{"cells": [{"cell_type": "markdown", "metadata": {"slideshow": {"slide_type": "slide"}}, "source": ["# Datascience et programmation\n", "\n", "Des ann\u00e9es 1970 \u00e0 la datascience d'aujourd'hui."]}, {"cell_type": "markdown", "metadata": {}, "source": ["* Une petite histoire de l'informatique\n", "* Le langage Python en quelques mots\n", "* Python en datascience\n", "* Comment partager son travail ? Package et Documentation\n", "* Les sujets fr\u00e9quemment oubli\u00e9s : tests unitaires, ing\u00e9nierie logicielle\n", "* Etre rapide avec un des langages les plus lents, mise en production\n", "* Ce cours"]}, {"cell_type": "code", "execution_count": 1, "metadata": {"slideshow": {"slide_type": "subslide"}}, "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": ["from pyquickhelper.helpgen import NbImage"]}, {"cell_type": "markdown", "metadata": {}, "source": ["## Une petite histoire de l'informatique"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### 1970\n", "\n", "* premier ordinateur IBM\n", "* cr\u00e9ation de Microsoft, Apple\n", "* MS-DOS\n", "* accessoirement, ma naissance"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### 1980\n", "\n", "* d\u00e9but de l'informatisation\n", "* les imprimantes \u00e0 aiguilles\n", "* le langage Pascal, l'\u00e9diteur Borland, Quatro Pro\n", "* applications des premiers r\u00e9seaux de neurones\n", "* accessoirement : mon premier ordinateur, un Amstrad 464"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### 1990\n", "\n", "* Les SVM\n", "* les imprimantes laser\n", "* popularisation des PCs\n", "* d\u00e9buts d'internet et des t\u00e9l\u00e9phones portables\n", "* premi\u00e8res versions du langages Python\n", "* Linux\n", "* cr\u00e9ation d'Amazon"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### 2000\n", "\n", "* premiers smartphone\n", "* les random forest, d\u00e9but de scikit-learn\n", "* python 3.0, d\u00e9but de IPython, l'anc\u00eatre de jupyter, pandas\n", "* d\u00e9but de Hadoop, l'anc\u00eatre du calcul distribu\u00e9 pour le traitement des logs de site web\n", "* cr\u00e9ation de Facebook\n", "* L'ENSAE passe du C au Python\n", "* Cr\u00e9ation de Git et GitHub"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### 2000 aussi\n", "\n", "* Accessoirement, ma th\u00e8se avec des r\u00e9seaux de neurones, gros pour l'\u00e9poque, petits pour aujourd'hui.\n", "* Premier pas chez Yahoo, en jouant avec des teras de logs et Hadoop."]}, {"cell_type": "markdown", "metadata": {}, "source": ["### 2010\n", "\n", "* le deep learning gagne ses premi\u00e8res comp\u00e9titions,\n", "* les librairies de machine learning deviennent open source,\n", "* elles sont accessibles en python, langage beaucoup plus facile d'acc\u00e8s que C++\n", "* le GPU prend son envol\n", "* les librairies impl\u00e9mentent les calculs de fa\u00e7ons distribu\u00e9s\n", "* calcul en temps r\u00e9el, loi RGPD"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### 2010 l'histoire s'acc\u00e9l\u00e8re\n", "\n", "* Python appara\u00eet comme un pr\u00e9requis sur de nombreuses offres d'embauche\n", "* Python est le langage le plus utilis\u00e9 dans l'enseignement\n", "* [Yolo](https://www.youtube.com/watch?v=MPU2HistivI)\n", "* [Kaggle](https://www.kaggle.com/)"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### 2020\n", "\n", "* Python efface R\n", "* l'exploitation des donn\u00e9es est une \u00e9vidence\n", "* le num\u00e9rique est partout, peut-\u00eatre un peu trop,\n", "* les r\u00e9seaux sociaux sont incontournables\n", "* il existe m\u00eame des [tableurs \u00e9crit en python](https://pyspread.gitlab.io/index.html)"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### 2020 aussi\n", "\n", "* les fakes news, les fakes tout court (mod\u00e8le GAN)"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### 2030 ?\n", "\n", "* t\u00e9l\u00e9phones traduisant en temps r\u00e9el d'une langue \u00e0 l'autre,\n", "* voitures automatis\u00e9es sur l'autoroute,\n", "* des drones survoles les champs pour \u00e9radiquer les insectes\n", "* ou autres champignons,\n", "* commande vocale de nombreux appareils m\u00e9nagers"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### 2040\n", "\n", "Back to 1984 ? ou l'odyss\u00e9e de 2001 ?"]}, {"cell_type": "markdown", "metadata": {}, "source": ["## Le langage Python en quelques mots"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### 3 concepts en informatique\n", "\n", "* la s\u00e9quence d'instructions\n", "* les tests - si...\n", "* les boucles - r\u00e9p\u00e9tition"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Les fonctions\n", "\n", "Les fonctions \u00e9vitent la r\u00e9p\u00e9tition du code."]}, {"cell_type": "markdown", "metadata": {}, "source": ["### La programmation object : les classes\n", "\n", "* On peut tout \u00e9crire sans classes.\n", "* Cela rend les programmes plus lisibles.\n", "* L'informatique : une complexit\u00e9 qu'on rejette au d\u00e9but et qu'on trouve pratique par la suite."]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Une syntaxe simple"]}, {"cell_type": "markdown", "metadata": {}, "source": ["* une des grammaires les plus courtes : [grammaire](https://github.com/python/cpython/blob/3.10/Grammar/python.gram)\n", "* indentation obligatoire : rend les programmes plus lisibles\n", "* documentation incluse dans le programme"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### portable\n", "\n", "* disponible sous mac, linux, windows\n", "* open source\n", "* un site d'extensions\n", "* de nombreuses extensions, automatiser, site web, calculs"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### langage interpr\u00e9t\u00e9\n", "\n", "* langage interpr\u00e9t\u00e9 :\n", "    * a besoin de l'interpr\u00e9teur Python\n", "    * il ne peut pas marcher sans non plus\n", "* langage compil\u00e9 (C, C++) :\n", "    * traduit en langage machine\n", "    * fonctionne sur une machine"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Interface C\n", "\n", "* Python est tr\u00e8s lent, voire le plus lent\n", "* Il v\u00e9rifie sans cesse ce que fait l'utilisateur.\n", "* **Mais** il est facile de l'interfacer avec le langage C pour \u00e9crire des calculs rapidement."]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Python outil id\u00e9al pour automatiser\n", "\n", "* Con\u00e7u pour automatis\u00e9 des t\u00e2ches r\u00e9p\u00e9titives.\n", "* Mail, Web, traitement de fichiers...\n", "* Un des langages les plus utilis\u00e9s, donc beaucoup d'aide disponible sur Internet"]}, {"cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [{"data": {"image/png": "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\n", "text/plain": ["<IPython.core.display.Image object>"]}, "execution_count": 4, "metadata": {"image/png": {"width": 500}}, "output_type": "execute_result"}], "source": ["NbImage(\"pp.png\", width=500)"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Conclusion\n", "\n", "* Python : langage simple\n", "* utilis\u00e9 par beaucoup\n", "* lent mais rapide pour qui sait le marier avec C\n", "* Pas le plus efficace dans tous les domaines\n", "* Mais pr\u00e9sent dans tous les domaines"]}, {"cell_type": "markdown", "metadata": {}, "source": ["## Python en datascience"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Avant Python\n", "\n", "* tableur : difficile de passer \u00e0 l'\u00e9chelle\n", "* [SAS](https://www.sas.com/fr_fr/home.html), [Matlab](https://www.mathworks.com/products/matlab.html) : on\u00e9reux\n", "* logiciel d\u00e9velopp\u00e9 en C/C++ par des \u00e9tudiants\n", "* R : naissance quasiment en m\u00eame temps que Python mais inspir\u00e9 de [S](https://en.wikipedia.org/wiki/S_(programming_language)) n\u00e9 en 1976\n", "* il fallait conna\u00eetre plusieurs langages pour tout faire"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### numpy, pandas, matplotlib\n", "\n", "* 1995 - [numpy](https://en.wikipedia.org/wiki/NumPy) : calcul matriciel\n", "* 2003 - [matplotlib](https://en.wikipedia.org/wiki/Matplotlib) : graphes\n", "* 2008 - [pandas](https://en.wikipedia.org/wiki/Pandas_(software)) : manipulation de donn\u00e9es fa\u00e7on SQL"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### jupyter, cython, scikit-learn, sphinx\n", "\n", "* 2001 - [jupyter](https://en.wikipedia.org/wiki/IPython) : interface m\u00ealant code, texte et ex\u00e9cution\n", "* 2007 - [scikit-learn](https://en.wikipedia.org/wiki/Scikit-learn) : machine learning\n", "* 2007 - [cython](https://en.wikipedia.org/wiki/Cython) : \u00e9criture simplifi\u00e9e d'extension C pour Python\n", "* 2008 - [sphinx](https://en.wikipedia.org/wiki/Sphinx_(documentation_generator)) : documentation"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Con\u00e7us par des informaticiens scientifiques\n", "\n", "* pandas, scikit-learn, jupyter, cython ont \u00e9t\u00e9 cr\u00e9\u00e9s pour combler un besoin\n", "* beaucoup d'outils \u00e9quivalents ont exist\u00e9\n", "* ceux-l\u00e0 ont gagn\u00e9 la **bataille de la simplicit\u00e9**"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Acc\u00e9l\u00e9ration avec Git / GitHub\n", "\n", "* Partager du code comme les id\u00e9es\n", "* GitHub : ET t\u00e9l\u00e9phone maison ([pandas](https://github.com/pandas-dev/pandas))\n", "* D\u00e9veloppement rapide (plus qu'un logiciel propri\u00e9taire)\n", "* **Possibilit\u00e9 de construire ses propres outils \u00e0 partir de ceux-l\u00e0**\n", "* Possibilit\u00e9 de partager ses propres extensions"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Contributeurs\n", "\n", "* Un bon projet attire les talents du monde entier\n", "* M\u00eame s'il est commenc\u00e9 par une soci\u00e9t\u00e9 priv\u00e9e\n", "* Les *core* d\u00e9veloppeurs d'un projet open source sont souvent une dream team."]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Cas du deep learning\n", "\n", "* le langage python est le langage du deep learning, [tensorflow](https://en.wikipedia.org/wiki/TensorFlow) (2015), [pytorch](https://en.wikipedia.org/wiki/PyTorch) (2016)\n", "* ces outils grossissent plus vite que n'importe quelle interface graphique qui pourrait en faciliter l'usage\n", "* \u00e0 la fois logiciel, \u00e0 la fois support de recherche"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Economie\n", "\n", "* Opensource : il est possible de recruter un d\u00e9veloppeur connaissant d\u00e9j\u00e0 pytorch ou tensorflow\n", "* Les logiciels sont offerts.