{"cells":[{"metadata":{},"cell_type":"markdown","source":"# Googld AI Open Images - Visual Relationship Track\n\nThis kernel just explore the dataset for starting the competition in kaggle :)\n\ntrain dataset is here:\nhttps://www.kaggle.com/mahmoudmohsen213/vrd01\n\n## Contents\n1. Description of the Competition\n2. Explore the Train Dataset\n3. Explore the Test Dataset\n\n"},{"metadata":{},"cell_type":"markdown","source":"![](https://github.com/seriousmac/img_link/blob/master/kg/visual_relation_19/0.PNG?raw=true)\n\nThis year’s Open Images V5 release enabled the second Open Images Challenge to include the following 3 tracks:\n\n1. Object detection track for detecting bounding boxes around object instances, relaunched from 2018.\n\n2. __Visual relationship detection track for detecting pairs of objects in particular relations, also relaunched from 2018. (This Competition)__\n\n3. Instance segmentation track [Link to be provided when launched on July 1], brand new for 2019.\n\n![](https://github.com/seriousmac/img_link/blob/master/kg/visual_relation_19/1.PNG?raw=true)\n\nIn this track of the Challenge, you are asked to detect pairs of objects and the relationships that connect them.\n\nThe training set contains 329 relationship triplets with 375k training samples. These include both human-object relationships (e.g. \"woman playing guitar\", \"man holding microphone\"), object-object relationships (e.g. \"beer on table\", \"dog inside car\"), and also considers object-attribute relationships (e.g.\"handbag is made of leather\" and \"bench is wooden\").\n\n![](https://github.com/seriousmac/img_link/blob/master/kg/visual_relation_19/2.PNG?raw=true)"},{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"cell_type":"code","source":"import numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\nimport matplotlib.pyplot as plt \nimport cv2\n\nimport os\nprint(os.listdir(\"../input\"))\nprint(os.listdir(\"../input/vrd01\"))\nprint(os.listdir(\"../input/open-images-2019-visual-relationship\"))","execution_count":29,"outputs":[{"output_type":"stream","text":"['vrd01', 'open-images-2019-visual-relationship']\n['classes-description.csv', 'challenge-2018-train-vrd.csv', 'attributes-wordtovec.csv', 'classes-wordtovec.csv', 'attributes-description.csv']\n['test', 