9S97pH8xIzL8THKF74FNiqKsA/6mOIZVCHFSUZTdwAeKokwF6gGPAyNvdBIhhNPtXExRFBtUUa0A1sXJgUVCCBNqLO0iRVF+RHVirwOL7uXNSSSScqXc/ImiKCGANWqCrj1q+FkCEF18yGLgf4qirEcVOy8AX97De5NIJGWLRYxHhBCnFUWJAN5UFOV14AHUCouyUICZkCs1EgCEEPuBscBnqDGh21DLFwKMAIJQb+jfUCt9/HmPl9wE6IB2wLzi3zsV27IBmInqzC6iLi+/eY/Xk0gk5UQ5+xNfYDmQDcQUn7uvEEJfvP9bYC2q6DkO/F68TSKRVEAsbDwyHGgJZAAfAkOFECn3aI/kLlFkWLFEIpFIJBKJRCKxZORKjUQikUgkEolEIrFopKiRSCQSiUQikUgkFo0UNRKJRCKRSCQSicSikaJGIpFIJBKJRCKRWDRS1EgkEolEIpFIJBKLpkL0qfHy8hJBQUHmNkMikVzFoUOHUoUQ3ua2406R/kQiqXhYoj+RvkQiqXjczJdUCFETFBTEwYMHzW2GRCK5CkVRLpjbhrtB+hOJpOJhif5E+hKJpOJxM18iw88kEolEIpFIJBKJRSNFjUQikUgkEolEIrFopKiRSCSVAkVR+imKMi8rK8vcpkgkEolEIilnpKiRSCSVAiHEWiHEeFdXV3ObIpFIJBKJpJypEIUCJBWL5PxkzmaepZ1/O3ObIpFIJADk6/PZc3kP+fp8fB188XX0xcfBB3sre3ObJpFIKjhFxiIyCzPVR0EmGYUZ5Bbl0synGbXcapnbPEkpIUWN5BoiUyJ57q/nSCtIY0b7GQyoM8DcJkkkkipKYl4i2+O383fc3+y7vA+9SX/dMa62rvg6+FLNsZoqdooFz5Wf1Ryq4WDtYAbrJRJJWaA36sksVIXJFYGSVZhFRkFGiXC5su/K8zx93n+er4lXEwbUHkDv4N642sqVfktGihpJCb/H/M70XdPxcfChuU9z3trzFgHOAbTwbWFu0yQSSRVACMHJ9JNsjdvK33F/E50eDUCgcyDDGwyna2BXvO29Sc5PJik/icS8RJLyk0jKSyIpP4njqcdJL0i/7rzONs7XiJyrhc8VMeRk41TO71ZSUVEUpR/Qr06dOuY2pdLzb4FSIkruUqA4WjviZuuGu6077nbuBLsGq8/t3HGzdSv53dXWFVutLVvjtrLq7Cpm7JvBzAMzub/G/QyoM4C2fm3RarTl94eQlAqKEMLcNtCyZUsha8GbD5MwMSdiDvOOzqOFbwtmd5mNoiiMWj+KzMJMlj24jECXQHObKSlnFEU5JIRoaW477hTpTyyLImMR+xP3szVuK1vjtpKUn4SCQqh3KF0Cu9A1sCvBrsEoinJb5ys0FpKcn3yd4Cn5mZ9Eqi71utc5Wjtes9Jzo5UfFxuX27ZDci2W6E+kL7kzTMJEmi7txisnxb9nFGaQVZClCpU7ECiudq64214rTK4WKFf2WWut79huIQTR6dGsPrua32N/J6swCx97H/rW7suAOgOo5SrD0yoSN/MlUtRUcXQGHa/tfI3NFzYzqM4g3mjzRolTuJB9gZHrR+Jp58nSB5fibONsZmsl5YklDkJA+hNLIKMggx0JO9gat5VdCbvIN+Rjb2VPO/92dA7oTKeATnjae5bZ9fVGPcm65BsKnsS8RJLykkjRpSC49vvR3soeXwdf7qt2H5PCJuFl71VmNlY2LNGfSF9y+xxKOsRbu9/ifPb5G+6/IlDcbN1ws3O7oUC5ep+rrSs2WpvyfROokyzb4rex+uxqdibsxCiMNPVuWhKe5mLjUu42Sa5FihrJDUnKS+LZv58lOi2aF1q+wOhGo6+bhTyQeIDxm8bTyq8Vc7rNwUojIxarCpY4CAHpTyoqsVmxJasxESkRmIQJH3sfOgd2pktgF1r7tcZWa2tuM0vQm/Sk6dKuW/FJyE1gW/w2bLW2PNnkSUY1GlWh7K6oWKI/kb7k1uTp85h9aDY/n/qZ6k7VebTRo3jZe6mixc6tRKyYQ6DcK6m6VH6P+Z1VZ1dxNvMsNhobutXoxoA6A2jj10aGp5kJKWok13Ei7QTPbnmWXH0uMzvNpHNg5/88duWZlby5+02G1x/Oa21eK0crJebEEgchIP1JRcFgMhCRHMHWuK1si99WMoPbwKMBXQK70CWwC408GllkONeF7At8cvATtsZtpbpTdaa2nEq3Gt0s8r2UF5boT6QvuTm7E3bz1p63SMxLZGTDkTzT7JlKWZRDCEFUehSrz65mfez6kvC0frX70b9OfxmeVs7czJfIafcqyKbzm3ht52u427mz+IHF1Peof9PjB9cdTGxWLItOLCLYNZhHGj5STpZKJBJLIrcol92XdrM1bivbE7aTVZiFlcaK1tVaM7LhSDoHdMbPyc/cZt4zNV1q8uX9X7L70m4+PvAxz299npa+LXm51cs08GhgbvMkkjIlqzCLTw5+wqqzqwhyCWLxA4sJ8wkzt1llhqIohHiGEOIZwtSWU0vC0xadWMT3x7+X4WkVCLlSU4UQQjDv6Dy+iviKUO9QZnedfdsx4UaTkSlbp7A9fjtzus2hQ/UOZWytxNxY4swqSH9S3lzOvczWeDWsbH/ifgwmA662rnQO6EzngM60829XqSuLGUwGVp5ZyVdHviKzMJPBdQfzdLOnZb7Nv7BEfyJ9yfVsubiFGXtnkFGQwbjG45gQOqHKhl+m6lJZd24dq8+t5mzmWWy1ttxf434G1h5Ia7/WMjytjCj38DNFUQYCfQAX4HshxKabHS8dR9lTaCxk+q7prI9dT99afXmr3Vt37Ijy9fmM/mM0CbkJLH1wKbXdapeRtZKKgCUOQkD6k7LGJExEp0WXCJmT6ScBCHIJKgkrC/UOrXL5d9lF2Xwb+S3Lopdha2XL+KbjGdVwlEXmEpQFluhPpC/5hzRdGh/s/4CN5zfSwKMB77R7h4aeDc1tVoVACEFUWhSrzq5ifex6souy8XHwoX/t/vSv3Z9g12Bzm1ipKBVRoyjKAqAvkCyEaHzV9t7A54AWmC+E+PCqfe7AJ0KIx292buk4ypZUXSrP/f0cR1OO8myzZ3miyRN3Hft9OfcyI34fgZ2VHcv6LMPDzqOUrZVUFCxtEHJVb4knz5w5Y25zKhWFxkL2Xd6n5sfEbSNZl4xG0RDmHUbXwK50Duwsv7iLOZ91nlkHZ7E1fisBTgG80PIFmW+D5fkTkGMTUAfsv8f+zkf7PyJPn8fE0ImMbTwWa82dl06uChQZi9gat5XV59TqaSZhItQ7lAF1BtA7qLesIlsKlJao6QTkAouviBpFUbTAaaAHEA8cAEYIIaKK988CfhRCHL7ZuaXjKDtOpZ/i6b+eJqswi/c7vE/3mt3v+ZxHU44ybuM4Gnk2Yn7P+XImspJiiYMQkP6ktEjTpbE9fjtb47ay5/IedAYdDlYOtK/eni6BXehYvSPudu7mNrPCciXf5mzmWe6rdh8v3fdSlc63sUR/UtV9SWJeIu/ufZft8dtp6t2Ud9q9IyM07oCU/JSS6mnnss7J8LRSotTCzxRFCQLWXSVq2gJvCSF6FT+fVnzoh8WPzUKIP2913qruOMqKvy/+zcs7XsbZxpkv7/+SRp6NSu3cG2I38OL2F+lfuz8z2s+o8rOQlRFLHISA9Cd3ixCCmKwY/o77m21x24hMiUQg8HXwLWmCeV+1++Qkxh1gMBlYcXoFX0V8RVZhVpXOt7FEf1JVfYkQghVnVjDr4CwMJgPPNn+WRxo8Igfhd8mV8LTfzv7GH7F/kF2Uja+Db0l4WpBrkLlNtCjKsvpZdSDuqufxQGvgGaA74KooSh0hxDc3MGo8MB6gRo0a92iG5GqEECw6sYjPDn1GI89GfHH/F/g4+JTqNXoH9yY2O5a5EXMJdg3miSZPlOr5JRJJ2WM0GTmcfJi/4/5ma9xW4nJUd97IsxFPhT1Fl4AuNPBoICct7hIrjRUPN3iY3sG9+fbot/wU/RMbzm+Q+TaSCktcThxv736bfYn7aFWtFW+1fYtAl0Bzm2XRKIpCiFcIIV4hvHjfi2p42tnVfH/8e7479h0dq3dkcrPJhHiGmNtUi6dMMjmFEF8AX9zimHnAPFBnQ8rCjqqI3qjnnb3vsOrsKnoF9eLd9u9ib2VfJtea2HQisVmxfH74c4JcgkoltE0ikZQ9Qgi2XNzCl0e+JCYrBhuNDa39WjMmZAydAjpRzbGauU2sVLjauvLSfS/xUL2H+OTgJ3x26DN+PfUrU1tO5f4a90vRKDE7RpORZSeX8eWRL9EqWt5s+yZD6g6Rn81SxlZrS6+gXvQK6kVKfgorz6xkcdRihq8bzv2B9zMpbNIt22xI/pt7FTUJwNUSPqB4m8QMZBRkMOXvKRxOPsxToU8xMXQiGkVTZtdTFIV3279LQm4C03ZMw8/JT840SCQVGCEEey7t4fMjnxOVFkWQSxAfdvyQroFdK2XTvIpGkGsQX3X7it0Ju/n44MdM2TqFVtVa8dJ9L8mBjMRsxGTGMH33dCJTIukU0Ik32rwhJzbKAW8HbyaETmBkw5EsiV7CkhNLGLp2KD1r9mRS2CSZv3QX3GtOjRVqoYBuqGLmAPCIEOLEbZ5PVisqJc5lnuPpLU+TnJ/MjA4zeCD4gXK7dqoulUd+f0Sd6emzDF9H33K7tqTssMQYeKi6cfC3IiI5gs8Pf87BpIP4OfoxKWwSfWv1rXKllysKBpOB8NPhzImYU5Jv80yzZ/C09zS3aWWCJfqTyu5L9CY9i44v4uvIr3G0duTlVi/TJ7iPXJ0xE1mFWSyOWszSqKXoDDoeCH6Ap0Kfkjk3/6K0qp/9BHQBvIAk4E0hxPeKojwIzEYt6bxACPHenRpY2R1HWbMzYScvbnsRW60tX9z/BU29m5a7DafSTzH6j9HUdKnJot6L5KxvJcASByEg/cm/OZl+ki+PfMn2+O142nkyvul4htYbKvM5KghZhVl8E/kNP5/8GTsrO8Y3Hc/IhiMr3f+PJfqTyuxLotOimb57OifTT9IrqBfTWk2rtILa0sgoyGDRiUX8dPInCo2F9K3Vl4mhEwl0lrlNYIbmm3dKZXYcZYkQgmUnlzHzwEzqutXlq25fmXXJeFvcNp756xm61ejGrC6zyjT0TVL2WOIgBKQ/ucL5rPPMiZjDhvMbcLZxZlzjcTzS4BE54VBBic2KZdbBWWyL30agcyAvtHyB+wMrT76NJfqTyuhLCo2FfBv5LQuOL8Ddzp3X27xOtxrdzG2W5Aak6lJZeHwhy08tx2gyMqDOAMY3HY+/k7+5TTMrUtRUQvQmPR/u+5BfTv9C18CufNjxwwoxWPnhxA98cvATnmjyBM81f87c5kjuAUschID0J4l5iXwT+Q2rzq7CRmvDqIajGNN4DC42LuY2TXIb7E7YzcwDMzmXda5S5dtYoj+pbL4kIjmC6bunE5sVy4DaA3jxvhdxtXU1t1mSW5Ccn8z8Y/MJPx2OQDCk7hCebPJklQ31L8uSzvfEVTk15jTD4sgqzOKFbS+w7/I+xjUex3PNn6swqyKjG40mNiuW+cfmE+waTP/a/c1tkkRSJUjTpTH/2HyWn1oOwPAGw3miyRNVsieKJdOuejvC/cL59fSvzImYw0PrHlL724Q9LcODyonKNjbJ1+fz5ZEv+TH6R6o5VuOb7t/Qvnp7c5sluU18HHx4tfWrjGs8jnlH57Hi9Ap+O/Mbw+oPkz7+X8iVGgvjQvYFnt7yNPG58bzV9i0G1BlgbpOuQ2/SM3HzRA4nH+b7nt/T3Le5uU2S3AWWOLMKVc+fZBdl88OJH1gStYRCYyEDag9gYujEKh+iUBn4d77NhKYTeKThIxaZb2OJ/qQy+JJ9l/fx1u63iM+NZ3j94UxpMQVHa0dzmyW5BxJyE/g28lvWnFuDtcaa4Q2GM7bxWDzsPMxtWrkgw88qCfsu7+N/W/+HVtEyu+vsCi0WsgqzGLl+JNmF2fzY50eZ4GaBWOIgBKqOP9EZdCyLXsaC4wvILsqmV1AvJoVNopZrLXObJillYrJimHVwFtvjtxPoHMjUllPpGtjVovJtLNGfWLIvySnKYdbBWaw4s4KaLjV5u93btPBtYW6zJKXIxeyLfBP5Db/H/o6t1paRDUcyJmRMpQ8plKKmEvDr6V95f+/7BLkG8eX9XxLgHGBuk27J+azzjFw/Ei97L5Y+uBRnG2dzmyS5AyxxEAKV35/ojXrCz4Qz7+g8UnWpdKzekWeaPUNDz4bmNk1SxuxK2MXMAzOJyYqhtV9rXmz5osXk21iiP7FUX7Itbhvv7H2HVF0qj4U8xqTQSdhZ2ZnbLEkZEZMVwzcR37Dh/AYcrB14tNGjPNro0UqbR1lhRY3sU3NrDCYDsw7OYmn0UjpU78DHnT7GycbJ3GbdNvsu72Pi5om09mvNV92+kj0xLAhLHISA5Q5EboXRZGRdzDq+jvyahNwEmvs057nmz1XoFVtJ6WMwGUrybXKKchhSdwiTwyZX+HwbS/QnluZLMgoy+HD/h6yPXU9d97q82+5dQrxkQ+yqwpmMM3wd+TWbL2zG2caZxxo9xqhGoypduGGFFTVXsDTHUV7kFOXw0vaX2Jmwk1ENRzG15VS0Gq25zbpjwk+H8/aetxnRYASvtn7V3OZIbhNLHIRA5fMnQgi2XNzCl0e+JCYrhoYeDXmu+XO0829nUeFHktLl3/k2E0Mn8kiDR7DWWpvbtBtiif7EUnyJEIKN5zfywf4PyC7KZnyT8TzR5IkK+1mQlC3RadHMjZzL1rituNm6MSZkDCMajKgQFXJLgwpb/Uzy38TlxPHMlme4kH2B6W2nM6zeMHObdNcMrTeU2KxYFkctJtg1mBENRpjbJImkwiOEYM+lPXx+5HOi0qIIdg3m0y6f0r1GdylmJLjauvJyq5cZVn8Ynxz4hE8OfsIvp35hasupdAnsIj8jVYS4nDje2/ceuxJ20dizMfN7zqeue11zmyUxIw09G/Ll/V9yPPU4cyLmMPvwbBZHLebxxo/zUP2HKnUoolypqYAcSjrE838/j1EY+bTLp7T2a21uk+4Zo8nIc38/x86EncztNpd21duZ2yTJLbC0mdXKFM56JPkIXxz+goNJB/F39GdS2CT61Oojwzcl/8nOhJ18fODjknybl+57iXru9cxtVgmW5k+gYo9N9EY9i04s4tuj32KlseKZZs8wvP5wi4zmkJQtEckRzImYw97Le/G29+aJJk8wtN5Qi6yiCBU4/KwyDUJKi1VnV/H2nrcJcArgq25fUdOlprlNKjXy9Hk8+sejXM69zNIHl1Lbrba5TZLcBEschEDFHojcipPpJ/nyyJdsj9+Op50n45uOt+gvH0n5ojfp+fXUr8yNnEtOUQ5D6w5lcrPJFaLUqyX6k4rqSw4lHeLdPe9yLuscPWr24OX7Xq6yjRglt8+BxAPMiZjDoaRD+Dr4Mr7peAbVGWRxYYoVVtRcoaI6jvLEaDLy+ZHPWXh8Ia39WjOr86xKWZbvUu4lRvw+AgcrB5b1WYa7nbu5TZL8B5Y4CAHL9Cfns84zJ2IOG85vwNnGmXGNx/FIg0cqTQy0pHzJKszi68iv+fnkzzhYOTAhdILZ820s0Z9UNF+SWZDJp4c+5bezv1HdqTqvtn6VTgGdzG2WxIIQQrAvcR9fHfmKyJRIqjtVZ0LTCfSt3RdrjWWIGylqKjj5+nxe3vEyW+O28nD9h3m51csW8+G6GyJTIhm3YRyNvRrzXc/v5Cx0BcUSByFgWf4kMS+RbyK/YdXZVdhobRjVcBRjGo+ptKU4JeVLTGYMnxz8hB0JO6jhXMOs+TaW6E8qii8RQrDm3BpmHZxFTlEOo0NGMzF0IvZW9uY2TWKhCCHYdWkXXx35ihNpJwh0DuSp0Kd4MPjBCh/CKEVNBeZU+ile2v4S57PP8/J9LzOiwYgqkeC5PmY9L+94mf61+zOj/Ywq8Z4tDUschIBl+BOTMLHg+ALmRswF4OH6D/N4k8fxsvcys2WSysjV+TZt/Nrw4n0vlnu+jSX6k4rgS2KyYnh3z7scTDpImHcY09tOl4UAJKWGEIJt8duYEzGHk+knCXIJYlLYJHoF9UKjaMxt3g2R1c8qIEIIfjr5E7MOqmFm83rMqxQFAW6XB2s9SGx2LN9EfkMt11o83uRxc5skkZQL2UXZvLbjNbbGb6VnzZ5MbTkVPyc/c5slqcR0qN6B1n6t+eXUL8yNmMuwtcMYVm8Yk8ImVYh8G8n1FBgK+O7Ydyw4vgAHKwfebPsmg+sOrrADTYlloigKXQK70CmgE39d/Is5EXN4aftLzDs6j0lhk+hWo5tFfeakqDEDGQUZTN81na3xW+kU0Il3279bJb9YJoVO4nzWeWYfns2OhB208G1BC98WhHmHyVwCSaXkdMZpnv/7eS7lXuKVVq/wSINH5CqlpFyw1lgzsuFI+tbqW5Jvsz5mfYXIt5Fcy+5Lu5mxdwZxOXH0q9WPF1q+UOGbq0osG42ioXvN7txf4342nd/E3Mi5/G/r/2jg0YBJoZMspky8rH5Wzuy/vJ9pO6aRUZjBCy1fqPKDmgJDAd8e/ZY9l/YQnR6NSZjQKloaejQsETnNfZtXyqIJFR1LDBeBihEyciPWxazj7d1v42zjzKwus2jm08zcJkmqMDGZMXx88GN2JuykpktNpracSueAzmX2fWSJ/qS8fUmqLpWZB2byR+wfBLkE8Xqb16tUBIek4mA0GVkfu56vI78mLieOEM8QJodNpkP1DmYfs8qcmgqA3qTn64ivmX9sPjVdavJx549p4NHA3GZVKPL0eUQmR3Iw6SCHkg5xLPUYepMegLrudWnu05yWvi1p7tscHwcfM1tb+bHEQQhUPH+iN+r5+ODH/HTyJ1r4tuCTzp/I3BlJhWFH/A4+PvgxsVmxtPVry4v3vVgmORuW6E/Ky5eYhInw0+HMPjSbAmMBTzZ5knFNxmGrtS3za0skN8NgMrD23Fq+PfotCbkJhHqH8nSzp2ldrbXZxI0UNWYmITeBl7e/TGRKJIPqDOKVVq/I8KrboNBYyLGUYxxOPsyhpEMcST6CzqADoIZzjWtWcgKcAsw+e1DZsMRBCFQsf5KUl8QL214gMiWS0Y1GM6XFlEpd2VBimehN+pJ8m1x9LsPqDWNy2ORSLblvif6kPHzJqfRTvLP3HY6mHKV1tda83uZ1glyDyvSaEsmdojfq+e3sb8w7Oo+k/CRa+Lbg6bCnaVmt/G9pKWrMyIbYDbyz5x0Egultp/NA8APmNsliMZgMnEw/yaGkQxxMOsiR5CNkFWYB4OPgQwvfFrT0bUkL3xbUcq0lRc49YomDEKg4/uRA4gGmbpuKzqDjnfbv0Duot7lNkkhuSmZBJl9Hfs3yU8txsHJgYuhERjQYUSr5NpboT8rSl+Tr85kbMZel0UtxtXVlasup9K3VV35vSSo0RcYiwk+HM//YfFJ0KbTxa8PksMmE+YSVmw1S1JiBfH0+Hx34iJVnVtLUuykfdfyIAOcAc5tVqTAJE+cyz3Eo6VDJI0WXAoCbrRvNfZqrqznVWlDfvT5WGlkX406wxEEImN+fCCFYHLWYzw59RqBzILO7zqa2W22z2SOR3CkxmTHMPDiTXQm7CHIJYmrLqXQK6HRPA25L9Cdl5Uu2xm3l/X3vcznvMkPqDuH5Fs/LvFGJRVFgKOCXU7/w/fHvSS9Ip0P1DkwOm0xjr8Zlfm0pasqZk+kneXHbi1zIvsATTZ7gqbCnZMhJOSCEIC4n7hqRE58bD4CjtSNhPmG08FFD1hp7NZZNP2+BJQ5CwLz+JE+fxxu73mDzhc30qNmDd9q9g5ONk1lskUjulX/n27x030vUca9zV+eyRH9S2r4kMS+RD/d/yJaLW6jjVofpbafLgiESiyZfn89PJ39i4YmFZBVm0SWwC5PDJpdpzniFFTWVrfqZEIJlJ5cx6+As3Gzd+KDjB7JyiZlJzEvkcJKak3M4+TBnM88CYKOxoal305K8nFDvUJnn9C8scRAC5hM1MZkxTNk6hQvZF3i++fM8FvKYDCWRWDyllW9jif6ktHyJwWRgWfQy5kTMwSRMTAydyOiQ0XKyU1JpyC3K5cfoH/kh6gdyinLoUbMHT4U+Ve5FR+RKTSmRXpDO9F3T2Ra/jc4BnXm3/bulmmQpKR0yCjJKCg8cTjp8TRnpRp6NSkROM59mVT4cwBIHIWAef7Lx/Eam75qOnZUdH3f6mFZ+rcr1+hJJWZNZkMncyLn8cuoXHKwdeCr0KYbXH37b+TaW6E9Kw5ccSznGu3vfJTo9mo7VO/Jq61dlKLqk0pJdlM2SqCUsiVpCvj6f3kG9mRg2kVqutUrtGlLUlDH7Lu9j2o5pZBZmyt4zFkZuUS6RKZEl4WpXykgrKCVlpFtUa0ELnxZ4O3ib29xyxRIHIVC+/sRgMjD70Gx+iPqBpt5NmdV5FtUcq5XLtSUSc3Au8xwfH/iYXZd20cCjAT/3+RmtRnvL11miP7kXX5JTlMMXh79g+anleNt783Krl+lRs4ccG0iqBFmFWSw6sYgfo3+k0FhIn+A+TAydSA2XGvd8bilqygi9Sc/ciLl8f+x7glyD+LjTx9T3qG9us+6Z5NP7SD6xnYYPTkJr62huc8qVK2Wkr4iciJSIkjLSNV1qqiWkiwsQVHeqXqm/oCxxEALl509Sdam8uO1FDiYdZHj94bx030uyK7ukSiCEYEfCDhLzEnmo/kO39RpL9Cd340uEEGy8sJGZ+2eSVpDG8PrDeabZMzK3TlIlSS9IZ+Hxhfx88mf0Jj39a/dnQugEqjtVv+tzSlFTBsTnxPPyjpc5mnKUIXWH8NJ9L1l+ToZeR/Lat/A8Og8tJpI13hR1e5eAdsOhEg/eb4bepOdkmlpG+lCyGrKWXZQNgK+Db0m4WkvflgS7BlcqkWOJgxAoH38SkRzBC1tfILsom+ltp9Ovdr8yvZ5EYulYoj+5U18SlxPHe/veY1fCLhp6NOTNtm8S4hVShhZKJJZBqi6V7499zy+nfsEkTAyqO4jxTcffVWSDFDWlzB+xf/DOnndQUJjebnrl6D9xYQ/54U/hkBPLGm137JoOouaRj6nPeS64tKDaw59jW72Jua00OyZh4mzm2WsqrKXqUgFwt3WnuW/zEqFT373+bYVlVFQscRACZetPhBD8dPInPj7wMX5OfnzW5bNKsTorkZQ1luhPbteX6I16Fp1YxLdHv0WraHmm2TMMbzBcthGQSP5FYl4i84/NZ8WZFSgoDK03lCebPHlH4f1S1JQS+fp8Ptj/AavOrqKpd1Nmdpp5T0toFYLCXNjyDmL/POKFF1+7PMeU8U/i42xHZq6Ov36cSddL83BW8kmuPwr/ge+AvSyAcIWry0gfTDrIoaRDJOQmAOBk7USXwC682fZN7KzszGzpnWOJgxAoO3+Sr8/nnb3v8HvM73QJ6MJ7Hd/Dxcal1K8jkVRGKoI/URSlFvAa4CqEGHqr42/Hl+hNeh75/RFOpp+kR80evHzfy/g6+paSxRJJ5eRS7iXmHZ3H6rOr0Wq0PFT/IR5v/Die9p63fK0UNaVAdFo0L21/qXL1njn3N2Lts5AZxyJDT3YHTeazR9vjZHvt7NL+qDNc+u0N+hVtQGflgtJtOo5txoIFr0KUJVfKSO9P3M/KMyvpUL0Dn3f93OLyLSrCIORuKAt/ciH7AlP+nsK5zHM83expnmjyBBpFU6rXkEgqM2XlTxRFWQD0BZKFEI2v2t4b+BzQAvOFEB9etS+8tEQNwA8nfiDIJYjOgZ3v5i1IJFWWuJw4vo38lrUxaxlWbxivt3n9lq+RouYeEELwY/SPfHroU9xt3fmg4weWX65VlwmbXocjS0i2qcFTOWOp1bwb7w9ugrX2xgO1Ar2RX9atp8GRGbTSnCTDpSFuQz5Dqdm2fG23MH49/Svv7HmH3kG9+bDjhxYVjiZFjcpfF//itZ2vodVo+ajjR7Sv3r7Uzi2RVBXKUNR0AnKBxVdEjaIoWuA00AOIBw4AI4QQUcX7S1XUSCSSe+N81nkcrR1vKwztZr7ErAGfVzXfNKcZ/0l6QTpv7HqD7fHb6RLQhXfav2P5vWdO/QHrnkfkJvO7y3BeSO7NxG4hTOle96ZJ7nbWWkYP6sfptp2Z9dMcRmTOw31hb3LrDcap73vg4l+Ob8JyGFZvGLlFuXx66FMcrR15s+2blaqYQGXGaDIyJ2IO3x37jkaejfi0y6eWH24qkVQyhBDbFUUJ+tfmVsBZIUQMgKIoPwMDgKg7Ofe5c+dKfp82bRp79uy5Zn9AQABLly4FYMqUKURERFyzv169esybNw+A8ePHc/r06Wv2h4WFMXv2bABGjRpFfHz8Nfvbtm3LBx98AMCQIUNIS0u7Zn+3bt144403AHjggQfQ6XTX7O/bty9Tp04FoEuXLte9v4ceeohJkyaRn5/Pgw8+eN3+MWPGMGbMGFJTUxk69HoN+NRTT/Hwww8TFxfHo48+et3+F154gX79+nHq1CkmTJhw3f7XX3+d7t27ExERwZQpU67b//7779OuXTt2797Nq6++et3+2bNnExYWxp9//smMGTOu2//tt99Sv3591q5dy6xZs67bP/2TuVTzD2DbhtX89MN8NP/6bg4PD8fLy4tFixaxaNGi616/fv16HBwcmDt3Lr/88st1+7du3QrAJ598wrp1667ZZ29vzx9//AHAu+++y5YtW67Z7+npyYoVK4Cq89m78ve6F8wqaoQQa4G1LVu2fNKcdtyIfZf38cqOV8guzGZaq2mMaDDCsgejeWmw4WU49it670ZM1bzMuhRf3hvcmOGtbr9ueL1qLjz/3Css3z2Q7M0zGXNqLfqzf6Dp8hLadpPByrYM34RlMrbxWHKKcvju2Hc42zjzvxb/s+zPUhUgoyCDl7a/xN7LexlSdwjTWk/DVis/2xKJhVAdiLvqeTzQWlEUT+A9oJmiKNOEEB/8+4WKoowHxoM68JRULvKLjKTlFpKaW8ToBfuxcoklL/o4ObHpaBQFrUbBSqtgpdEw5ecj+Pr4EBuRQEKGDivtlf0arDQKF9Py8MOaihDxJFGR4Wc3ICotilHrRxHgHGD5vWeEgBO/wfoXoSCL9JbPMfhoK5LyBHNHNqdrA5+7PnVSdgFfrdhMx5jP6Kk9RIFLEHZ9Z0K9XqX4BioHQgje3/c+P5/6mWebPcuTTSucjr+OihB+dqeJvXDv/uRYyjH+t+1/pOvSebX1qwypN+SuzyWxfIwmQVJ2AfEZOuIz8onP0OHuaEP3hj74ucpB7+1Slv6keKVm3VXhZ0OB3kKIJ4qfPwq0FkI8fSfnrWhjE8ndkZCpY03EJVZHJHAyMQeNAh3qetO3iR/OdlZk6fRk6vRk5uvJ0hUV/9Rf9bOIvCLjf55fUcDFzho3B2vc7K1xdbBRf9qr29Sf6rYrz12Lf9paWU5IekWhwoafVUTy9Hm8uO1FPOw8WNx7MW52buY26e7JSYTfX4CT68C/Ocfve59Ra3Ow0igsn3AfTQPc7un0vi52vDu2H5tO3Mdzvy3l2cz51F72EIba3bF64CPwqphhheZAURSmtZ5Grj6XL458gZONEyMajDC3WWbhThJ7i8NHHlcUJbys7RJCEH4mnA/2fYC3vTeLH1xMiKfsMVHZuZFo+eenjkuZOgym6yf/3lgFoQGu9AypRs9GvtTxcZIrsBWHBCDwqucBxdskVYSMvCJ+P3aZNRGX2H8+HYBmNdx4q18j+jT1x9v5zlbeiwwmsnT64ocqfDLzVTGUpdOTlV9UIowydXoupuWVHH8D91GCo42WVsEe9A/zp0ejatcVapLcGfKvdxVCCN7e8zbxufEs6LXAcgWNEBCxDDZOA0Mh9HiXjc6DePaX4/i72fPD2FbU8Cy9RqE9Q6rRtvZzzN7QHeXgPJ4/txLNnDZo2k2CTi+CrXOpXcuS0Sga3m3/Lnn6PN7f9z5O1k5VtWnjIuArYPGVDcWJvXO4KrFXUZQ1VxJ7y5oCQwHv7XuPVWdX0d6/PR92/NBy7//KiF4H5/6Ck+shLwVsHK99WDuAjRPYOBRvcyre5ojRyoGUIisS8rRczBXEZRlvKVp8nG0JcLcnLNCNvk39CHB3IMDdngB3e/zd7InP0LE5KomNJxL5eOMpPt54imAvR3o28qVniC/NAt3RaKTAMSMHgLqKogSjipnhwCPmNUlS1uQXGdgclcSaiEtsO52CwSSo7e3ICz3q0T/Mn5qejuqYKD0W4k5D5gWo0Q4CWtzy3DZWGrydbe9YDJlMgpxCA1n5ejJ1RSUrQJnFQig5p5At0ck8vzwSW6tjdGvoQ/9Qf7rU98HOWq7i3Cky/OwqVp5ZyZu73+SZZs8wvul4c5tzd2TGwdrn4NwW9Wbt/yWLTml5e10UYYFufP/YfXg42pTZ5Y9czGBm+HYGpX/PQ1bbMDr4oO31LjR5CDSyBC5AobGQyVsmczDxILO6zKJbjW7mNumGlHO4SFvgLSFEr+Ln0wCuxLzfbrUiAFdXV9GsWbNrtt0sabHQWEhh40L0bfSMrTuW1a+uRuHaAWllT5hdsmQJgYGBLF++nK+//vq6/WZJmDUZQJeBpyaHFf31oM9n2lbBnksKCBOYjCBMBDgJlg5W+0BN2VBAROK1YSL1PDXM66eGiY1fq+NUmsCEBpOiQSgaQqo7M2NYCFZ2Tkz+4SiXs4pQNFpQNKDR0rZJHT547hGEtQODn5lJSlI6irUdGld3FCsrWnfoTJO+49gUlcSK9yZh1BdirdXg7miDh4MNDw8ZwLSXXwIqX7L2nST2lmH1s5+ALoAXkAS8KYT4XlGUB4HZqCu/C4QQ793BOa8UMXryzJkzpW2ypBTRG03sPJPK6ogENkUlkV9kpJqLHQ+HODAgIJdgLqGknoa0s5B6GpF+nsIMDTnx9uReVgWK1sUJbfW6aGs2RuvuhdbNDa2bK1pXV/V3V/V3jYsLShmMY0wmwZG4DNZEXOL3Y5dJzS3CydaKniG+9A/1p30dr/+sTFsVkeFnt8HZjLN8sO8DWvu15vHGj5vbnDvHZIJDC2Dzm+pKzYOfYGoxjo82nubb7afo2ciXz4c3w96mbJV/sxruLH6uH/N3NGbYn78zPW8RTX6bgDjwPcqDM8G/2a1PUsmx1dryRdcveHLTk7y47UXmdp9LG7825jbL3Nx1Yi9cm9xrZ3f7jU6zCrOIyYrBs8CTBfcvoJVnK9aw5q7fhOQeKcyBnMuQnwYFWaovc7aD0Ccx1u/LxYTfSMg4QKHBSKHBRKHBxCGTO/UKnsOBApIMCygwncdWC3ZasNUKbF282dd4GJ42Bgr3/QKGFLTChNZkBGHE0dEZX2dbKMrEqigbpaAATEaE0YhJr5CvS+bi1NMUpFuTF3OZIuMV0RSLxsEBxc6eQX1688jI5lxe7kViWjbp+UWk5RaSnF3A7D9PEx9wmJ4hvhhNAq1cwSlVhBA3jOMVQqwH1t/lOStsESOJGlVz6EIGa49cJPJYJF4FFwixSeQHz0zqWyXinBuDciQDjhQfr7GlwBBEziV3sk9ZoU/JBY2CfdPGaIzZGJPiKIqIxLj/OCb9Te5PRUHr4oLWzQ3NNaLH7VoB9C9RpHFyuqkY0mgUWtT0oEVND97o24i9MemsiUzgj+OJrDycgIejDQ80rkb/UH/uC/KQq8A3Qa7UADqDjkd+f4T0gnRW9F+Bl72X2Wy5K9LOwZpn4MIuqNUV+n1OoXMAU389ytrISzzapiZv9Q8p9y/TC2l5vL7yKNXOr+Q1m19wFVkozR+F+6eD061rkVd2sgqzGLtxLPE58XzX8ztCvUPNbdI1WGJiL9yePzEJE98e/ZavI76mrntdPuvyGTVcbr8KoKQUSY+B6HVq7l/cfkCARy1o0Bca9uOMdX1+PXyJlYcTSM0tBP4JD/snLOza8LA7DdswFRZSGB2N7uhRdEePoTt6FP3Fi+pORcG2Vk3sGtTGvm4N7Gv7YYreTN6OreQlWqNLswWTQLG1xaFFcxzatsWxbTuUunXZez6TTVFJbI5KIiWnEGutQptanvQMqUaPhr5Uc719AW7pVITCI3eKuccmkmJ0GZB6lsvnjnLhdARFiafwN8RRQ0nCRrlqZdbRB7zqgVcdhEcddMlWZB+5QM72vRgSk8DKCsfWrXHu2RPn7t2w8izuXi+EOn46MB9xYi3GQhNG37YYg/pgdG2IMSsbU1YWhsxMTFlZGDOzMGZmYsz656cpN/e/7ddoSsSQ1tUVjZsrVm5uaIpFj22dOtiHhmLt63vNywoNRrafTmVN5CX+jEpCp1dXofo29aN/mD9NqrtWyTw+2XzzFry1+y1WnlnJNz2+oZ1/O7PZcceYjLB3Lvz1HmhtoPf7EDaSrAIDE5YcZG9MOi/3bsDEzrXM9sEXQvDbkQRmrzvIY/rljLHaiMbGEaXrq3DfE6C1NotdFYVUXSqj/xhNZmEmC3strFCV9ipS+NmdcDv+ZMbeGSw/tZy+tfoyve107K1kFatyQwhIOq4Kmei1kHxC3V6tCTToBw37kuVUlzXHLhN+MI7I+CysNApdG/gwrEUAnep531OsuTCZKDp/Ht3RoxQUi5iCU6dArwfAytcX+6ZNsWvaBPumodiFhKB1crz+RNmXYPdXGPcuRHfJRJ6+AXlJthTGqAuOGldXHFu3xrFdW+xbt+GE4sKm6CQ2nUgiNjUPgNBAN3o28qVXiC+1vSt3oQEpaiQ3xWSEzIuQegbSzkDqaUg9gzHlFNr81JLD9EJLsnV18KqLd3BjbHwbqELGsw7CypG8/fvJ2bSZnC1bMKamotjY4NihA849e+DctStaV9eb25GTCIcXw8GFkHMJXAOhxRhoPhqc/rtarNDrMWZn/yN0MrOuEj2Z/4ifK+KoWBiZ8vNLzmFVrRr2TZtiH9pU9UEhIWgc1Pzn/CIDf0YnF+cLJaM3CoI8HegX6k//UH/q+lad3GUpam7CH7F/8NL2l3iiyRM81/w5s9hwVyRHw+qnIeEg1H8Q+nwKLn5cytQxZuF+YlPz+HhoKAObVYxmgel5Rbz3ezQRR/bxocOP3GeMAO+G8MCHUKuLuc0zKwm5CYz+YzRGk5HFDyyuMCsG5SxqrFA7gHdDTew9ADwihDhxp+e+HX9yKv0UR5KP8HD9hyv1QLLCYDKqqzAni4VM5gVAgRptoGE/aNAHo2tNdp5N5deDcWyKSqLIYKJBNWeGtghgYLPqeDndXZ8gQ0oKumPH0EUepeDYUXTHjmPKyQFA4+iIXZMm2DdpUiximl43W3pL8tJg39ewbx4UZmHw7UyeXRfyzqSRt2cPhsuXAbD298ehXVsc27QluW5jNl3Ssykqici4TABqeTnSI8SXno2q0SzQrdKFmEhRIwHUENPUM9eJF9LOgbHwn8OsXYmlOpE6H84JP4RXPRqGNKdjq5Z4uzmVHGcqLCRv125yNm0i5++/MWVloTg44NSpEy49e+DYqfONJyVuhdEAp9bDgfkQuw001tBogDoZW6ONWse5FDAVFFAQHa1OsEQeVVeJrzTC1GqxrVu3WOiEYh/aFJtatcguMLLxRCJrIi+x+1wqJgENqjmXCJxAj9IrBFURkaLmP7iYfZGH1j1EXbe6LOy9ECuNBaQYGfWwczZsn6lWFXtgJjQeAopC9OVsxi48QF6hgW8fbUG7OhUvjG7X2VReXXmUepk7+NDxJzz1l9VBTbPRENRerVxUBYnJimHMH2Owt7Lnhwd+oJpjNXObZFGJvVcjByIVBEMRxG6Hk2uLq5YlqwODWl2gYV91MsbJh5iUXMIPxbPycAKJ2QW4OVgzINSfYS0DCfF3uSPRacrPp+DEiZIQMt2xoxguqaICrRbb+vXUAUITdTbUJjgYRVtKeYYF2eoAaM8cyE+FGm0RHf5HkVVd8vbuIX/PHvL27ceUnQ2AbYMGOLZpQ1FoC7Y7BLLhbCZ7zqVhMAm8nW3p3lCtpNautmel6GUhRU0VwmSC7ARVsBQn6KuPs+rqxxUUDbgHgVc99O51OF7oy4ZEJ1ZedCDF5Extb0cGhlVnQFj1ayq2mvLzyd2+g5xNm8jdtg1TXh4aZ2ec7++Kc48eOHbogOYOcitvScppOLhArSpbmAU+IXDf49D0oTKp7mpISysOhT1KQWQkuqPHSsLbNE5O2DVpjH1TVeTk1arPxkt61kRe4tCFDADCAt3oH+pP36Z++LhUvhBXKWpuQJGxiFHrR5GQm0B4v3D8nPzK9fp3xaUIdXUm6ZgqZB6YCY6qcNl9NpUJSw7haGvFwrH30dDP5a4uIfR6MsPDKTxzFo9x47AJKP2VngK9kS+2nGHR9pNMtt3ABM0qrIw6dcAT2Bpqd1Fzg/ybgcbyv8xvl6i0KB7f+Dhe9l788MAPeNh5mNUeSxuEyIpFFYDCXDj7p7oac2YTFGaDtSPU7aFOXtTtAXau5BToWXf0MuGH4jl0IQONAp3reTOsZSDdGvrc1iBeGI0Unj17TRhZ4Zkz6oAKsA4IuCqMrCl2DRuiKY8O8UX5cGQp7PocsuPVsLqOL0DD/ggBBVFR5O3eQ96ePegOHULo9WBtjUNoKNr7WnPCrz5ri9z4+3QaeUVGnGyt6FLfm54h1ehS3xsXO8sM2bUkfyJ9yV1g1MP2j+HUH6qQ0f8TVoWtC3jVLc53qQue6u9615rsjM25rnJZ/zB/BoT508jvn0kNY04OuVu3qkJmx05EQQFad3ecu3fDuWdPHFu3RrEpu8quABTlwbFwOPAdJB4DG2cIHa4KHJ+GZXZZYTJRFBtbvJITie7oUQpPnYbioiXW/v7Yh4VSWKche+38+DnDnqMpBSgKtAn2pH+YP71DquFehpVvy5MKK2rM6Tg+2v8RS6OX8kXXL+hao2u5XvuO0RfAto/UL0lHb+j7KTToU7J71ZEEXgyPJNjLkUVjW+Hvdudf3EIIcv/eSvInn1AUEwNaLYqNDd5PP43H6EdRrEv/i/RkYjavrDhGdFwynezOMtTtLPeZIvHIjlYPsHOF4E6qwKndVU0eruQcSjrExM0TCXYN5vte3+NsY744WUsahFyNnF0tZ/LS4PQfao7Mub/UEBJ7D3UlpmE/dWXG2g6TSbAnJo3wQ/H8cfwyBXoTtb0dGdYykMHNqt9yRlEYjaoY2LuHgsij6KKiEMXx6BpXV+ybNLlGxFh5mHdSAEMRHPsFdn6mDvI860KH59XZ3eJcQpNOR/7hw+oqzu49FERHgxDqbGzL+0is04S/nWoSnmJFap4ea61C29pe9GzkS49Gvvha0CysJfoT6Utuk4zzEP64Gg4f1FEV8p51ikVMPTUXpVicmEyCwxczWF1cvjg9rwgXOyv6NPVjQFh1Wl1V3cuQkUHuli1kb9pE3p69oNdj5eODc/fuOPfsiUPLFihWZoiwEQLiD6orsydWgrFIfd/3Pa4WOCmHXGGTTkdBVBS6iMiSVZ0roa5YWaHUrst5nyC2anzYYV2NJGdvOtVXe+B0b+Rr0U0+K6youUJ5O46/L/7Ns38/y8iGI3ml1Svldt27Im4/rJ6sLt02GwU9Z4C9O6AKka+3nWPmhlO0qeXBt4+2xNX+zm+mgqgokj6aSf6+fdgEB+Pz4ovYNahP4rszyP37b2zr18fv7bewDwsr5TendvPedCKRv08ls/10KonZBXiQzSDXM/RxOkUj3SHs8otvVLeaqrip1VUVOw5mHrSUETsTdvLMX8/Q1Ksp3/T4xmxJ7JY4CAE5ECkXsuLh5O/qisyF3SCM4BKghpU16As12oJW/dK8mJZP+KE4VhxOICFTh7OdFf1C/RnWIoCwQLdbhpcVxSeQtXIlmStXYkhMRLG2xrZRQzX8oqmaD2Nds2bFzY0yGSFqNez4VF1ldw2E9s+p/tz62nvbkJFB/r59JSs5+ji16IDWx4fCJs2J9KnLr8KPyAJ1xjUs0I2exXk4dXycrrt0RcIS/Yn0JbfBiVWw5llAQP8vIGTQDQ87laiuyKyOuERCpg5bKw3dG/kyMKw6nep5lazO6pOTyfnzT3I2bSb/wAEwGrGuXl2tWNazB/ahoWXSK+auyUtVV2YPfq8WOnDyLS4s8Bi4lm9Osz45+ZrcnIJjx0oKERTZO3HKPZCjzgHEeNfEr3VLerStT5f691Z4xRxIUXMVl3MvM3TtUKo7VWfpg0ux0VbQ5biiPPhrBuz9GlwDoN/nUOefJo1Gk+CtNSdYsvcC/UL9+WRY0zuOu9YnJpIy+3OyVq9G6+aG19OTcX/ooZJVGSEEOX/+SdKM9zAkJ+M2/GF8nn8ercvdhbbdCiEEZ5Nz2X4mlR1nUtgbk0aB3kg9bSIjvM7RxeoENbIPodXnAooannZF5AS2Aqu7SySuiGw8v5GXtr9EW/+2fNn1S6zNUCXOEgchIAciZUbKaTU/JnotXCpuAOFVX12NadgX/MJKZmPzCg2sP6aGl+2LTUdRoEMdL4a2CKBXSLVbfomaiorI3bKFzF/DyduzBwDHDh1wGzoUp65d0JR1mElZIASc2Qw7PoG4feqqe9vJ0PJxsLuxTy2KjydvT3E+zp69GDPUmHlqBhMXHMJfjjVZQzXyre2p5e1Iz0bV6BniS1hAxSs0YIn+RPqSm6DXwcZX1VyT6i1g6AI1P+YqEjJ1rIm4xOqIBE4m5qDVKLSv48XAMH96hlQrWS3QJySQvWkzOZs3oztyBITAJji4RMjYNWpUcSctrmAyqqG3B+ar97migQYPqoUFgjuXWmGBO0EN0T2H7mgkBUePkh95lMKzZ1GKQ3QTHL2I8aqJNqQJDe5vw33dWmNjX/FXf6WoKUZv0jNuwzjOZJ7hl76/VJgqU9cRsw3WPqsu6d73JHR/85pkNF2RkWd/PsLmqCQmdKrFy70b3NEXmCkvj9T580lfuAiMRjweG43nhAlonW8c6mTMzSPli8/JWPojWk8Pqr36Ks69e5e5kynQGzl0IYPtZ1LYfjqV6MvZWGGgo8NFHnI/w32mo3hmHkURRrB2gJrt/xE5Pg3N4kRKk5VnVvLm7jfpWbMnMzvNRFvO+UWWNgiRcfCljBBw6fA/PWRST6vbq7co6SGDV92rDhfsj00n/FA8vx+7TH6RkSBPB4a2CGBw84DbCostPHOGzPAVZK1ejTEzEyt/P9yGDMFt0CCs/f3L6p2WL1d6YuyYpYbr2blCqwnQ5qmbrj4Lk4nCU6dKVnHyDx5EFBSAVktuUD0ifOryu1UAJ9xq4ObmSI9GvvRs5Eu72l7YWJl/ZtvS/AlIUfOfpJyG8LFqafZ2z6i956zUiYb0vCLWH7vMmohL7D+fDkCzGm4MDKvOg0388HZWJx8LY2PV0subNlFwQi1yadugAc49e+DSsye2deqY572VBumxcGghHF4CunQ19PS+xyF0BNi7mdU0U14euuMnyIuI4NLegxiPH8chR50s0Wu0ZFWvhUvzMGp2aIVDWCjWAQEVTlBKUVPM54c/Z/6x+XzU8SMerPVgmV/vjinIgs3T4dAiNX+k/1dqRbCrSM8r4vEfDhARl8mbfRsxpn3wbZ9eGI1krlhByhdfYkxNxeXBB/H+3/9uuxiA7vgJEqdPpyAqCseOHan25nRsAgLu5B3eE8k5Bew6m8r20+pKTmpuEU7kM8TjPH2dThJScBiH7Bj1YKdqaix/rS6q0HE2fzWxu+GHEz/wycFPGFx3MG+1fatcnYslDkJADkTuCaMBLu4uFjK/q4nuilb1Qw3U0sv/DqlIyNSx4lA84YfiuZiej6ONlj5N/RjWMpCWNd1v+Zk15eWRvWEDmb+Go4uIAGtrnLt1w23oUBzbtim96mQVkYTDqrg5uU4tqNByLLR9GlxuXbjGVFSE7kgEeXt2k7dnDwXHjoPJhMnWjriAevztFMR+jzqkeQfQuWE1ejbypUt9b5zNVGjAkvyJnCC5CRHL4PcX1NDJQd+qxT+AnAI9b6+NYtWRBAwmQR0fJwaG+dM/VK1cJoSg8PTpEiFTWPx3tWvaFJeePXDu0QObmjXN+c5KH30BRK1SV2/iD6iTr02GqpPVfk3NbR2gTkblxydwaOMu4ncdwOZMNLUz4rAzqn27hKsbTmGhxb1z1JDfsorWuV2kqAF2X9rNxM0T1cFhu7fK9Fp3xelNsG4K5FxWv9S6vnpdvPXFtHweW7ifS5k6Ph8eRu/Gt1+xLXfHTpJnzqTwzBnsmzfH9+WXsA+98w72wmAgY9kyUmZ/jjCZ8Jo0Cc+xY8qkkMDNMJkEJxNz2H4mhR1nUjgQm0GR0URNq3RGecfQ1foEQdkHsCpQZ4rwbvjPKo6FlY6eEzGHbyK/4dFGj/JiyxfLTdhY0iDkaqSouUP0BRDztxpWduoPdWbRyg5q36+uxtTrfd0Kgq5I7ZMQfiieXedSEQLa1vJkaIsAHmhSDQebmyehCiEoOHqUzPBwsn9fjyk/H5vatXEbOhTXAf3Nn+Rf3iRHqwUFjoWrFR/DHoH2U8Dj9ietjNnZ5B84ULKSUxSjTvDoHF2I8KrDXvfaHK9Wj7pN69EzxJceDX3LtdyrJfoT6UuuojAHfp8KR39Wk+IHf1civo/GZ/L0siMkZOp4tE1NhrUMoFFxBdaC48fJ2bSJ7E2b0F+4CIqCQ4sWamhZj+5Y+1lA5dnS4FKEKm6OhYNBBwGt1NC0RgPAuuKEfOUVGvjz+CX2btlPzpFI6qadp0l2PH6ZiSXH2AQHl/TNsWvaFLt69cp1DFjlRU2qLpUha4bgYefBsj7LKlb3cCFg42uwd4468B4wBwJaXHdYZFwmj/9wAINJMH90S1oG3d6XfsGp0yR//DF5O3diXaMGPi+8gHPPHvc8MNYnJpL03nvkbP4T27p1qfb22zg0b3ZP57wXdEVG9samseN0KtvPpHA2ORcFE+2dLjPc4xytRCTe6YdRjIX/Kh19P/iHVejS0UIIPjrwET9G/8jksMlMDJ1YLte1xEEIyIHIbWMywa7ZsP0T0OeBrSvU66Xmx9Tpfp3wF0Jw+GIm4YfiWBd5mZxCAwHu9gxtEcCQ5gG31fDNkJFB9tq1ZP4aTuGZMyj29rg8+ABuQ4diHxZW4cIcyp30WNj9hZp4bDJA46FqxTTfRnd8Kn1iInl79qorObv3YExVu7Inu3hzwKMOEd51MIW1oGOLOvQM8aW2d9kWGrBEfyJ9STGXj8KvYyAjFjq/Ap2mgkaLySRYsCuWjzacxNvJli9GNKNFoCu6iAhVyGzerPaJsrLCsVUrVch074aVV8XroVdu6DIg4idV4KSfAwdPtUVHyGB1XFKBiiBk5heVNPmMPBlPnfQ4OuiTaF14Ge+4s4gMddJYsbXFLiSkuEmo2ijUys+vzPx5lRY1RpORCX9OIDI5kp/6/EQd9woUpykEbHod9nwFrcarlc1ukOz+98lkJv14GE8nG34Y1+q2vnwMKSmkfPEFmStWonF2xuupibg/8kipJ9jm/PUXie/OwHD5Mm4PPYTPC/9D6+paqte4Gy5l6th5JpVtZ1LYdTaVzHw9thQx1Duefk6naFx4GKf04mb1dm5qNbUrKzl3MDtaXpiEiTd2vcGac2t4pdUrjGw4ssyvaYmDEJADkduiIAt+ewpO/a7mx7QcC0GdSuLiryYxq4CVR9TwspiUPOyttTzQpBpDWwTQJtjzlvl8wmQif+9eMsPDydn8J0Kvx65pU9yGDsHlwQfROlXsql1mIfuy+r1wcKEqOBv0hY7/U/OZ7gIhBEVnz5K3Zw+5u9UmoOjyMaFw1q06Ed51SarbhOAu7egeVoPQMig0YIn+pMr7EiFg/3ew6TV18D1kPgR1ACAtt5AXfo1k66kUejby5aN+9dEvXkBmeDjGlFQUGxsc27dXhUzXLmjd3Mz6ViocJhPEblMLLZzZBIYCcPaHkIGqwAloWaHygpNzClh/9DJrIi9x+GImCEFXdyODbDJokh2PcvIEBVFRiKIiALTeXsXhaqrQsWvcBK1T6UTIVGlRM+/oPL488iVvt3ubwXUHl8k17pq/ZqjNqlqNVxtp3uAD/PP+i7y26jgN/ZxZMOY+fJxvvkxp0ulIW7iQtPnfI/R6PB4ZgddTT5WpQzHl5ZHy5VekL16M1t0d32nTcOnzYIWZdTWaBMcTsth+OoUdZ1I5fDEDg0lQ3SaPR33Oc7/NCYKz92OdW9zp2D3on944wZ1KSmibG4PJwNRtU9lycQsz2s9gQJ0BZXo9SxyEgByI3JKkKFg+Ui0/2vM9aD3hOt9ToDfyZ3QSvx6MZ8eZFEwC7gtyZ1iLQB5s6ndbPQ70iYlk/fYbmeEr0CckoHV1xaV/f9yGDsGufv2yeneVi/x02Pct7PsGCjJVv9TxBXVgeQ/+Vej16I4dJ2/PbjJ27EJ/7Cgao5EijRUnPIM5E9AAp3btaNm9DW3r+pRKoQFL9CdV2pfoMtRm3yfXQd1eMPBrcPQE1GbfU5ZHkKnT83qfhgx1yuHytGkUnT2HU7duuDz4AE6du5TaILbSU5gDpzaoPW/O/qn2vXENLBY4g8C/eYUSOHHp+awrFjjRl7PRKNCmlicDQrzpapON1amoktLSRefPqy9SFGzr1MYu9IrQCcW2Tp27ypmssqLmUNIhxm0cR6+gXnzU8aMKM8gGVDHz1wxoPhr6fn7dkqMQgs/+PMMXW87QuZ43c0Y2v+lAQphMZK1eQ8rs2RiSknDu2ROfF/5Xrol3BVFRXH7zLQqOHcOxXTuqvfUmNjUqXoW5nAI9e86lsaO4dPT5tHxA0NYlgxFeZ2ktjuKTth+lKFcty+gXdlXp6NY3nM0uL4qMRTy95Wn2Je5jVudZdK/ZvcyuZWmDEJncexscC4c1z6jVFIf9ADXbluwSQnAsIYtfD8azJvISWTo9fq52DGkewJAWAQR73XqAIvR6crdtI/PXcHJ37ACTCYe2bXAbOhTn7t3R2FaesuvlSmGOOqO7+yvIS1bj8Tu+oIYLlsL3mikvj/yDB0nfvpO07buwi4sFIMfanijfuhQ2bUFQry607RSGi/3d+T9L8ydQhUXNxX2w4nHISYTub6mlxxUFg9HE51vO8NXfZwn2cuSrIY3xWrWUtO/mY+Xlhd+Md3Hq2NHc1ls2BVlwcj2c+E2tjmjSqz36QgZB48FQrWmFEjhnk3NYE3mZtZGXiE3Nw1qr0KmuN/3D/One0Bc7XS66Y8eKe+dEUhB5FGNWFgCKgwP2ISEluTkOzZph5e19y2tWSVGTWZDJ0LVDsdXasrzvcpxsKlCIw+4v1bCzpsNh4Nzr8jn0RhPTVh4j/FA8w1oE8P7gJlhr/3umLG/vXpI+mklhdDR2TZvi+/JLOLS4uzCFe0UYjWT89DMpn32GMBjwemoinuPGoVTgvhIX0/JLCg7sPptGTqEBG8XAEN8k+jufpnHhEZxSjlSY0tH5+nwmbJ7A8bTjzLl/Du2qtyuT61jiIASq8EDkZhj1qs/Z943aGHPYopKKgCk5haw6ksCvh+I4nZSLrZWGXiHVGNYygHa1vdDeRhhSYWwsWStWkLlqNcbUVKx8fHAdPAi3IUOwCQws4zdXhdDr1HybXV9A1kXwbayGpTUaWKp5gYbUVDJ37eH8pq2Ig/twykoDINnBncTaTXBp345mA7tTLej2mwtaoj+pcr7kSp7dXzPU/njDFpaEPCZk6pjy8xEOnM9gWIsAXqunIf2N1yk8fRrXQYPwnfaK2atiVTp0GWoVyuMrIWar2uTYo7YqcEIGgW9IhRE4QgiOJ2Sz9ugl1kZe4nJWAXbWGro19KV/qD9d6ntja6VFCIH+4kV0VzcJjY4GvR6Pxx7Dd9ort7xWlRM1Qgie+esZdl/azdIHl9LI886TLMuM/d/B+qnql9CQ70u6bl8ht9DApB8Ps/10Cs91q8uU7nX/c4WpMCaG5Jkfk7t1K9b+/nj/73+4PPhAhei2q09KIun9D8jZuBGb2rXxe/stHFpW/O8zvdFEZFwm28+ksv10CkfjMzEJqGZbxKN+cXSziaJ2zn6sM86pL7hSOrp2V/VnOZWOzi7KZtyGcVzMuci8HvMI8wkr9WtY4iAEquBA5FbkJMIvj0HcXmgzGXq8DVprjCbBrE2n+HZ7DEaTICzQjWEtA+jb1B9X+1tXsjHpdORs2kTmr+HkHzwIWi1OXbrgNnQITh07oljdOkRNcpcY9eqq285P1f5BHrXVggJNHy71lWQhBAWx5zn5+1+kbd+B26mjOBbpALgQ2IBem1beVhSEJfmTKrnqm5sMK8erlRBDBqkNv+3U/NiNJxJ5KfwoBqOJ9/o1pP2+daR+8w1adzf83n4H5/u7mtn