\n", "* Pas les ressources de calculs."]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Apart\u00e9 : Kaggle\n", "\n", "* [Kaggle](https://www.kaggle.com/)\n", "* [XGBoost](https://xgboost.readthedocs.io/en/stable/)"]}, {"cell_type": "markdown", "metadata": {}, "source": ["## Comment partager son travail ? Package et Documentation"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Cas d'un indice \u00e9conomique\n", "\n", "L'INSEE calcule l'[indice des prix \u00e0 la consommation](https://www.insee.fr/fr/metadonnees/source/indicateur/p1653/description) tous les mois. \n", "\n", "* C'est un processus qui est ex\u00e9cut\u00e9 tous les mois m\u00eame pendant les vacances.\n", "* Il existe depuis 1914."]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Processus et manuel\n", "\n", "Reproduire un m\u00eame chiffre p\u00e9riodiquement, c'est suivre la m\u00eame liste d'instruction avec des donn\u00e9es actualis\u00e9es\n", "\n", "* Collecte des donn\u00e9es\n", "* Nettoyage\n", "* Fusions des bases\n", "* Calculs des indices"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Transmettre un code ou un manuel\n", "\n", "| pour                                                    | contre                                                       |\n", "|---------------------------------------------------------|--------------------------------------------------------------|\n", "| le code est pr\u00eat \u00e0 l'emploi                             | le code est compliqu\u00e9                                        |\n", "| le code est pr\u00e9cis                                      | le code ne marche pas toujours sur une autre machine         |\n", "| il n'est pas besoin de tout conna\u00eetre pour l'ex\u00e9cuter   | tout le monde ne sait pas coder mais tout le monde sait lire |\n", "| c'est rapide                                            |                                                              |"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Comment transmettre un code ?\n", "\n", "* envoyer son code par mail\n", "* envoyer les instructions pour le r\u00e9cup\u00e9rer\n", "* utiliser le m\u00eame moyen que tout autre librairie Python ? --> les packages\n", "\n", "**La simplicit\u00e9 c'est souvent utiliser ce qui marche d\u00e9j\u00e0.**"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Package et documentation\n", "\n", "* Python propose des outils pour simplifier la cr\u00e9ation de package\n", "* [cookiecutter](https://github.com/cookiecutter/cookiecutter) : mod\u00e8le de packages"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Quelques exemples pour le cours\n", "\n", "* [td1a_unit_test_ci](https://github.com/sdpython/td1a_unit_test_ci) : un mod\u00e8le pour un package python simple\n", "* [td2a_plotting](https://github.com/sdpython/td2a_plotting) : un mod\u00e8le pour une librairie graphique bas\u00e9e sur matplotlib\n", "* [td3a_cpp](https://github.com/sdpython/td3a_cpp) : un mod\u00e8le pour \u00e9crire une extension avec cython\n", "* [td3a_cpp_deep](https://github.com/sdpython/td3a_cpp_deep) : un mod\u00e8le pour \u00e9tendre pytorch"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Les entreprises en pleine mutation\n", "\n", "* le partage de code n\u00e9cessite un langage commun, les m\u00eames outils, le m\u00eame langage --> uniformisation\n", "* un temps d'apprentissage non n\u00e9gligeable\n", "* un code grossit sans cesse : il faut communiquer sur les nouvelles fonctionnalit\u00e9s"]}, {"cell_type": "markdown", "metadata": {}, "source": ["## Les sujets fr\u00e9quemment oubli\u00e9s (\u00e0 l'\u00e9cole) : tests unitaires, ing\u00e9nierie logicielle"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Les tests unitaires\n", "\n", "* Au d\u00e9but, on les oublie car c'est ennuyeux \u00e0 \u00e9crire.\n", "* Ensuite, on les \u00e9crit tout de suite car on est paresseux.\n", "* Un logiciel sans test unitaires est comme un avion en plein vol sans avoir \u00e9t\u00e9 test\u00e9"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### La documentation\n", "\n", "* On l'oublie toujours quand on l'\u00e9crit apr\u00e8s.\n", "* Ce qui est logique pour les uns ne l'est pas toujours pour les autres.\n", "* Ecrire des exemples est fastidieux mais utile.\n", "* [gallery](https://matplotlib.org/stable/gallery/index)"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### L'int\u00e9gration continue\n", "\n", "* Indispensable aujourd'hui"]}, {"cell_type": "code", "execution_count": 4, "metadata": {"slideshow": {"slide_type": "fragment"}}, "outputs": [{"data": {"image/png": "iVBORw0KGgoAAAANSUhEUgAAA74AAAEiCAIAAABRJ5lHAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAAFiUAABYlAUlSJPAAALYYSURBVHhe7J0HXBTH28ffv9LtvdfYe+8NkQ6CFGvUqDGWRNOM3dhiTDSa2CvNbuyASMceQYogChY6duDg7uhw+z57M7cue0VEVNDn+3kymX32mdnZ2Znd363L7v8xCIIgCIIgCIKUApTOCIIgCIIgCFIqUDojCIIgCIIgSKlA6YwgCIIgCIIgpQKlM4IgCIIgCIKUCpTOCIIgCIIgCFIqUDojCIIgCIIgSKlA6YwgCIIgCIIgpQKlM4IgCIIgCIKUCpTOCIIgCIIgCFIqUDojCIIgCIIgSKlA6YwgCIIgCIIgpeKdpLNIJEpLS6MLCPKhePHiRXFxMV1AEASpGMhkMppDEOTTpezS+fLly6tXr3ZEKga2R3qN/a/152MO3j12Xl5Px2JJrl69unz58ukfj/EuQ2yC29qGtLUJbiNPK0fe0bfXnhu/004sCUz2FStW0N377Nl8Zu7Pt5r8HNz05+AmiyBV5Nn0Pft337e7/eoUPSolqVzHaFpYP8EItCk5Giumf/n9iYGvztIeL0nQ1RsLf1lhMnaCkZUj2vs2h8if+kZM6BcxoYKnA+9M/vrhan/Rf3SUKAHTdunSpXRWfLrUlVOvXj2SAT5wfsCAAYcPH6adXh6UUTp7e3uDXHNAKgw2h3tZ32z9udmOIKF6Bt08bdo0mKuClPBh/HLp3AautfJU2Sq0X1k9BwUFff/999zeET7nPJXOxEDRcnm+vTf/L8HNbzx3pcdGQaU7RtNCWelcGW1cSLtLL47QflcQeOX6aEsHtA9mDpE/9g2f0C98Qt8weVrh874ZN+lY4XHlyhWYCmReKM+UT8lPNezHphzVcxml86pVq+yRisTYQz2trrf63MzeqzsdkQqWLVs29WNjd77n2FutK6k5+PSkXalg6dKldMcQOZtOz/npv8Yf0XbdG0ePjYLKd4xu9xEMvEpky+9PoP2u4Nuflhma240yHydP7ZRS9Jez3z7ihz63x/cNHV9Z0hkxK+lY4bF48WI6HT516lQMBgwYQLv+nSmLdM7Ly/v666+pZEMqBmPdelpdbWV1rbU8bfX55OmgVEBn6kfF9lIXwbW2chntSgV0r97EZotvzg1ZcLvnz8QgDx667tPiz1Pf/CiXsB8rXR3Wix4bBbRllYcvS0pna16ebxXT/1X4QNrvcmQy2WgLu1Fm40aa2UI6SikFv+PUbxynzv5q7g+QGluP5/zq4tGv2W8f/kOfEMe+IeNfp7d5+Yrn7397gowRPgdPJ8NnAJWuderUrl2b5j5Snnb9O1PGu852pebnn39eVRK6AilXrF16WF5p9RkaHZEKvvzyyylTpnzc1OZiJ6ubLa3+a8WmN+VppcrTrlSgYU9Jusl8dlzHxZmtlisb+GGthrKVMf3j39k/3Gj4w82G8rTRR8nTY6NAQ2srZjoluPfrsac8Giu8n/a7ghEmY1WakaX9vB+X/bPXdZfTUb4tX7fJfNwkQTBa6c0u9Pvetxz63HKoRCkdKzw0z5FPKa1VYaBd/86UUTqPKx2BgYG0AA8nJye6Gik/rJ17WAS1sgxqRVJifM+n6qejSgHM0o8Olc6V1mhXKqB7pQbv/guJSo7vsPjY8G/XWM0idm7wAvCQVecGf0ejPwmodP6oRo+NAtqySkRwL8Goq1xG+13BcGNrZRs7ftrfe1wEoplv075ZKCiCVkqzu72w938Off5zqEQpHSs86Fz4DKC6tQJAu/6dKaN0ti0F27dvp9EMExwcnJCQQPInTpygEZ860igXv9RS9dW7Y3Wwu0VAy4pvPgEtpcwji4c3BH4NJs1Y75PNCJyckUHFMVkN2feOBD1Ru1YzqZcnR2dD2SN0mQfzJOjIvWy6oGCsZ0fLGy0qr9GuVED3ShVENz9ru/TgqLnUVRLwc+qZusoDld3+wfj95Kzvrzf4uEaPjQLasvdAEMwdJnXy4Wi6XE5MutVTMOoql9F+VzDUyFJgVg5fatbNxGZ9+7Og4FtZwiXLUAls/aDA/7a2w8iSkdwZeilZ4PcwspQwyUN33xH4S2mSiIMeySo6J3S3ZQIjHWrkI/CX3uyCF/a6Yd/7hn3J1E7JU4H8dKzwoJPhM6BmhYF2/TvzvqTzN998I5VKSfDKlSuJEwQ0LFYu6fwu8ldyx9knmS3rYmvLSKNs/VOI/31geaC7uX9LC/+WFTy95N9Swjy0uH9dQ4wglbxYfymTUbeWjDEOOlOVkN49HJDyFqcqkMlMdvTky6mQTw6cHCWFsofJKj5MSsDhu1K6oMDas6OF/PpqcV2evs/8QbiKF96wfPGoDGXV5WlXKqB7pcSfpl8T3bzY5ivqUsUay5lEPUOGut4Zld3+wfj9xKyF1+uDfl0oV7EfJU+PjQLaslLwtlI4AOYOkzzZLYoulxr+DFIGpHMZRmbZ8j43WsDPdcuEG2Uoqy5P+13BkNEWAtu0fT8o42Vr/zS2dvx5xfrf/tq5euPfnE2YPgeMqOdxk2YIypbe4r0sQsSw9QMC/9va9tEWjDh8iFeiwH9htIWYSRyyK1zg12AJlyxAzQ8ZfRDy4rADFxJVdE7ILot4RjJk9CWBv/Rmd2tBr2t2va/by1O795ffcs2OKYzq9TylDGUFeTpWeNDJUH5k3zui7i5P+ZJ6eXIq2344nZSKGnJAuZI088ryK5lMjRrLBf7YwzViGfC7CfxcShD4oRYm8zLUKPBDSuB7aNe/M2WUzjZvIjo6mobygo8fPw6LkNJlNTApflIo5RRJlz8IKf42pLUEvxTa7JGm1hZGA0n+bZHccQLpDBlrG5uBI01tLIyI/31gsa+7mW/Lim9evqx0Nrt3XeDXYJLn670yGYGTM3rAFEyUM2nSJJIBSF4Sdcg/hVH2E5Tz4ydOHG5iM8neCvLJARMjJVD2kHw9CxfPpPgfipKQPED8Vu7tza81+zBmeq1Z34ONzb0aCfzvYrQrFZBdU4Y8j3Fg5By6rJ4jQ+dBZFyHxXT5nRF0+wdmw/GZC67WUzYmm7756GY2wzkfxNVjmEziJzxnYO1hzs8Pfv683oNCWFzFedQZqYqDtqwU2E2caDLccqKjDV1+E/4wd5jkia6RdLnUkBk00c6SLgu42UMw6t6fjbnWbMgfjcwDGgv872K03xUMGmU6eJQZpIPk6ZezviWyeM73i/l+ft52wnQS8+vvWwQx7qNMGSZ5kFcS8buPMmOYJNCjg0btJzGS8AO32cUD8RdNQToPGrVPuf63ym+DLYrDB11MFPjPjzIF6TxoZ5iGsoJ872Fmffr2GTTKBPLisH3nE6F5wpiQnWYgnQeN8hL4S5+3vfldzyvjel0Z977THlfGtd1v1tPTWENMKVM6VnjQuSCHfw3i81Z+uMxFSaDa15cqjlLWEwhnVyZ1YlCq5ni4Jiaz7YfTA0VzfPWSgHS+nMlUr76cLiuIPVQdpHP16q50uXRALUQ602WN0K5/Z8ounceOHctPV6xYcezYMZDFkAYEBNA4OQcPHoQY6MH4+HhYhBhBWX7qB9KZgYPC+MrFq4bI8k2TfG1ui5mBA7cNGDBg28CBjOS2c5RUQ3xpUpDO3kkfaC8s9nYz825p5tOSTYlVyLyXt0I6l/RryFPprOQnefkQew2ZrsoQ6UwX3hIinVWej1RquA8pnd+H0a5UQPeqJH+YzAI1DOqZLr+JZ22XQvwv1tPo8rvxcaXzb8dnCIQsWI5oVWQ+Y7VxudXGGtnPl58TZRL/DP96VhtrgZMYk+Z2LDHz3NW6TOFVp8RMX/AwsQueP4DIm1frPWEezDjgtuBqXVJWg9Fjo4C27D1QZumsmQk3uwtGXeUy2u8KBo40GTTSVJ6aQPrbXzuJLP5u0Uq+n5+3GDeJxICNsbTnx/QdacLEe2wLk5D4kJ0mWWKJmBEP3BlBYpjES1tviweNHBPvaQLSeeDIvcr1v1X+H6hTHDbQM1HgPz/SBKTzwB1hGspqyItD951PYJT9ITtMWek88qLAX/q87Y3vel4e1+vyuEqU0rHCg06G8gMuc+ouVaUE5jsIsIkByXRZDcrSWTNUtyr4fKUzaDUBdIUaXrx4wT2/AUqallGFNMoZJOztv7clM+KxY305JylLgIrGHoyEdVImBTIkhlt0hsZIIyOjUuSBzFj/FFib4v+6hcQjIMl3bIQE9ssJ8lZjxyZ4mcAQHDvWGTYNIh6cpP6UqEhSCWmDvCiLyvqJdIYMaRLxk3oi/aEe2iH89jApvsQJred2P/IgVE6DI6Vsq4ifj/mebqaXWvBNkv76cyH707PAs/9SC6bghmnqIxIgWOTivaUM8YAlpdJ/mowoAOd64iQFlf1gjJT9hwUgiQH/ceL0Bj/ziPjD07NAOptGXyerImJgVRZZRRpJjGsM1CN5ft1L9LpJAiNhHBPUwP4cZ0cE/ceQwCc0MjVoQrQU8ocFi7AMwROCUsGZHDABBsOECYdIDPOE/vErRDIp7M984uewuNDO7GpTMFN5CuZ9tSnsdbLoOikoYRizuOvgP3C1KVNIncB+URaJB4uIa8r1TEQhY3Z1LTilorXEAxwQZUH9pAaz5w8hn/y86Z1CRio6RgLg1GZ29RhpAzSAYR4S/x1RlhQOgbwBxLh2kjwJ46B7VZIzA78FKXy59/eHh84F2z/ym/m2U+g6BavMvyJrwUK6/wjxR4bOo+smTAiEnmRSoX7oxQlu0QIncDhaSpwEdd1OjhS3ijuUpa8/2m2ClJFyMaRCwdb5rD/61bdX6n53pS6Xnr1Slym4cuJJ5rdX6oAn4EydZ0zst1cO82MgffC4rpTJnBt55QbEMw/mJcZ+DZm0Q2fTMmEtk334y0goxdYgqF85JbvAQVumBIxnEsD1DPSAlEklO0tGO/OE7T0SoBzvB3OHSZ7gGslWJ+8r7lzE7yLl3ibdSGaQMuNvdBOMOhi9MM6ZbDodYCTDmAc/f+gCMNRJPKSSktOB+JlsOv4BEkxmH8w4bivcHIHTCzdHSsw40cNkBubXMW5bxEie3UpJBgwfwzdOEw8zshCs4tvS1RtJ2OQZ8wSrHruPBE08YPheyIvD9p5JYJj4CwM8E2HxwvAxDJPYzyMB8nGeY4JBU4d7kWZA1IDhXqSGbRAmiSD+2xJaFTEmSRj/DzjFYQM82Tr5dg62ziQkhIWReDHE74ggq8They8kvd5xbjHh4hhuc+LQPecSaMztHdBsOPOxBIclxTHiAcM9ySooQvzAgItJxAkV3mbP/uwu8DdEzPbatz0CbXoG2pL0r0AbpjAqOINe9IEeT1PIWnGGC/EAwYVMj0Bn4g9+ZMsFc/7gRzZ8Z89AF1Izqe0s1MakBD99fe3mtnI2EGqDrbMEZ0ghrOejKH4LISVr+dDJoApuJgJkopEJRTwqT3QE7lKlLp4PfysQc3jCYa4IAFdJOOty10oA4mFLsE24JsL1RX56YOFOzmQtcfKpVq2agYEBl4ouLwXpzDy4TEq5PmCIP8bNIIaBvAvkYw8ZkLUAxFZbcpmrIfZQNbqCYZZdyVxqYADSeellEVnLPHAVxPNTWuydKaN0tlaCrngT2dnZs2bNomWUcIJ6JHf2hYv3WVjCTHaKkhL/cGPLAXK2DRjAiG97JjDWFsbe1tYSJsl6fwSJ4RYPspVEwKQbMMATilibjU72s5YvXoBF+A9OmNZ+yaQUR5KPNUhna2soXWJRcucgyF/wQP1QMBxOCmwrBkhCQdyD3xdWqaufK0uaZO2TxNUTIYX4nRCf4DGAqwd0M3uOkJOk8LN9Io3cFioh/ogdA5yihI0HzHd3M73YgrOI+y0kTFafS1f7/FofLOnZGtP7N/ZdbMHkXTeNf0hi+IuStPVwGu3z6xE2+EF905RH4ExKgUoYUknY1fr701jnfihVcGNrShapWZLyK6ueL66HVYzk+DkRVML6E8PrQ5+YXjxO4s+9on4m8yjUaRp1HeIVjbwG/q2wKv/a/rQs4XYv1YfrGSudeXvHNzqwFMAsHT9+PJcSIA/SGdYO904ZNmyY5/DhcO2B8xH4OVlM4sni+PFucO5hJJETApIFMaCpEqGe4Z5QT4o31MOAhiNludTifDvTK03g+sqlXuy19kFYAdNn48o+G+tKEleynXPlmPHVpn32NwRPn411zm1k5ZQpiOArTfZDfOG1rYmZ4Ie1Yf/V2S96GPG4Cdtd/ldIfNLL1aaPr+2DyPxrpilsqcQU9oq+/WUWKcWkHfbOhq3Q2s6lwdblfvER9hA8uMZvIT9l+5GHYO9ISqSwwIK7/wgCGtYuspp2t/PPgrVgl/otIDUEjB8Pp9mdYdLhw4fvHTYsOjVy/PhD8nP93Z3hUujbsB3QtxIQXiReGn0Iun3YMA+IT/YeBq1yk0tnOCRwrt95h60HSknv7Lwrv0i8Vf1jHSYwiV6H77LjAeq86zY+kZE6WHuRPVVOWel8uQ7oVy49faUOk3/l9MtM4rl2uU42SOfHV0AH8yOz01cGpDHf+tb6F+Jzr4Bivnm5Dhz0uQmxz57V9clivvWsxY/npSXqgZQcGg6V7YTrHFzhh+8Ig/2N3DP8UDSrYv1gLjDJ411YKZzsD73E7Ixke8naeFhqkKPmeOgZ+I0xbHsExO8cNoyRRhySX9RV9jZ/BkGMIB1/vZtg1JHRu/U5O+a3wijNv0JG72iYIxsbkDEf4l9XwmTCrz7wg9KVT4erZFQnvfwV/EnPmsTDqWPjYTpHJEdhlsHsk8DMeniN24ryHCECncw4dusFV2G8mV45qjw7SMp2Oo/+w4w4s3b4kgji9Zt38P3KNnvBIhIJGcGq4H+MQBL33x4OeSbhYu8L8Vm3d11IZDcUst3oMSPpP8gd8nEeRrD1TaGS9l16gjHxZ0nMP1BKEr4xhPrFIVtAzvYftpetLdHr3wSG+OPO9oTTfv9hXn+DXxzW3yMBAvh2DvxMAqjz9l22sPUEbyHxsIrIYi6SW4QmQXz/YXv4zgtsPUkbg8VQyUZopzg0jq3HE1YlXDSCfPsuZ2HVv7CKSeh/MYmUhV1rfz4R/L379gQP32yvzu8RYNMzwIakfwXYMHkRUE+b3zxb/zbw0Z2BoPJ7PoyEte1PWoOnzW8DIWVEF85mMzS+IHJDooT4b90Y+FdG8haFk8SDc0tGMqm5R1ISlDoDeSbpVgFsZTfZCmyxZ4AfG1MYeSIN/PKtiD3hF1jPmAh+CyGVj5QScDOCwM0ObubChIKrVUBKEviZ1ADPZDgHDgMnk+wJihYiyTnNM4n6JeE7yfWL+uXxOyFeEkbi+VshtwzIfIeYZEmY44QAD2PQJ4nDLyZCbbbmw8L2j5PCedItWrGt6IPhcL2z5aQz+EE3h0nhmginhGFMkic5A5OtECAPmpUPSGfYfV23GF1dXRddXYaJMTgUC35OOkNeX98A1hKYWBeQ2gYGIJJBNxuAMtb9JQj8UNY1NoaVzqIgkM6wNvPK0qBMRk93CVQAi8qQnn93yiidrZSgKzRy69Yt6EdaQBV3DlglMBKLMR4WVlbwo/OOBDbkRNfJAXEJpwBTw38g721lJWESrfZHkFXc4kEIk0TsCRVbWZmCH8rDGdzQPYEswn8wJ32ThbuQ6GMFWtnKCkpb+bK7k7w3HGqwkESw8hecUD8MqTEXoR5ondXOMRYg4q38kjTUz5UlTbLySVSu56IpNDvB6kBEpHzfTQ0vgBPYY2pF6idlYZjK22U1xtzKYowhieFjtrOriUcLzsIiQVNmXhIzZHGYa3OTM033erBa2ST+IXFyi/sgk399WPgDE49mbPDxFibHm0CGyTq+PT4TCkJ+5JkWY1xZZ1Jyi5A8ZozrGsiDubs2h8uSyb3rEfdaxDNZhnuPEP/O4y2Y3KsmyQ8F/pCbzeDiZxJ5HfKM+HX9YzxaxEc3jciHBq8XbJd5dfSiiO6IstGxpcBRDZJIt1AJ42hzAPI2jo4JnuZRUgg+lOzvCOcaR0c3EsYtwjIcVkf/ZL7zEDilUeaeCfI6HA/YQEyoW6REXvQ15ufamlxuzLeLlxuDQhoZ+8DkciNY3H66ESgnk2cP+TGGlxuLE1dfkjCQ3wvx+VdBXptcPgqLI30bjzndIOwBVJJ5ScoGgA073cjEtwGJNEmGmhsnJjcG3TDm9GESEP+4YTj05+U1EY/Bn2l4hPpDQhuyh+DBVbKobLQrFdC9Ksm+4bOftV0a12Hx4SFzwbz6fkceyQD/T5ZfPpXnYS34SQDkIeBvw1mkeEqgIxwOe6u9kLd3dLQys4L/R7k6JjBSeysvcFrZOcI1PiCF3bqGbufXA3hZ2UuZBEfXqLeqH0jwsUlh9zQA8tK7h7wTIc9uSyXrjkyfH1Sbb3OCakufrIgtYOYHrYTFbNFKuPLNj73Cj7kRVBvGwKzHsZCH+Ds3a5JnnX3SmBvuNZ/AqsOuTDb7Aw+Y/+wBv6yykTAO2rKSMCkBe8Mkjnaw7+z+2luZQ8YP5gKT7OgSCXkY2KCESIcAb4yHizcXAL2a4G1FJpHK3ubPIBVc6yIYdfzROwZG6c0GiTBK5eOfM5gIIHb3pYO/scrpAJVAb+wTsQGs80hDk8sNYfax5yj5gFc3RxKfNQ7JBf8qcMLW42Nh63T2qTTS7Rx9hxhyZmztQAQxGN+vbPN/WkbCvv7uZ8GqvwYZMuKQvh4J54cYPmYS+w5wv7V1ZLBY3HfIHri+3BLDFndCWJwHrGX69DtPSj2+MBxkK8TEexreFIP/H+I/3W+kmInruy0sZBvES3r1PkP8mwfAVoL7eiZuHQKZUNgc8XN2DvxMQo9zcX2HjITFTf1GknjIi2/vBlnMRXKL0CR5G3bznfz29IF2nhsCorPvEM+/2e2GdzsD9Q+HVb2GGGbd2n5eXgTKQpG+A7ZDXtlsLs/r7ju2h+9Ykm72HQsC948EcQ9fc/B0PDuWeXXhrzQJPwbSG8HmcQzE+JJ4aAPkwd/Zc2y3Y0Z/KZzdfX0hvpPn2O7HjEhk98Qk8JyGPJPUKSahh68VlDp1dmwWdM6DO7cejH3ESLo4X+C2kgVbuRfBbZekdKzwoHOhJPLT3V1uosFpyMbM/K4rTEOpzb5QEhO6zwYUraPrXXJOszHzIH64zJFLFd9PpmqkfKqSMEKkiyMDlQSwMxRizMfA5mxgvoP44aatE5wnJRGH7kohD3VCG+wPREAeTh2gSeD0AGdMCZNitQcaBnWwDUuGOh392cI89EsC0jlIxOjrLYG8nr4+KGOXWAbyMW76IJ319V3kUa+JcdMj/qUQLApaclkkL8f+p6enR5wgnTPl1erpLialVEK7/p0pu3S2tLQkog2AvL+/v1QqXbhwIWRoEA9YtX//fhqtVJbmQCMm+9ACciSwoQN3wE9ipJFOcKI3GnaeiE6BdL5kacmTzuEW3qxOBQ5YWsKxpzUqAEUraANIZ7pOzoUExmLMMPBz8hfqF0P9+8JJPKeGNdTPlSUxRDoL6uH2Iny/pXyPSwBFzC0t5f+adpt4YAKQnxNQFlIC5E13dDV2b27i3gJSY/cWI043D/FqRK7NwL5XWeDf6w5a+ZpJ3EMSs9e9OUhn47iHXuBnHnJ+Us8+WJt/fe+zTK5O4k+Ia07q5IDf9yCFwyLBTzfHAXUmJjSHo2bsfozU43m6hRi2FXkd8ozin005wvKYfe7rBNsVP1kL0pnfBn6ellRAZ6oScE6B3zNkegOcGuYyAr9K6QznA/5pBepikn2VpbPZubbG8usrl8LFW8z+wLhKPETvGsv1rkS0hjZdDlzIjS+vgYv3ln0NQF4TZ0Qhe3Uf4ds4xP/135ztE2VBbZx0hjzIgtBciFxNtpKQ3DhMXlvik8ZwCEAHEL+nL6skjOVKgngEKamfg+6VEnPHTqY5OdPtJpwZOB9Sopsv9/oe8nSdHP6iVP5MHtfthOQAejomi5I7zn7sIWNPw+q6HQ4NbaUCmEKg896qfsAb5Dho7sAUuBIwTIqND8h0taw9PG2eXL/yU/tTtRmRK2lDijgzlcmc5+PGj3n6rPadXGa++wrimeZT22JDDfMN1WYdrgaK+WhiZuyjWslMpvkG1+TIakSF8+sXpGRDHLRlPOQXYPZCSpcVCKTzHbaXnCFfmngm2Y9uTwHpZJW9zZ9Byjhc60LGGzHIw+gN441eUMYgnY3l4pXJPkq3J8dLDKOaldGC6QCRQz0ax0e+viVPfp3S2aeQzmSOQB6MzBGYXJK0NSRD/KH32K0T6SxoJ8nQDSjoM3gU3zjpbGxlL1jFt/Wbt5OwcRO/EqzqPXgUE+8OujPecxQI5T6Dd54ZOErMxPfZFsYkeP4VIu4zeASExbnTtaQULIJs7TN4F2RoyxTAWbrPP6HB/4AfsiXo456wFTYnDoUMqYezs4NHgTrs8/dtssgPE9/edTbh9Y5zi1wb+E5x6G7OCXbr71EgYfsM9iAV0nYooEVu7zoTDxl2N5XNJmhedx/rHj5jSbrZxxoEbre4RM7DZHr/9VwC+eDY1487Alkgl2PvdPOx3rDThMmlV9XgQqa7j1N3n7G/7TTmO3v4OJOaWensM/a0j3UWk9Q9Mpxs5YwPSOckkMiPk63hB0x3H2/iP+0B/kRWOvNaCCmplg+dDCUhE0cwE0HmSkCSyqchf1F4TlNIZ/DTbSiQh5VQtPDzWBK2l65mmEPRrD6GigQnW5Dd5LYC/HL2SmSIoIdTB9lupCvEv+5eABYcXWk7OUCzgsblUtHlJazGJdJZT48s6usvJRJZT4+Vzkv09JjMIFopwxA/p5L5tbGRIhqp6xrD+VWmJOzdKaN0Bq2mzMyZMyFVls7x8fFklWZ8WE2Z1G9rSD8FTMJ5n2S6rWQ/UMbMqK0hlpYmxEO0suX+cMGiXKeGW/okEj9Z7OceTyuVYzKqH1nLkehjGSJm+vX7RxAgiThwSa6DoX4G6vdNIn5uKxrq58rym6Su2SCd4xlJv37naRVyLE1GQQyo537DjGARGsdIQg5GSklZPibbuhifb863Ef8277WiYa8V9YMv1pfAReju9b3nQStfM378kARwi7shk/vaT8wInOnH9j7N5DvBEh43D85leq34FWomNnBFfeN/m4TdaR7PZPZacYTzgxkfaRwaCtIZKjlGioffbS5msozvXIM8SOctcVDkdfyILfWNzjcTbJfJPMZKZ8WiwOggU+CgBkmkK/zqcHBwI4vJ/g4gGhwcXLmMwA/LjOSOg38y3+nMcwJQF6gN10gJWeQwPdtmTGAj46BGYwIbytNGF4PgZ8wD49QHxL8nsCHo3TFJDyIeNRJDp/ld6bmhdq8NtcWJK+D6PSZwNcSMDmzYa18D8G/ZUJvJv7JPlAllR/g06rmhDkQG+9WGYzrm0bW9ULNchcO2EpIbsVo5aDXZbkISuwi1hUY3Yg9B0BHiD5NvdEzMVdI2eUrbSfK0KxXQvSod+4bPZu83t/+FLqshyc+O3+2EOwcdkhnoTF+yyKT4w68dyGjodjg0oRJm6NAdQxUYDR3qMM7yreoHxsIICTsQkMKkBDrAzwyHsd7Er5I1h6bODaw5N7AWpPPkKcmbH6ppvsEALPVxtehcZm7gci7mdGBNpuDKsSeZ/HiSf/a01p185pt/lz9JrUlKRUfXesJkzg08xMUob4seGwW0ZTzsYAeTfZQHJ+w8SGEH5zuQ5w/+0sTDL5YdYRLa0XIsjYZCUZW9zZ9Bythf7UxHXRAddTB6E+E3dtBROkpj2UE7JvBI0rNG8QzTa8Ohnhtqwchn0g5dzIJRzcYM927Ua0Nd8JPpYPzoGviHusMcYWdTfGStJIYZE3QUZh9IZ+PYqzDCX88R+XYTkhqSOZKQwGY4f2JqQxgGsHXSNjZVtFOewnQuQa+BI3oPHNFr4HB5OoJ7o/O6TdsHDh/D+fkxs+b/RGLAxliMU455dH4IyE3xbc/NweJeA4f1GDiCiTt/zj0BtGyvv0JITNyFESCdew3cQcqSxd4Dd0LmZhbTtsPmth26EuvQoWvv/gNv/T3iESNu2+EM5wfr3bf/FqhcfLvXhXhBG86AHzbnnkD8JKz3hQTIZ4XsBGnLxWeF7Dgbz/YD1wZ+zOMLw0q0031EHAN+D1Jh29Nx/Pb06NlVuX5B/4wNnNvtklX3S1Yk3XzJCgTu5pcSzsOIvDc/l7D+gsjfEiSt1vQHu+HdP4sRd4+OgBiwVtvHgPO3Nf2ZvNub06VQqmtJ51/pUlJzt4REWHv6khVo4m7REWQrZBFq+++O1WNG3O2SN/HfigG/hGyFeEhKxwoPOhlKQiaOYCbecYZpKCHTkL8IUxgkLHdOg8scmYnghwE8dKgHnahDh5rBiZE9z5XAyNKOrE3wHAqS18H1LlQE0pk/bb3HgichIBD0c8pYnwTi5LZLWjJ0eyipByBnYBLGQZ674CBaWVd3CVlkYl2cY2FRN8ZNFySyrq4LrAA1DE5tOTGu2nz/kssiUhAANcwFQxj0KnkOBPwkQADp+XenjNLZQomjR4/OmDFj69atsBa0Ml9AwyoapBFJ5MFQMWNhtocuW1iE7DROYsQWFj539lvAj5mh7vEWZqPpOguLfRYWjDjs4B0JWWRSfEBbW+wLPwB5+V1n4jezsIj3HA2HGeohnmRfC1CfJM+R6G0RLoEYKF0CIn8hcwmqZRJhXFpYHIRFaeRBOKFbGHtqqJ8ry28S1ANaGdrJhvIWR5tZMAkePilsPEGa4mOx/w5UekeaTDZK6jG7lEAC+Jj802XM2eZjzjUnaVJi872vskg+LIIVr2NOHBt9jlXDXnDhkfslL9fB0Rnz+CH4xUlr2cvM2eOk7JikRxATH9MUunRM1HXIR0Q13/vyEazdcQgE99W9L7MgD/69EF9wfcy548NPNGdeHvUSM8TP1i8+DmV3gj/3GvHvPducKbguhjojrkFM8HWoP5PUD2thE15its387SYlseLYM4O2WTmVD7HX2MuBuUoyAMlLItk/FnEITIG8v709nAUORkjs7cdFHLSXwjnINYrzwyG2t3dxgbz8wg/+JD97cj4aB/VEHIRTiINDAPild92gTjjHQZ5AtmV6pjVcYvnmyerRBxEF7EUaFiUZq9l/Nb5wOCymIVGx4PSSx5ALOeQjCh+SYKKzRyc8SHracG8GiAm2QijIHlPvI5wKByenA0gMt7jTuyGTe8UrG/Ksn5EeZQ+BfKMqTd6RryG7BpC946cEvie4G/sM9NaRMwV+Qby3NXR1gttdKeTd7O2jUqH/3Q6Og2tGODgh5q4rnK+l1j4JkAdVB92eyrbEH+JJt7vILy37oR5JKKkHIqGqFGkUHJ23qp+UDT9gB8KaSUndHwYH1Jrz81MCK50DalIDLRtQ83RAzdjsK3MDV0A+9iH7JMbU2Fh+TMzDmslM5uyTriSe818PqJnNxE69cRny0Xdrsoo5wI1q6IDlJIYfz+XJoeGgLSvZ2gSfsXA5tHdhxzbsr9vdFPD7whhmku2d74CTDGwY7aWMD91nDZMFAkg8xPinsp0p6O27qXeht/kziMSTlHKFlc4Kk0vnJHbgkRG+FwZh4TWPNBixDcBPVCz4Ix5BTKZnJjuSVU4HmFljnsLEYZ1QMDSXHf+egax0JgO+5Bx5LZ09LkDNsYqtN2IKrnEb5RnbTmKk2zl6DRjWa8DwnvIU8rPm/wiCGFKQzn/8s3fuD0uMzMeRmP7DjL6cNX/l+s2cboYYflkuf2vLMHKHuOf5eOLPCmG/MvYYNtfvAomJk0vSngN2kLJxF4azCnXAjk19hjHiW1tCJcS/ZcDweHFYrwGeXfsNY+LOnU2AGLotcYJHr7/DtgyA+Ns9L8CGSrThDPiZeFDwvQbsBL84dOcj2Hqf85B/fH4oyF+oE/LBf7N/OQBKF/JcGyAPelruHH4a2sPEkfacY+tMZPdigDts5dHZAexPI3k9sDbeY5g87HVZ4hf0z9iAud28rLp7WZF0sxcrcNlnLbx8wROXAuKV6RYaRvygoUkkI/EFf/e7EWe8rIILkrt5OXFluz5OBOetguTuXk4QSZzdHifSDEhnL6vTXnLpfDeC1EYWobY/zlsxuWFnpLD1EluBPPFQvxL8GcHl4SqTcMmSm4kBDjDLUixs7dkHzFLpTGdS/b0SGXtbC27qgZN//SInRi4eVt2VpspPkCxkWymB9tyZE66DRIs7sdM2wi2KPbvKY9lToSTsALSWfSjLej/xw6kDBAOcHkjDouS3pcDPnYHlRVlIPNWtCvjSOegXkLMiXVdW73LSmTwATZwAaGviZ2NcdUSQVzzr7BIbw+lpHXatNrdWJaTn350ySmdzJcC5ZMkSyBw5cgQ6CzIvXrwgweCRh2jiANQgidgTJjY3N6Uuc/M9puaMOPRAJJy6vUlVBBhV5vvvQFzINkPuIYeQcImEiTffF76frSrc3DuR1mJuPmKMeeKFviQM8ExmTA2H0HUKEr3NQTqbm0PpEkgi9nslsfvrZW4uZhITw0JIJbDVEVtCzM3HwCp19XNl+U2CekArQzshL1jsO2IMk3Se1sIwQ7ySzE0NYQPh2+CnP93N84mM+ZgRpCwf462djc4052zwoeaSp6tJEWBw6EOjM03Bf3YH+/dqxHkbhljuVaPHsKp5/wPNJHG/Er97JmN0qDGpJCG0PnHCBcbwwBpwGp5pvmVTEyb/GvFDswYHXiOV94TKM44QP9D//kOj400g/izEKzYanJQFlzGjcCjSfNjx5rc94SIE5xkW2MSIHavJdsWKxsAFTBx3FaQz+FUaCeOwUwNI59BEGDr00Rr3RMbW3BD85jZ2kjD271EAGFcJYRFw6rGzc3Gxs4MLv51fEsSQm2rghLyhuS2T6E7iw6FAYihoOHYDPExOtzIKaMA3z4AGsNeJqfQBDLgeDva7YhRQf9ilBuLElcQJrRMnXg3LY4wCfh0R0CDsRl3uH6PPihij83UHn28geUGDgcF3HkANewIawC8ZoyTIg8hoQIqTjXKLhgENzu6pB8qAFAx+lilmYo1irpIwZSNhHHSvSgd54nm58Zd0WQ1Wdnbug03kOy2fznvv2NnZ2oLTEM4krBPYHiaxszYh8eq6HYpshyLScLIK1pjsgqqs3rZ+4IANHHE4BzC2+8OpSw2/uk75xr/GHIVB/mv/GvdCqpPjlc0wZr6X55yvxo9hpG6Bacwc/+oknu93Tcic48X6p3jVYNLoIx9TYmK4GMG2SIaEcdCWlcTE2i7RczAJuCOFAb8PnL4wF5gkOyfokBIDG3hjPEyWcPaUCz1KY6xM9oBfZW/zZ5Ay4y53FIw6GK7BuTAH6BwJzmNGuK0APwx7Jo0+Ah5WwDBpV0A6E7/ydOh/sj4joW/XhkM57AY7y8jsIwNe3RyBGRfs93rGBac+AC1uFHCEhCkbCePo0W8I3/oNNfx7j8vfu52HGpqu27QN9LGh2VhuLYhpTjeDOX45k1vFt019hjBZwSBVe/Q+Rzz/bR4EHXQqDjY3gHgeXxgCOrVHv+2Cxe79hmzo0o8Rs98gA+CE0/GPEFKqdbcBTPxp4gfan0vo0affX/2GMOKQHhfiST2cnek3JIuJjwv+jwTDmbrjhv9IPR17D2HiaD2gWbP+CwalC35+k0DrE2fXfkNubOgjbwjLjf/iHzPiHv0ukHoen+5E/MCpBKZ7zx78sirN2m9OVw/Lbp6WXT0sIN3kYcnkhf/Hvv2WXijbhsd3O2vUxcMShgDnvPEkKYsRd4sK7+Rh+d+VUdxIPpHBdD0+srNnSWc60+3EqE0eFlBz1/hE2MppDwvQyl0jw8l2T3tasko6KryLh8WJf2ArsHWWG08k0GngJ23j2knW8qGTQQmYiZKw1x9jdrrDnuu2m4Aopw+3gIq1MtkOTph6odtg6rE7mAz/U1y/6IlREQ8ciIBKwP2aS9Z23MkQMPRMtLMxh4iES6ycA/xTaQvD99tC3fwa4NQB0ll+erAbDA1Loidn4FIS+KFdJdApiShoMY2Wo+Uao6OjDX6QxSCRtbWd5Xn2FjIhJjaG80OcKOgXuoJhFgeJFmtrM6JAyPDXLr4s0tZm6xRASr075S+dCX5+r5+KK410Bh3cd4ihQNFyTtCUfXkM6dsXNCUEGJqay5dYDA2