'VRD_sample_submission.csv']\n","name":"stdout"}]},{"metadata":{},"cell_type":"markdown","source":"# Train Dataset"},{"metadata":{"trusted":true},"cell_type":"code","source":"df_train = pd.read_csv('../input/vrd01/challenge-2018-train-vrd.csv')\nprint('shape of train data frame:', df_train.shape)\ndf_sample = pd.read_csv('../input/open-images-2019-visual-relationship/VRD_sample_submission.csv')\nprint('shape of sample submission data frame:', df_sample.shape)","execution_count":30,"outputs":[{"output_type":"stream","text":"shape of train data frame: (374768, 12)\nshape of sample submission data frame: (99999, 2)\n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"df_train.head()","execution_count":31,"outputs":[{"output_type":"execute_result","execution_count":31,"data":{"text/plain":"            ImageID  LabelName1        ...             YMax2  RelationshipLabel\n0  fe58ec1b06db2bb7   /m/04bcr3        ...          0.627778                 is\n1  82d16a22f703df5c  /m/04dr76w        ...          0.950450                 is\n2  b54d41beabcfd900   /m/01mzpv        ...          0.778557                 is\n3  4b6a08cc110d7275   /m/01mzpv        ...          0.999167                 at\n4  0144cfbb726f4c72   /m/01mzpv        ...          0.673333                 at\n\n[5 rows x 12 columns]","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>ImageID</th>\n      <th>LabelName1</th>\n      <th>LabelName2</th>\n      <th>XMin1</th>\n      <th>XMax1</th>\n      <th>YMin1</th>\n      <th>YMax1</th>\n      <th>XMin2</th>\n      <th>XMax2</th>\n      <th>YMin2</th>\n      <th>YMax2</th>\n      <th>RelationshipLabel</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>fe58ec1b06db2bb7</td>\n      <td>/m/04bcr3</td>\n      <td>/m/083vt</td>\n      <td>0.00500</td>\n      <td>0.033125</td>\n      <td>0.580000</td>\n      <td>0.627778</td>\n      <td>0.00500</td>\n      <td>0.033125</td>\n      <td>0.580000</td>\n      <td>0.627778</td>\n      <td>is</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>82d16a22f703df5c</td>\n      <td>/m/04dr76w</td>\n      <td>/m/02gy9n</td>\n      <td>0.61200</td>\n      <td>0.735000</td>\n      <td>0.418919</td>\n      <td>0.950450</td>\n      <td>0.61200</td>\n      <td>0.735000</td>\n      <td>0.418919</td>\n      <td>0.950450</td>\n      <td>is</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>b54d41beabcfd900</td>\n      <td>/m/01mzpv</td>\n      <td>/m/083vt</td>\n      <td>0.37250</td>\n      <td>0.399375</td>\n      <td>0.706413</td>\n      <td>0.778557</td>\n      <td>0.37250</td>\n      <td>0.399375</td>\n      <td>0.706413</td>\n      <td>0.778557</td>\n      <td>is</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>4b6a08cc110d7275</td>\n      <td>/m/01mzpv</td>\n      <td>/m/01y9k5</td>\n      <td>0.17125</td>\n      <td>0.255625</td>\n      <td>0.557500</td>\n      <td>0.749167</td>\n      <td>0.20750</td>\n      <td>0.683125</td>\n      <td>0.611667</td>\n      <td>0.999167</td>\n      <td>at</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>0144cfbb726f4c72</td>\n      <td>/m/01mzpv</td>\n      <td>/m/04bcr3</td>\n      <td>0.85500</td>\n      <td>0.950000</td>\n      <td>0.561667</td>\n      <td>0.609167</td>\n      <td>0.82875</td>\n      <td>0.999375</td>\n      <td>0.568333</td>\n      <td>0.673333</td>\n      <td>at</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"df_train['RelationshipLabel'].value_counts()","execution_count":32,"outputs":[{"output_type":"execute_result","execution_count":32,"data":{"text/plain":"is                194142\nat                111493\non                 31604\nholds              20986\nplays               8932\ninteracts_with      3756\ninside_of           2392\nwears                836\nhits                 593\nunder                 34\nName: RelationshipLabel, dtype: int64"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"numerical = ['XMin1', 'XMax1', 'YMin1', 'YMax1', 'XMin2', 'XMax2', 'YMin2', 'YMax2']","execution_count":33,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"df_train[numerical].hist(bins=15, figsize=(20, 10), layout=(2, 4));","execution_count":39,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 1440x720 with 8 Axes>","image/png":"iVBORw0KGgoAAAANSUhEUgAABJoAAAJPCAYAAAA0Sr4gAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzs3X+43nV95/nnS+KPlIqAbc9FCbthx0xdalYLWcG10z2VCgG7Dde1lmKZEhzGzK5onWlma5jttbSKDs6MtUAtvaKkBAcnMkwdMoqmKXKurtduIiCMKaBDxFiSAbGGH41U3dj3/vH9HLk5nJCTc+5zn3PfeT6u61zn/r6/n+/3vD/m+OXc7/vzI1WFJEmSJEmSNFcvWugEJEmSJEmSNBosNEmSJEmSJKkvLDRJkiRJkiSpLyw0SZIkSZIkqS8sNEmSJEmSJKkvLDRJkiRJkiSpLyw0SZIkSSMgycVJ/myh85CkqXw+HV0sNKkvkvx4kj1JLu6JvTzJXyV5a5Ibk1SSNVOu+0iLX9qHHC5M8v8keSbJxFzvJ2k0LJLn079J8lCSv0ny1SSXzPWekkZHv55TVXVzVZ0zw595UpKtSf5ru8fyPnZJ0ohYoOfTW5J8McmTSR5L8vEkL+9rxzSvLDSpL6rqAPBPgD9I8pMt/K+Au6vq1nb8X4AfvblKsgS4EPh6n9LYD/wBcHWf7idpBCyS59N3gf8FeAWwFrgmyf/Up3tLGnIL9Jz6O+DzwP86y+slHQUW6Pn0CuAq4KeB/x44GfjXs7yXFoCFJvVNVW0DPgtcm2Sc7uHyzp4m/wn4+SQntOPVwFeAxyYbJPl7Sb6Q5DtJ/jrJzUmO7zm3P8np7fink3y7/Syq6s+r6hbgv85vTyUNm0XwfLqyqr5aVX9XVTuB/xt4w7x2WtJQ6dNz6tIkX+w5riT/WxtR+WSSjyZJ+3nfqqo/Au6a145JGnoL8Hz6ZFV9vqqeqaongI8Bb5zPPqq/LDSp3/4ZMA7cCvzzqnqs59z3gNuAi9rxJcBNU64P8C95tnp9CvC7AFX1deC9wL9N8mPAnwCbq2piHvohafQsiudTkqXA/wjc34c+SRotc31OTeeX6Z45/wPdm8Nz+5WspKPKQj6ffgH/bhoqFprUV63ifD/wY8CfTtPkJuCSNgrgfwb+45Trd1fV9qr6flV9G/j91m7y/MeA3cBO4CTg/5yXjkgaOYvo+fTHwH8Gts2tR5JGzVyfU4dwdVU9WVV/BdwJvK5f+Uo6eizU8ynJm+mWHfi/Zpu7Bs9Ck/oqyT8ElgN/Dnxo6vmq+iLwk3RvwD5TVX875fqxJFuS7EvyNPBvgZ+YcpuPAa8Brquq7/e/F5JG0WJ4PiX51+38hVVVc++VpFEy1+fUIfSOOngG+PG5ZyrpaLMQz6ckZwGfBN5aVf9ldplrIVhoUt8k+SngI8A76BaMuzDJP5im6b8F1jP9cMoPAgWsrKrjgH9IN11l8mf8ON2C3zcAv5vkxL52QtJIWgzPpyS/B5wHnFNVT8+5U5JGSp+eU5LUdwvxfEryc8BW4B9V1R1zvZ8Gy0KT+ukPgf9YVXdW1aPAbwMfS/LSKe2uBd4M/MU093g5cAB4KsnJwP8x5fw1dDsc/GO6Ben+ePJEkmOSvAxYArwoycuSvLgfHZM09Bb6+XQF8OvAL1XVd/rRIUkjpx/PqSPS/m6avP9L27EkTTXQ51OS19DtivnuqvpPc7mXFoaFJvVFkguAn6fnjVdVfZxuB7jnzKetqv1Vdcchpo38HnA68BTdG7Ufzf9NsoZuB4P/vYV+Czg9ycXt+DeAvwWuB/5Be/2xOXdO0lBbJM+nDwL/DbA7yYH29S/60T9Jw6+Pz6kj9bd0BXSAr7ZjSfqRBXo+raebhndDz99NLgY+ROISEZIkSZIkSeoHRzRJkiRJkiSpLyw0SZIkSZIkqS8sNEmSJEmSJKkvLDRJkiRJkiSpLyw0SZIkSZIkqS+WLHQC/fYTP/ETtXz58sO2++53v8uxxx47/wnNs1HpB4xOX0alHzD3vtxzzz1/XVU/2ceUZiTJzwCf6gn9d3Tbr97U4suBPcCFVfVEkgDXAOcDzwCXVtWX273WAr/T7nNVVW1u8TOAG4GlwO3Aew63levR9nyC0emL/Vh8hvX5tFiN2vNpGPIchhxhOPIchhxh5nn6fHqumT6fYDh+F8yxf4Yhz2HIEebh+VRVI/V1xhln1EzceeedM2q32I1KP6pGpy+j0o+qufcFuLsW+JkAHAM8Bvy3wL8CNrT4BuBD7fX5wOeAAGcBO1v8RODh9v2E9vqEdu5LrW3atecdLpej7flUNTp9sR+Lzyg8nxbT16g9n4Yhz2HIsWo48hyGHKtmnqfPp9k9n6qG43fBHPtnGPIchhyr+v98mtPUuSTHJ7k1yVeTPJjkDUlOTLI9yUPt+wmtbZJcm2R3kq8kOb3nPmtb+4fa6IHJ+BlJdrVrrm2jDiTpSJwNfL2qvgmsATa3+GbggvZ6DXBTe37uAI5PchJwLrC9qvZX1RPAdmB1O3dcVe1oD9ybeu4lSZIkSUetua7RdA3w+ap6NfBa4EG6UQJ3VNUK4I52DHAesKJ9rQOuB0hyInAlcCbweuDKyeJUa/OOnutWzzFfSUefi4B/116PVdWj7fVjwFh7fTLwSM81e1vsheJ7p4lLkiRJ0lFt1ms0JXkF8AvApQBV9QPgB0nWAOOt2WZgAngvPSMGgB1tNNRJre32qtrf7js5YmCCNmKgxSdHDHxutjlLOrokeQnwK8AVU89VVSV5wTWV+pTDOrriOmNjY0xMTBz2mgMHDsyo3TAYlb7Yj8VnlPoiSZI0SuayGPipwLeBP0nyWuAe4D04YkDS4nEe8OWq+lY7/laSk6rq0VbofrzF9wGn9Fy3rMX28WzhfDI+0eLLpmn/PFW1EdgIsGrVqhofH5+u2XNMTEwwk3bDYFT6Yj8Wn1HqiyRJ0iiZS6FpCXA68O6q2pnkGp6dJgc4YmAQRqUfMDp9GZV+wEj05W08O20OYCuwFri6fb+tJ/6uJFvopvE+1YpR24AP9kznPQe4oqr2J3k6yVnATuAS4Lr5744kSdL8SvLPgH8MFLALeDtwErAFeCXdAIPfqKofJHkp3VqVZwDfAX6tqva0+1wBXAb8EPjNqtrW4qvplmA5Bvh4VV09uN5JGoS5FJr2Anuramc7vpWu0OSIgQEalX7A6PRlVPoBw92XJMcCbwb+SU/4auCWJJcB3wQubPHb6Xae2w08Q/cHFa2g9H7grtbufZPTfIF3AjcCS+mm9DqtV5IkDbUkJwO/CZxWVX+b5Ba69S7PBz5SVVuS/DFdAen69v2JqnpVkouADwG/luS0dt3PAj8N/HmSv99+zEfp/kbbC9yVZGtVPTDAbkqaZ7NeDLyqHgMeSfIzLXQ28ADPjhiA548YuKTtPncWbcQAsA04J8kJbdTAOcC2du7pJGe13eYu6bmXJL2gqvpuVb2yqp7qiX2nqs6uqhVV9UuTRaO229zlVfX3qmplVd3dc82mqnpV+/qTnvjdVfWads272vpzkiRJw24JsDTJEuDHgEeBN9ENLIDn79w7uaPvrcDZ7b3bGmBLVX2/qr5B92He69vX7qp6uK3xu6W1lTRC5jKiCeDdwM1twd2H6UYBvAhHDEiSJEnSUKmqfUn+DfBXwN8Cf0Y3Ve7JqjrYmvWunfuj9Xar6mCSp+im150M7Oi5de81U9fnPXMeuiJpAc2p0FRV9wGrpjl19jRtC7j8EPfZBGyaJn438Jq55ChJkiRJOrw2w2QN3cZPTwL/Hli9AHkc8Rq8MBzri5pj/wxDnsOQI/Q/z7mOaNIIWb7hs7O6bs/Vb+lzJpK0sKY+D9evPMilM3hG+jzUqNu176kZ/X9hKv+/IQ2NXwK+UVXfBkjyp8AbgeOTLGmjmnrXzp1ch3dvm2r3CrpFwQ+1Pi8vEP+R2azBC3Ddzbfx4S9+d0ZtJw36+TQMa6AOQ44wHHkOQ47Q/zxnvUaTJEmSJGmk/BVwVpIfa2stTa7Deyfw1tZm6jq8k+vzvhX4QpvJshW4KMlLk5wKrAC+RLdcyookp7blVy5qbSWNEEc0ac4cCSVJkiQNv6rameRW4MvAQeBeupFFnwW2JLmqxW5ol9wAfCLJbmA/XeGIqrq/7Vj3QLvP5VX1Q4Ak76LbEOoYYFNV3T+o/kkaDAtNkiRJkiQAqupK4Mop4Yfpdoyb2vZ7wK8e4j4fAD4wTfx2uo2iJI0op85JkiRJkiSpLyw0SZIkSZIkqS+cOjeiZrtukiRJkiRJ0mw5okmSJEmSJEl9YaFJkiRJkiRJfWGhSZIkaYCSHJ/k1iRfTfJgkjckOTHJ9iQPte8ntLZJcm2S3Um+kuT0nvusbe0fSrK2J35Gkl3tmmuTZCH6KUmSjk4WmiRJkgbrGuDzVfVq4LXAg8AG4I6qWgHc0Y4BzgNWtK91wPUASU6k2378TLotx6+cLE61Nu/ouW71APokSZIEWGiSJEkamCSvAH4BuAGgqn5QVU8Ca4DNrdlm4IL2eg1wU3V2AMcnOQk4F9heVfur6glgO7C6nTuuqnZUVQE39dxLkiRp3llokiRJGpxTgW8Df5Lk3iQfT3IsMFZVj7Y2jwFj7fXJwCM91+9tsReK750mLkmSNBBLFjoBSZKko8gS4HTg3VW1M8k1PDtNDoCqqiQ134kkWUc3HY+xsTEmJiYOe83YUli/8uAR/6yZ3LufDhw4MPCfeaSGIUcYjjyHIUcYnjwlaa4sNEmSRtbyDZ9d6BSkqfYCe6tqZzu+la7Q9K0kJ1XVo2362+Pt/D7glJ7rl7XYPmB8SnyixZdN0/55qmojsBFg1apVNT4+Pl2z57ju5tv48K4j//Nxz8WHv3c/TUxMMJP+LKRhyBGGI89hyBGGJ09JmisLTZKkRc+CkUZFVT2W5JEkP1NVXwPOBh5oX2uBq9v329olW4F3JdlCt/D3U60YtQ34YM8C4OcAV1TV/iRPJzkL2AlcAlw3sA5KkqSjnoUmSZKkwXo3cHOSlwAPA2+nWzfzliSXAd8ELmxtbwfOB3YDz7S2tILS+4G7Wrv3VdX+9vqdwI3AUuBz7UuSJGkgLDQtcof7FH/9yoNc6if9kiQNjaq6D1g1zamzp2lbwOWHuM8mYNM08buB18wxTUmSpFlx