4KkBemtr8+MRvELsdhAm86hULnMHg08DcFpZgMgkOXshgbeQl1h+7TFpeEc52VvQKqUb/UH/a1fbE6qqJelNhIYXR0WhcXbENvnVOc5UTNUuiljDzwMxyS6i+bQ4vVkM/6veBh36Af3WJT84uYOyiA5xMzOH9QY15+L4bh24Z0tNJ/eorMpb/gsbeHs8J4/EYPbpChnbkbN1K0jvvor90CdehQ/CdOtWiEgYz84vYfS6N7adT2H46hUtZBQC0cs9jhHcMbcVRfFP3oujUmUx8Gv2Tj1OzXZmWjk7VpTJmwxjSdeks7L2Q+h6lm6dgSYOQq5Gi5iou7FarFhXmQP8v1R4JQFa+nmd/PsK20ykMaR7AU11qUcfnxs13/01BVBSZ4eFkrV2HKScH65o1iksxD8Dax6cM34zkOkxGNedh+yeQeBRcqkO7Z9WwZptbV6O7q0saDJzaeZAz6//CpDcw8LPpt/U6S/QnVcaXnPtbFTSF2dD7Q2gxBhSFAr2RD9ZH88OeCzSu7sIXrVzgo7cpjIrGpW9fqr3+mkV9n1caclMgejWcWAXndwJCHXtcETheFacglsFoYte5NNZGXmLj8URyCg14OtrwYBM/+of506KG+x0XJalSouZE6glG/TGKjtU78nnXz8s8j0av1xMfH09BQcHNDyzKg/w0sLIHR6/rlgz1RhNpuUWYhMDD0QY76+tDCYQQmHJzMeXmghBoHBzQuLhUiJWZmyFMJkw5OZjy8kCjQevigsahbL5wyxq90UShwUSh3kihwYRJgAI4WBlxUAxYiyI0xkLAhF32eQKyD2Ld/zM1h6EMuJx7mdEbRlNkLOKH3j8Q5BpUaue2xEEIVKGByM0QAvZ+rYaceQTDw0vVUEngdFIO4xcfJD5Dx9sDQhjZ+tZ5d8bsbLLWrSMrfAW6+HiMk55CqVsXxdGxQk6mVEn0Beqg1FCgNky1dQZbJzUvsByxs7MjICAA63/1rbBEf2LJvuS2xiZCQGEWFGSrk6wOXiWTrXqjiYy8IoqMAmdbLU7GQkw5Oep3uKsrGvsK1BqjKmMygj5ffRgK1W1aGzVU3sYBNDdeMf+v+7QsKdAb2XoqhbVHL7ElOokCvQl/Vzv6hvrTP9SfEH+Xe171rVSiJqcoh4fWPoRRGPm136+42pZ9aeHY2FicnZ3x9PT87/8MXQZknAcbJzVM4F8iJK/QwPm0PBQUgrwcrmtcJ4TAmJWFISkJodejdXbGytcXjV3Fadh0O5h0OvSXLmHS6dA4OmLt72/RAyKTEOQXGsgpNJBbYECnNwJgrQF3qyJMuekUnvmb4Mu/w8hfr2umWlrEZsUyZsMYbLQ2LO69GD+n0mlmZomDELDsgUipUJirrgifWKmWAh4495owkv8tj8DexoqvRzXnvpv0uxJCoDt4UG2QuWEjorAQ24YNMbzwPzzq1cPT27tiFV+RqBTmQm6iujqnaNVJNEfv6yIDygIhBGlpaeTk5BD8rzASS/QnluxLbjk2MRSp4xJ9nlqu2aU6aNSch4x8PZcydWgUCHSywiYlEZNOh9bFBWt/fxlaWlExFoEuU33o89Rt1vZg5w72biUtQ252n5YXuYUG/oxKYm3kJbadTsFgEoxrH8z0frdOF7mZL6k0n0whBG/veZvLeZdZ1HtRuQgagIKCAoKCgm4iaLIg44IahuRR6zpBk19k4HxqHlZahSAvR2ytrl2hMeblYUhUHYrGzg7r6tUttq+Dxt4em1q1MGZkYEhMovDsWay8vbHy8qrwq003QqMoONlZ42RnDa7qzFZeoYGcAgPphRr01r7kOocStGca/PIYyvAfy2RgEewazLc9vmXchnGM3zyehb0X4mVf9ZqbXRUHb25TzEfqGVg+Ss2z6P6W2pVeUTCZBJ9vOcPnW87QNMCVbx9tgZ/rjUW2ISWFrNWryfw1nKILF9A4OalJ/0OHYh8SQnR0tBQ0FRlbJ7CtA0X5qrjJTYK8FHXg6uhTptUbFUXB09OTlJSUMruG5Pa46dhElwWZFwChVtZyUCc3jCbBpUwdGflFONlYUR0dprh4hEaDTWBghehBJ7kJWhtw8lEfhkK1DLwuE3IuqQ9rB7B3R7FzM/t96mRrxcBm1RnYrDqZ+UVsOJ5IXd97H9tWGlETfiacjec38lzz58okafpm/OeXe0G22oHX2r54heZawaLTG4lNzUOrVajl5YS11bWJU4akJIzZ2ShW1qqYcXOz+IGEoihYeXigcXbGkJiIITkZY2amupStKKBoUDSK+rtGo75fjebGz292rKKU+9/KWqvBzcEGNwcbNblWb+JQogOv68fy3pkF5C9/HIfhC0u1StEVGng0YE73OUzYPIGJmyeyoPcCXGyqViUaIcRaYG3Lli2fNLctZiF6rdpQ08oGHv1NLVyBWsL8+eWR/BmdxJDmAbw3qPF14a3CYCB3504yw8PJ3boNDAbsW7bA76mJuPTqdV2oiaX7oSqBjYM6kaYv+EfY5KWqA1gnH7Aqm5V++dmoOFz3fyFMkH1J/SxY26s92Yo/B/lFBuLS8ykymPCz1+KSlYwpPx+tswvW/n4o5RimJCkFrGzByVd9GApVcVOQAdkJkJ2AYu2oruZmXwaX0onuuFvcHGwY3qp02n+UuqhRFKUW8BrgKoQYWtrnvxGnM07z0f6PaOffjnGNx5XHJW9NYQ6kx6gOw/N6QVOoNxKbkodGUajl5VgiaITBgCElBUN6OigKVj4+FruScTM01tbYBAZidHPDkJKCSadT43tNAoQJIYT6/J4uUix0FA1olNsUSHdy7H+LKUVRsLfR4ulkS9OBz/PJ2kKmnv6RmEVPEjxmfpn8fzbzacbsLrOZ/NdkJv85mW97fIuDtWXmLknuAKMB/npXLcVavQU8tFgtxwrEpOQyfskhYlPzeLNfI8a0u3bmVp+YSMby5WSt/A1DUhJaT088HhuN25Ch2NYyT1iCpJSxtgP3mmplxtxkNbczP00NSXH2LbOwWEkFw1BQHG6mU8MRXfxB0SCEIDW3iMTsAqw0CrWti1ASUxEaBeuAALSurlKoWjpWtuq97uyrfg6uhKjpMuDTnmqZ/8aDodEAdcLDgrktUaMoygKgL5AshGh81fbewOeAFpgvhPhQCBEDPK4oSnhZGPxv8vX5TN02FWcbZ97r8B6ack6KvCGFucWCxhY861yXqFVkMBKTqsY7Bns5YmOlRZhMGNPTMaSkIIxGtO7uWPn4oKnksyNaZ2e0zjdOohdCqHXzhUAU/6R42w33Xff8KoH0731GI0IIRPEx15y3FMWUISWVPkoqac9+TPhCE0Mv/sSfn0OzJ+bi6Vz6M6XtqrdjZqeZTN02lee3Ps+X93+Jjbbi9giS3CO5KbBinFris+U4tXJRcdz03yeTefbnI1hpFJY83op2tf8JSSw8d460+d+TtW4dGI04duyA7+uv4dyli5yRraxY2YJb4FXiJlWdubV1UbeVYaVGiZnJT4esOEAB91pgr4aRGYwm4jN0ZBfocbdW8M5LReTno3F2VnNepS+ofFjZqfe7czVILoIu09T8y/VT4Y+X1D58jQdDw/5qPp6FcbsrNYuAr4DFVzYoiqIF5gA9gHjggKIoa4QQUXdqxKlTp+jSpQsAr7/+Ot27dyciIoIpU6Zcd+z7779Pu3bt2L17N6+++iqxWbGk6dKo51GPoXOHMnv2bMLCwvjzzz+ZMWPGda//9ttvqV+/PmvXrmXWrFnX7V+yZAmBgYEsX76cr7/++rr94eHheHl5sWjRIhYtWsSbb76J5qpZ9zo1q6PNOEdyvkKGXoEr/UyKqVWnLjGpeRTlpGNtLOBCjoIwGBBFRShCEOzhgU21aiSmp5MTE3PNa7VaLVfyBeLj48nLy7tmv7W1NbVq1QLg4sWL6HS6a/bb2toSFBQEwPnz5yksLLxmv729PTVqqEuAMTExxMbGMn78eFq0aMGRI0fw8/Nj48aNLF26lC+//JKioiJq1KjBzJkzsbe354033sDT05MjR46QkJDAe++9x+rVq4mIiKBp06Z8/fXXVKtWjU2bNvHSSy+h1+sJDAzk/fffx9HREXd3d3x8fDAajZw9e/a6v72npydeXl7o9Xpi/vW3AfD29sbD05OioiJiY2Ov2+/r64ubmxsFBQVcuHDh2p2Kgl9AAC4uLuTn5REXF4eAawSOv7c3jvb25OXlcelKLKoo+Qd/d3fsrazI0elIys4GIKVAx+MDB6JxceHrb+dxLAV0WxbT9os/sPcOxt3hH8Fxp5+9f7N+/XocHBw48/sZjD8YWZq1lPV266nlVgsFha1bt173GokFE38QfhmtzroPmAvN1PL1Qgjmbj3HJ5tO0bCaC98+2oJAD3XFThcRQep388ndsgXFzg73hx/GY8wYbAJuv4liZeb8+fM88MADdOjQgd27d1O9enVWr17N0qVLmTdvHkVFRdSpU4clS5bg4ODAmDFjsLe358iRIyQnJ7NgwQIWL17Mnj17aN26dcl9umnTJt58800KCwupXbs2CxcuxMlcuZFaa3Ctroal5Keowjj1tFrIxrma+rOKz8xXmvw8kxGy41VRY+MIbkElOVW5hWq4mdFooqamEJs0tS2BJYS6V4n7tDzQWkOXl9VHcrTa5PPESlj3PPw+FYI7qWWiG/Yrybuq6NzWsoYQYjuQ/q/NrYCzQogYIUQR8DMw4HYvrCjKeEVRDiqKclCv19+2wVeTpksjVZeKn5NfxcgfMBkh/Zy6MuPofd0XgxCC2JQ8DEaBu6MNGgSmggJEYSEoChp7e2yDgipUVbMLFy7wyCOPsG7dOlxdXVmxYgWDBw9m5cqVrF69mtq1axMe/s+iXEZGBnv27GHatGlMmjSJxx57jHXr1nH69GmOHz9OamoqM2bMYOHChaxcuZLGjRvfcIBuVq4OI9NoSh4aOzu0jo5oHBxQrKzUh7UVirU1irU1Vu7uWPv6qjlDtrZobG1RrK2xqVkTka/j8vNTcD9uT7ZrK/xFCtlJF4hJycVoKv0KhF72XtRwqUF6QTrns86X+vklZkQIOPA9LHxA9TWPbyoRNPlFBp5edoSPN56iX1N/VjzVjgB3e3K3bePCqEc5P3wEuoMH8Zo0iTp//0W111+TguZfnDlzhsmTJ3PixAnc3NxKfN6BAweIjIykYcOGfP/99yXHX/F5n332Gf379+f555/nxIkTHDt2jIiIiBKf9+eff3L48GFatmzJp59+asZ3WIzWCpz91K7kLv5q3H3aWVXg6LLufcXaghFCrBVCjHe15MR4o179v8xPVwWsZ12wUvM9k7ILiE3JxcZkpHZROjbpKWgcHLCpUwcrd/cKLWiuUGXu0/LCpyHc/xo8fRAm7oQOU9RwxbXPwid1YekQOPKjGrZWgbmXnJrqQNxVz+OB1oqieALvAc0URZkmhPjgRi8WQswD5oFaNvHfs8hhYWE3nVn2D/HH/ll7hnoM5fte32P1rxCv7t2707179/98fb9+/ejXr99/7n/44Yd5+OGH/3P/mDFjGDNmDNHR0dSvX1+NU009ow4yPOvgY2WLT7V/kq8MJhOxKXkUGkwEeTpgl6NHLwRYW2PtXx2tp8c1eRb+/v7/eW2AgICAm+6/suLyX1xZsfkvatWqhUajITg4mIEDBwLQoUMHzp8/z/Hjx3n99dfJzMwkNzeXXr16Ub9+fRwcHOjRoweKotC7d+8S5wHQokULcnNz2bt3L1FRUYwdOxaAoqIi2rZtq/4Ni9Fqtdc8/zfW1tY33W9jY3PT/XZ2djfd7+DgcNP9Tk5ON93v4uKCi4sqsk0mEzujozHm5pK+YAFpCxfRXK9nWaf2+Pgf5Q1tZ3a49OGjYaEEBnoCt//Z+y8mTZrEpEmTAPg68mvmRszlgYYPIISwiC8ryU3Q62Dd/yByGdTpAYPnlcygXUzLZ/ySg5xOymHaAw14sl0Ncjas5/L87yk8dQorPz98p72C29ChaBxlqNF/ERwcTFhYGKD6rf/yeVfo168fiqLQpEkTfH19adKkCQAhISGcP3+e+Ph4oqKiaN++PfCPz6swaLTqoNfBG3TpalGBjOJ8UCdfsHev8is3FoUQcGghGALAy08tUmSnfh8VGUzEZeSTV2jAj0KcstJQACt/f7QWImauUOXu0/JCUaBaE/Vx/xtwOQJO/KY+Vk+CtdZQp5va5LP+AyWfrYpCqRcKEEKkARNL+7xXU2gs5MXtL2KrteWjTh9dJ2jKHX2BOsOlaNSiAFbX9l4xmgTnU/MpMJgIclSwjr+AvqBA7Tfj54fGpuLmPNhe1UdGq9Wi0+kYM2YMq1atIjQ0lEWLFl0jPq8cr9FornmtRqPBYDCg1Wrp0aMHP/30U7m9h4qA1skJ72efxX3ECFLmzCHz13ByIvz5X73lODV1YPh3Op7oEMwLPevfsPHq3TKx6URyinJYErUEZxtnJoVNKrVzS8qZ9Fj45VFIPA6dX4HOL5eUiN95JpWnfzqMySRY+EhTmkRuJ6b3BPQJCdjUqY3fhx/g2qePzJe5Daqsz9No1Bh6B081gTg3SS37m3O5WPR4lHsjT8kdUpAFa56FqFXQZw14NyhpI5Ct0xOXkY/WaKR2USYaXb7aL6569Qo9Bvkvqux9Wp4oCvg3Ux/d34aEw2p42onf4PQG0NpC3R5qiFq93mo5eTNzLx4qAQi86nlA8bbbRlGUfoqizMvKyrqjC886OIuT6SeZ0X4G1Ryr3dFrSx2TQRU0oBYF+FeZTJNJcCEtj8LCIoKNOVjFX0QYjNjUqIF1jRoW6UxycnLw8/NDr9fz448/3tFr27Rpw65du0pyZvLy8jh9+nRZmFkhsfL2xu+tt6i1bi2OnbuSetyZYSt/4eu8dSzYdpb+X+3keMKd3Q83Q1EUXmz5IoPqDOLryK9ZfGLxrV9kodytP7EIzmyGeV0g8yI88gt0nQYatXLR/B0xjF6wjyBrAyucTuI3/mGSZszAyseHgLlzqbVmDW4DB0pBcw9UKZ+nKKqA8W4A7sHqSk5WHCRFqQUGTEZzWyi5EQmH4JuOamn37m+VNFw1CbX3zPm0PNz0+dTMSURTWIC1nx82QUEWOQb5L6rUfVreKAoEtIBe78GU4zBuk1qcJuEQrHgcPq4Nyx9VBU9R3q3PV0bcyxLHAaCuoijBqGJmOPDInZzgbvpK/HnhT346+ROjG42mc2DnOzK41MmMU528hzd41VVLZ16FSQgupuej5GQTXJiFYjRi5emJlY8Pirb0e5WUF++++y6tW7fG29ub1q1bk5OTc9uv9fb2ZtGiRYwYMaKkUMGMGTOoV69eWZlbIbENDibgy6/Q7d9F8rSnqLlpK2t8jjG3wQAGzcllSo/6TOhUCyvtvc+MKorCm23fJFefyycHP6GdfzvquFt4AuwNqJR9akwm2D4Ttn4Ivo3h4SXgoZZaLtAbeWXFUXbujmJG5kGab9iGQafDqXNnPMc/iUOLFmY2vvJQJX2eoqhdyO1c1RYFuUlqj4vcJHXA7Oh1XWVPiZk4tAjWv6iuqI3bAIGtIDqaQr2Ri+n56AuLCCrKwrogH42Dg7o6Y2t7y9NaGlXyPjUHGg3UaK0+er0PcXvVIgNRqyF6jdrks15vdQWnbo9yLRuviNtIBlQU5SegC+AFJAFvCiG+VxTlQWA2aknnBUKI9+7GiJYtW4qDBw/e8riE3ASGrR1GTeeaLH5gMdZl0J39tsm+DAsfILr1xzRs3l5tdHYVQggSkrOxz0jGwVCIxt5eLZFoL3sCVCWio6Np2LDhTY8ReWnkvtWLlB3ZFGZqSfKvxazgnli3aMmnD4UR5FU6+Q9FxiL2Xt5Lp4BOt3W8oiiHhBAtS+Xi5cjt+pMKT346/DYBzmyC0BHQ59MSP5OQqeONz9cQtmsd3eKPoFHApc+DeD7+BHb1y/ZL+HY+05JKSmGuKmoKs9VQNEdvtdjAv3IxbvQZsUR/UuF9ib4A/ngRDi+G2t1gyPySHLtDkcew9aqBsz4fn/xMECasfX3RenpaVO6MpOwodV9uMsKFXepqTdRqtTKnjZOaexMyWM3Fsbp3MX0zX3Jb0yxCiBH/sX09sP4ebLtt9CY9L21/CSEEMzvPNK+gyU2Bxf3VrryO3tcLGpOJ9LjLeORkgqJg7eeH1sNDOhLJDVEcPXGevg6n+b3IOp6D1eksZu76hiMxDZkY3ZdRI7sxsnWNe/782GhtblvQSMzM5Uh1KT/7kipmWo4rGTgeXLeVk7Pn8GL8cUw2tniOfATPMY9hXV1WMZOUMbZO6qMoXxU3ep0sImAusuJVH3HpMHScCl1fBY2WvEIDb6w+Tp/qBpo7pGFXmK9OqgYEVMrVGUkFQqNVy0AHd4IHPobzO9QcnOi1cOxXtSdWgz7qCk6triXlxUsTs64d30kt+L8u/sXRlKN83PljAp0Db3l8mZGfDosHqA5l1ArI/1dRgNxcdPEJOBj06B2ccAqsLhtYSW6NczWUMWtwW/gALnUSyXB8kuZLVhC2+RO2nPyLZ/qM5I1xXfF1qTjlviVlRMQytU+Ag6caShLQEiEEuX9vJfqzOTifPk59W0c0Y56k7oSxWLm7m9tiSVXDxkENg6zCZZ/NSux2+HWsWob74R+hYV8Ajidk8exPRwg4uhuvMW2wK7JXV2e8vOSkqqR80VpB7a7qo8+nELOtWOCsg8ifwM5N/dyGDILgziUFLe4Vs4qaO4mB7xXUCz9HP5p6Ny0Hy/4DXSYsGaQWBhj5C9RsB9HRAAiDAX1iIsbMTEwaK3