HQBz4SSmBvgR1S8KAMSP6Ui8PCIDt8IU7AbZP4onGHbErnFTCtYGgsn6uLL9JIIVJO+UhwkX+nnK7wN9NlY0HjLd0NjrVzOhUcy7tv69ZzyX1ei6pD6mRa2PiH3EKnI2Jf9j2ej3/bGLk2kgQP2J7Pa6ewa7gZGsY9md9o1NNid/wVPOef0IlrL//kvpGR6FyGt9zexPih5SreQT4lzSi9exr2H9JI6OjTUj8sKNQf0MSP+xP2C44WX//fU1J/OAl9fvvazxiO2yd9SundJApoDNVCdBe5iaDBxuaD5ZjZTKYrpCvIk7DwYNNoCsNYRUrsyCYSCs4kREniYczBYkHJxQgEpyP8elWo/3Z6+to//rylErnwRFXe6yv1fO3Wv3W1zS6VI/E9DtRDzzgHwz+E3WH7alFSg31qt/ztzokfoRrTVIPKIMe62uSeBDTEAmyuMfueiQPYmLo7lqgFUgNg843kNdWD/Ij/Bv0/K0uKRsaUQdkwWivw6ROeUrbSfK0KxXQvSodce1/KY10BuDkCt1Huh1O2QInIOhYdd1OjhRZpbKqUtZP6uEHq4OVzn7VQcV+4y9P5fkpXtVN1xuYrtcbt17/Gy9WN/NjTF2qTXfV58dz/lknDDi/qStUomcKNZwDZ4n6BXl6bBTQlikBQ5fsLDeAoQcMB5uQfRQMbEBzPADqWd6jJWIA5d4m/cn/ccIHpLNg1IGKDc1jerqtIqOUG8mjYYS7kjlSE8Yz5Ee4sXME/Iq5w8aPPl+X1NPDjc4RbpaNgMj1dUazeTJHYDZBzex2B52vP1QxR/gzLiG+zu1cZrT/KlKnPKXthO1CSvtdQfe+gwXmMGUmyOKFv6wU+MFWrN/E6eaFi1cJ1nLWre/gVp37dmjXqXvf/sTTqffgVu26d+kKHhrTodfgTp1hcaDyIineql0nsHZQSe++JAasVdf+xA/WvVdvLpjk+dal7+B27Xp36PGGeqCR7Xr0JQ3jt6Fdj4Fca0njSXynHr07tOvO7RcUIX4wLp5fVtmsfb/p6m7Rzd2CpJvcWYHb4WFiq5VDW67s12plv67HRpK1HU9btlw5FDzg7+gysu3Kod1Ojxb4O+/tR+rpcNqi1cphpIZOe/uBp4u7Rauto+W1WXZyt/hi5ciup0eTmju7W7RdObLbaSOSb7lyJKntv5ChjxhJt9NepE4upWOFB50MquCuSgA5U5EJRTz8qcdNSZh+clVEV/HjAeXrFP9kCHBTlTsJcFfJO052CYzE1vwiWQQgBkrK62DhTs4A/9rKQZ67AOVKMlpaYFpVFSjc2lrsf2xC8lwMZDi/HB3iB0hRiIAAUj9kOT+B2y5Au/6dKaN0NlOCrpALZVgUSGcSU6nxMjMD6Wy2N5wuVzDG/NV59Mlmn6HRQaZg3Lhxtra2Hzcd828LQ796hqBrIfVjU0//+iCdDe9fFfhpnp++H3/4k3qGD9mt7/GrzxRcdU9jDP3qqounXalAw54qpxf7sO97dhv0jYaYyp6ucpk827fabL9q8rT6R8nTY6NAQ2srZmob2EEw6hKS6oN0NvRfJfB/sHx4Bmz6COQvssr4wY7UzNdzRCme9ruCrr0HKpv95K+4h575/hXrqHSes/AXvh+t9GblPbvLefOu581Juum8OUjnLo8TOc8HTm8lmXe5ewfybEsK7vz7kul6fowgho4VHprnSMVJmRQ/r0RorZWGGM0paN8KAu36d6bs0tnU1JSfLl68+JdffoF02rRpAo8gspKmXqamIJ1N94RpiPmIqdHmzqNPNAMzlKefT56OSAUwSz86VDrzzNOvnlw6XxH4P5jtOFePKXj9T+HDXZYLAvhGwjjoXpWOpWOmgHT+a8RXdPlThErnj2r02CigLatEBLYXjLqEpHqsdPYj0vkjmMdx9sXqpD/PpDEjj9cSBPCNhHF06TVApfUeOPyrOQv/3uPMdy5asX7BopUDR4zhO9Heyqwuze5yzrzrOXOSbjqnkM4KzwdO/zgMipk+NnYjj+m462/lGLKWD50LFZsoZ1sJI7Xcc5sulwmqWysAtOvfmbJI57y8PEdHR5PPDEMTk0G9h5mMGUmXKxij/+xkeKzZZ2h0UCqwqQCMPtl8pG/dUb51R/rIU9+6w3zr9llXa5RnbYFfmL43/3Dfut131e62tkb3dTUG7aoxyreOhnjalQroXpWa3UNm0twnygqnibN8DL72MfhY6c+Xv6DHRgFtWeVhbEC7EiPQt+7As3VVjMwPmA6FObKuVrd17BwZ6gwtURtpf6UL7Xc5Mpmsa68BnXv069SjbyelFP3vw2958evOZ0y7nDHrfJpNO50xbbFpVGe34ayH5/9g+c5nzFpsGtliWZ8WS/t02NSn8xljQUy3M+YyRkZHjAI6GSo2plY2IweOtLEypctlomrVqlWqVKkIKe36d6aMd52/++47KtmQisHoPzqNOtJ01JFmr9OjvPwn6jdz7khHpIIFCxaMHTsW5upHTI3cWo/0YRUAm/rI08qTNzsjlGUw2TXv7+eWstLZW3+Wj748Nfjw+U0hJvTYKKh0x2isfzt1I7Di5xeGCN+WMPHLrzp179Oxe5+3Sfuiv8x+S89ZnU+ZdgHNWknSyYE/0rHCY/78+Spnx6eXgmatCHTpUuJH77tQRul85swZkGvGSIXB8PeOIw83/dxsk8dyOiIV+Pv700+TfzzGbO08wrt2JbVtQSU+0QJ4e3t/9dVXsF9wBiQ7CHzO+eUHJ8y8pPex7GvvaoFJ++ixUVDpjpG13xeCgVdZbJRPvfPJTrTfFVz08u7QtRfaBzML95mdTpp0/teETU/K04qdv5RMn5fj4+fnR6fDp87/KgZOTsKZW2bKKJ0BDw+Pb7/9lui2MWPGkIwA9BM+gN9wQ8eRbk0/HzM90FFZNxNAPX/33Xd0yn4MxmztJLjcVgozPd1WWTcTQJl93C6tUCw7OF4gZz+Y/RlsfCP19fsf+VSuY2Tl11Yw/CqFLQyx9H5ynPZ4SS5e8h4/eVqn7n3ad+mJ9r7N4sLMTidMOp8wqeBpt3/NpwT86J18lY4SJXx9fefNm0dnxacLla4fj06dOh08eJB2enlQdumMIAiCIAiHTCZ8nhVBkE8PlM4IgiAIgiAIUipQOiMIgiAIgiBIqUDpjCAIgiAIgiClAqUzgiAIgiAIgpQKlM4IgiAIgiAIUipQOiMIgiAIgiBIqSi7dL569eqyZcumVwZmzV9Y77cbaGWwwXvDj955QQ95Sa5cubJ06VLaxR8E2/3XBc2rIDZwT/jhiOe0X5QICgr6YDOl345bdddDL12HtK5Sin70Q2rlFnUsUvWkvnbtWmU5q38atNl8S3AyKV/TcAJHEKTMlFE6g2yaVnmYOe+7OuuvwWVDnl7D/NvmjyidfD/KALDZW8b2f5i8SvUMunnhwoXQeLhMkr0A3l++33ZWOqOhabb6v10/ePspHaMKKtdZ/dOg9aab6s4n5ZhXPoEjCPIulFE6L126dOrUqXT2T5tWwfMz5n5XZx17BsG0bOnAPWH0wCtYsmQJ6VvSycAHyI/de1VzOz9uOmC3sJcAMlM+GH23/cc2Bg3tTWblFkXHqAIYq9xcIxkA84T3lGelMxyON51b3jEdpHQCRxDkXSijdIY5/+WXX1aW9Ks539ZeexXtXYweeAUaevv9pdZ7rghaVdGM9g4PaPmHpM+2/2qvY39gyNOrvDw/RT/6r3X6+xYdowpg/GiYfZi+j7TVnzcE55D3ZPQYIwhSHpRROsOcr0TwpDNfe2H+LfL0wCugPfth4Unnt27/h8nT3uExZcoUaPkHS3v/faPWmiu11l5hU2VDPzH0y42OUQUwhJAPTMs/rgvOIe8pT48xgiDlQRmlM1ykKxHTvplXc/XlWmsuQ1pTnmL+bfP0wCugPfthsdoVpLJtFSdPe4cHbfqHgkpnNLRSGB2jCugYQj4gLTZe03A+Kcc8PcYIgpQHZZfOkydP5qeEiumfNpuVzmjvYvTAK+D6lqC5/8vLbymXzhXZaO/wIO3/YPTaer3matJLmJY2dVsdlMlk1/TO0BDzSaZ0jCpQN+/Q//78LX5npfMHMHqMEQQpD8oonSe9H45MmsRkRx+5l02Xy4mpX8+t8WvQ+7bM+MfL4wsFzvdnzKunNcKzBc73Z/TAK6A9+34IgmHApE66nEqXFVjsDBS0qqIZ7R0e7AXzA9Jz6/UarDBiGyNM0a/G7waTl8mscfGFwK8u/pPx0zGqgE6zD07q5Ump7NZh3r9f3mVD6k5KQPa9I9HZUC1cvt6aZhuuwuH4AEaPMYIg5cFHls6Cc9khOD1Jog5FSchieTH16zk1fmVVV3V5SvIgPenOMMzlXKbGr4+VY9x+DWSYbBqkAPQxP4bLZz56sPxRnrIfUkE9coUtjCF5fmQsA616qnJbkDLPk2rczlRZz+VLcKIsJJUARGGrrAfkPonJhLBLGSpjSJ6EcdCefT/4s1ep5EmByXRZgcUOaIyKtqnLx4YHkSOrLmY57FduBtkjdWOAy8OAUfw0Uh+jxERVHJ44EX4i0gg5cOmVu18TfWgif8xMDEqlKyZOTA2aqBxP6PHX1eqrApVt+arXe+r2ihGszYx/RFZBJ1Rf9Uiwlhh/vgDVw7OJ3w1WKdopKM5VC6OrOowuuVNzS8CgFDvFlPzvz1xho0xmdc8XAr864+8yAPO0+qqnZBW3ywT+jnAdqLzXseF05PCr0tCxYKSIclWkFL8eDUYq56DTTAmQhjRCDrRp0uFouq48SA6clMxuHeb9++VdNqTupATIL1lQLVy+3ppmG65oOJ+UY54ePARByoMySme4SE+YMOHd0+SAiXAumzjRn3gOTZjASCLhPKQuvmzplJnfVFsZUG1VAJvKbdmqgJicV3oTIvQm/AbGPItwfckIYiBvsDJAb4IbiQFzmfCbiHlV7VwqP4bLZz6KXfYwT9kP+cteAS5P8rlKGCajWphUEAOp68oAhpEuiaaRzJMIuGRWW/mIH8PlmeeJ1UIylf2QxoYGLInOI5UEuf2WyeRXu5QhiAGDy3wMw5BOiAmCMKaalyJMqU564BVo6O13T+0mThw+3GKinaXAb7bdX2XblPOx4bTBl7OZaqtiOb8gnnn5xOUZ9IC8t9WMAUghjNSm7vhyeRLGB1quzCGQzpLI4RcThyuQ3tnJBiv0MYhjWBx+MYms9Rw+HC7bh6OzyVqYNZESqBmqESKXzgFyYSRM9b4+qTdhgyjaz/U56KrX/suXAjKZQj23BFgLnSBXe8KykMJPtSXRuRAj764N1X9xA7/bqoBYJltvwiPiZ57dUYi8gMuXQI+San/LjPZXiDm2NnUtgZQV1jnp1YNBZ5fwv9fUdRX0ACedNUWSFOJFTCbpMdg7A+iNVefJWvj97PIERpSil37YTf3xj4JywO/P9sYDf4WkZmuDVdA58mm4wXXCb26vxOAHBRzLSOUdy9YDx4Wnntm1mYxU9OSVvANpPSRlXj2JeZIfy+RWX5XE96tM5SP0NTB+BDOOpCANwyTM8OE7yWhkkjyjpXAGYH+5qYx/21RwFXh/aXLABNjQhAl+GmLUpX5wbWKSJwQkK6+F/iHzUeAvTdr0t8uCc8h7ytNjjCBIeVBG6QxzXiXS6MNwVmWe0DtqgU/YyED2pJM6ISiVxADMk0BYFe0GfoimQACRznfhJzyTyjlpmQkTUoMgnsL5SeXR0RDPVgX/TXCLJqs4qHQuafqzTlZbeZ7kRQ+iL+fwRKoaY5XWE6baUigoXAWW+ZAnnUuawdIAg/m7SX4p1CN9tuxxPreWM1DYIpC5Cmkec/M8XFOrrUzhAvj2Wjormf6igGrf0825LoWTZka10CxuLbFlUENO+tLoXNIJSxdBqxKWxRXwY/hGup2D9qwSKgcAHGgpI+WOy2E4ZPI4kolmC5Q43MoDhkCls8JAc/B/WkDvgVYj6h96ADQHiDNWOq/kSWeeyXu7wOB0AllU9PYTLoAzvfnnl4AuhP5Rc3w5I3vBhza9JGScw2WYLk+YYDzWURK+kygS0i0grCeMsyBrx02YELbDQspISB+CAoBLtbwaId02XzFYyfaSPPVXSgNED2NAb/H9ZFQbLD0B+Zib56ATDNhOoPFcJIy3pezul/Drg01wM1h5jviDTp+PYWBEpUA+NtQ/SAqTJYL1n/OW+xNJKZIqtwTSy17+MUxBtUWPMuMf0h6UA7/rDORHFgK4f1FZGl9ISsFvTpCSl8MyuFXw01ReJ/wchXh6kuHiBf6g+EKQzgaez8l+cZsm44HEy1PaTheI4cVzfkgzH8a4PIdSJfxLYVs5aUtjoAe8wePy/aVYmJLymqEZIkai75ZA+hD602A+9CdbEDoW5iapIeg0HJdcA/ZUwK6CQwY/sDMf3JZv6/XWY8MCQKAHnb4Rw8CR4ve2sJ0kJbvJQceQEkQacuMt0XdcChscQBZVnpZVngcIZHgTP0x7mO5E0Uqj6XumySzgIgXnB7h2kDy/Ts4JCIoHwrJ8cxDPSWdYS64+XCUq9wKAHaFeOfw5y0H6h3lyl8SQOtVd9eiCgibr5dKZfw5R/FAH+Oc3OLjUq/BrOAEqGy2JIEh5UEbpPF4Nkig3WDvMN3nYsGEew4YxTOKhaKkt+MMO3JVCqUMQc9dtvISR2u4Ps7Yfn3hxTCLEs7HDxtuau40fD5ICzm7DhnmDRxK2E06Z48cHkFJwChy2PQz8YduHQX68213w+0ERJjlSCkX2wiom0YPbEMfkGbMNVvgZrPBXmS5dwV7YXJ4yBiu81MVAevmiHygt/VMJ6mJACoC2ULeWS6EeBhTbrQzltfpL/JnU+9CBS+MKwcO8eALS1mDFWeVIdu2zRIOQTGW/II0Jg2szY3DhmcDP7nX2c3mDqYd5GsdKGV4MP5Uf9tfQnlVC5QCAA80kegWk0lKpQePhoDtYewkOd+JFOKwMHFZyTMf7J5N4DtNt0JjXrXJZ6MUwrwzCJMST+fihSypjsOQEFwNHJCgb9iiGX4pLgzxApeUZLIkjnqALfpmgTi6+VI6EdCn0AO0uFWu5lHQOH2g2XCAFKSed+f6Igw4gjmH377pNSIRfGuN8+GudoA8lEYfuSiGf7D8eLtXjx7txa7m026bLbEtAGLHtkVvJPPSJyzP5UVb4YSAtfQDDjM0HXQBRyHVCibIgnekuMYzrSzJOhDGZcQ9BLhusuAv5IHdSVZrAz8WzLaHjTW5yPz2IK8/pLbykO3492JIJ6+FHCzmyoJtBHOi6pVJ/TiqZKaBlQV2DatQdf0N3wvqYoPXsD4AVT1jNmpsBKhPig9zWM3C4QVLD0ZT7oU52E1DP8/vsHPF4Dm2IDYNRAf4I1p/K3us1WPGQv4+Qd4HeYMU8hdRJYmCnqFeub0jZ13NNHgPFWeXtxXZy0EXYXH5mHP0OJbs7JbdFMtAtXAdCC2FPDRbeoIdSEePKVivVP5EQ5HGJHdggnZXqEeTJRjlg/KgEJjUZb2SRSQ04ECYZPx7O6+xchjlLTsvc/AU//zwAp2sGBnZQKvjhvMxI73okw5RnSQ7bOd4tCsYzBO+NkoKHO2lAsPLlAMJ2RrJh/DqBYea2bHWqirPN2xMJq2zNh8GGoDa4aMAVBXTtsItJ4IRIdXsBfpDExE9qVj4pAconPWiYuqseKcLReF2g4ByiO+8cGfku42HEvloaVwB+uGrIR34CO/LHr4/JhstQisvCS4oTIFtWcQI8zq+NS6GFCIKUF+UvncPgJGt7APJw7mASvQNS2GBOFkBeevdQBMTYO0GeO5dBHiAnO+68rCgVBXk4X3slQilryPOlGJFZ1t6J48c7wGKir628QlZtc0ya8bU+e/rw018uTxV5V1brwHmSAYGlP++1nx/D5dmbc6mMPntiUh0DVzKiRNXVExNGpdWS6Dz9n11VxgSdOsswuSQMLsIGC+8rx5A8kc7qtgV50CukHt2gV3A+VY4BrXw5h9Ff/pD4oUK5qBLWQ/KkKg7as0qoGwBwXFIUxwUOpTccyvG2gsPtbQuHMnF8YLJa6fyPL79tetDm1IcuL9g2u0CekejdfMVvMxwRIp0F+0LyMbf92B5ekUj8QR7w0yhPH/QTL4bL86SzsB5+nvYOD0dVuDk6wo47+ifTZTm+do4SJsnRJTI5wBHGsKOjH10hxx6KJPuChIF8sr8jSBl5NUK6/hkEjZGbryJTwkQP78NR5haXwpHNean/XwaJd1kMnZABncAFcPXozD2r47gezNmRvaK7vHxdCbGYUCjL6B1P0F/uBYt6i/1cHHcxTBZ0AkhkvRnriJ8zovwE7WSep8BBlOepH+YdHES9uTTPzp3F7GAGi7l2HEQkDGAXdnxm6l2DguwmXH6ClrzSv/gy6CIc4nz9BewkgvbA71IYLZAnfr25EaQemHQiJpfsNfMidUl0rv7ys6/9F9PkYa/bqQe94ehKeiMmcB3soH6olKzSXeBF/Isd1zHZCUG5bPPAL4qJIE2FekRxD6GjyOZgEELxJQ/zdRzXyTs2g6tKYb78joU9FTFSef6iogNpJLkVrb/4GIxkkM76yxO5VXIr0c/E2AHKg04zJYg05OMWBXOWBebyXpi/8tPyAZi/kjCYv5Dnnwfor74oVs6Sn8225h6QB8ytHaAsXAXIb2nw8E8agvNDkp8DF8avk49yca55ALncBIyHHQfdnDze1pz41e0F36/upASoOOnJr03qrnp8QDqrO5/Q8xv785I9O8EPP/3bWSRGfyZMgbPsCSf1oat8SMuvZRJ9+dwR1EPy9MghCFIelFE6O/CACzbNOThIIl3vSGCtK+TBL7njDL/6HRzcLMc5ELEb4ODAMCmWlxIcHOwgJtnfAc5lDg6+JB6KMZI7rpESti4HhzvODhJGAmIC8qAbfJNfbxcW/eSLUBLOaA7+ycTPr5Bj4vRZest8lU13ma+Og4uOwzpRjC9c5PSWPRAEcBYT6svei5oTLvDzDUTJkod5AiffdH+Eza1b7LCOyUlZElcoWAsW5OkrYgp1XFMgDAyuyrBRvWWpgjBizLMEveBMgZNvOt9dhEqcYXOgKkKlgrVgi+ecgas7PaJylsTmCmI4oxEKoFf5x50DBgCoOjIA2EXFADgwFo5sqNtd6V1XOKbSsftCYS053CAHSbAzWYySkmMKypL4OYz/9hG0KujfMzEMtDkVDlCQlNH7MYS/Fo4Iq7qW3ec7OQs6D8cUyibQRXfo/Dw99+dcAN+WQA9kP9d8fMFo7/CAXlJGpXSOdAXpLIXRfscJpDNI5BLS+RAUkUa+UTp3+TNITy6S5Kmvcp5IZ76fefp46aN8kg+6CJ2Qr8eqOhVlSV4XxFygK4xM/eWpnB/6H3ZWN/CV3uJjJB4EH8ToOD4EURhzc70gHmJ4LaH1u0ApRqz3E6s1aUzcA3bezQgHpQge+HVHepWDPb7L78sFZab++adcPeyzyB7Pya8j/eUJxC96EEW2KPC7LIb4DIgnPyTgB5LAz7WH+Pl53Z98mdRwosgFMUHHjxHlDXndBReZp+GkzTHPC2Dg6S+OY2PkgxDEPeTBRNEhUBVkuHpIx+oEvoIfDOBhXqYuZvUxm+d+BUGeHf8gzRfcgHyQB9TJSmd5DSrazOVJezjILINRRDIcMKlDJczQoTuGyoH/MdJwmMtuME+lkbSwApi/pAh3HuCf1VWen4kTBjxZ5E71gvMDv6x8FbsNsgrOKtzjNwC/OBcD+wU1kADYHYexB4ifzCzi54C9IH6uOHehUdk/cNWD+UgX33TVI5B6Gq4JEJxD4CrA/yNvdsTCteMXX9Hd69QFJ2rF5UPzCZBvtCSCIOVBOUhnPuQkQs6YAJPi78OexdjzRYLP2BQ4fwQGgFhzGOtNAgRnUsHJjkhnB+c7kFeWzmSRO6MRv8pTs1w6E9WlOg064yk/ASWrXLtkmQ+Tk77kLgScUV7LpTzprDaGpKKYcPY+nJI/5jYr0PV+eUw8Lj/6MMwrvVCxciSkCuks9AtS+HnApD5YAtpI1VqdGcd0HNbKFfZadlv/ZSjHkJQeeAW0Z5VQNwBgBEgiDkZJGendgHAIGHcQ1goOt7/8UNr5JQmOKYdcOpdo1ZJffBkp/A4pYG8WxsAB8uSv5UnnEqVISrWy50viibnto+h8YSSk7Bh4LZ2Fa7mU9g4P2vSSkB0X7KD0rhvpmYNwxZWEOwSm0BVyouQ/OewOhEMeBjm/k/l0/iNAdyk0xgdSXXkqyIsesHqL74eBBKOR5GNC2E7QXfRYZVkuHxMsD1uWQPwxoeyOa4Nu/tGZi4Ejsvhuru6yM5BfAiNZGq8bKibxJIZrCVc/1AOXf92lUSQGFkWwFdd43aVnaJ3PEqBObYe1YDBu2XTGWt2lni5QGyPSdX9O6nFZSheDzkM7oQ20