Refnj5ytUZyR3gXhNGr0azoDeehh9wW/ErqT+v5/7FS+g0L5LFWzvSaOoz9OkgQ38qJYZC2DANDn4PQR1h6EKErRvZa9aQ+t18is6cQWfvxoFuIxk54xnc3C24h4akciC/38oXIWDPHNg8XS1MNPxH8KqLySRYtPs8c1Yd5Jnjq2h9/jCGJzpgW7t2hep7J6miaK2hbnf10fczOPd3cZno1XBkKdh7qA0+m41S88HuAYvK8jOroCnMgR+HQtIJGPET1OqibhcCQ3o6hqQkhNFEuq0LeHri7+4gBY3kzvGqC6NXwaI+aH8bie9TG/B4dBQxMz/lwT9+J3/iHhZ0Hsigd5/H3aNi1YeX3ANZCfDLaEg4CO2nYGo7lczfVpO+cCH6S5dI9PBnafPhhDw6lCm9GqHRSN8ikZQVdxJFUm4U5cHqp9XZ7ob9YeBcsHUmLj2fl3+JwHrnX3wbtRbHony8p0whxctLChpJxcPKFur3Vh/6Aji3RS0ycCxcLXRRlUSN2SjKh2UPqzW6H1qsVnMACk6fxpCWht5kwmhrT7yDC47OjgS420tBI7l7qjWBkeGweCAsGYT1mN+p/9nH5I1/nENvvE/bLT8TtWcDVo9PoNXERy26kp4EdVk+fBwYCjE88A0ZB9PJeKMnxsxMjCFN+bxeX3Z71+eTh5rxQBO/W59PIpHcExWukmLaOVg+ClJOquWa209BAD/vPMver5fw+Km/8c9NwbZhA/w//BC7+vVJKY4ikUgqLNZ2ao5Ngz5qfp6h8J5PadZOWhbRV0JfAD+PgIt7YMh30LAvJp2O5FmziB08BKE3UOTlS4y9J/ZODlLQSEqHwFYwYpn6ZbZ0CBRk49iwAZ3CF2P8dC7ZTu64fPkRe7v2JnXTn9xOFUNJBUMI2DkblgxEb/IgUfcIZyd+TOqXX2EfFsbZ12YxqMFjnKoVxsrJHaWgkUiqIqc3wryuahPUUSugw/NcupTK3IlvE/j0SCYe+oWAAC+qz55NcHg4dvXrm9tiieTOsbZXS8jfI2YVNUKItUKI8a6uFTQ23FCkdvCO2QYD5kLjIeRs3UpM336kfTcf1/790bt7csFgjbOdNYEeMuTsavR6Pc2bN7/r11+6dImhQ4eWokUWRq0u8NAPaiWsn4arK4ZA4we70uXPNewZPZW8vAJSnn2GE8NGkH/kiHntNTMWMUlyBSFg3fMUrniXS9GNObtUR8Zvf+DSswcBv61iXu/JPHNC0KaWJ2uebk/9as7mtlhyG0ifJyk1TCa1P9Wyh9R8y/Hb0Ls0YcdLb3Opd0/u37Ycq1q1Cfj+e+r9tgKX3r3kqv1tIu/TyotZRU2FxqiH8LFqj4i+n6H3u5/4Z58jfuJTKHZ21FyymOMjnyFDZ8DR1oqaHg5opKC5hp07d9K+ffu7fr2/vz/h4eGlaJEFUv8BGDwPLuxWcy4MRQDY21gx7tXHcVwWzpI2w8k+E8OFEY9wcfIzFMbEmtlo81DhJ0muQh8+jYQvVxHzhw/Zp/JwH/EIdTZuwPr1d3h8WzpL9l5gQudaLBrbCjeHytPxu7IjfZ6kVNBlqhEiWz+A0BEU9fye2E++52SXbnisWU5sUGNs5y+hzcplOLdvJydT7xB5n1ZepKi5ESYTrHoKTq5D9PqQ9CgrYh7sQ+62bXhPmUKt31ZyyC2Ip5cdwdpKQ5Cno9kSd8+fP0/Dhg158sknCQkJoWfPnuh0OgC6dOnClcZhqampBAUFAbBo0SIGDhxIjx49CAoK4quvvuLTTz+lWbNmtGnThvT09JLXP/fcc4SFhdG4cWP279+PyWSibt26pKSkAGAymahTp07J86vZsGEDDzzwwDXbjEYjY8aMoXHjxjRp0oTPPvsMgLNnz9K9e3dCQ0Np3rw5586d4/z58zRu3LjE5gEDBtClSxfq1q3L22+/DcD06dOZPXt2yflfe+01Pv/881L661YQmgyFfrPh7GZY+aTa4KqYtvV9mfbNq6x98UsWN+hF2vYdxPTrx+Xpb6JPSjafzZIbYszNI3nqaM69uYqcy454TphAnb+2UO21V4kWjvT/aidHEzL5YkQzpj3QEK0sCHAd0udVAZ9XlUmKgu+6wtk/KWj0EgnbHTjbZxA5v63kzxotODLjWwavXkytDhW7j6m8T+V9ahaEEGZ7AP2AeXXq1BEVim0zhXjTReQvflnEDB4iouo3EBcef0IUXrgghBBif2yaqP/6etHrs23i+PETZjU1NjZWaLVaceTIESGEEMOGDRNLliwRQgjRuXNnceDAASGEECkpKaJmzZpCCCEWLlwoateuLbKzs0VycrJwcXERX3/9tRBCiClTpojPPvus5PVPPPGEEEKIbdu2iZCQECGEEG+99VbJMRs3bhSDBw++oW333XefyMvLu2bbwYMHRffu3UueZ2RkCCGEaNWqlVi5cqUQQgidTify8vJEbGxsyTUXLlwoqlWrJlJTU0V+fr4ICQkRBw4cELGxsaJZs2ZCCCGMRqOoVauWSE1NveO/Y1kRFRVVeifb9aUQb7oI8dskIYzG63ZvPH5ZdHl1hZjZ5wlxvFFjER0aJpI++0wYsrPv6nLAQWFG/3C3jxYtWtzV+y1LTAaDyPj1V3GqVUsRVb+BiB/aVhTFxZXsDz8YJ+q+tl60+2CLOBafaUZLb02pfqbvAunzKrbPE+LGnxFL9Cfl7kuOhQvTu9VE7vN1xYVH1LFHRJMw8XbfiWLkh2vFmaTb9+XyPpX36a0w92fkbrmZL7GYPjXlxpnNiC3vkZzQkvRf1qL19KD6Z5/i3Ls3iqJwND6TcQsP4O9qz5LHW5MaH1Py0rfXniDqUnapmtPI34U3+4Xc9Jjg4GDCwsIAaNGiBefPn7/lebt27YqzszPOzs64urrSr18/AJo0acLRo0dLjhsxYgQAnTp1Ijs7m8zMTMaNG8eAAQOYMmUKCxYsYOzYsdedPyEhAQ8PDxwcHK7ZXqtWLWJiYnjmmWfo06cPPXv2JCcnh4SEBAYNGgSA3X+UoezRoweenp4ADB48mJ07dzJlyhQ8PT05cuQISUlJNGvWrOSYSke7p9XS4ts+VLt69/7wmj4RPUOq0XxaH6atrMHKAx14/uJfNPnmWzJ/Xo7XpKdwGz4cjY0MZSpv8vbuI+mjjyiMjsbeS4/vozWwf2k9WNthMJp4f/1JFuyKpU0tD+Y80hxPJ8vp+i193j9Inye5a4wGxKbp5IR/T9q5ahQkFmF0u0R4WD9WBbbmyQdDWdSpFlbauwuukffpP8j7tHIjSzpfTdo5RPjjJJ0MJiPyEm4PPYTPi1PROqtJuqcScxi9YD+uDtb8+GRrvJ1tSTWzyQC2tv8MgrRabckSr5WVFSaTCYCCgoL/fI1Goyl5rtFoMBgMJfv+HaurKAqBgYH4+vry119/sX//fn788cfrbNqwYQO9evW6bru7uzuRkZFs3LiRb775hl9++eW2l2RvZAvAE088waJFi0hMTGTcuHG3dS6LpcsrqrDZOwdsXeD+167Z7eVky7xHWxDeyJe311ajVo2OvHHpL4zvf0D6D4vxnvIcLn36oGhk5GlZU3ThAkkff0zun1uw8vHEv0M+Ls2ro4xdCdZ2ZOQVMXnZYXafS2Ns+yBefbAh1nc5aKlqSJ9XhXxeJceUnkD2uyNI23mJohwPtAE+7OzXm5nUpk6AJ8uGhdLI3zJ7ksn7VN6n5Y0UNVcozIXlo0iNsCEjsgCPsWPxeenFkg9nbGoeI+fvw9ZKw49PtMbP1f66U9xq1qK8CQoK4tChQ7Rq1equk9qWL19O165d2blzJ66urlxJwn7iiScYNWoUjz76KNobVFzZsGED77777nXbU1NTsbGxYciQIdSvX59Ro0bh7OxMQEAAq1atYuDAgRQWFmI0Gq977ebNm0lPT8fe3p5Vq1axYMECAAYNGsT06dPR6/UsW7bsrt6nxaAo0Os9KMqB7TPVFZv2z/3rEIVhLQNpW9uTqb9GMsyhGo837sHDR1Zz6eVXsAsJwbZWLTO9gcqPMTub1K+/IX3pUjTW1niPH4WHbj4aR1d4dCXYuxGTksvYRQe4nFXAx0ObMqxloLnNviukz/sH6fMkd4IxN4/MebNIX/oThnywDQok66mJPH/Jg+R8PZO61OaZ++tiY3XvEx3yPv0HeZ9WbqSoAbW86pqnSdtxkdRIF1yHDL5G0MRn5DPyu72YhODnJ9pQ09PRzAbfHlOnTuWhhx5i3rx59OnT567OYWdnR7NmzdDr9SU3KkD//v0ZO3bsDZd3jUYjZ8+epUGDBtftS0hIYOzYsSWzNB988AEAS5YsYcKECUyfPh1ra2t+/fVXNP9aTWjVqhVDhgwhPj6eUaNG0bKlmihpY2ND165dcXNzu6ETq3QoCvSdrQrxzdPB1hlaXj8LFODuwLIn2rBgVywzN55idcun+HiMFQ2loCkThMFAxi+/kPrlVxgzM3EdPAjvx4djvXoE2Gjh0d/AuRr7YtIYv+QQWo3CT0+2oUVNd3ObXmmQPk9S0TGkpZG+ZAkZi3/AlF+AQ3UFj9df5hPRjJ8OxFHHx4aVj91HaKCbuU0tM+R9Kikz/ivZpjwfZk/s3TlbZDzqK6LqNxBxzzwrTAZDya6kLJ3oPPMv0fjNDTdM4LXURKvb4epkvn9z4MAB0aFDhxvu27Fjh5gwYUKp2rJw4UIxefLkG+4zGo0iNDRUnD59ulSvWRqU6efDUCTEjw8J8aarEJHLb3roqcRs8eDn20XNl9eJU4m3l2yKBSb2CjP5k5zt28XZPn1EVP0G4vyjo4XuxAkh8tOFmNNGiPeqC5FwRAghxIpDcaLOq7+L+z/5W1xIzbv5SSso0uddj/R51yILBVxP4cWL4vLbb4vopqEiqkEDEdetpsj/sJfYf+KMaP/hFhH0yjrx3u9RQldkuPXJbgN5n16PvE+vxVI/IzfzJWZdqVEUpR/Qr06dOuYz4txfZC/8gMsH3HFs1w7/Tz4uaWCVnlfEqO/3kZxTyJLHW9O4esXvf1EefPjhh3z99dc3jFcF6NChAx06dCgXW6Kioujbty+DBg2ibt265XLNCoPWGoYtgh+HwW8TwcYRGtx41querzO/TWrPzrMp1POVjRxLi8Jz50j66CPytu/AukYNAr76Eqdu3VD0OlgyCNLOwshwhF8oszef5vMtZ2hby5NvRrXA1cHa3OZLbhPp86oWpTk2KTh5krTv5pP9xx+g0eDW0BqPwCQ0vZ/lg4LBLFh8iiBPB36d0JaWQR73bnwVRt6nEkUVPealZcuW4krN8nIl4zx5b91P3GZb7Bo3ocbChWgc1dCy7AI9j3y3lzNJuSwcex/tanvd8BTR0dE0bNiwPK2WWBDl8vkozIHFAyHxKDzyC9TuWiqnVRTlkBCiYjdDuAHl4U8MGRmkfvkVGcuXo3FwwOupp3AfNVKtLmfUw/JRcHojDFtEYf1+vLLiGL8dSWBoiwDeH9SkVOLkzYX0eZJbcaPPiCX6k7v1JUII8g8cIO27+eTt2IHGwQG3BzviYbUOaxsd59p/whMH/IhNzeOxtjV5+YEGONiU7hyzvE8lt8JSPyM38yVVN6emKB/dZw8Tt8UGm+AgAr/7rkTQZOn0jF24n1OJOcx7tOV/ChqJpEJg6wwjf4VFfeHnR+DRVVCjtbmtqpSIoiLSly0jde7XmHJzcXv4IbyfeQYrj+IZViFg7XNwegP0+ZSMoAeZMH8/+8+nM7VnPSZ3rSO7f0sklRRhMpH711+kfvcdBZFH0Xp64j3lOdzr5KPdNQOTRy2+q/4lH2ww4edqYtkTrWlXR44vJJLSomqKGiEo+O5xLq7KxMrbh8CFP6AtrpyRlF3AYwv2cy4lly+GN6NrAx8zGyuR3AYOHmoi+sIH1HC0MWvBL9TcVpUrZRnOKoQg96+/SJ75MUUXLuDYoQO+L7+E7b/DCv58EyJ+hC7TOB88nLFf7yYhU8cXI5rRP9S/1O2SSCTmRxQVkbV2HWnff09RTAzWAQFUe3M6rn16otn8Muz8layavXgsYywR+02MaBXIqw82xNlOhqBKJKVJlRQ1RWs+5OJ3h9A4OFNjyU9Y+6jCJSYll0e/309mfhELx7SiQ105gyKxIJx9YfRqWNBbzecY+wd41ze3VeWGKKNmvgUnT5L04Ufk792LTe3aBM77FqdOna4/cPeXsOtzuO8JDtR8kvFzdwGw7InWMlZeIqmkCKORmAEDKYqNxbZBA/xnfYJLr14o2Rfhx/6IpBPsqjmJMWfa4+lkzcKxTelaX06WSiRlQZUTNfqDa7j4zgJQbKnxw0/YBAQAEBmXydhFB1CAn8a3oWmAm1ntlEjuCrdAeGyNKmwWD4Rxf4B7kLmtskgMKSmkfPEFmeEr0Lq44Pv667g//BCK9Q1mVyN+gk2vQ6OBrPZ/jhfn7yfA3Z4FY+4jyMsySsBLJJI7R9Fq8XhsNNbVq+PYoYMaXnpmM6x4HKOAt5zeYsmpugxuVp03+4XIAiESSRli1mxVRVH6KYoyLysrq1yuZ7hwgrinX8RYZEXgd/OxrV8PgB1nUhjx3V4cbLSEP9VOCppSwsnJyWzX3rt3L08+efcT9mvWrOHDDz8sRYvKEc/aaiiaPh8WD4Dsy+a2yKIwFRaS+u08zvXqTeZvq/AYPZramzbiMWrkjQXN6U2wejIiuDNz3F/iueXHCKvhxspJ7aSgqWJIn1c1cR8+HKeOHVGEgG0zET8OI8WqGt1z3+GPghC+fbQFnz4cJgVNBUHep5UXs67UlFW4yI0wZaUR99hwirIVAj97F/uWbQBYHZHA1F8jqe3txOJxrfBxsStrUySljNFovK6x1R9//EHv3r3v+pz9+/enf//+92qa+ajWGEatUEXNkoEwZj04eprbqgqNEIKcDRtI/mQW+oQEnO6/H58Xp2IbHPzfL4rbD7+MxuTbmNdtprFsy3kGN6vOB0OaYGslm61Jygbp8yogBVmwcgKc/oNttl2ZmDqa+5vU5N0BjfF0sjW3dRIzIO/T8sdy64reAaaCAuIe6UtBkp7qL4/DsddQABbuiuW5nyNoVsOd5RPaWqSgOX/+PA0bNuTJJ58kJCSEnj17otPpAOjSpQtXylGmpqYSFBQEwKJFixg4cCA9evQgKCiIr776ik8//ZRmzZrRpk0b0tPTS17/3HPPERYWRuPGjdm/fz8mk4m6deuSkpICgMlkok6dOiXPb4QQghdffJHGjRvTpEkTli9fDsDkyZNZs2YNAIMGDWLcuHEALFiwgNdeew2ApUuX0qpVK8LCwpgwYQJGoxFQZ1peeOEFQkND2bNnz3XX3LJlC927d79m2+XLl+nUqVPJ+9mxYwcAGzZsoHnz5oSGhtKtW7eSv9HTTz8NwJgxY5g4cSItW7akXr16rFu3DoBOnToRERFRcv4OHToQGRl58/+w8iSgJYz4GTLOw9LB6peu5Ibojh3jwshRJDz/PzROTtRYuIDAuXNuLmiSo+HHYRidqjFRTGNZZDrPd6/HrIdCpaApQ6TPkz6vwpEcjZjXFdOZzbxjHMOUoqeYOaINcx5pXmUFjbxP5X1qDiq9qBEGAwmPDyX/XCZ+ozvg/NhLCCH4eONJ3l4bRa8QXxaPa4WrveUuC585c4bJkydz4sQJ3NzcWLFixS1fc/z4cVauXMmBAwd47bXXcHBw4MiRI7Rt25bFixeXHJefn09ERARz585l3LhxaDQaRo0aVdLc6s8//yQ0NBRvb+//vNbKlSuJiIggMjKSP//8kxdffJHLly/TsWPHkps3ISGBqKgoAHbs2EGnTp2Ijo5m+fLl7Nq1i4iICLRabcl18/LyaN26NZGRkdc100pNTcXa2hpX12ubpS5btoxevXqV2BIWFkZKSgpPPvkkK1asIDIykl9//fWG7+H8+fPs37+f33//nYkTJ1JQUMDjjz/OokWLADh9+jQFBQWEhlawimPBHeGhxZB0HJY9DEX55raoQqFPTOTSyy9zfthDFF24QLV33iZ45Qoc27a9+Qsz42DJYAwaGx7VT2NrPHz2cCjPda8rSzaXA9LnSZ9XYTAUYVg8mMzMdB4qeJULtUex6fnO9A/1r/K+QN6n8j4tbyp1oQBhMnH5+YnkHjqHby9/3F6eh8Fo4rXfjrP8YBwjWgUyY2ATtJpScjx/vAKJx0rnXFeo1gQeuHn8ZHBwMGFh/2/v3sOqqvI+gH8XiIKC5A0rufoipqKCII0aqJm3yVt4KccbKnjX1JrRJht608mKxnybykuTkdiYl7HMMc2sYZQ3JwFrtMlrgolZosRFQeTye/84eF7xHOAAh3P25nw/z3MeOWvtyzprr/XzrL322TsEABAWFobMzMwaNztw4EB4eHjAw8MDnp6eGDlyJACge/fuOH78uHG5iRMnAjCM/PPz85Gbm4sZM2Zg9OjRWLx4MTZt2oTp06dXu6+UlBRMnDgRzs7OaN++Pfr374/U1FRERkZi7dq1+O6779C1a1f88ssvuHz5Mo4cOYLXX38d7733HtLT09G7d28AQFFREbwq7lTn7OyMsWPHmt3fgQMHMGTIEJP03r17Y8aMGSgpKcGYMWMQEhKC5ORkREVFIaDijHzr1ubvUjVhwgQ4OTmhU6dO6NixI06dOoXx48dj5cqVSEhIwKZNmxATE1NtPdhN0FAg+m3gbzMND4WcuBVo4phnD28rLyzEtU3v4tpf/gKUl6NNXBzazJ4FZ0uutb5xDdgSjdLiAkwqjcdptELSzDA82NEBL+9jzDOLMc9xlKgmWFg8H2fK2mDu2IcwLsxbe4MZ9lOz2E8bn0Y7qBERXHnhOeR99r9o29sFrV/ehZtlggV//RoHT/6MRQ8HYsngIO0Fnzpo1uz/v6A6Ozsbp3ibNGmC8vJyAMDNmzerXMfJycn43snJCaWlpca8u+tHKQUfHx+0b98eX3zxBY4ePYr3338fFy9eNAafOXPmYM6cOTWWu0OHDsjNzcX+/fsRFRWFnJwcbN++He7u7vDw8ICIYNq0aVi9erXJuq6uribXqt62b98+LF261CQ9KioKhw4dwt69exETE4OlS5eiVatWNZbz9ue++33z5s0xePBg7N69G9u3b0d6erpF27KL4Gjg1g3g8/8G8rIMNxNwYD+tXIW8Dz+Ex7Bh8Hr6KeNdEGtUfB346wSU5WRi6q1n8LNnIHbF9EbHdvb74akjYsyrjDHPflycnTB5whPo2K4F7r/Hzd7F0RT208rYTxteox3UXH3zDeR8sAutHihG21d3Ia+8OWITv0LahV/w36O6YVpff+vvtIazFrbm7++P9PR0REREYOfOnXXaxrZt2zBw4ECkpKTA09PTOG0aGxuLyZMnY8qUKXB2doaPj0+lazjvFBkZiQ0bNmDatGnIycnBoUOHkJCQAAD41a9+hbVr1+KLL77AtWvXMG7cOIwbZ/jN06BBgzB69GgsWbIEXl5eyMnJQUFBAfz8/Kosr4jg+PHjxrNDd7pw4QK8vb0RFxeH4uJiHDt2DM8++yzmzZuHjIwMBAQEICcnx+wZkR07dmDatGnIyMjA+fPn0blzZ2M9jBw5EpGRkRYHIbvpNQXoOgpw9ax52Uau7ZzZuGfcWDQPC7N8pdJbkO1TIZeOYc6txSjx/RV2TQlH6xZNG66gWseYZ3YbjHmORfPPtGM/NbsN9tPGp1EOanKStuDqG2/B078Q7Vf9D664BWDqhiM4f/U6/jwxFCN6OMaTvZ9++mlMmDABGzduxKOPPlqnbbi6uiI0NBQlJSXYtGmTMX3UqFGYPn16jdO7gOGHdkeOHEHPnj2hlMIrr7yCe++9F4AhqBw4cACBgYHw8/NDTk4OIiMjAQBdu3bFqlWrMGTIEJSXl8PFxQVvvvlmtYEjPT0doaGhZmfgkpOTkZCQABcXF7i7u2Pz5s1o164dNm7ciOjoaJSXl8PLywufffaZybq+vr6IiIhAfn4+1q9fD1dXw00lwsLC0LJlS4vqQRM4oAEANPXzQ9Nq2pGJ8nKUfTgXzt9/juUlcWjeYxTeGNeDNwTQGMa8yhjzSIvYTytjP7UiEbH7KywsTKwl96OP5LvOD8gPD/tJ+afxcu5KgfRd/bl0fW6fHD6TbbX93Pbdd99ZfZta0b9/f0lNTTWbl5qaKg899JCNS1SzlStXytatW626zWnTpsmOHTvM5l26dEk6deokZWVlZvP13D4ApIkG4kNtX9aMJyIiUl4uN/f8ViS+pbz8+zj504HTUl5ebt196Iie23RNGPMM6hPzRMy3ET3GE6vHEhtiP9UWvfRTPagulth1pkYpNRLAyMDAQKtsr+CLL/DjM79H8/a30GFKKI53Xojp649AAfhgVh909+ZZamt46aWXsG7dOuPdPrRkxYoVNtvX5s2b8eyzz2LNmjVwcmr0NxJ0WLmfvYJ70jYgsWwY/uux5zA23MfeRSIbY8wzYMwjLWM/NXDkfqoMgx77Cg8Pl9v3LK+rG//6ChdnxaGZ5y34jmqKr4bvQuz2c2jdoimSZj6IgAZ6svfJkyfRpUuXBtk26Z+e24dSKl1Ewu1djtqyRjy57cLB9fBLWYa9eAitJyeiT2DVtwd1FHpu02Qb5tqIHuOJNWOJrbGfUk302kaqiyWN4jc1RSdOIGvePLh4CHwG5uOfEe9j1l/PINDLHZtnROjyoZpEZF9pn25B6JfL8ZVzKDrHJSHwPvO31CQiIiL70/28VPG5c7gYNwvOrgLfflk43OMPmLGvEGF+rbB9Th8OaIioVkQEH+3egeAvF+N7lyAELvgbBzREDkQpNVIptTEvL8/eRSGiWtD1oOZW1iX8MDMWQAl8+2Ti2H/9BtPTfDC0W3u8NyMCLV1d7F1EIrIRa3wRKSkrx/+8/yEePrYIuU3vhe/Cv6NNawd8qCaRAxORPSIy6+4nvxORtul2UFOanY0fZs5A+Y0C+PTNQua9ofhNxlBMjPDFW5PC4OrCW60SOZL6fhHJv1mC3739MX5zdglUM3d4zfsErp5eVi4lERERNQRdDmrK8vLwQ2wcSn++Ap+Hb+BG65aYcC0O8x/ujBcfC4azk+l9wEk7Ll++jCFDhtR5/bS0NCxatMiKJSJHdzGnEDPf+ARPXl4Gz6bl8Ij9GE6tfO1dLGokGPOItI/9VP90d6OA8sJCXJw9B8Xff4/7x7RGkyZZmFb4PJaO/hWm9vG3d/HIAvv378fQoUPrvH54eDjCw3V1Ex3SsG8u5mJR4mGsK3sePk1y4TzlY8BLf3eEIe1izCPSPvZT/dPVTE35rVvIWrgIRcePo+0TPeDZ5Bs8UxqLuU885tADmjVr1iA4OBjBwcFYu3YtACAzMxNdunRBXFwcunXrhiFDhqCoqAgAMGDAACxbtgwREREICgrC4cOHAQCvvfYaZsyYAQA4ceIEgoODUVhYaLI/f39//O53v0P37t0RERGBc+fOoaCgAAEBASgpKQEA5OfnV3p/p/3792P48OGV0m7cuIFHH30UPXv2RHBwMLZt2wYASE1NRd++fdGzZ09ERESgoKAAycnJGDFiBADg+eefx5QpU9CnTx906tQJb7/9NgBg6tSp+Oijj4zbnzRpEnbv3l3XKqZGav+3lzFl4yGskVfQVV2A8+ObAd8H7V0sqgFjHmMeaR/7KfupzVX1VE5bvix5am95aalcfHKxfNf5AbmweolIfEt57w9PSMrZ7Do8j9R67P1E1rS0NAkODpbr169LQUGBdO3aVY4dOyYZGRni7OwsX3/9tYiIjB8/XpKSkkTE8DTepUuXiojI3r17ZdCgQSIiUlZWJpGRkbJr1y4JCwuTlJQUs/v08/OTVatWiYjIe++9J48++qiIiMTExMiHH34oIiIbNmww7uNOpaWl0rNnT5P0nTt3SmxsrPF9bm6uFBcXS0BAgBw9elRERPLy8qSkpET+8Y9/GPcZHx8vPXr0kMLCQsnOzhZvb2+5dOmSJCcny+jRo43b8vf3l5KSEkur1Wrs3T7qAzp8ArhYGk/Ky2XDP89Jx+UfS8qLvxaJbyny9V/rVlEOxt5tmjFP2zFPxHwb0WM8sSSWaBX7KftpTezdRuqquliim8vPCg4cQMH+/VCTx8Pr5ltIQ1eEzvwzuvu2tXfRjF4++jJO5Zyy6jYfaP0AlkUsqzI/JSUFjz32GFq0MDxcNDo6GocPH8aoUaMQEBCAkJAQAEBYWBgyMzON60VHR5ukOzk5ITExET169MDs2bPRr1+/Kvc7ceJE479LliwBAMTGxuKVV17BmDFj8O677xrPTNzpq6++woMPmp4J7969O5566iksW7YMI0aMQGRkJE6cOIH77rsPvXv3BgC0bNnSbFlGjx4NNzc3uLm5YeDAgTh69CjGjBmDefPmITs7G3/7298wduxYNGmim+ZODezl/aex/p/n8F77HeiXlwIMXgmETLR3sXSHMY8xj7SP/ZT91FHo5vIzj2HDkP+HF9Cy5H3kqZZoN+MDTQ1otKhZs2bGv52dnVFaWmqSd3f62bNn4e7ujh9//NGYNnToUISEhCA2NtaYppQy+btfv37IzMxEcnIyysrKEBwcbFKmffv2YdiwYSbpQUFBOHbsGLp3744VK1bghRdesPhz3lmWO99PnToVW7ZswbvvvmucuiYCgIcf8MKWwGT0z9sN9F0E9OOPOxsDxjzGPNI+9lP20wZT1RSOLV4ARgLYGBgYWON0U/J3lyTlub5SHN9Grp3+sn5zV1Zk7+m79PR06d69u9y4cUOuX78u3bp1M07xduvWzbhcQkKCxMfHi4hhijc1NVVERLKzs8XPz09EDFOhQUFBcvr0aRk8eLDs2LHD7D79/Pxk9erVIiKSlJQkI0aMMOa9+uqrct9998lbb71ldt0+ffpIfn6+SfqlS5ekqKhIRET27Nkjo0ePNpnizc/PNzvF27NnTykqKpKrV6+Kj4+PXLp0SUREfvrpJ/H19ZWIiAiL6rIh2Lt91Ad0eLmIWHrJyNG/GC452zVHpLy8TvXjqOzdphnztB3zRHj5mRawn7Kf1sTebaSuqosldp3zEpE9APaEh4fH1bRseFEKWjh9i6Jf/xmtg/rYoHT60KtXL8TExCAiIgKAYZo1NDS00nSupZYsWYL58+cjKCgI77zzDgYOHIioqCh4eZk+q+OXX35Bjx490KxZM2zdutWYPmnSJKxYscI4BXyn7OxsuLq6wsPDwyTvxIkT+O1vfwsnJye4uLhg3bp1aNq0KbZt24aFCxeiqKgIbm5uOHjwoMm6PXr0wMCBA3H16lU899xzuP/++wEA7du3R5cuXTBmzJha1wU1cs08gAdGAKNeBxRvAa8njHmMeaR97Kfsp3ZR1WjHli+LzoaUl4tkmP9xmD3pdaRbH35+fpKdbf4GDTt27JDJkyebzUtKSjKeRbGW+Ph4SUhIMJt348YN6dixo+Tm5lp1n7Wh5/YBHZ5ZldqcXeUMTZ3ouU3XFWNe7XCmxv7YTytjPzWl1zZSXSzRz6+TlAL8q/5xGNnfwoULsW/fPnzyySdm8ydPnmyzshw8eBAzZ87EkiVLUNcnzFMjxxkaqifGPCLtYz91HMow6LGv8PBwSUtLs3cx6uTkyZPo0oUP6iPz9Nw+lFLpIqK7J4npOZ7ogZ7bNNmGuTaix3ii51jCfko10WsbqS6W6ObuZ0REREREROZwUENERERERLrGQQ0REREREekaBzVERERERKRrHNRQvSUmJlZ6yu/d/vWvfyEursZHEVXp448/xksvvVTn9YmIrIkxj0j72E8dDwc1VC9lZWU1Bo59+/Zh2LBhdd7HqFGjsHz58jqvT0RkLYx5RNrHfuqYOKhpBNasWYPg4GAEBwdj7dq1AIDMzEx06dIFcXFx6NatG4YMGYKioiIAwIABA7Bs2TJEREQgKCgIhw8fBgC89tprmDFjBgDDU3SDg4NRWFhosj9/f38sW7YMvXr1wtatW5GWloZJkyYhJCTEuI87ff7553jkkUcqpV2+fBlRUVEICQlBcHCwsQz79+9Hr1690LNnTwwaNAiA4WzLggULAAAxMTGYM2cOwsPDERQUhL///e8AgKioKHzzzTfG7T/00EP497//XdcqJSINY8xjzCPtYz9lP7W5qp7KacsXn9pbd2lpaRIcHCzXr1+XgoIC6dq1qxw7dkwyMjLE2dlZvv76axERGT9+vCQlJYmISP/+/WXp0qUiIrJ3714ZNGiQiIiUlZVJZGSk7Nq1S8LCwiQlJcXsPv38/OTll182vu/fv7+kpqaaXTY7O1sGDBhgkv7qq6/KqlWrRESktLRU8vPz5cqVK+Lt7S3nz58XEZFr166JiMi7774r8+fPFxGRadOmydChQ6WsrEzOnDkjHTp0kKKiIklMTJQnn3xSREROnz4tWmlT9m4f9QEdPgFcdB5P9MDebZoxT9sxT8R8G9FjPNFSndYW+yn7aU3s3UbqqrpY0sTOY6pG5acXX0TxyVNW3WazLg/g3t//vsr8lJQUPPbYY2jRogUAIDo6GocPH8aoUaMQEBCAkJAQAEBYWBgyMzON60VHR5ukOzk5ITExET169MDs2bPRr1+/Kvf7+OOPW1T+AwcOYMiQISbpvXv3xowZM1BSUoIxY8YgJCQEycnJiIqKQkBAAACgdevWZrc5YcIEODk5oVOnTujYsSNOnTqF8ePHY+XKlUhISMCmTZsQExNjUfmIqO4Y80wx5pHWsJ+aYj9tnHj5WSPWrFkz49/Ozs4oLS01ybs7/ezZs3B3d690HerQoUMREhKC2NhYY9rtQFWTqq5ZjYqKwqFDh9ChQwfExMRg8+bNFn8upZTJ++bNm2Pw4MHYvXs3tm/fjkmTJlm8PSJqHBjziLSP/ZQaCmdqrKi6sxYNJTIyEjExMVi+fDlEBB9++CGSkpLqtK28vDwsWrQIhw4dwoIFC7Bz506MGzcOn376abXreXh4oKCgwCRdRHD8+HHjGZk7XbhwAd7e3oiLi0NxcTGOHTuGZ599FvPmzUNGRgYCAgKQk5Nj9ozIjh07MG3aNGRkZOD8+fPo3LkzACA2NhYjR45EZGQkWrVqVac6ICLLMeZVxphHWsR+Whn7aePFQY3O9erVCzExMYiIiABg6DyhoaGVpnMttWTJEsyfPx9BQUF45513MHDgQERFRcHLy6va9W7/QM7NzQ1HjhyBm5sbACA9PR2hoaEmZy8AIDk5GQkJCXBxcYG7uzs2b96Mdu3aYePGjYiOjkZ5eTm8vLzw2Wefmazr6+uLiIgI5OfnY/369XB1dQVgmK5u2bIlpk+fXuvPTkT6wJjHmEfax37KfmoPyvCbGytuUKkWAN4CcAtAsoi8X9M64eHhkpaWZtVy2MrJkyfRpUsXexdDk1atWoXAwEA88cQTVttmTEwMRowYgXHjxpnk/fjjjxgwYABOnToFJydtXFmp5/ahlEoXkXA7l8Gh4oke6LlNNzTGPANzbUSP8UTPsYT9tGrspwZ6bSPVxRKLalcptUkpdUUp9e1d6cOUUqeVUueUUrdv1h0NYKeIxAEYVa+Sk66tWLHCqkGjOps3b8aDDz6IP/7xj5oKGmSK8YQaK8Y822M8odpiP228LL38LBHAGwCMv5hSSjkDeBPAYABZAFKVUh8D8AZwomKxMquVlAiG+8KbM3XqVEydOtW2haG6SgTjCZFFGPNqlAjGE7Iz9lNtsGhQIyKHlFL+dyVHADgnIucBQCn1AYDRMAQQbwDfwMKZoNOnT2PAgAEADCPoRx55BN988w0WL15ssuyLL76Ivn374ssvv8Tvzfz4be3atQgJCcHBgwexatUqk/wNGzagc+fO2LNnD/70pz+Z5CclJcHHxwfbtm3DunXrTPJ37tyJtm3bIjExEYmJiYiPj680+g4MDISzszOuXLmCX375xWT92z8c++mnn5CXl1cpTymFoKAgAIbpyrt/4Obs7IzAwEAAQFZWFm7cuFEp38XFBR07dgQA/PDDDyYPm2rWrBn8/f0BGB6AVVxcXCnfzc0Nvr6+AIDz58+jpKSkUn6LFi3g7e0NADh37hzKyir/n+Dh4YH7778fAHDmzBncfWmjp6cn7r33XgCGY363Vq1awcvLC2VlZTh37pxJfps2bdC2bVuUlJTg/PnzJvnt2rVD69atcevWLWRkZJjkt2/fHvfccw9u3ryJCxcumOTfd999aNmyJQoLC3Hx4kWT/A4dOsDd3R3Xr1/HpUuXTPJ9fHzQvHlz5Ofn4/LlywAMx3nu3LkArN/27vbJJ5+gefPmeOutt7B9+3aT/OTkZJM0e2joeHL27FljPLltxIgRePrppwHAJA8w3Ipz3rx5KCwsxK9//WuT/JiYGMTExODq1atmLy+YO3cuHn/8cVy8eBFTpkwxyX/qqacwcuRInD59GrNnzzbJ11Pc8/PzMznjyLjnmHHv9nG1p4aMJ99//73x72eeeQZHjhyplO/t7Y0tW7YAABYvXlzpIYsAEBQUhI0bNwIAZs2ahTNnzlTKDwkJMT6UcvLkycjKyqqU36dPH6xevRoAMHbsWFy7dq1S/qBBg/Dcc88BAIYPH16p7cfHx6NVq1aNtu0B5v/PvZOfnx9cXV2Rm5uLn3/+2SQ/ICAATZs2RU5ODrKzs03yO3bsCBcXF1y9etWk7gH9x73brN32gLr9n2uN7yj1mQvrAODOVphVkbYLwFil1DoAe6paWSk1SymVppRKu/s/ET25/cAforuxbdSK1eLJrVu3GrakDo7tmqqjkbZR53hyZyy5+4ugnrCfUnUaa9uw+EYBFWdC/i4iwRXvxwEYJiKxFe+nAHhQRBbUthB6/jFeRkYGPDw80KZNG7N30iDHJCK4du0aCgoKjA/s0puG/GEv44l+MeZRVaqLe3qMJ3qOJeynVBW9fz+pLpbU55bOlwD43PHeuyLNoXh7eyMrK8vs1CU5NldXV+NlK1QjxhOdYMyj6mgk7jl8PGE/pepopJ9aXX0GNakAOimlAmAIFk8A+E1tNqCUGglg5O3rBvXIxcVFlyNdIo2pdzwh22DMIx1w+HjCfkqOyNJbOm8FcARAZ6VUllJqpoiUAlgA4FMAJwFsF5H/1GbnIrJHRGZ5enrWttxEpFMNFU+UUiOVUhvv/kEmETVeDRFPGEuI9MnqD9+sCz1ft0rUWGnhYXl1wXhCpD16jCeMJUTaU++HbzYUng0hIiIiIqL60sRMjVIqG8AFAJ4Aqhvh1DXfXLolaW0BXK1mf9ZU02ez5vqWLFvdMrXNuzvN3DKs69rlWZpenzbtJyLtLFxWM+6IJzWxdz8AbNvuq1Lf/mCN7empTwGN77jVdVu1WU938aQWsQRgu7xNb+3Slt87q0rncavdelXHktv3MtfCC8DGhsg3l25JGoA0rXx2a65vybLVLVPbPDP1am4Z1nUt8ixNt2eb1vrL3v1AK8ejvv3BGtvTU59qjMetrtuydtvR84vt0vptwhbt0pbfO3ncGr4Mdr38zIwqH65Xz3xz6Zam2Up9912b9S1Ztrplapt3d5o969ka+9dCXVuabu+61jJ79wOtsHa56rI9PfUprbBmueq6La