nczzZLJFgT/IE+LzIZ6NefoYjrKyHzxcOwV5tgNZfSP0B7lDcfbHGPFrf3OGtDnmxg5W5SyLAr/zL9ASmHRsSyBG0Cdcx+r+4Ax1LoFtZb+UD6vXsAdiaaro0QO6zEM3VMpvj3KeximgY0gJwaTm5rKrgxsM4x1hEiKpCXaWRoIi/Gmu8vwscAqkM3d+4IfxVxHdPNQziTSASfRQWRyAGiQMs2N7KPyMvytXt+AkYcp7ISiu7qQEvO1Vj0/DNf5wRLhzCAw50M06gRnyuydr2V9lcukMa+W3QthzdUzgWvZfITzTwb/kF5hZ/BPgRX5t/JQeYwRByoMySmd7NUgiXeAkYm/vAvkoF3s4o1n7JJBV+63t4YQFZfeHSeztrYkz4qA9iGN7l0iyCMXgbOUSCQEsd5zla53vQD50n7WUjYyCPNQMMsJ6XyjkfaEIk2Tvl8QWsLdP8rNPYpsH7tdMmDZDZ6k3XCrkqTekcIYKyi3UXRpL/MyLFLiY6SyNhLzzUm9o9uI4WEvjY0K94fqk8+09UpZfDz8Pl73FD+CiqyJGFBe7JK6Q5OX1SxbH5pIYWEUufpAPOgercuE8SMqSVTo/PFBZJ1ywdYNFyn5ImZcpOnDVlOeDPKHOAt3/MiC/BFblpru8hD0tEU/8zqmwLSeV24KUHHcO2rNKaBgAXhbjGIa9yW1xCTzjwCM43KAdYbW9tbfgmHKM2aKibTHX2DdnsX347T2+H1LRg3uB2dC39zk/VWOe6ZDXWeTDpN5zeUF6gx0DbA8sPQ152vNLU7htLYYeyH6u7vhyeXnflIA23d4eLpNcSq7KZAeJB/Ydgi3lPQNdk3DJgi0rV8+wlty12gMTx9YCPDDISSeTtQSS77yRlc4a7LU4U1jgv2dEoCM90yFP9S5I1aWsAIXDx8ovUIGePkvi2Aznh74iYaI4VrQReUcCiIlio8jAhjzVoOeelghQagm79fuwdU/IKzTEK91FR7kAeVPzSFNJjMtLdp5yWpn4ucXFcIjlwoIEsxPhxivIEz9MBMizejQnnR0V8uIxN85wzYb2sFJ+UTgpLmIKyaZj4h5wbSDqllQL23V5SXuJVEuKQz7oJcTDRFbs1zWIZ3cTDH7Zkh8V8o6VkKrAVHas9ldHyS8HMNFdX3bEOsCPhzM633py/kDXtfDDQNshXHBElA3q50MGkjL8SQ34s9Mz5WCEZJz9OBioUigoPy0DMIxh0EKGX4Q/zclVwD+VbisF4l2iBCdtIp0hIzg/8MP4q5L97cl5A/LytiWrLA6QGsbZ+uwYbcE1g0w35b0AP/xCAOErr5XOUOWTEvC2Vz0+DVb78c8hMSHya438nC+/Uoid2TnCXoMUVxDvwHPeMGF13J+Rc07MtaNQP1uKvUKpPi8pH2sEQd6FsktnOzs7MvkBkocUTiI0Qs5gz0R7a1Oy1hbWRhyEH/3j9odz8Ra29kziRRIMZysXOzvuZAdrOelMIhM9DUkkMNozCWQE+DmZReokJ0e5+3Wrxk+dobOEPYOwqdwWLwIlBGdFCggs7ek+Oks8IMZ5CZywRCCSSLx8UQp6QmfJaVoDrx6+B4QaW0rJD2ngaQ+GyaMbY5hf7ubrzN/OlZKrtATIay/yDnTZzkWCH65/6rYL0lkHpLOSH9KYG6dJDQStwHTQxOCX679ncC4mwWTXSMwv9/N1vllD/KzxtyhPSRgHdCy/h7lUMACGeCbZWZuwofLrBpN4CS5ydtaXYBHiybWNhjJMIsOYDvGws7MSHFMuNdpy6XWriMHx+h7a9oqoXs4PVxpaqZygHNjlWNYfDFcdBq46pOwvs08z2fEkBsaAzjcwBlg/PShLEui+v3w9VIDFj9lrGIl8ncqNRvCAZitDxjmNkBMuZQYP3g59RQJMrO1Ctw0GIUVXs1MpydbckKxN8rODS7W8GiGdfvfXWcL2krY85ecFewF9or0khl378yXR3WvEyXbC/Gsk3nnJJZgI2rfYMcbGPAgmMQAMYO3Z7GiRx7xuJCDvtxTYrvZ8DyaVPqIAsIPw5yOkPcot0VkSA6NRxGTpnGbLsjEvSsSwR80jTftnMkHyiTMQCs6+BvHysiKdC3BYacsVi95O0w8zTDqJ/+Vuns73MBHYGLn/FfEH3s8XMa9IvNb30BtUZ9yHCTgdJqA7+AMvQJ15JOaX+e5MzjMSAw3TconX+f4g+KHfYq47ET/g/BSKQy95gD/wJPzAKwCnPD6Fi2drm32KyaGvfGbPDN87gV8+N4Udq8127OtjCqOUzGVSD+cPvADyC05ECQK/cp5WrYCdpXJgIPFTwaQGdoRJYDTCWlNr+yTPIdQLk1Yxu4malE9xe/5ZHfSoB8hW+U9oGm/jzWliskUinSHPaV/ih2HPndv5shgmC5MIJ1iWZPgvMYlKZ952CVwN0Ax3Q3NoBgwyUMZQg2AviBA3NLflag5n/+AwgUhnrmdIquGqR/S34KoHKQHyrHSGY6E4k2jD2Uwxa+CIM6ki9hDTsxwdxoA2mU3yUs4/KE6AgisF77wEeVoSQZDyoIzSeZwaJHecQyWsCBgsZ5yVMV0hB9ZeTGTGjTWjy3IGG1txwTawaGhmYzaKrDIbO27U4FFcvLHVOBIJcDVbQZHBxtwixBjDWtb9Gscvv9JefElgVacdrmq3hpjWtDXai91fr7I7qD1/u7pFdUaks8BJTOsnqGQ7tzl+bVrz3bXAs/iUciTfr2zMs3iQNQInMa2FUAndFpj2woPcKthrED2vF+0O0pg37SA98ApozyqheQDARXFfmIQ7Os7gkdzZHiYhwdyBExxTjtF/eQlaRYw9QLydAhP0AHd8wc/26k+HuUhuGPDHgOCgVJ19iqsKTENf0d7hAVdHZeBn5GBDc7LXBHPDwXL3a8xt7AYPNqSrBw/mVDUAV3rleELH3/3YloByVUrZvbCX74I81Zou31/5Wthf4mc7YYm8E+T+qvYH2cFTMuZ1DyhiOD9bA6xdIu+3klvk6hH4ISUtCfS4FCjN1/45VHUM1ABHbYl8gtjLJ0jJ1mrBGPhZflhJa6Hl8njIcy1km61oA9+vNX87P/51b/D2BbarZb/9dZ3T5cOGxPC3+7184MnbTMck8f8M/u3K8YLa+C1k56agB2BA8kpBO9lN8DzU/zP0BtRDW64hpWNUAZ1mSowys6GjUAF3igb4p2VuCkMRM0PIwxmd/Y9/Viezmx8vOGnDFcHKGBaFBflhglXcRYSt19hYZXGAXwNpBreoci8ArmYzs1GwRvmkBAj6RxCj8qrHUW+VD3s4eMadbWCoVJ39+szGnavB+Kd0skpwAlQ2eowRBCkPyl8635HAWhBFQvxAJzEpVt4JdPkD4vjldK3FrOqSp17vKQ/S+RdWOpelbBnyRDprjinHPD3wCmjPKqFhAEQ5j5MwUpt94XRZIZ2hAF1+E6M3l739HyZPe4eHra0ttPyDpR1+89X6xasymuhRjPNTRuuXCwI/2vszOkYVwBBCypc3XvVAOms4n5Rjnh5jBEHKgzJKZ7hIqwSEUIQE1jrTZR7SKJfbsMpyL13+gNhPmVZ10UWtXy5CWlWekjw/fXd/1Xnnq8795/3VL/BXmXVSa8EBZb8gLS8/PfAKaM8qoWEAMCl+FxMZW2tTuiwPYiQRIJ3p8psw3OTJted972/Z/LR3eNCmfygqr3SuOu+C1terBU6092p0jCqgYwgpP9541au70lvD+aQc/fQYIwhSHpRdOtvY2Cino8zGmo4caGs7VnntSFNr01GDbGysBf4PkNpPZqUz2rsYPfAK1PW2hqM8aIylhRGMjdeesba2A0eajjUbJYhUl46SS+eKbLR3eECzPyTt1vtUWQS95FnlZ3mKecyryTdfJ/xMhubZh2kZ0jde9Yh0/gBGjzGCIOVBGaXzt99+O7byYDfpyyo/e8AF43/ylOT5Kfo1+7tsukwPvIL58+fTzv2AjPwDGsO252P1g2Z/lz+FvQTATIELJDT+w6Tt1nmzrZLLIxWGfmLo/9lz1O7/6BhVULnO6p8GtZd7qTuflKO/q9IJHEGQd6GM0tnPz49O/crAuElf/u8nOIl4QApnE8y/bd45pMTrDoCPMgBG/uGuuZ0fN+8UzP59vwBvb+/p06dbW1tD+yElvL/8F2svsY1BQ9NoVX/23HWdvuaCo3Kd1T8NQDrD4SjzOaeUeRelEziCIO9CGaUzAOfZ+fPnW1lZwQW7gqe2E6b834/u//eTO5sqG/qJqfJ3/jNIWTcTfH19586dq7nnyzcdvvEC26qP0Q+sqfd33Bh48FYS7RclQD3DTIFd+ADIpTP7AwNTTNWlI3fecFUzqf39/SvLWf3TSGstu6h8PhF6iJXJz57Ag1E3I0g5U3bpjCAIgiAIgiCfFSidEQRBEARBEKRUoHRGEARBEARBkFKB0hlBEARBEARBSgVKZwRBEARBEAQpFSidEQRBEARBEKRUoHRGEARBEARBkFKB0hlBEARBEARBSgVKZwRBEARBEAQpFSidEQRBEARBEKRUoHRGEARBEARBkFKB0hlBEARBEARBSgVKZwRBEARBEAQpFSidEQRBEARBEKRUoHRGEARBEARBkFKB0hlBEARBEARBSgVKZwRBEARBEAQpFSidEQRBEARBEKRUoHRGEARBEARBkFKB0hlBEARBEARBSgVKZwRBEARBEAQpFSidEQRBEARBEKRUoHRGEARBEARBkFKB0hlBEARBEARBSgVKZwRBEARBEAQpFSidEQRBEARBEKRUoHRGKiAyWUFukSglP/l27j337NuHpDf2SIK2iH3WZXouFZ1bmHHy6/TDk9Nc7NKcbNJc7dMPT8o4Ni3jxEzRqTmiswtEF37MurhMHLhJGuyUE30hL+F64avY4uw0pqiAVo8gCIIgCFImUDojH5uigmLpq4KnUXmPg3KiTktv7RcH/AHaN/PCj6Iz34r+nZ1xYgarjI9OAYmcfmh8uptDmsu4NGebNCfrtINWbOo8Ns3ZNt1lXLqrXbqbPQRAGMRnHP8q49+vRWfmZZ5fmOn+c5bXSnHAn9JbTrmxlwqe3ZXlZDDFhbQNCIIgCIIgpQClM/LBkRXL8qWF6fH5ibdyoy9k3zooCdqUdWlVpvtPotNzQSWz4tjJKu2gZTmb01ioOeP49MxzC7O8V0kub8kOccm975mfHFKYkSgryIGG0RYiCIIgCIKoAqUz8kGQFRdnZxSmPS5ICc176JcTcVJybQfIZdGpOaxQFmjcD2XphyaIzszL8l4tubErJ/J0fsINaGFxbhYjk9FmIwiCIAiC8EDpjLxHZEUFxTmZRaLkgmd3c+95gFzO9FiccfRL9kELJSH7kc3JOvPsd5LrO3Mf+Ba8iCnKeirLz8b70AiCIAiC8EHpjLwHiotkhXmyfElhekJujLc48M/0o1PYJ5IFarVimottxqk50pu7C5JDiqWv2Ac5iovofiEIgiAI8nmD0hkpb4qLikRJufc8snzXZZyYyf7Rnovte3l2+f2Zk3W667j0Q+NFZxdkhx4uevUYbz8jCIIgCAKgdEbKieLCIlFKboyXJPBP9uUYJ2eB9ExzqiR3mtWZi23G8elZnkuyQ1wLn99DAY0gCIIgnzkonZF3RCYryC54GpUTeVocuCnz3IL0Q44V8VHmdzEQ0CdmiH3W5oQdhT2V5YnpriMIgiAI8pmB0hkpI7Ki/CLx8/zk2zlRp8UBGzNOzmLftSwQnZ+SOY3NOPql2P/3nLvnCl/cl+VJaEcgCIIgCPLZgNIZeWtkBTlFWU/zU8KyI05kei5lv+r3id1m1mjpR6dIAjflx10tlrzA19ghCIIgyGcFSmek9MiY4kJZvrTgyR3Jjd3sbWYlWfnZmFXGv19nhx4ulqbJivJRQCMIgiDIZwJKZ6S0yArzCl89klz9h31vhss4+XewBYLyczIn6/RD4zMv/JSfHIIPbyAIgiDIZwJKZ+TNgGgueHZPGuwkOvd9+pFJleYNze/bnKzSXe1EZ+az7994+VCG799AEARBkE8dlM6IJmRFhQXP7+dEnMzyWZNxYsZn9Uxzqc0q4+QscdDmvLjL7AcIEQRBEAT5dEHpjKhGVphXlB6fG+MtufwX+1jzZ/54xpss3c0x6+Ky3Fjv4txMfP0zgiAIgnyqoHRGlCguKs7OKEgNk97cm35sKj6eUVpztsk4PTc31od980ZxIe1MBEEQBEE+IVA6IyWQFRUUi5/l3nMXnf1OKA3R3mhOVmluDrl3zxVlPYVfILRPEQRBEAT5VEDpjPCRFaSEiv03pB+eUOm/of3RzCr90PjsEOeijETaqQiCIAiCfCqgdEZYZMWFRVnPpCGumRd+TD8Euhn/HPBdzCrj2FTJ1W0FT+/S/kUQBEEQ5JMApTPCFGen5cVdFQduzjgxM83FVkkIopXFMo5Nk1zZWvA0ivYygiAIgiCVH5TOnzdFBYWvHuVEnsq6uCzN2UYg/tDe0TKOfgnqmX1yo6iAdjiCIAiCIJUZlM6fLTJZQXbhs2jp9Z0Zx6cLNB9aeVnGkUnZwU5FmakyfOcGgiAIglR+UDp/lsiKZfnZBanhmReXpbnaCdQeWjmbk01O1Jki8XN83zOCIAiCVHZQOn+OFOeIcu97ZpycKX9IA/8i8L1b+uFJufc9oNvpAUAQBEEQpHKC0vkzQ1Zc+OqR9L/9ct2Mr5/7UOZkLTozH36uyPIk9EAgCIIgCFIJQen8GSEryMlPvCW5vCXj+FdCbYf2vs3ZJstzSd5Df0aGn0pBEARBkMoKSufPA5msODsj72FAls9a9nMnAlWH9kEs3dVe7Le+8GUsPSgIgiAIglQ2UDp/BhQXFWen5d73Ep1biK9t/riWcXy69OZe9rEN/JNBBEEQBKmEoHT+1AHdLHmZe98z/cgU/IvAimAZJ2YWJN6SFeTQA4QgCIIgSOUBpfMnTlHW05zwY+lujvht7YpiLuNEZ78rykjGG88IgiAIUulA6fwpU/jqsfTWQfkXT1A3Vxhzskp3tZPKv5NCjxOCIAiCIJUElM6fLIXP70lv7GZfQieQbmgVwDL+/TrvoZ8sN4seLQRBEARBKgMonT9FiosKnkVLrm3POIEvoau4JvbfUPDkDiOT0aOGIAiCIEiFB6Xzp4asILfwRYw48M/0I5MEWg2tQln6oQnZYcfwxjOCIAiCVCJQOn9SyIryC1/EZnmtTHcZJxBqaBXQxD6rC1JC6cFDEARBEKTCg9L5kwJ0szhgI35hu7JY+pFJ0lsHZMX4fUEEQRAEqRygdP50KHz1UHpzD34ssDKZ09isi8sLnkTSQ4ggCIIgSMUGpfMnQpEoJTvEOePEDKE4Q6vYlnFsmvTmPrzxjCAIgiCVApTOnwCy4uz0nIgTotPzBLIMrRKYs03m+R8KXz2WFRXQ44kgCIIgSEUFpXMlR1Ysy5fkxlwSnZmP3wuspJZxZHJ2iEtxbiY9pgiCIAiCVFRQOldqZLJ8aUFKWMbJr9Oc8E8DK605j804PqNIlISf5kYQBEGQCg5K50qMrDC/4Old0b+z05xthGoMrXKZs03eI38Z3nhGEARBkIoNSudKi0xW8DQqy3u1XDfjoxqV3JysJVe2FKYn0IOLIAiCIEiFBKVzZaUoI0ka7Jx+CF9F92mYVcbJWfkptxkZvmoDQRAEQSouKJ0rJbL87Nx7HqKz3ykpMLRKa05jc6POFmen02OMIAiCIEjFA6VzZUT+qIbvujQna6H8QqvMJgn8o/BZND3ICIIgCIJUPFA6VzpksjyJ5Nr29COTBcILrbJbxr9f5z7wwfdsIAiCIEiFBaVzZaO4KPe+l+j0XPzTwE/QnKyyg52Kc0T0WCMIgiAIUsFA6VyZkBUVFGYkZV74Od3VTqi60D4JEwdsxGc2EARBEKTCgtK5EiErlryQ3Nidfmi8QG+hfTKW6f5z3kN/esARBEEQBKlgoHSuNBTnZuY9Csw4Ng0/HPgJW8bx6dlhR+ghRxAEQRCkgoHSuZIgKy54Fp3lt06gtNA+NXMZJ7nyt6wwjx53BEEQBEEqEiidKweyPHFu9AXQVUKlhfbJmdhndVFmKj3wCIIgCIJUJFA6Vw4KUsPFvusFGgvtk7TMCz/lJ96kBx5BEARBkIoESudKgKwwJzv8GL7I+TMx0b/fZEf8S489giAIgiAVCZTOlYCCJ3eyfNbitwM/E0s/PEl8eQs99giCIAiCVCRQOld0ZIV50hCXjBMzBAIL7ZM1Z5tMzyVMcREdAQiCIAiCVBhQOldsZLLClw9BSOEt58/KRGe+Lc7Nwi9yIwiCIEhFA6VzhUZWlJ99yynj+HSBtEL7tE10ak5hWpysqICOAwRBEARBKgYonSswxYXF4uegovCW8+dmopOz8pOCZYW5dCQgCIIgCFIxQOlccSnOycgJO5p+eKJAV6F98pZxfHru/YuyfCkdCgiCIAiCVAxQOldUZLLCjETR2QVpzrYCXYX2yVvG0SnS227FeVl0MCAIgiAIUjFA6VxBkeVJ8uKuyj8faCXQVWifvKUfniC+vKU4J4OOBgRBEARBKgYonSsoRRmJ0ht7BIoK7TOxdFf7rIvLiqVpdDQgCIIgCFIxQOlcIZEV5yeFiM7MFygqtM/FnG0zTs0tlryg4wFBEARBkIoBSueKSHGeOOfu2TQnfFTjczWnsRnHphaJn9MBgSAIgiBIxQClc0Wk4Fm0JPBPoZxC+3zMyTr98KQi8TM6IBAEQRAEqRigdK54yGS59y9m/DtbKKfQPiOzSndzLMp6SocEgiAIgiAVA5TOFY5iaZr05l75uzUEcgrtM7J0l3FFmU/omEAQBEEQpGKA0rnCkZ8cknVppUBIoX1ulu5iWyRKZmQyOiwQBEEQBKkAoHSuaMiyQ49mnJwlEFJon5252BamJzCyYjouEARBEASpAKB0rljIigrFARvTnG2EQgrtczNn28KXD5niQjoyEARBEASpAFRQ6SxjGEleYUJ6dnCiyOPu84P/Je+4mrDjSsKuq4l7rifuu5F44GaS861k1+DkQyEpR28/OXPn2fX49Li0bHFeoazy/hu3TFaYHpfluQS/IIgGP58KX9yXFRXQsYEgCIIgSAWgokjn3IKi1Mzc8JQs/wev/o14AuJ4s//jNV4PFl+4/92puzOP3pl2OAJs+uGIr45EzDgSAZ5Zx+58fezO7GOR3xyPnHcyatG5e6u9HmwOeLz/RtLZyGdXHqfdfyZ5Ic7LK6w8/+QtK8594Cs6M0+ootA+QwPp/BylM4IgCIJULD6+dH4lzY9IzTof9WzX1YT13g9/OX/vmxORE13CbA/cttkf8rY27sDtSa5hoKSXud//0+8xSPCT4U98Y19GPslKz84vKq7YN6SLiyTXd2Uc/VKootA+Q3OxLXz1GIYEHRsIgiAIglQAPo50lsmY3MLip1m5kU/EZyOfgWL+6kiEQASXlzk4hc4+Hvm776Nzkc/CkjOTM3IkeYUVUULLZLJ8aabHL+n4Wjo08nI6UQq+YQNBEARBKhQfWjoXy2Q5BUUvJfmRT7JcgpPnnIwCaSsQu+/P5pyI2nk14Xpc+pPM3KzcwoIiaA5t2EdHVlRQlB6fceIrgYRCE5qTVZrLuHRX+3QXW4WTeOzSXN7055VO1lAKysr/ELMCP1DuZJV+aAJ+TRBBEARBKhofWjpn5hQEPUr7zefhZLfwcQdu2yqp2/dqsDnY6Hjn0IWnow+FpDx6Jc2vME9CF4OYv++ZcWSyUEWhCczNXuT+Y5bPr6Jz31GPi63owg+ZPqtEZ9/wmHj6kYmZbOTqjNPfQCnB2gpkzmPTT8woEj+ngwNBEARBkIrBh5POuQVFVx6lr730cPbxyImuZXyUuVwMBLSDU+i0wxHzTkZt9H3kGf08JSPno7+Xo1jyQhL4V7qbg1BFoZW0jH9n5sZ6F6Y9zrlzgnjSj0yGXx2FL2Ozww5xYSot0/2nHIhMi5MG74dSxAl9nnF6DljF+YJjOvwYOPNtseQlHRwIgiAIglQMPoR0zissfvwq2y04ZdG5exNdwgRC9uPaJNewBafu/hXw2Dvm5UtJPm3xR0BWmJGQ8e/X+EbnN1rGqdn5iTeKc0Qgl6nn2Jf58deKc9Jzo89xYSot03NJXvwVKJsdfiTj6BTqdF+Un3IbLOPf2RXkKY50V7tMj1+Kpa/o6EAQBEEQpGLw3qVzRnbBzfiMfy7Hzzh6x/aAULlWEANB/9PZe863Um4nibJyCz/8E9Cywrz81DD5IwT4Ruc3WPqRieKgTdkhLmLftcRTeumccXKGOPAPKJvltTTdzZ44xX6/kaOQef579kHqkkU+iqW7OWR5/1qcnUYahiAIgiBIBeH9Suc0ab5f7KtfL8baHfxoj2eU3sa7hEFTL9178fhVtjT/g74UrDgnIzfGS6CfPkFzHsveVneyFvqVjf1bQFvWSqFlSy+dVVpFlM6HxksC/4RRQRqGIAiCIEgF4X1J52KZLCu30CP6+fdnogUKtYLbBJewrYHxwYkiUU7BB3uJXVFmqjTYWaCfPhWzArmccXRKxslZovMLRRd+EJ2Zl3FiRsaRSaySFgbLnzw+Nk10em6mx8+Z7j9DBhbZR8A5Ues8Nv3IpIzj09IPTyAetdLZ2Sb98ER2WydmsG/VAMnuYkvLHnJk37Yh35bk2g5yFLJ812acnAmytUQlH8OgkdIbe4pzM0nDEARBEASpILwv6ZydX+gWnDLz6J0P/A6NcjHbAyFfH7uz70ZiiiiX7s97pvDlA3HA7wL99ImYi03Gv1/nRLsXiZ+z38YrLpTlS9k/8gs/mnFiujDY2VocsCE/8WZxjoj9GkhxkawgGxaz/NenudqRGCiVfedkfvItafAB6lEjnUWnZmffdilMjyt89SDTexXIZdG577IjjuYnBUuu/Q2qWhz4R178tSLJC3IUisRPCtMeSa/v5FfyUSzj2NTsyNOyPAlpGIIgCIIgFYT3Ip1fiPMOh7C6uVI8p6HSoOVfHgpfdP7+qYinadL89/30c0FqhOjc9wL99CmY81jRhZ9g74pzM0ENF76MzU8NLXgZC/ninIy8uKBMj0WKYKs013HZ4UdBVYNkhLUFz+8XPLtbnJ0hyxMXvLgn/W8vufGccWp2XvzVYumr3HvupKxK6Zx59tvcqLMQVpT1VHx1S/rRKWlO1pmei/MeB4IzO+xQxtEp4sCNefFXuHfAFWWmwG8YyfUdXCUfyzJOzsxLuCYryCENQxAEQRCkglD+0vm5OO981PNvjkeO+3ivnysvs3cKnXcyau/1xNgXkvyi9/YGaJksP/56+uGJAv1UQezxNpMHfxsn7zIT+EtjIAGzb7uCFC5Mjwftm+W1HLRy1qWVkhu7i7KeFGenSa/vJH+ul+46Lst/PShXUNV58dfFgX9mXVyWdXGJ5MqWgqeRUEN+UrDo9DyQvxmnv8lPugV6OjeWPh2uLJ1F577Ljb4AmyjKSJDe2pdxbCp5OCTz4hJWkuaJsyOOQamME19levySE36cHAfpzb1Z3qtE/35NKvmIJjr1TWHaI/YmPYIgCIIgFYlyls7p2QWX7r/86dw9gQat1DbtcMS2y/GhyZnZ7+dvB2UFObn3L5bqj+c+lD3fb/HwH2Of5UN2zur5s1X770y/2DGjZ+Sm0YKwN1qm+895cUGyovy8+KsgB6nfySr96OTsqNMFz6JAT2ccn8oK4uPT8hNvyPIl+SmhWX7ruZf0gbAG8c2KYFGS5PIWUMBvkM7ONqKz83LunYciha8eSkOcoGbuvSUC6UycFe/PBK1EZ78rzstiZBXlez0IgiAIghDKUzrnFRZffZy+8mLs2Mp/v1lgE13C/vB79F9CRlZuId3b8qNI/Dw79LCSfvoI9uqA5aNtJjfWjTz904A/p3SdNLR5h6bV9XW1qlT5X9+2dXbN7CmIf6Nlev6Sl3BdVlSQnxqa5bWc/cM+8gsBBO75BZIrf2VdWg4SNt11XKbHz8W5WUxxYXaIS/qRSSUq8ViUE3U6L+6K5ApIZxtN0jnmoujstznRZ4pzM4oyErKpbuZVVSmks7NNpucSpuJ8Ix5BEARBEAXlKZ0fvZL+7vvoE3hOQ6XZHby9+MK9q3FpEviJUK6yhn2+9uo/Qv30Ae3FfovkXWYxW8cEbxh14JveX41q1blZDS2tKv/jUV1Py7pvk5cHLARlNZvozLycqDOyonymuDj3ga/YfwP7eg32BRfj5feVqU4FrQwyWlaYK8uTSIL+4opTc7FJP/YlKOb0Y1NB2qqVzrmigtTbOXfPMMXscw6gtkWnZr+uRG6VQjqzf78YtIk0CUEQBEGQCkW5SeeiYtmuawnTD0cIFOenZLYHbs88dsfj7nNxXnnee85PCs66tEqgnz6kPfrH5MxPA2cbtW7dyEBft2rVKlVKqGYFA9vVjd9h8uqAsLgmc7bJvLi0MO0xIytib6MWFxVLX+Un3ZTc3JXx7yz5J2DYsIzj06WhrqCwC9MfZ/mseV1clamTzuzrOECjKx5yyE+4nuW1jCtFrFJI54wTX0mDnUiTEARBEASpUJSPdM4vLA58mLbw1N3K+0qNUhrs4IwjEc7/JSdn5JTXnee8B36is98K9NOHseDfRv1q32l453qtG1SrV0NHR1uNapbTs1WtK6uHP9/3VjeerdLd7EVn50uubct7HFiU9VRWmCcryC7OTivKTMmLv5p5aQUIaPnb686zz3Ukh2R6LlaqpISplc6gzYsLi6Qvc6LPFedkFOeJcyJPpR+dzBUEqxTSWXR6Xk60O2kSgiAIgiAVinKQzoXFstTM3OUesRNcwgRC81O1r4/d2XU1IeqpuFy+mZJz93zGsRKP5H4wWz+hc+82tQ10q1J1zKN2NR3Drg0mD23erJ4+0dMdm9Y4/F3fJ3vMBZW82ZytM45OyTy3MMtnteT69pyo0/mpYbJ8aXFuVu4DX9DKGSdmgJaVFeXnPwnLvLhUWBzMyRoqIc9Jq5POoJvZR1+ubBWdnsu+gS4no/DVA5DsfDVcKaRzpvvP+UnBpEkIgiAIglQoykE6i3LYt2pMcg0X6MtP2746ErE1KC40OfOd1bMsJ/w4+607JQn1Xu3lAYs7m0ab925cTU9LLpVLUF1fa3jnekcX9PNZPmTcgCb1auiAs1UDg42TuiTvfou31GUcm5rluTjT/Sf64UAnK/YDfqe+yfJbn/fQrzg3s0iUJL2xM/3oFMmNHbLCvEJRYpbfOn4NYBknZ4mD/pTe2p/ptZx9F4c66ZwvgTrJtwCzfFcXPI2CmPzEG+wdfYUgrgzS2SrLZ01RZippEoIgCIIgFYp3lc4yGROXlr3a64H9p/6ohrJNdg370+9RzHPJO/3RYFFBdogz9wdzH8ye7jXfMbNnhybViVYW0Ll5jd8mdnkpf6x5w8QuXZrXBGezuvpLbdon7TLl16PZsi4uzXvgnRt7KePEV1Q9yy3ddZw44HcQiOzXSUKc0t3sM72WFeeKivMk0pt7S3wK28lacmVL4YuYoqwnklv73vCGDcV7ndlX2oUfK5Y8L5K8yL5zQv7ZbbaH3yCdL/xAXwDy8Szd1U5ydRv70DaCIAiCIBWPd5XOOQVFN+LSHZ1DK+MHt9/dph4K3xzwOE2aX1TWe8/FOSLpjd0C/fS+DTTxo23GZr0b1qqmTbQyHx2tqo6DmoX/YUiCnef1GdS+Lvhb1jdYN77zW30bRez/W2F6nCxfKrm6NeP49DQXG1abuthksLeZdxaJnxWmJ0iubYdfDiBk2U+fFOTkPQ7K9Fqe7uaY7jIOJG/GyVl5D/1lBdkFz+9levys4a6z8GuC7j/lPfIHDVokfprpuQgkKetUKZ1918mK2L/7zPL5lf0wjeKV0h/FMk7MzFZ8ogVBEARBkIrGu0rn+LTsvdcTBYLys7IpbuG7riVkZJfxw29FomT2Sx9KEuq9GshfzyVDGtbSJVpZQNtG1UAic8Fnfx5g2K0B+Ns1rr5nds/UPW8hnUVnv8u578HIioskL6Q394hOz0k/Ogm0r+TqtqKsJ0xxQd7jQPItbpC27H3orFRZcWF+Sqg48A/R+e8lV7fkJ4cU52YV52Tk3ndPc2HvW5dSOqc5WUmubGUrLMyDCsk3AlVK56yLy4pEKQwjy4sLklzfJjqt+HTLx7Asj8X5jy/TwYEgCIIgSAXjnaSzjGGuxqXNOxkpUJOflY07cPtLt/DTEU9fiMvyj+yFz++L/TYI9NP7trA/DKePbFld1VPOgMOgZpeWDeGCD8zpPaBdHfB3blbz9E8Dn+59iz8TZL914rW04Old0K/F2elFmanwU0H+nEYaU5yf/yQ8y28dfc7bySr9kCMo18KXMexfEEpfFmU9LZa+khXkFEuegyYWnZlH6iytdAb/v7Oyw4+wL60ryJHc3J1xbJpK6Sw6Mz/3oQ/oe9hukeS59L99XA0f3sSBmwpfxNLBgSAIgiBIBeOdpPNzcZ5bcIqDU6hATX6GNu/fu76xL0U5b33vOT8pJOvSSoF+eq+Wusf87M+DWtQz0Kqq4kV09WrorJ/QJW67CRe/cXKXri3YZ517tqp9dc2It3w5nWX6ofFZF5dlhx3Ke+RXkBpW8PxuwZOwvEcB2aGu7PcFS344kP2zQt81sCo39lJe4o28R/45kf9KLm8WnZ7LvQSa/WJI4B/SEKcs39XU42YvDvhdGrxf2JMutqKz32YHO2WHHc7yWZV+dHLGia/EARvYsl5LoBQJS3dzzHT/KTv8aO59j5zoc1neH/RwlDSr7GDn4txMOjgQBEEQBKlgvJN0/i8hY9XFWIGI/GxtrfeD20mit33oGURkpvvPShLqPVroRsMlNh3Uvb/ZqHuDMz8PfMWL/9mqXasGBrCq3xd1YrYav+0HBVlzsgbNCvIUZLHYf4PYd02mxyL2E9mqnipOdx0Hq0Tnvwdhnenxc8ap2eluDiVinMeC4GY/SXh4AvU4WVMP/+8LibnYZhz7EraecWxKuss4WEw/MlEe6VjiLwKdbTJOzhSd+050dn76kRKvgv6Qln54Uk70eaa4iA4OBEEQBEEqGGWXzjkFRYdDUj7tzwe+lU10DTtwM+lJZi7toNKRE31BdObDfQ/lxX6L49/3Jw9gKKOtVWWNY6c7m0bzi0wb0aJeDR2tqlWGd67/dK8FX1Wjla+Jzv+Qx37bBUEQBEGQCkrZpXNCevbvvg8F8vEzt4Wn716Iep5fRL8FXRqyw49mnJwlkFDvz2L/Nv7VvpOulop7zlWr/K9BLd1ziwaBvObiXx6wsOjTSE+najU9Lau+jTk/2vswSdBfhc/v05GBIAiCIEjFo+zS+fKjtF/O3xdox8/dDtxecTH20Ssp7aNSIA12Sj82VSCh3p/9+8MAs56NqFguiY5WVbNeja6uHcEFvzpgGb/DZEjHerC2fk3dqSNacqvQ3ofl3GU/IU5HBoIgCIIgFY+yS+dTEU/nn4wSasfP3qYejth2Jb641E88S//b98Eern2+z2K5bceGNVW/k85AV2uNQ+e7m424+Kd7zY8s6Ee+h9K6ocEi63bcKrRyt4xjU9nvb+ODzgiCIAhSgSm7dN59LXEaPuisZHYHb3936m7kk6y8wlI9tiG9vlPFH7e9H/tv/UjHQc1UvlijSpX/1amu47lkMP8720m7zL4e3bpRbT0IGNyhrtv8vtwqtHI2Jyux/4bCtMd0WCAIgiAIUiEpo3QuLJat93mIr6VTaVPcwrdfiU/PLtVrniVX/qYvNn7/tufrXv2/UP0Hgvo6VXu3qR252Yh7gcbz/RZ3/hzdq1UtWKWjVWX84Obhf5T480G08jQn69x7F4qz0+iwQBAEQRCkQlJG6ZwmLVh8AR90Vm12B29/dSQi6mlWbiluPIuDNnFvLH6v9nyfxZwxbRrLbyEr07CW7myjVvzXOcfvMD26oH/dGjqwtmkdveW2Hfl/PohWniZ/uV7hi/uyorJ8VQdBEARBkA9GGaVzZGrW92eiBZIRjTPb/SGuwcmleFGdTOy/ocQLht+PvTxgeXeL0cgu9bXUvM+5XePqznN7p/Ke1gj7w/Dr0a2qyb84aNi1weHv+nGr0MrX0l3tMy8uK5a+ooMCQRAEQZCKShml88V7L+bh3whqtK+PR4YlZ2r+a0FZUaHYd51ASL0Pe7bPfP+cXp2b1SBCWZk+bWvf3zLmJe++sv+qYW0b0i8OzjdpE7ZxFLcKrXwt49jUnMgzsjwJHRYIgiAIglRUyiid911Pmnn0jkAsovHN7uDtfyOevJTk0S5TRXFBdpbPGoGQeh+Wstvsy+HNG6h5t0bd6joTBjd7ztPNj7aZ7JzZS0ebFc4Naulund4dxDe3Fq18TXRqTlF6AlP01l9xRxAEQRDkA1NG6bzKM3aKW7hALKIJ7Hffh5GpWbTLVFGcI8ryXiUQUuVuz/exf/DXrUVNXe2qVCyXpHOzGr9N6MIv4rN8qP2gZmTt6G7sp7n5a9HK0dIPjRf7bZAV5jGyt/uEO4IgCIIgH54ySufv/r073hlfr/EGm30s8tK9F0Xq3/FcLH2ZdWmlQEuVu8VtN3GZ11fdLWdgRJf6Z3ni+PE2k01fdmteT5+sXTy2w+3fDbm1aOVrojPzc+950gGBIAiCIEjFpozSec6JSEeUzm+ycQduH7yZlCZV99oEWZHkRdalFQItVe4W/ofh1BEtqsv/4E+ZqlX+ZzegWeSm119C8Vgy2G5AU/mqKs3q6h9d0D951+s/H3xHe3WANYHzfRhsJXWPeeJO0yd7zN9qi8/2WSTvNkvZbcZ/8vuN9vKARVneQOJsm+m7viA9kY6IMlFcXFxQUFBYWEiX3wTEQ7AMb3IjCIIgyNtTRuk8/98ovOtcGlvj9eB2UibtNSWKJa/e911n0HN+K4a2bWSgraX63RrVdLXmjGnzYh+VfaAal9p0aFpHD6Kr6WnNNGwV8nvZ/0AQBCXo19i/je/8Ofr276P++23k1TUjrqweEbxhFAj66L+MknaavpVC1WzP9pk/3m4C1Yb9YXh97cgLvww+urC/++LB19aOuLNpdMIOU83qNnWP2d3NRn4rh536caDnkiHwg+GlRs3NqvPdZrB3EX+Ovrp2ROCvw29vNIzZOob/1LiyQSnQ5dAbN9aNvL7R4s7pPzIy0umAKAUgfPPy8rKysl6+fJmSkvL48ePIyMjY2Ni0NE3vhAatLBKJUlNTHz16BPHBwcH37t17+vRpTk4OjUAQBEEQpBSUUTovOH13ggtK5zfb7OORJ8Ke0F5Tojg7PevS+33W+fE2k71f96YyWRVtGlVbO74zF++5eLBxj4bg16papWUDgxvrRjzbW/Y/EEzaZRawatgSmw6mPRt2a1mjQU0dXa0qejpVW9TXH9KhzvSRLU7+MCB+h6mgVNkMZHHEH6O3Tu/21cgWgzrUbVBLl7wepEqV/xnoVLXs3chlXh/QtYJSnL06aOm/cuj0ES2b1tGHIu2bVI/5e8wbpPZuM98VQ5bYtDfs2qBRbb3q+lo9Wtf6yfKL+1vHQG2CYM5AN/uuGNqteQ340QItHDJkyMWLF+mA0IhMjlQqvX///tmzZ3///fepU6f27du3Tp06M2bMCA8Pp3GqAJV8+PDh6dOn9+nTx8DAoGrVqi1atFiwYMGtW7doBIIgCIIgpaCM0vmHM9EonUtj9k6hf1+OL1DzuPMH+DPBy6uHzxjVUi6SVTO0Yz2nub1JMCjFOWNaN5F/NqVFPYPVDp3id5iU7fmK278bbprSzaJ3o7YNq9WvqVtdTwsUMyhFUKVgOlpVDHS1alfTadOwmuOgZoe/7atZpGo2KBuwaugv1u37fVGncW292tW0oXKimwmwxRr62qDXvx7dOujXYYLiybvMAlcNm23UukvzmtAkba0qkFr3afx0r7kGBfx4u8mR7/p1bFK9QU3danpa2lWrwM7BPtavoXtkQX91vwegM0M3jjbp0bCmPvv6El1dnYkTJ2q+YQzEx8fv37//xx9/HDt2LGjfdu3agfBt2LBh7dq19eWsW7dOLBbTaB7FxcWZmZkHDhwwNTVt2bIliQfdDH2ira1dt25dKyurc+fO0WgEQRAEQd5EGaXzj2fvTXQJE8hENJW29tKDJ1mqv40iy8sSe/8qUFfla8e/79+jVS2iIFXiMKiZ97KhEPl8n8WFRQMHtKsDuramgbZF78ZhGw3fVtG+PGCZsMP04Jze9gObdWpWo5aBturHRBSA3AStadyjIcj3Mrz/LnWPuc/yoYtt2kMNrRsYgHKl9cqBTROlTher/K9ZXf2vRra6uX4k+T2QstvMe/mQJTYdRnSu36SOHvdMS9tG1VbZdeS+Sa5scdtNXOf3HdyxLl+gE2D5e4sv1D3lcm+L0Zap3UCaV5UXBBG8YcOGoqIiOiBUkZGR4erq2r17d5DLtWrVAslLNsTRqVOnQ4cOgUqmBRS8evXKw8Nj+vTpPXv2rFmzJvurRYl69epNmjTp+fPnmtuAIAiCIAihjNJ50fn7E11ROpfKFl+4fztZ9ePOsnxJls9qgboqRwNl+ff0HjX0hWKLz0LzL+5uNgKJfH/rmGkjWzSspQsKq0+b2jtn9nzb+82gRK+sGb7SruPQjvVAfEPloA9BPfdqXXvikObzTdo6DmrWqWkNZQkHMWO6NwxcNQxqENSpzqDBEX+O3jqt+7j+TVspiebGdfSMujf82qj1T1btF5h/YdO/SYNaukSttm5YbalNh6Rdpn4rh/4xuavdwKYglAUCuP8Xdc4tGqhu92HTJ3/ob967kY6a3wUOg5oGrRbe2wZ7utf89E8Dh3euR+P+9z8LCwsvLy86GtQQEhIya9YsWkAVDg4Oly9fptEKEhISyM3mGjVqkNvMKgE93aVLF1DY+NAzgiAIgpSGMkrnJRfuT0LpXDr79t+7F+4+px1XEtl7/iRK1GajRdbtqUpSRXU9rd8ndX2y1/zBP8a7Z/VqUZ99jKBBTd25xm2i/3r9zo3SGHsHd9mQ+SZtGtfWrSpXx/Bfi/oGDoOagcC9unr4g7+NQRxDQN3qOmTrfEBqL7PpUMqNQoOvrxuxxrFT5+Y1yLY4dLWrdmtRc45xm2ML+0duNkrZYxa3w8Rr2eBJQ5s3qcM+iAIiu0vzGtu/6jF1REsQzcp/PQk1WPVprOGpaOjVb03bVtNTq0cnDW0GPyEEpUCIX183coFZW12FytfV1f3xxx+TkpLoaFBFYWHhwYMHe/ToQYooA9p37dq1cXFxtID8kejHjx//888/Q4YMUXmnWUCzZs1WrlyZman2j1kRBEEQBOEoo3Re7hEzGaVz6eyrIxG7r6l++5isMFf8PqXzpWVDJgymXzZRSYcm1Q980zt5l9mZnwZ2b1lTR6uKVtUqpj0bHv++v6AqzfZ0r7n/yqEzDVtV030tKOtU0/lqZKvg30byI48t7Nf/izo0ggcI1r5tal1bO4IfrNJS95hfWT0CNChodFpYAcjizs1qbJ/RA34JCB5Tvrx6hFmvRkQog57Ug58IalRlo9p6oIz5ZfkG1brM6zOs0+s7xwKg1h8t291WemDj/hbjVXad2jQ0oHH/+1/Lli337NkDSpeOBlW8ePHi+++/B5FNyyhRrVq1U6dO5ebSJ4KKioqSk5NBTHfp0oVGvIkGDRrMmDEjPf0t3vKBIAiCIJ8tZZTOqy7i1wRLaw5OoSs9YmnHlURWlC/2XSvQWOVoB+f0HtxehU7lAJV8btFAn+VD7AeyL3IGalfX/mNK15Q9b/ci51u/jZw4pLkBTzeDgrQb0Mxz8WBBpN+KoZOGNqdBPKpW+V9tA23fFexT1xoMlGvgr8PG9m9ioOqmb8emNf6e3v254kV7AvvFun1T+Y1nzfRrW2fHjB6CspxB5TMNW9eSP46iEtjxTVO63St5+/zFAct147t0alqDBskZN25cQEAAHQpqcHd3NzY2pgWUqFq1ar9+/W7evEmCi4uLQQF//fXXjRs3phGloH79+tOnT0fpjCAIgiCloYzSeb33wy8PoXQuldnuD1lw6m6GNF/FZwVlxWLf9XyNVb62bnznZnXpRwFVMte4jev8vj9YtuO04OzRrS+vHl76p5xf7LcI/8PQYWDT+jV0+fdxh3So6za/r/KzyxF/jF4xriMNKknVKlWOL+yfqv5xZ6jt2MJ+wzvXY//6kLctQvsm1Zbato/bbqLutRhbp3XvqfEvJgm2/Zv6rVDxpDIx72VDRnSur+6WNXi1tapcWDwolffbAzoTNt27TW3Bh9B///335ORkOhLUsHjx4ubNVfzSIOjo6Hz33XcPHjyAyMLCwujoaAcHh4YNGyr/KaEGoP5169ZlZWn6YjyCIAiCIIQySuf9N5NmHr0j0Iho6uyb45F3UrPyCoXvQADtLA78M83ZhpNZ5WiJO03nGbfVKynXBCyz7fidWdsO8ruhOlpVxnRvcO7ngUml/nbgywOWj/4x+WZMKxDofDVZtWqVDRO6RG1W8eBy2EbDZTYdaJwSp34c+FTNa6ShVZ5LBw/tWK+Gvoq/zqumq/Xl8BY31mt63uPgnN5QnBZQA6jbBWZfgP4WlOVs18we3VvWpNFK6OtU7dm6ZvCGUeTtHM/3sX/L+NuELn3a1BZ8zbFZs2Znz57Ny8ujA0EJmUwGmtjCwkLD0xp6enrHjh1LS0srKioKCQmZPXt2vXr1qlatCtK5R48e06dPB0+jRo20tFR/SJLQvn37I0eOZGdn0w0jCIIgCKKeMkpn75iX805GCQQimjr76kjEuchnkjwVn0qWXN2W5ubAF2flZTfWjbQboOlBZz2dquMGNO3Xto62VhXQfN1b1XSZ1/fRNrWqUdke/G28c2bP1g0MQHbTSv/3P+2qVTo0qX5xyWCVD05cWT3869GtaCgPEN419bW8lg5R+QG/Z/vMA1cNmzi0meDGLceQDvUOzOmj7lENYnu/7jWwfV1aQA2tG1bb/GU3DTfdl9q0b9OwGo1Wom51nXnGre7L/8QwdqvxyR8GzDFu07lZDUGzq1SpYmtrGxoaSgeBKkANHzx4sGvXrrSMEqCPW7RocefOHdDfoJu///77pk2bQs01a9a0s7Pbu3fvf//95+vrO2LECFDYtIwqunfvDpH5+eo+F48gCIIgyGvKKJ1jXkh+PBstEIho6myyW9jfQXEZ2QW0+3hIb+5LPzKZL87Kyw59229Yp/pUHykBUtdAt2qTOno19bVB+HZsWn3dhM4JO0w1fAREYEm7TM/8NHBIh7o6WiV0oYEu+/nusI2GgnhiF34ZZN67EQ3lAeJyUPs611X9mSAI2ZANo5badKhpoPpj4rWr6awY1/HOn6MFBQX2x+SuXZqXeNpYGeMeDUHvCgrybcaoVo3ln4xRCfTn1mndPJcMPv3TwN8ndjXv1aiavoo7vlpaWv/8809qaiodBKoAQTx16tQGDRrQMkqARDY3N3/69Ondu3cXLVrUqlUrqLZ+/fqTJ0/28/Mjb8yIiYkxMjLS11f70A5I7YEDBz579gzf64wgCIIgpaGM0lmcV7jcI8b2gFAjoqk0B6fQ789GPxer+Nd5aYhrxvHpAn1WLrbaoVOHptWpRFJP1Sr/a1lff65x66TdZqV/xBkir64Z/o1Ra1qLgipV/tegpu6RBf3id6i4e/1iv4XzvD7dlR44hlL1auiuG9/p/hYVr4R7vM1k05Ru7Rqrvdc7rGO984sGCUop2+Kx7TU/+Q18Z/rFf7+p/poJsakj2Fdf02glYN9/GdvecVCzLs1rVlfzOu2qVavWrVv31q1bGp7WACH7/Pnz7t27a3hquXnz5hs3boyOjgbd3KZNG9DNjRs3njBhwsOHDwsK2B9phYWFN27caNiwoYYHNkBVW1hYKH9OBUEQBEEQlZRROgObAh7jBwVLabb7QxydwpLSs5XfRJZ952TGv7MF+qxc7MvhLWpXe/Ofi1XTY58SVvlcsgZ7vs9i2/Qe9WoK39CsK3+tcuxWY5Uq/NE/JivtOiq/HENfp2rvNrXC/xyt/MTFq4OWznN7D1f/MjgAfiS88ZYz2KzRrWBnaRlVaFWtsu2rHvy/8FO2PyZ36VjyRRlvS40aNUxMTDT/gWBWVpa7u3uzZpqet+nUqVNAQMCCBQtatGgBi/Xq1QPd/OrVK+7+8cuXL52dnUmwOlq1arV48WISjyAIgiDIGym7dD4UkvL1MfxLwdKa3cHb95+JC4qEt/dyoy+IzswX6LN3NJCtD/8xMezagHw/TwNVq1RxHNTs4pIhmp8SVraAVcOmjmip/BnqRrX0QKEm7jQVxBPbPatn/3Z1BG+n0NWuMqRjXY/Fg5/uNVd+XOTGuhH2A5uqk7ywg/VraJOygoICS9lj5jComeb+AE184vsBmm+9B64aNqKz2sdgSkPTpk23bNmi+WVwqampP/zwQ506al8sqKur27Fjx0WLFoFu1tbWBt08a9asmJgY/nMXkZGR8+bNowXUMHTo0DNnztACCIIgCIK8ibJLZ5+Yl/i4c+lt3IHbYcmZOQXCJ0rzYn0yz/8g0GfvaC/2W3gtHdKnTW2qj9QzpluDQ9/2Vad01dnL/RZ/Tu4meEsxoWUDgzWOnVR+T/vK6uHjBzcTfBVcR7vKqK71nef1Sd5lpvIx698mdO7UTO1dXpDUDgObhqp5rppvEX+ONu/9hrcd2w1o6r9S7WvpiCXtNP1zclfoW+WfDRzwg6Rd4+oju9TvrXQIqlSp0qlTJxC1Gp7WAO7fv9+rVy8Nf94Hcrl27dpt2rSBTM2aNWfOnHnjxg3ynAaHt7d3v379aAFVaGlpTZw4UfPnDBEEQRAE4VN26XwnNevXi7ECgYimzkA6X4/PUH7JRt7jK1keiwX67B3t2T6LPbN6qpS2HKD8ujSvsWtmT5WPF2u2278bTh7WwkBXxZ1gkIw7Z/ZI3VPiHvCzfeYBq4bNNW7TuuS7KbS1qgzvVH/HzJ4qX+vx8oDFw3+MLXo3qqnmoWGgYS0957l9NLxLjjP3XwZp+AQg4Vf7jqCwBQWVLXTjqG1f9bDo3bgm78Mo8BugQU3d7i1rmvRoOGdMm63Tui+ybq/86wVkrpWVVXZ2tobHi6VS6aVLl6pXr161qtoXC4IEJ2tBXoP89fHxycnJoeXlQCX79+8HeU3iVdKsWbNVq1YJBDeCIAiCIBoou3R+lpW7yf+xQCCiqTOQzv4PXmXmCqVzfvLtrEurBOLsHe3pXvO1jp00vEMNdDPovF/tO73tI87EDs7pM7Cd6re8dW5W4/jCftzjEy/3Wzz4x/jUjwNmjGrZtI4e9+5naEDzevqgMvfM6qVOu6fuMTvxff/2TdT+paO+jla/tnXubjZ6sf/NT5v8NqGzhmeUoV26WlWOLexXyhvwsX8bb5navZHiS+At6hnAviwwa7t5Srd/fxgATYKO/cW6fb0awmfBv/jiizVr1tBjr4akpKTNmzfTAhrR1tY2MjLy8vISi8W0sIKHDx8uWrSIxqlh5MiRJ06coAUQBEEQBCkFZZfO+YXFe28k2TuFCjQimkoD6ewZ/SJd6f10hS8figM2CpTZO9qTvebfW3yh4W0S1XS1BrYD0Tmm9K/U4OzVQcsfLNq1rG9A6ypJz1a1fJYPfb7PAuRs/A6T27+P2je716D2dbUVr5WDTJ1qOl1b1PzWtM3l1cPVfTsQGgbydPKwFsrqkwN2cIHZF298yhmqStltZtmnsYavZ0OrmtfVv7Z2ZGk6BGLC/xi90r5TzWrsfXc9naozRrXiP+kBXXR8Yf/R3VS8V27YsGG+vr702KshIiJixowZtIB6yOPO7u7uIpGIluTh4eFhZmZGQ1VRpUqVOXPmREVF0QIIgiAIgpSCsktn4Fzks2+ORwo0IppKA+l8KuLpS4nwwxPF0jTp9Z2c6ioXS91jPm1ESw3vUKtTXcduQNO3+voJZ8/3W5r2bKTuI4V929YJ2WAIcjZm6xinub3HDWjC175Vq1RpUkdv4pDmAauGar5V/Gyf+fW1IxrW0tPwl4592tS+8MvgN95yhh8SfiuHabh7DdTQ17bt37Q0T2uApe42/31StxaKHw9dmtc49eMA/pdcnu2zmGXYqm51oejX0tKaMGHCGz95ffny5VGjRtEyagDh+8UXXxw6dEj5fjMgk8k2b96s4QvegL6+/tatWzU/co0gCIIgiIB3ks6RqVkbfR8JNCKaSht3IORwSMqzLCWlUlSQHeKa5jSWE17vbqm7zaz7Na5TTe392vo1dL8c0eLx20tn0KnX144EfSx4SwZHl+Y1nef2mTOmDcQ0raMPkpT7czrQ6w6DmrrN7xu9ZcyTPSpepsG3qM1Gaxw6qXycmmNE5/r31bwFj2+PthkvNGvbsKbaHxJA/Zq6v03s/OAfY0FZlbZjRs+B8vvo0AmQ/vNVD/L5QM5Ofj8AApT/jrBt27Zr165940uU3d3dW7VS8cFFPqCbV6xYASpcZW1PnjyZM2eOhndCAwMHDjx37pzy2xIRBEEQBNHAO0nnzJzCE2FP8JmN0pjtgZD9N5NSRLm0714jy4k6k35kCl97vaOl7DYb1qmehncYN6yl982Y1qX56zqBgeQ9OKd3hyZqHxoGrdyrdW0QzbrarHIE+diwlu7obg1+smx/cG6fy6uHPy7dRq+vHTGufxPBdwr56GhVsezT+Nmb3qkHWj/491FdmtfQ01FbFdC8noHP8qEqXwwisDM/DTDt2Yi8J6SWgfb4wc1u/TaKe7Uf6HjooslDmzdQpdSNjY3Pnz9PD7sa8vLyXF1dNXz/D6hRo8akSZPu3btHyyhx8eJF2BaNVsPcuXPDw8NpAQRBEARBSsc7Sedimex6XPqC03cFMhFN2UA6b7+SkJhR4jUIhNwHfqIz33Li7N0teZdZtxY1NejOJrX1frRoF7/j7d5JBwbicq1jp1YNVD/ozAGquXk9/RGd6880bP37xK4nvu8fsmFUyu433Gnmm++Kob3b1FK+cctRu5r21OEtBKWU7dE24wNzesOvCHW3yQFQ4T1a1Xq0zeSlxmc/nu41v/XbSMfBzYgs1tep2rdtbY/Fg/mCG3r+3KJBnZrWkP9wEPLNN988fPiQHnU1PH/+/I8//qAF1NCvX7+9e/cWFgr/5JRj7dq17dq1o9GqqFq16p49e16+fEkLIAiCIAhSOt5JOgNxadl7riUIZCKasoF03uT/GLqLdhyP/KTgLK+VnPx6R3t1APSiSesG1TSIxWZ19VfadUp4y9c5gyXtMlto/kWTOmrfNwyCskOT6lZ9Gi+ybu86v+/t3w3feGNY2V7stzjz88AmdXQ17ELL+gY/WbYTFBQYKPVra0dMHtpcgwQHauprG/dsqPmjMEQ3/2D5BXnwA9R2l+Y1Vjt04j9p/fIA+/boSUOa11b194g6Ojpr1qwRvEJOmTt37nz77be0jCqqVKkya9asyMhIWqAkMplMJBI5OjpWq6b2/SpQQ5MmTXx9fdWJb6ikqKgI1uLjHAiCIAgi4F2lsySvMOjhK0enUFslsYjGN5DO67wfPnwppR3Ho/D5fcnlLZwCe0cDqRq60bCp+tdrAC3qG/w2scvbfgnlyR7z/34b6TCoWb0aap8bblxb7xujVlGbVHxSu5T2fL9F5Caj3yd1BXlKK1VF1xY1N07qIigrsOTdZiDfm9fT0yDBgYa1dKeObKHhzw1hFXTpGsdO1fVZEQ7WqoEB/IR4+E+Jh0/gdwUo/lpsjIrtNWjQYPv27fSQqyEzM/PUqVMWFha0jCr09fU3bNiQna3iNxgAkvf69esDBw6k0arQ1tY2NzePiIigZUoCclkqlcbExGRkZGi4sY0gCIIgnyfvKp2BR6+k35+Jtjt4WyAW0fgGPy1WeMbGvpDQXuNRlJkqDXbii7B3sZTdZj4rhoKEpUJJFc3r6a926PRW0hnk4/W1IyYPb96lRQ0Nf73XoKbu9JEt31aUc/bqgOW9Lcar7Ds1Uv96EMKg9nUPfNNbUFxgAauGQWNoAfWADl48tr066fzqIPtV801TuusqXipSU19rgVnbqM3C13Hc+m3kt6ZtSYwy3bt3P3z4MD3kqgDVC7p56tSpmp+1aNGixcGDB2kZJfLz89etW9emTRsarQoQ32vXrk1MTKRlSgJyOTQ01NbWFrT1G++RIwiCIMjnRjlI55eS/CO3U8c74x8LajLbAyGrvR48UHXXWZYrzo06J9BhZTaQrad+GNCwlibpDML6O7O28TtK+2eCz/ZZnPlpkFUf9q0dbRsZaJDOoC9BW0f/NYb/srbSW8iGUXON27SoL7+7qxHDbg3+/XGAoDjfknaa/jaxSyONPyEInZrV2DK1+8sDqqVz7N/G68Z3ade4GmmRrnYV0MdX1wxXvq3uvnhQ/y/Ufr3P2NjY09OTHnIlQDfv379/8ODBoJtbt25Ny6iiV69eGr5jAmLXxMSkRg1NH5KsXr26j4+PyrfaQTMCAwNBNzdp0iQ4OBilM4IgCIIIKAfpnFdYfP+ZZPrhCFCHAr2Ixhl0zgbfh49fqZDOTHFR3qOANJdxAilWNkveZea1dLBmyVivhs7koc1L+XI60OJ7Z/c26dGobnUd7apVureoRd4voZIqVdgbz8cX9k94y79BfLbX3G/F0CnDWpRGNwNDO9Vznd9HUAnfDn/X16h7A/bhiSrsV0s0PLPRq3WtQ9/2VSmdo7eMWT6uY9cWNUExQw1Qy1cjW3kvH6L8JZeU3WZ7vu6l4asrlpaW3t7e9IjzkMlkL168+Pvvv/v37w+itlGjRi1atKBlVGFkZHTx4kVauCTFxcUvX77s2LGjlpb63za6up06dXr8+DGoZFpMARQHST158uR69erVrl371q1bKJ0RBEEQREA5SGeZjBFlF6zxejDRJUygF9E4A+n8V8DjeFV/JgjkJ93KOP6VQI2VzZ7tMw/ZYNi0jqZnnavpaQ3uUPf+ljd8TRDWPvzH+O/pPUZ2qV9TLgp1tKoO71RP+WMffPR1tMYPanZ93QgNTw8LLG67yakfBzgMatawpm4pZDNLtxY1/5jcVVAPZ74rhkwY0qx+TV0D3aptGxoM71xPg9zv3rKW05w+Aun8fL9F9F9Gy207dGlegzx1DT1g3bexx5LBKh9HCdtouMhK04MWffr0OXbsGD3eCkCtPnr06K+//urRo4eeHvtrp0GDBk2aNCFFVOLo6HjlyhVaviS5ubmRkZHNmjWjoaoAWTxlypRXr17RMgpAwQcEBEyfPr1x48YQhtIZQRAEQVRSDtIZyCssvnT/xdyTUfjHguoMpPP2K/EqX04HFDyNyvT4RaDGymbkDRttG1XTcKNVqyr7VT+f5UPVfQcbDCR45KbRf03t3rtNbX35S5FBQTarqz/TsFVzjX+DCNutV0MHdG3UZiNBncr26qAlKHjXeX2s+zTmHiauU11Hw1fECQ1r6c4Y1Qo0t+CFd0/3mv+3fuTXo1s1r8feve7YtPoPFm1/svpCw234LxpVX+PQmf9mutQ95jfXj1zj2LlNAwNt+X1r2CPz3o3OLxqYvEt1j11cPtJhqNoHnYE6deps3LiR/+d9+fn50dHR4OzSpQuJMTAwaN68uWbtO2fOnLCwMFpFSTIzM8+fP9+oUSMaqorWrVvv2rVL8LRGYWFhcHDwtGnTiGrX19fv27fv/fv3oYU0AkEQBEEQOeUjnYtlMlF2we8+j/DGszoD6bzvhspPorAUvnokDtosUGNltuTdZv2+qK35IyCghleM63jnz9Eqbzw/2WsevGHUxkldQWGTxyf0tKuCHJ8+ssXhb/t1a1GTVKIBaMDuWT0f/mOs4d7z830W97aM2TWz57CO9WgxuW4269Vo2ogWGh6xIPRoVevId/3gdwLU8/IAK5of/G3st3LYt6ZtyUfIofGzjVqHbTRc69hJwwv16tfQdRjY7O5mI9jrJ3vMH/xjDD8qFll3qCmXzbD39Wvqju3X5NyigereS53uMu7EcsvRA7vTGtUwefLk69evv3z58tWrV3FxcZBftGgR9yd9oJsHDRo0YsQIzc86L1++XN3LoV+8eLF169Z69V53pjKdOnW6dOkSdzu5uLhYIpFERUVNmTKlQYMGEADN6NWr15YtW1Q+DI0gCIIgnznlI50JQY/Sfzl/XyAZ0YiNOxDiFpzyNFO1dC7KTM0OcREIsjJbym4zy96NaldT+4gCAIK4c/MaZ34eBGKRXxaU9PP9FtfXjfjB8gsdbapeQUF2aFp9uW2H1N3m97eMGdS+7ptkLUuftrU3Tekau3UMqOeXByygZpCerB2whEXQu1GbjVbadezU7PXftOloVZk0pLnX0iGnfxqoWfoDEPxFY4MDc3pDPY+3m95YN3LT5G69Wtcit9u1taqMH9z82poR8TtM/5zcVcNtbOiKpnX1F1u3u7Z2+JU1wyF4eCeqPqGi6vpa00a2vLpmBL+XSpqV6Mx813VfD+jXj5RSR82aNUePHv3PP//s2rXL1ta2UaNG7A1tOVpaWr179/bw8Ni4caPmZ51B1D59+pSOm5IkJycvWLCgVq1aNFQVoNT379+flZUlk8lAN4M+DgoKMjU1Jd8vrFq1KjRj586dtEYEQRAEQUpSntJZklfo9F/yFLdwgWpEAwPpfDzsyXNxHu2skhRnp+fev6ikycpoT/aaf2fWVvOrnQFd7aojutRzntsnSfHw7sv9lldWj1hm02Fguzp1qulw932Hdqy3c2bPx9tMQPWCDp46vEX9GpoedyaA9m1aR29M9wZ/fdnt7E+DgjcYPt5m+uBv46trhh9d0O9bk7a9W9euX0OXPKcBm9LVqvqdaduAVcNS95hfXDr4i0YGRASrA9aBem5US69to2pfNK7esr5B/Zq6eoqnPmC7h7/r+3SveeJO0+0ze7Sor+kLiNpVq9SprtO8nkHzevpQib7iFSL1augsselwc/1IqIfrXoFlHJ+RG3vp4K6/O3XqREqpA4SpgYFBgwYNGjZsWL16de6P+SA/cuTIK1eugKLdvn275lfL7d27Nz09nY6bkjx69MjGxkbDx1AAkO8mJiZ+fn5xcXEBAQGLFi2CZkMRIuKHDh3q5OSUmZlJa0QQBEEQpCTlKZ1lMiYiNet330cC1YgGNu7A7bORz15JVD88KivMzU8OSTtoJZBlZTNQt+6LB/VpU1uj8mSpaaDdo1Ut2/5NZxu1nm3UCjKgkls3qMa9fk5bq4pN/yau8/rEbB3D1X/6pwGGXRuU5u/5QPvW1Nfu2LRG37Z1hnWqb9St4ehuDYZ0qNerda1mdfU5mVu1apWGtXSXjO1wZfVw8l3rkN9HQXu4p581A7sp2NPuLWuB1n/wjzFUlbzbDJR6m4aaBKUyIMpbN6z2+8Sut34bBT9FuH0vYU5j049Mzr3nXiR5ceTI4T59+tDCbwPI6PHjx1+7dk0ikchksn379nXs2JGuU8XBgwfVSVtQw46OjiDEaagqQK/XrVt34MCBo0ePhrRly5ba2vRfJ4yMjFxdXdXd0kYQBEEQBChP6Qxk5RZ6Rj//5nikQDiigXS+GP0iPbuA9pQAWXFh2uOMEzPSnMcK9dnb26uDlvE7TGYYtnrjh0UI+jpV69fQrV9DR1eLe4KA1Y5N6+hNH9kShPKjbawG5Sxuu8kq+05fNHo7MaqOanpaPVvVWmbb4fbvo7ibuwk7TU/9OADktXYp37ihANR21+Y1N03uxj7GLa/q2T6L4A2jurWkD3KUhhr6Wv3b1flraveozUZqP4voZJ1x4qvsiGPFkhdMcZG7uztIT1q+dEBft2/ffv78+YGBgdyr4lxcXHr06EEjVAEB6p5CLs0DG8pAM6pVqzZ27Njjx48/efKE1oUgCIIgiCrKWToDj19l77qWgB8XFBhIZ7/YV5k5aqQzw4ACEwf+me7mIJRoZTXX+X2Gdar3dsJTQe1q2gPa1fnJot31tcNVvoXDd8XQL4e3JG/eeBca19Yz7dlw05RugjflQT5mq7H9wKZ1qr35yRAOA12t7i1rbZjQhV8bZJ7utTDv1UjD++n41K2uY9itwT9fdU/ZY85vUgljdfMM6X/7inPSQTfD4bt58+a0adNoFaVAX1+/f//+q1atun37tkwmI2MAOHr06IABA2iQKg4fPiyVqno7OMO8fPly27Ztmv9MUICenl6bNm2mTp3q7e0tEoloRQiCIAiCqKH8pXN+YXFosuibE5EgFgXy8XM26I0bcRmSPOF3KDhkeeK8B74ZR6cIVVpZLfZv4+8tvqhfQ/et1LO2VhWQsxa9G+2b3evJXgt175R4ts/CZX6/vm3rVNPTKvXN3NdACR2tKi3rG0wb2fL8okEq7+ym7jE78UP/3m1ql/KxDdDNvVrXXje+cypIXqXaltp0aFuK2+Sgm2HfXeb10fhSaqv0o19Kru8oEj+jB09+x3fjxo2anzMmaGlp1alTZ9iwYW5ubs+eva6BcOrUqeHDh9NQVRw7diw3V/Ufm4Kkvnz5crt27bhnMDRTo0aNnj17Ll68GBqv/IUUBEEQBEGUKX/pDLyU5B8PS52An+bmmd3B2/eeifOLimkfKVNcWCx+nnHy6/J64hnMZ/mQL4c31yvdQw+ggEE3N6un/4t1++DfRgmqUraEHaYnvu8/uENd0KylV88QWLVKFT2dqq0a6G+b3uPeX68foVY2kNQ/W3do27i65h1gW161CojszV92e7FfWAmxsz8PHNapnoZ6YI2OdpWJQ5pfWjZEUFZoLuMk13YUiZLogVMAsnXw4MGvH3lRAlaBbq5fv769vX10dLTKtyZfuHBhzJgxtEBJqlatCsVBW2t43TKo6i+//JK8Zk4DUJWOjs6oUaMOHz6ML29GEARBkNLzXqRzYbEsNTP3x7P3HFE9y23cgZAph8JSRbmv/2FeFbLC/CyftemHHIVaraz2ZI/5jXUjVo7rWOtNH7fW1a76ReNq35m2vfDL4Ni/jZ/tU/tCCc5eHrBI3mUWssFwmS17Q/cNG1BQy0B7cPs6qx06X1k9PG67icabu+xD2w/+Mfl9UteerTU9v9u0rt48k7aeS4Zo+Pp3ym6zP6d07dFKdT3Q+Po1tOE3A7RK8LY+oTlZS65tL3gRAz916GFT8OLFi0OHDtWpU4d7dYaAevXqWVhYQMyzZ89AsPKf0+Dw8vKytLSkBRSA5q5evXrPnj0nT54cGhpaXKz2B1hRUdHNmzdHjx5Nvk2oEm1t7SFDhmzdujUiIkIsFqtsBoIgCIIgKnkv0hnILSwOfJj27b93bfGxjf0hjk6hP56JfqHmzXSvKS7KvvOv/MazkmIrq4EQvP274T/Te5j0bNimYbVqildnEPR0qraor2/asxFo2XOLBoX8Pippl5naB3xVGWjfiD9HO83tM2NUy/5f1G5RTx9UuEBEa1VlH4To3rLmtBEtt33Vw2f5kMhNo5/tVf8kMc9APUdtNto1s6dVn8Yt6hlAVYQqVdgnNNo1NpgwpNmBOX1u/ca2XFBWYJGbjECF921bW1/39RMg+jpVOzWrMWdM6yPf9QvdaJi6R2MlTmMlV/4ueBr1/+3d/1MTdx7H8Z/7J/QH7+pd56be3FznrtPpzbWed9Nre71TQazWL7RarF9Qrzp69Oz5jepYPb/jd4shkS8qX5RUxAMBJaB8DyR8DQSEECAEsrvZbDZfNrvLfQKM7azfIvKd12PePwFJ2OgPzyyf3Y8sPGFXSEEQurq6NBrN22+/rVi58eqrry5ZsuT8+fMVFRV9fX3DD3iS/Pz8iIgI0sqvvPLKnDlz5s+fv2XLlri4OK1WW1xcXFtbO3RL5uGffgz5Fsdxubm5UVFRil0JyRPOnj07LCzs9OnTRUVFVqv1aQs/AAAA4GnGKp0leYDxCFf13dFptYqOnIGzXKP/b76Z4p/3l3FZFmz1zpvblcX2ckPq1nzyw4xt75yJenPvJ29Ef/iLFe/OIhW79e+vxy6Zc2r1b69vf4e07AsVs2LaT390d897iZveImVMnnPLx7+MnDtr+buzov7yGnmVPYvfOLzyN5fW/75g93vmuA8Vjw1lGo//VfuvP5APADvCfrVq3s9Wzfv5Vx+/vm/pr8+teTPnP3/sOPNRSBUe/w/9ob+oN75F3oS177+28r1ZGz6YvSvijYtrf1cUO9d2Mbhpi+IhP45qAZWykrt/hvwDyf4fN9NWEEWRYZj09PR9+/atW7du5cqVGzZsiI2NvXDhgk6ns9vtzz3F29DQQAp727ZtR48eTUpKysvLI7lMHhgIKE9yPwPP82VlZefOnduxY8eaQTExMYcOHUpISCgoKCDPhpXNAAAAIzNW6TzEynjjSzqjUgyKlJxp83lidWKFlfU+v35kn8tVcJBSL1Km2yhN9/mP6468X7h3bsm+eS0ng1tYK37g5YeUevOJD4q//VPhnrkVB/5MXsU+Sq9C0rbz7N9K9s8r3T/PdOKDkf3y5EnIm0AamuSy8fD7Q7eRfs4khNFpX3IPzouMRRafeo+UnyIBTSK4vLzcZDK53e5nLLFQYFnWYrFYrdaXr1tBEPr7+0l5Ezabzed73h89AAAA4HnGNp2Jpl7ulK59uUavqMkZNeTDwx1Tn9sfUgzx+mQ6NUpZb5iJGnU4k76er1BL7v7hfyEAAACYqcY8nQljN7svpyU8XhmUM2fWXzPW97h8gZBOPfrbS9jsb5QBh5mQUS2g09d567TPWKQBAAAAM8d4pLNXEPWdzi3X62fsnZ43pdU6eL8ohXQrA9HVyxXFKRsOMxHj/GGbr00nedngLvMAAAAw441HOpPoYL2B4lZq/TXj4pm3y2DEpcqvtY2CFGp8yaLgMaRSKSsUGYcZz6E0i105e/2WMsnrRDcDAADAkPFIZ0KSg/WsrbVFp9bOtHPPnyVWn7j3cPiNCI2/o4TN3qGIOcw4TcJCOnWNu/iU0KV/4k3oAAAAYMYap3QmJFlmPEJCWee6q0ZFXE7v2ZRWd7O2d/hdCI3IdvOVGseY3WcD87ShkpY5b/2b16cE+s2Dfy8BAAAA+NH4pfMQK+M5f79jVXLNzLlqcGdWU6ONGz7+EMmS31JGZ0Qrwg4zhpOwkL66iis45G9/gJPNAAAA8ETjnc5EJ+1RlVqWa/Thj1Xm9JtFlyqPFLT5hFBv6/tI8MRzVaLj0nxl4WFGf+Y71OH01dWeOq3EPWurPwAAAJjhJiCdA5Lc4/RmGm2RidP/3POXV41X9N0juMxMlgSht4G+8rlDtfCx1MOM5tApkZzuuOgwyz7XgIRt9gAAAOCpJiCdCVLPvS5fblPfxrS6T1RVitycTrMzq6m4jR4+7Bck8bSnJpVKWqZIPcyojWYxe3untzFbZCwDwW0CsbgZAAAAnmVi0pkYumowz9Qfe7s58nK1ojinzZzSPbTQI1w4K4uCyFhdBYeo5OXK5sO8zKjmU5qI4OWAFWq/pULiqeF3HAAAAOCZJiydCXlgQAhIJQ/p07r2jWm10++Wz18kGzKNNo/wEmsAZEnorGRv76I0i5X9hxnBqBZSySucWTHc/TPBvU44O1ZoAAAAQOgmMp0f6WK8GTU92zMbVkyjawcXxVceKWit73ENH+SIybK3IduZ9TV1eYkyBDGhz2A0M5n/5IpO+jtK5YAPG50AAADAi5oU6Uz4RamxlzuU17pMrSfROQ0COjKxWtdK+QIvfG+Nx8mS6OsoZXP3OtThuOfGC49q8AYaV1ZxD84Ktjo5uKYZAAAAYCQmSzoHF2+IUr/bf8/cv/uWaYVGryjRqTUk/b8vsXTSntE6sSkHfAFHK69PoZIjHaoFyjrEPG1INKet5SsvB/qaJa9TFv2D/9cAAAAARmKypPMQSZZZb8Dc777dYD+Q2xKZOCUvH1ym0R/INZvs3Kiccn6EZJ/o6vV3VvHlCcyNzZQmQpmJmEejWshkbOAKj3kbsgRbHXnfEM0AAADw8iZXOj9C84Kxm80w9HyXa16TYpgqVxAuU+tjtI3JVV3Gbpc3MCbXn8mSKDIWX5vOXa5i/7eLvhaFk9CPhnycYNLXu/IO8FWJPvPdgL1J8tAoZgAAABgtkzSdCdI7Ll/A0MWmVncfLWiN0TasTq6JmKwNvSShamNa7bG7bfnN/Q63f/gYxg4paJfN31HqqUnldMdJQzPXo4N3gJ6R+6dQmsV02pds9jdcUZynLlPoNkpelrxFw+8VAAAAwCiZvOn8U3bOV9RKnb/f8fUPjeuvGVcl1Xyq1i+a6IxeFF9Jfo2oFMPm9LrY283XDT3tFC+P830byOsJHqG3wVOvdRUec97czqStpVNWUupFwcvjHqvMaTKqBZQmghwmnbqGubGZzY3l9SlCtyFYzAAAAABjZmqk8xBRlvvd/qJWR0Jp565bpi+uGJaq9UtUVYsvVZKKVXTtGE34YDGTVySvuzq5Zne2Kb2mp9nOj+6y5hGSJcnt8HeUusvinZlfUUnLgwGdEDa4omPqZzT5JJCw0KEOJ9FMpUQ6s2LcpRd9rfdEtgs3zQAAAIDxMZXSmZBkmUQq5xMZj9Dt9FZb2Uyj7cS9h1uv15OMDotXlu7oTnh85TK1/quM+hP32nKa+sz9bqdH4P2iQKJ++BecaLIkB3yyj5N4Wuht9NTf5AqPObVbg/sRTukl0QkLqSufsTm7+UqN72GxSLdLHpocphzwDu5pgtXMAAAAMB6mWDr/FMlojyA63P5O2mOyc5UWZ3aDXVViOZDbsu1GfXRq7erkmk/VVSM7IR0eX0FafN014zc3mw4XtKpKO7W1Nl0rZehyNvVy5BVJuwvipF5NSxpa4imRsQTsJqHb4Gt/4G3I4isvD66N3slkRDsSl07SnlaH0SmRzsytrjv73Q/OeQxpPvNdcggBR6vossk+14AUGD5IAAAAgHE0hdNZQZRkp0ew0J7ablfJQ1pnduQ195OYJsmbVt2dVGG9VGI5W9x+/G7bwTzz/pwWMt/dMR/ObyVfiSt8SL518UGHqtSiKbcmVVrJQ7Lqe++1OCosTIONI09L8YJ3MqzKGClZFCQPIzq7An0mv1Xva9N5G297DKnusniu8Cibs8ep3Rq8X8flT8avpxPCqMSl9NVVTMYG583tbO63XNFJvkLtqb3ubc71t5cIPcaAo03i7LLAY/M/AAAAmHDTJ52fZmiNB6nqXpePFHCz3U3auqrTScbQxdbbXCa7u62fJ9/qdnrtnM/BC06v4BVIik//VpMDXhKmgb5mv6Xca7rDG9Lcpd9zxac43XHu3hGu4KAr/4Drzj42Zy97eyebvYPNinH+sM2p3cLc2MxkRDPpa+nUKNK+dEoklbyCTPDSvSufBb9y7Qs6dQ2dtpZJXx/8yRubyKOcWTHkeVx3vnUVHOQKj3H3z7jLVXz1FU+91teS77dWBUOZpwawdhkAAAAmpemfzvDCZEkW/VJwwTQlsjaRtgT6zIKtXujS+ztKg6erW/K8jdmeuky+5hpflegui3c/OMsVxXFFJ0kNcyUXyFf4Cg1flcTXXPUY0711md7GW8E47iglzyPSHRLvkAM+nEgGAACAqQXpDAAAAAAQEqQzAAAAAEBIkM4AAAAAACFBOgMAAAAAhATpDAAAAAAQEqQzAAAAAEBIkM4AAAAAACFBOgMAAAAAhATpDAAAAAAQEqQzAAAAAEBIkM4AAAAAACFBOgMAAAAAhATpDAAAAAAQEqQzAAAAAEBIkM4AAAAAACFBOgMAAAAAhGBg4P9B9HQudlI6xgAAAABJRU5ErkJggg==\n", "text/plain": ["<IPython.core.display.Image object>"]}, "execution_count": 5, "metadata": {"image/png": {"width": 600}}, "output_type": "execute_result"}], "source": ["NbImage('sklearn.png', width=600)"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Version, d\u00e9pendance\n", "\n", "Et les conflits de d\u00e9pendance. What is [deprecated](https://scikit-learn.org/stable/modules/generated/sklearn.utils.deprecated.html) ?\n", "\n", "Que se passe-t-il quand on construit un outil d\u00e9pendant d'une dizaine d'autres et que tous \u00e9voluent en parall\u00e8le ?"]}, {"cell_type": "code", "execution_count": 5, "metadata": {"slideshow": {"slide_type": "fragment"}}, "outputs": [{"data": {"image/png": "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\n", "text/plain": ["<IPython.core.display.Image object>"]}, "execution_count": 6, "metadata": {}, "output_type": "execute_result"}], "source": ["NbImage('req.png')"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Design\n", "\n", "* refactoriser or not refactoriser ?\n", "* La programmation sert \u00e0 optimiser des usages pr\u00e9cis.\n", "* Les usages \u00e9voluent.\n", "* Il y a peut-\u00eatre des romans intemporels.\n", "* Il n'y a pas de code intemporel."]}, {"cell_type": "markdown", "metadata": {}, "source": ["## Etre rapide avec un des langages les plus lents, mise en production"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Vitesse de conception\n", "\n", "* Ce qu'on gagne en vitesse de conception\n", "* On le perd en vitesse d'ex\u00e9cution"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Grosse machine pour apprendre, petite pour pr\u00e9dire\n", "\n", "* Estimer des mod\u00e8les n\u00e9cessites beaucoup de ressources et des calculs sur de grands jeux de donn\u00e9es\n", "* Utiliser un mod\u00e8le veut souvent dire calculer une pr\u00e9diction tr\u00e8s souvent"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Mise en production des mod\u00e8les de machine learning\n", "\n", "* Obtenir le m\u00eame r\u00e9sultat, avec le m\u00eame mod\u00e8le mais un code diff\u00e9rent\n", "* On n'optimise pas les calculs de la m\u00eame mani\u00e8re sur toutes les machines"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Aujourd'hui\n", "\n", "On fait des choses \u00e9tonnantes avec le deep learning comme la traduction en temps r\u00e9elle d\u00e9crite dans [La nuit des temps](https://fr.wikipedia.org/wiki/La_Nuit_des_temps) (Barjavel). Les progr\u00e8s sont \u00e9troitement li\u00e9s avec la capacit\u00e9 \u00e0 calculer rapidement."]}, {"cell_type": "code", "execution_count": 6, "metadata": {"slideshow": {"slide_type": "fragment"}}, "outputs": [{"data": {"image/png": "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\n", "text/plain": ["<IPython.core.display.Image object>"]}, "execution_count": 7, "metadata": {"image/png": {"width": 500}}, "output_type": "execute_result"}], "source": ["NbImage('nvidia.png', width=500)"]}, {"cell_type": "markdown", "metadata": {}, "source": ["## Ce cours"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Objectif du cours\n", "\n", "* Programmation en Python\n", "* Introduction aux principaux outils pour la datascience"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Ressources\n", "\n", "* [Manuel de programmation avec Python](http://www.xavierdupre.fr/app/teachpyx/helpsphinx/index.html), [antis\u00e8che](http://www.xavierdupre.fr/app/teachpyx/helpsphinx/c_resume/index.html)\n", "* [Exercices de programmation](http://www.xavierdupre.fr/app/ensae_teaching_cs/helpsphinx3/td_1a.html)\n", "* [Exercices de datascience](http://www.xavierdupre.fr/app/ensae_teaching_cs/helpsphinx3/td_2a.html)\n", "* [Exercices de datascience - actuariat](http://www.xavierdupre.fr/app/actuariat_python/helpsphinx/index.html)\n", "* [Ce qu'on demande \u00e0 des \u00e9tudiants de savoir faire en deux heures](http://www.xavierdupre.fr/app/ensae_teaching_cs/helpsphinx3/questions/exams_1A.html#seances-notees)\n", "* [Introduction au machine learning](http://www.xavierdupre.fr/app/papierstat/helpsphinx/index.html)\n", "* [Jeux d'algorithmes pour les petits et les grands](http://www.xavierdupre.fr/app/code_beatrix/helpsphinx/index.html)"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Ressources (2)\n", "\n", "* [cette pr\u00e9sentation](http://www.xavierdupre.fr/app/jupytalk/helpsphinx/index.html)\n", "* [machine learning customis\u00e9](http://www.xavierdupre.fr/app/mlinsights/helpsphinx/index.html)\n", "* [Sujets math\u00e9matiques avec python](http://www.xavierdupre.fr/app/mlstatpy/helpsphinx/index.html)\n", "    * [KMeans](http://www.xavierdupre.fr/app/mlstatpy/helpsphinx/c_clus/kmeans.html)\n", "    * [R\u00e9seaux de neurones](http://www.xavierdupre.fr/app/mlstatpy/helpsphinx/c_ml/rn/rn.html)\n", "    * [ACP et NMF](http://www.xavierdupre.fr/app/mlstatpy/helpsphinx/c_ml/missing_values_mf.html)\n", "    * [Courbe ROC](http://www.xavierdupre.fr/app/mlstatpy/helpsphinx/c_metric/roc.html)\n", "    * [p-value](http://www.xavierdupre.fr/app/mlstatpy/helpsphinx/c_metric/pvalues.html)\n", "    * [distance d'\u00e9dition](http://www.xavierdupre.fr/app/mlstatpy/helpsphinx/c_algo/edit_distance.html)"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Ressources (3)\n", "\n", "[ONNX](https://onnx.ai/)... ce sur quoi je travaille"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### En guise de conclusion\n", "\n", "La programmation\n", "\n", "* Automatiser des t\u00e2ches r\u00e9p\u00e9titives\n", "* Traiter un volume d'informations humainement impossible \u00e0 traiter"]}, {"cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [], "source": []}], "metadata": {"celltoolbar": "Diaporama", "kernelspec": {"display_name": "Python 3", "language": "python", "name": "python3"}, "language_info": {"codemirror_mode": {"name": "ipython", "version": 3}, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.9.5"}}, "nbformat": 4, "nbformat_minor": 2}