3WjT2wXRrorV3a8nunJduzF70dN7M0cfmZViml0kRnP2zUK9a1bbCetYXHQ5943EiL2C71icfNerQ2U6M1G+1dAAfCurYN1rO28HjoE48baRHbpT7xuFkJZ2qIiIiIiEjXOFNDRERERES6xkENERERERHpGgc1RERERESkaxzUWEgp1VEp9Y5Saqe9y9LYKaXGKKXeVkptU0oNsXd5GjOlVBel1Hql1E6l1Fx7l8fRMc7oE2MWaRHjif4wltSPQw9qlFKblFJXlFLf3pU+TCl1Wil1Tim1HABE5LyIzLRPSfWvlnX9kYjEAZgD4HF7lFfPalnXJ0VkDoAJAPrZo7yNHeOMPjFmkRYxnugPY4ntOPSgBkAigGF3JiilnAG8CWA4gK4AJiqlutq+aI1OImpf1ysq8ql2ElGLulZKjQKwF8Anti2mw0gE44weJYIxi7QnEYwnepMIxhKbcOhBjYgcApBzV3IEgHMVZzhuAfgAwGibF66RqU1dK4OXAewTkWO2Lqve1bZdi8jHIjIcwCTbltQxMM7oE2MWaRHjif4wltiOQw9qqtABwMU73mcB6KCUaqOUWg8gVCn1jH2K1uiYrWsACwE8AmCcUmqOPQrWCFXVrgcopV5XSm0AZ2psiXFGnxizSIsYT/SHsaQBNLF3AfRCRK7BcI0jNTAReR3A6/YuhyMQkWQAyXYuBlVgnNEnxizSIsYT/WEsqR/O1Ji6BMDnjvfeFWlkfaxr22FdawuPhz7xuJEWsV3qD49ZA+CgxlQqgE5KqQClVFMATwD42M5laqxY17bDutYWHg994nEjLWK71B8eswbg0IMapdRWAEcAdFZKZSmlZopIKYAFAD4FcBLAdhH5jz3L2Riwrm2Hda0tPB76xONGWsR2qT88ZrajRMTeZSAiIiIiIqozh56pISIiIiIi/eOghoiIiIiIdI2DGiIiIiIi0jUOaoiIiIiISNc4qCEiIiIiIl3joIaIiIiIiHSNgxoiIiIiItI1DmqozpRSTexdBiJqHBhPiMgaGEscFwc1ZJZSyl8pdVIp9bZS6j9KqQNKKTelVLJSaq1SKg3Ak0qpQUqpr5VSJ5RSm5RSzZRSvZVSuyq2M1opVaSUaqqUclVKna9IX6SU+k4pdVwp9YFdPywRNSjGEyKyBsYSqg5Hs1SdTgAmikicUmo7gLEV6U1FJFwp5QrgLIBBInJGKbUZwFwAbwAIqVg2EsC3AHrD0N6+qkhfDiBARIqVUvfY5NMQkT0xnhCRNTCWkFmcqaHqZIjINxV/pwPwr/h7W8W/nSuWOVPx/j0AUSJSCuB7pVQXABEA1gCIgiGIHK5Y9jiA95VSkwGUNuSHICJNYDwhImtgLCGzOKih6hTf8XcZ/n9m74YF6x4CMBxACYCDAB6qeN0OHI8CeBNALwCpvAaWqNFjPCEia2AsIbM4qKH6OA3AXykVWPF+CoB/Vvx9GMBiAEdEJBtAGxjOnnyrlHIC4CMi/wCwDIAnAHdbFpyINIfxhIisgbHEQXEESnUmIjeVUtMB7Kg4m5EKYH1F9lcA2sNwVgQwTOneKyJSsewWpZQnAAXgdRHJtW3piUhLGE+IyBoYSxyXEhF7l4GIiIiIiKjOePkZERERERHpGgc1RERERESkaxzUEBERERGRrnFQQ0REREREusZBDRERERER6RoHNUREREREpGsc1BARERERka5xUENERERERLr2f1cIK3K2Ff8CAAAAAElFTkSuQmCC\n", "text/plain": ["<Figure size 1008x216 with 3 Axes>"]}, "metadata": {"needs_background": "light"}, "output_type": "display_data"}], "source": ["fig, ax = plt.subplots(1, 3, figsize=(14, 3))\n", "for i, ncol in enumerate([10, 50, 100]):\n", "    piv = df[df.ncols==ncol].pivot(index=\"nrows\", columns=\"name\", values=\"average\")\n", "    piv['numpy / scipy'] = piv['numpy'] / piv['scipy']\n", "    piv['numpy-lower / scipy'] = piv['numpy-lower'] / piv['scipy']\n", "    piv['onnx-py / scipy'] = piv['onnx-py'] / piv['scipy']\n", "    piv['onnx-rt / scipy'] = piv['onnx-rt'] / piv['scipy']\n", "    piv = piv[['numpy / scipy', 'numpy-lower / scipy', \n", "               'onnx-py / scipy', 'onnx-rt / scipy']]\n", "    piv.plot(ax=ax[i], logy=True, logx=True)\n", "    ax[i].plot([0, max(piv.index)], [1, 1], '--', color='black')\n", "    ax[i].plot([0, max(piv.index)], [10, 10], '--', color='black')\n", "    ax[i].set_title(\"ncol=%d\" % ncol)\n", "ax;"]}, {"cell_type": "markdown", "metadata": {}, "source": ["## Test with a new operator CDist\n", "\n", "The final question is: *should we introduce a new operator into [ONNX specifications](https://github.com/onnx/onnx/blob/master/docs/Operators.md)?* The function [pdist](https://docs.scipy.org/doc/scipy/reference/generated/scipy.spatial.distance.pdist.html) is not necessarily often used for a big number of observations as the square matrix it produces will even bigger. It seems reasonable. We showed that a python runtime based on *numpy* would not help, the implementation must be done in C++ or directly used the *scipy* version. The experiment was done with a [GaussianProcessRegressor](https://scikit-learn.org/stable/modules/generated/sklearn.gaussian_process.GaussianProcessRegressor.html). The following section tests with and without a new operator ``CDist`` reusing *scipy* implementation."]}, {"cell_type": "code", "execution_count": 34, "metadata": {}, "outputs": [{"data": {"text/html": ["<div id=\"M65c98b71616f47d3921b78cc1ad239f8-cont\"><div id=\"M65c98b71616f47d3921b78cc1ad239f8\" style=\"width:;height:;\"></div></div>\n", "<script>\n", "\n", "require(['http://www.xavierdupre.fr/js/vizjs/viz.js'], function() { var svgGraph = Viz(\"digraph{\\n  size=7;\\n  orientation=portrait;\\n  ranksep=0.25;\\n  nodesep=0.05;\\n\\n  X [shape=box color=red label=\\\"X\\ndouble((0, 4))\\\" fontsize=10];\\n\\n  GPmean [shape=box color=green label=\\\"GPmean\\ndouble((0, 1))\\\" fontsize=10];\\n\\n  Sc_Scancst [shape=box label=\\\"Sc_Scancst\\nfloat64((112, 4))\\n[[5.8 4.  1.2 0.2]\\n [6.  2.2 4.  1. ]\\n [5.1 2.5 3....\\\" fontsize=10];\\n  kgpd_Divcst [shape=box label=\\\"kgpd_Divcst\\nfloat64((1,))\\n[1.83052462]\\\" fontsize=10];\\n  kgpd_Mulcst [shape=box label=\\\"kgpd_Mulcst\\nfloat64((1,))\\n[3.14159265]\\\" fontsize=10];\\n  kgpd_Divcst1 [shape=box label=\\\"kgpd_Divcst1\\nfloat64((1,))\\n[1014.92199795]\\\" fontsize=10];\\n  kgpd_Powcst [shape=box label=\\\"kgpd_Powcst\\nfloat64((1,))\\n[2.]\\\" fontsize=10];\\n  kgpd_Mulcst1 [shape=box label=\\\"kgpd_Mulcst1\\nfloat64((1,))\\n[-2.]\\\" fontsize=10];\\n  gpr_MatMulcst [shape=box label=\\\"gpr_MatMulcst\\nfloat64((112,))\\n[-0.0439394   0.00606057  0.00606066 -0.0439394  -...\\\" fontsize=10];\\n  gpr_Addcst [shape=box label=\\\"gpr_Addcst\\nfloat64((1, 1))\\n[[0.]]\\\" fontsize=10];\\n  Re_Reshapecst [shape=box label=\\\"Re_Reshapecst\\nint64((2,))\\n[-1  1]\\\" fontsize=10];\\n\\n  UU000UU [shape=box label=\\\"UU000UU\\\" fontsize=10];\\n  UU001UU [shape=box label=\\\"UU001UU\\\" fontsize=10];\\n  subgraph cluster_Scan2173239081280_2172776124464 {\\n    label=\\\"Scan\\n(Sc_Scan)\\nbody=node {\\n  input: 'next_in'...\\nnum_scan_inputs=1\\\";\\n    fontsize=10;\\n    color=black;\\n    B_next_in [shape=box color=red label=\\\"next_in\\ndouble((0, 0))\\\" fontsize=10];\\n    B_next [shape=box color=red label=\\\"next\\ndouble((0,))\\\" fontsize=10];\\n  \\n    B_next_out [shape=box color=green label=\\\"next_out\\ndouble((0, 0))\\\" fontsize=10];\\n    B_scan_out [shape=box color=green label=\\\"scan_out\\ndouble((0,))\\\" fontsize=10];\\n  \\n  \\n    B_cdistd_1_Identity [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Identity\\n(cdistd_1_Identity)\\\" fontsize=10];\\n    B_next_in -> B_cdistd_1_Identity;\\n    B_cdistd_1_Identity -> B_next_out;\\n  \\n    B_cdistdf_1_C0 [shape=box label=\\\"cdistdf_1_C0\\\" fontsize=10];\\n    B_cdistdf_1_Sub [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Sub\\n(cdistdf_1_Sub)\\\" fontsize=10];\\n    B_next_in -> B_cdistdf_1_Sub;\\n    B_next -> B_cdistdf_1_Sub;\\n    B_cdistdf_1_Sub -> B_cdistdf_1_C0;\\n  \\n    B_cdistdf_1_reduced0 [shape=box label=\\\"cdistdf_1_reduced0\\\" fontsize=10];\\n    B_cdistdf_1_ReduceSumSquare [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"ReduceSumSquare\\n(cdistdf_1_ReduceSumSquare)\\naxes=[1]\\nkeepdims=0\\\" fontsize=10];\\n    B_cdistdf_1_C0 -> B_cdistdf_1_ReduceSumSquare;\\n    B_cdistdf_1_ReduceSumSquare -> B_cdistdf_1_reduced0;\\n  \\n    B_cdistdf_1_Identity [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Identity\\n(cdistdf_1_Identity)\\\" fontsize=10];\\n    B_cdistdf_1_reduced0 -> B_cdistdf_1_Identity;\\n    B_cdistdf_1_Identity -> B_scan_out;\\n  }\\n  X -> B_next_in;\\n  Sc_Scancst -> B_next;\\n  B_next_out -> UU000UU;\\n  B_scan_out -> UU001UU;\\n  Sc_Scan -> B_cdistd_1_Identity [lhead=cluster_Scan2173239081280_2172776124464];\\n  X -> Sc_Scan;\\n  Sc_Scancst -> Sc_Scan;\\n  Sc_Scan -> UU000UU;\\n  Sc_Scan -> UU001UU;\\n\\n  kgpd_transposed0 [shape=box label=\\\"kgpd_transposed0\\\" fontsize=10];\\n  kgpd_Transpose [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Transpose\\n(kgpd_Transpose)\\nperm=[1 0]\\\" fontsize=10];\\n  UU001UU -> kgpd_Transpose;\\n  kgpd_Transpose -> kgpd_transposed0;\\n\\n  kgpd_Y0 [shape=box label=\\\"kgpd_Y0\\\" fontsize=10];\\n  kgpd_Sqrt [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Sqrt\\n(kgpd_Sqrt)\\\" fontsize=10];\\n  kgpd_transposed0 -> kgpd_Sqrt;\\n  kgpd_Sqrt -> kgpd_Y0;\\n\\n  kgpd_C03 [shape=box label=\\\"kgpd_C03\\\" fontsize=10];\\n  kgpd_Div [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Div\\n(kgpd_Div)\\\" fontsize=10];\\n  kgpd_Y0 -> kgpd_Div;\\n  kgpd_Divcst -> kgpd_Div;\\n  kgpd_Div -> kgpd_C03;\\n\\n  kgpd_C02 [shape=box label=\\\"kgpd_C02\\\" fontsize=10];\\n  kgpd_Mul [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Mul\\n(kgpd_Mul)\\\" fontsize=10];\\n  kgpd_C03 -> kgpd_Mul;\\n  kgpd_Mulcst -> kgpd_Mul;\\n  kgpd_Mul -> kgpd_C02;\\n\\n  kgpd_output02 [shape=box label=\\\"kgpd_output02\\\" fontsize=10];\\n  kgpd_Sin [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Sin\\n(kgpd_Sin)\\\" fontsize=10];\\n  kgpd_C02 -> kgpd_Sin;\\n  kgpd_Sin -> kgpd_output02;\\n\\n  kgpd_C01 [shape=box label=\\\"kgpd_C01\\\" fontsize=10];\\n  kgpd_Div1 [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Div\\n(kgpd_Div1)\\\" fontsize=10];\\n  kgpd_output02 -> kgpd_Div1;\\n  kgpd_Divcst1 -> kgpd_Div1;\\n  kgpd_Div1 -> kgpd_C01;\\n\\n  kgpd_Z0 [shape=box label=\\\"kgpd_Z0\\\" fontsize=10];\\n  kgpd_Pow [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Pow\\n(kgpd_Pow)\\\" fontsize=10];\\n  kgpd_C01 -> kgpd_Pow;\\n  kgpd_Powcst -> kgpd_Pow;\\n  kgpd_Pow -> kgpd_Z0;\\n\\n  kgpd_C0 [shape=box label=\\\"kgpd_C0\\\" fontsize=10];\\n  kgpd_Mul1 [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Mul\\n(kgpd_Mul1)\\\" fontsize=10];\\n  kgpd_Z0 -> kgpd_Mul1;\\n  kgpd_Mulcst1 -> kgpd_Mul1;\\n  kgpd_Mul1 -> kgpd_C0;\\n\\n  kgpd_output01 [shape=box label=\\\"kgpd_output01\\\" fontsize=10];\\n  kgpd_Exp [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Exp\\n(kgpd_Exp)\\\" fontsize=10];\\n  kgpd_C0 -> kgpd_Exp;\\n  kgpd_Exp -> kgpd_output01;\\n\\n  gpr_Y0 [shape=box label=\\\"gpr_Y0\\\" fontsize=10];\\n  gpr_MatMul [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"MatMul\\n(gpr_MatMul)\\\" fontsize=10];\\n  kgpd_output01 -> gpr_MatMul;\\n  gpr_MatMulcst -> gpr_MatMul;\\n  gpr_MatMul -> gpr_Y0;\\n\\n  gpr_C0 [shape=box label=\\\"gpr_C0\\\" fontsize=10];\\n  gpr_Add [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Add\\n(gpr_Add)\\\" fontsize=10];\\n  gpr_Y0 -> gpr_Add;\\n  gpr_Addcst -> gpr_Add;\\n  gpr_Add -> gpr_C0;\\n\\n  Re_Reshape [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Reshape\\n(Re_Reshape)\\nallowzero=0\\\" fontsize=10];\\n  gpr_C0 -> Re_Reshape;\\n  Re_Reshapecst -> Re_Reshape;\\n  Re_Reshape -> GPmean;\\n}\");\n", "document.getElementById('M65c98b71616f47d3921b78cc1ad239f8').innerHTML = svgGraph; });\n", "\n", "</script>"], "text/plain": ["<jyquickhelper.jspy.render_nb_js_dot.RenderJsDot at 0x1f9ff2733d0>"]}, "execution_count": 35, "metadata": {}, "output_type": "execute_result"}], "source": ["import numpy\n", "from sklearn.datasets import load_iris\n", "from sklearn.model_selection import train_test_split\n", "from sklearn.gaussian_process import GaussianProcessRegressor\n", "from sklearn.gaussian_process.kernels import ExpSineSquared\n", "from mlprodict.onnx_conv import to_onnx\n", "from mlprodict.onnxrt import OnnxInference\n", "\n", "\n", "iris = load_iris()\n", "X, y = iris.data, iris.target\n", "X_train, X_test, y_train, __ = train_test_split(X, y, random_state=12)\n", "clr = GaussianProcessRegressor(ExpSineSquared(), alpha=20.)\n", "clr.fit(X_train, y_train)\n", "\n", "model_def = to_onnx(clr, X_train)\n", "\n", "%onnxview model_def -r 1"]}, {"cell_type": "code", "execution_count": 35, "metadata": {}, "outputs": [{"data": {"text/html": ["<div id=\"Mc6cabaf623814a55ae3716d7e305cc01-cont\"><div id=\"Mc6cabaf623814a55ae3716d7e305cc01\" style=\"width:;height:;\"></div></div>\n", "<script>\n", "\n", "require(['http://www.xavierdupre.fr/js/vizjs/viz.js'], function() { var svgGraph = Viz(\"digraph{\\n  size=7;\\n  orientation=portrait;\\n  ranksep=0.25;\\n  nodesep=0.05;\\n\\n  X [shape=box color=red label=\\\"X\\ndouble((0, 4))\\\" fontsize=10];\\n\\n  GPmean [shape=box color=green label=\\\"GPmean\\ndouble((0, 1))\\\" fontsize=10];\\n\\n  kgpd_CDistcst [shape=box label=\\\"kgpd_CDistcst\\nfloat64((112, 4))\\n[[5.8 4.  1.2 0.2]\\n [6.  2.2 4.  1. ]\\n [5.1 2.5 3....\\\" fontsize=10];\\n  kgpd_Divcst [shape=box label=\\\"kgpd_Divcst\\nfloat64((1,))\\n[1.83052462]\\\" fontsize=10];\\n  kgpd_Mulcst [shape=box label=\\\"kgpd_Mulcst\\nfloat64((1,))\\n[3.14159265]\\\" fontsize=10];\\n  kgpd_Divcst1 [shape=box label=\\\"kgpd_Divcst1\\nfloat64((1,))\\n[1014.92199795]\\\" fontsize=10];\\n  kgpd_Powcst [shape=box label=\\\"kgpd_Powcst\\nfloat64((1,))\\n[2.]\\\" fontsize=10];\\n  kgpd_Mulcst1 [shape=box label=\\\"kgpd_Mulcst1\\nfloat64((1,))\\n[-2.]\\\" fontsize=10];\\n  gpr_MatMulcst [shape=box label=\\\"gpr_MatMulcst\\nfloat64((112,))\\n[-0.0439394   0.00606057  0.00606066 -0.0439394  -...\\\" fontsize=10];\\n  gpr_Addcst [shape=box label=\\\"gpr_Addcst\\nfloat64((1, 1))\\n[[0.]]\\\" fontsize=10];\\n  Re_Reshapecst [shape=box label=\\\"Re_Reshapecst\\nint64((2,))\\n[-1  1]\\\" fontsize=10];\\n\\n  kgpd_dist [shape=box label=\\\"kgpd_dist\\\" fontsize=10];\\n  kgpd_CDist [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"CDist\\n(kgpd_CDist)\\nmetric=b'euclidean'\\\" fontsize=10];\\n  X -> kgpd_CDist;\\n  kgpd_CDistcst -> kgpd_CDist;\\n  kgpd_CDist -> kgpd_dist;\\n\\n  kgpd_C03 [shape=box label=\\\"kgpd_C03\\\" fontsize=10];\\n  kgpd_Div [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Div\\n(kgpd_Div)\\\" fontsize=10];\\n  kgpd_dist -> kgpd_Div;\\n  kgpd_Divcst -> kgpd_Div;\\n  kgpd_Div -> kgpd_C03;\\n\\n  kgpd_C02 [shape=box label=\\\"kgpd_C02\\\" fontsize=10];\\n  kgpd_Mul [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Mul\\n(kgpd_Mul)\\\" fontsize=10];\\n  kgpd_C03 -> kgpd_Mul;\\n  kgpd_Mulcst -> kgpd_Mul;\\n  kgpd_Mul -> kgpd_C02;\\n\\n  kgpd_output02 [shape=box label=\\\"kgpd_output02\\\" fontsize=10];\\n  kgpd_Sin [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Sin\\n(kgpd_Sin)\\\" fontsize=10];\\n  kgpd_C02 -> kgpd_Sin;\\n  kgpd_Sin -> kgpd_output02;\\n\\n  kgpd_C01 [shape=box label=\\\"kgpd_C01\\\" fontsize=10];\\n  kgpd_Div1 [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Div\\n(kgpd_Div1)\\\" fontsize=10];\\n  kgpd_output02 -> kgpd_Div1;\\n  kgpd_Divcst1 -> kgpd_Div1;\\n  kgpd_Div1 -> kgpd_C01;\\n\\n  kgpd_Z0 [shape=box label=\\\"kgpd_Z0\\\" fontsize=10];\\n  kgpd_Pow [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Pow\\n(kgpd_Pow)\\\" fontsize=10];\\n  kgpd_C01 -> kgpd_Pow;\\n  kgpd_Powcst -> kgpd_Pow;\\n  kgpd_Pow -> kgpd_Z0;\\n\\n  kgpd_C0 [shape=box label=\\\"kgpd_C0\\\" fontsize=10];\\n  kgpd_Mul1 [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Mul\\n(kgpd_Mul1)\\\" fontsize=10];\\n  kgpd_Z0 -> kgpd_Mul1;\\n  kgpd_Mulcst1 -> kgpd_Mul1;\\n  kgpd_Mul1 -> kgpd_C0;\\n\\n  kgpd_output01 [shape=box label=\\\"kgpd_output01\\\" fontsize=10];\\n  kgpd_Exp [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Exp\\n(kgpd_Exp)\\\" fontsize=10];\\n  kgpd_C0 -> kgpd_Exp;\\n  kgpd_Exp -> kgpd_output01;\\n\\n  gpr_Y0 [shape=box label=\\\"gpr_Y0\\\" fontsize=10];\\n  gpr_MatMul [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"MatMul\\n(gpr_MatMul)\\\" fontsize=10];\\n  kgpd_output01 -> gpr_MatMul;\\n  gpr_MatMulcst -> gpr_MatMul;\\n  gpr_MatMul -> gpr_Y0;\\n\\n  gpr_C0 [shape=box label=\\\"gpr_C0\\\" fontsize=10];\\n  gpr_Add [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Add\\n(gpr_Add)\\\" fontsize=10];\\n  gpr_Y0 -> gpr_Add;\\n  gpr_Addcst -> gpr_Add;\\n  gpr_Add -> gpr_C0;\\n\\n  Re_Reshape [shape=box style=\\\"filled,rounded\\\" color=orange label=\\\"Reshape\\n(Re_Reshape)\\nallowzero=0\\\" fontsize=10];\\n  gpr_C0 -> Re_Reshape;\\n  Re_Reshapecst -> Re_Reshape;\\n  Re_Reshape -> GPmean;\\n}\");\n", "document.getElementById('Mc6cabaf623814a55ae3716d7e305cc01').innerHTML = svgGraph; });\n", "\n", "</script>"], "text/plain": ["<jyquickhelper.jspy.render_nb_js_dot.RenderJsDot at 0x1f9ff86f880>"]}, "execution_count": 36, "metadata": {}, "output_type": "execute_result"}], "source": ["model_def_cdist = to_onnx(clr, X_train,\n", "                          options={GaussianProcessRegressor: {'optim': 'cdist'}})\n", "%onnxview model_def_cdist"]}, {"cell_type": "code", "execution_count": 36, "metadata": {}, "outputs": [], "source": ["oinf = OnnxInference(model_def)\n", "oinf_cdist = OnnxInference(model_def_cdist)"]}, {"cell_type": "code", "execution_count": 37, "metadata": {}, "outputs": [{"name": "stdout", "output_type": "stream", "text": ["5.02 ms \u00b1 237 \u00b5s per loop (mean \u00b1 std. dev. of 7 runs, 100 loops each)\n"]}], "source": ["%timeit oinf.run({'X': X_test})"]}, {"cell_type": "code", "execution_count": 38, "metadata": {}, "outputs": [{"name": "stdout", "output_type": "stream", "text": ["396 \u00b5s \u00b1 12.9 \u00b5s per loop (mean \u00b1 std. dev. of 7 runs, 1,000 loops each)\n"]}], "source": ["%timeit oinf_cdist.run({'X': X_test})"]}, {"cell_type": "code", "execution_count": 39, "metadata": {}, "outputs": [], "source": ["oinfrt = OnnxInference(model_def, runtime=\"onnxruntime1\")\n", "oinfrt_cdist = OnnxInference(model_def_cdist)"]}, {"cell_type": "code", "execution_count": 40, "metadata": {}, "outputs": [{"name": "stdout", "output_type": "stream", "text": ["396 \u00b5s \u00b1 13.3 \u00b5s per loop (mean \u00b1 std. dev. of 7 runs, 1,000 loops each)\n"]}], "source": ["%timeit oinfrt_cdist.run({'X': X_test})"]}, {"cell_type": "markdown", "metadata": {}, "source": ["It is 10 times faster for this dataset so it is worth it. For bigger datasets, we should expect a lower gain but still significant."]}, {"cell_type": "code", "execution_count": 41, "metadata": {}, "outputs": [], "source": []}], "metadata": {"kernelspec": {"display_name": "Python 3 (ipykernel)", "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": "3.10.5"}}, "nbformat": 4, "nbformat_minor": 2}