1zlJkiRJkiT1hYUmSZIkSZIk9YWFJkmSJEmSJPWFhSZJkiRJkiT1hYUmSSMpyfFJbk3y1SQPJnlDkhOTbE/yUPt+QmubJNcm2Z3kK0lO77nP2tb+oSRre+JnJNnVrrk2SRain5IkSZK0mFhokjSqrgE+X1WvBl4LPAhsAO6oqhXAHe0Y4DxgRftaB1wPkORE4ErgTOD1wJWTxanW5h09160eQJ8kSZLmTZKfSXJfz9fTSf6pH9ZJOhIWmiSNnCSvAH4BuAGgqn5QVU8Ca4DNrdlm4IL2eg1wU3V2AMcnOQk4F9heVfur6glgO7C6nTuuqna0rcdv6rmXJEnSUKqqr1XV66rqdcAZwDPAp/HDOklHwEKTpFF0KvBt4E+S3Jvk40mOBcaq6tHW5jFgrL0+GXik5/q9LfZC8b3TxCVJkkbF2cDXq+qb+GGdpCOwZKETkKR5sAQ4HXh3Ve1Mcg3PfvIGQFVVkprvRJKso/uEj7GxMSYmJg57zYEDB2bUbhj0qy/rVx6cezJzMLZ0Zjks9n83f7ckSUfgIuDftdd+WCdpxiw0acEs3/DZ5xyvX3mQS6fEprPn6rfMV0oaHXuBvVW1sx3fSldo+laSk6rq0faJ2uPt/D7glJ7rl7XYPmB8SnyixZdN0/55qmojsBFg1apVNT4+Pl2z55iYmGAm7YZBv/oyk2fDfFq/8iAf3nX4/2TuuXh8/pOZA3+3JEkzkeQlwK8AV0w9N4gP62bzQR3M/IOhXoP+0GIYPigZhhxhOPIchhyh/3laaJI0cqrqsSSPJPmZqvoa3dDvB9rXWuDq9v22dslW4F1JttCtJfBUK0ZtAz7Ys6bAOcAVVbW/LY55FrATuAS4bmAdlCRJml/nAV+uqm+144F+WDebD+oArrv5thl9MNRr0B8SDcMHJcOQIwxHnsOQI/Q/zzmt0ZRkT9sx4L4kd7eYOxJIWgzeDdyc5CvA64AP0hWY3pzkIeCX2jHA7cDDwG7gY8A7AapqP/B+4K729b4Wo7X5eLvm68DnBtAnSZKkQXgbz06bg+5Ducn3aVM/rLukvdc7i/ZhHbANOCfJCe394DnAtnbu6SRntfd2l/TcS9KI6MeIpl+sqr/uOZ7ckeDqJBva8Xt57o4EZ9LtNnBmz44Eq4AC7kmytS0aN7kjwU66N4Kr8c2cpBmoqvvonitTnT1N2wIuP8R9NgGbponfDbxmjmlKkiQtKm0DlTcD/6QnfDVwS5LLgG8CF7b47cD5dB+8PQO8HboP65JMflgHz/+w7kZgKd17O9/fSSNmPqbOreHZYZKb6YZIvpeeHQmAHUkmdyQYp+1IAJBkckeCCdqOBC0+uSOBDyJJkiRJmgdV9V3glVNi38EP6yTN0JymztGNQPqzJPe0BdvAHQkkSZIkSZKOSnMd0fTzVbUvyU8B25N8tfek24fP3eF2TZjNzgqLlduHLz6j1BdJkiRJ0vybU6Gpqva1748n+TTwetw+vK8Ot6X3TLfcHgZuH774jFJfJEmSJEnzb9ZT55Icm+Tlk6/pdhL4S9yRQJIkSZIk6ag0l6EwY8CnuxoQS4BPVtXnk9yFOxJIkiRJkiQddWZdaKqqh4HXThN3RwJJkiRJkqSj0Fx3nZMkSZIkSZIAC02SJEmSJEnqk9HYrmwILD/M7nGSJEmSJEnDzkKTJGmgLLxLkiRJo8upc5IkSZIkSeoLC02SJEmSJEnqCwtNkiRJA5ZkT5JdSe5LcneLnZhke5KH2vcTWjxJrk2yO8lXkpzec5+1rf1DSdb2xM9o99/drs3geylJko5GFpokSZIWxi9W1euqalU73gDcUVUrgDvaMcB5wIr2tQ64HrrCFHAlcCbweuDKyeJUa/OOnutWz393JEmSLDRJkiQtFmuAze31ZuCCnvhN1dkBHJ/kJOBcYHtV7a+qJ4DtwOp27riq2lFVBdzUcy9JkqR5ZaFJkiRp8Ar4syT3JFnXYmNV9Wh7/Rgw1l6fDDzSc+3eFnuh+N5p4pIkSfNuyUInIEmSdBT6+aral+SngO1Jvtp7sqoqSc1nAq3AtQ5gbGyMiYmJw14zthTWrzx4xD9rJvfupwMHDgz8Zx6pYcgRhiPPYcgRhidPSZorC02SJEkDVlX72vfHk3yabo2lbyU5qaoebdPfHm/N9wGn9Fy+rMX2AeNT4hMtvmya9lNz2AhsBFi1alWNj49PbfI81918Gx/edeR/Pu65+PD37qeJiQlm0p+FNAw5wnDkOQw5wvDkKUlz5dQ5SZKkAUpybJKXT74GzgH+EtgKTO4ctxa4rb3eClzSdp87C3iqTbHbBpyT5IS2CPg5wLZ27ukkZ7Xd5i7puZckSdK8ckSTJEnSYI0Bn+5qQCwBPllVn09yF3BLksuAbwIXtva3A+cDu4FngLcDVNX+JO8H7mrt3ldV+9vrdwI3AkuBz7UvSZKkeWehSZIkaYCq6mHgtdPEvwOcPU28gMsPca9NwKZp4ncDr5lzspKOOkmOBz5O9wwp4B8BXwM+BSwH9gAXVtUTbdTkNXTF8GeAS6vqy+0+a4Hfabe9qqo2t/gZPFsIvx14T3vOSRoRTp2TJEmSJE26Bvh8Vb2arij+ILABuKOqVgB3tGOA84AV7WsdcD1AkhOBK4Ez6dagu7JN8aW1eUfPdasH0CdJA2ShSZIkSZJEklcAvwDcAFBVP6iqJ4E1wObWbDNwQXu9BripOjuA49tmBucC26tqf1U9AWwHVrdzx1XVjjaK6aaee0kaERaaJI2kJHuS7EpyX5K7W+zEJNuTPNS+n9DiSXJtkt1JvpLk9J77rG3tH2pDwCfjZ7T7727XZvC9lCRJ6qtTgW8Df5Lk3iQfb5sWjLWNBgAeo1trDuBk4JGe6/e22AvF904TlzRCXKNJ0ij7xar6657jyWHfVyfZ0I7fy3OHfZ9JN6T7zJ5h36vo1ii4J8nW9snc5LDvnXTrC6zGxXYlSdJwWwKcDry7qnYmuYZnp8kB3bpxSeZ1TaUk6+im4jE2NsbExMSMrhtbCutXHjyinzXTe/fLgQMHBv4zj9Qw5AjDkecw5Aj9z9NCk6SjyRpgvL3eDEzQFZp+NOwb2JFkctj3OG3YN0CSyWHfE7Rh3y0+OezbQpMkSRpme4G9VbWzHd9KV2j6VpKTqurR9jfS4+38PuCUnuuXtdg+nv2bazI+0eLLpmn/HFW1EdgIsGrVqhofH5/aZFrX3XwbH951ZG9x91w8s3v3y8TEBDPtz0IZhhxhOPIchhyh/3k6dU7SqCrgz5Lc0z4VA4d9S5IkHVJVPQY8kuRnWuhs4AFgKzC5hMBa4Lb2eitwSVuG4Czgqfa31jbgnCQntKUKzgG2tXNPJzmrLTtwSc+9JI0IRzRJGlU/X1X7kvwUsD3JV3tPDmLYN8xu6PewDLGdien6cqRD2heDmQ7FX+z/bqP+uyVJ6ot3AzcneQnwMPB2ugEKtyS5DPgmcGFreztwPrAbeKa1par2J3k/cFdr977JEeLAO4EbgaV0o8EdES6NGAtNkkZSVe1r3x9P8mm6rXUHOuy7/fwjHvo9LENsZ2K6vly64bMLk8wcrF95cEZD8Qc9/P5IjfrvliRp7qrqPrr1Kac6e5q2BVx+iPtsAjZNE78beM0c05S0iDl1TtLISXJskpdPvqYbrv2XOOxbkiRJkuaVI5okjaIx4NNdDYglwCer6vNJ7sJh35IkSZI0byw0SRo5VfUw8Npp4t/BYd+SJEmSNG+cOidJkiRJkqS+sNAkSZIkSZKkvrDQJEmSJEmSpL6w0CRJkiRJkqS+mHOhKckxSe5N8pl2fGqSnUl2J/lUkpe0+Evb8e52fnnPPa5o8a8lObcnvrrFdifZMNdcJUmSJEmSNH/6MaLpPcCDPccfAj5SVa8CngAua/HLgCda/COtHUlOAy4CfhZYDfxRK14dA3wUOA84DXhbaytJkiRJkqRFaMlcLk6yDHgL8AHgt5IEeBPw663JZuB3geuBNe01wK3AH7b2a4AtVfV94BtJdgOvb+12t23KSbKltX1gLjlLkvpj+YbPHrbN+pUHuXQG7SRJkiSNhrmOaPoD4LeBv2vHrwSerKqD7XgvcHJ7fTLwCEA7/1Rr/6P4lGsOFZckSZIkSdIiNOsRTUl+GXi8qu5JMt6/lGaVyzpgHcDY2BgTExOHvebAgQMzatcv61cePHyjWRhbOn/3HrSZ9mWQ/26zMejfrfk0Sn2RJEmSJM2/uUydeyPwK0nOB14GHAdcAxyfZEkbtbQM2Nfa7wNOAfYmWQK8AvhOT3xS7zWHij9HVW0ENgKsWrWqxsfHD5v8xMQEM2nXL/M1dWT9yoN8eNecZkAuGjPty56Lx+c/mTkY9O/WfBqlvkiSJEmS5t+sp85V1RVVtayqltMt5v2FqroYuBN4a2u2Fritvd7ajmnnv1BV1eIXtV3pTgVWAF8C7gJWtF3sXtJ+xtbZ5itJkiRJkqT5NR9DYd4LbElyFXAvcEOL3wB8oi32vZ+ucERV3Z/kFrpFvg8Cl1fVDwGSvAvYBhwDbKqq++chX0mSJEmSJPVBXwpNVTUBTLTXD/PsrnG9bb4H/Oohrv8A3c51U+O3A7f3I0dJkiRJkiTNr7nuOidJkiRJkiQBFpokSZIkSZLUJxaaJEmSJEmS1BcWmiRJkgYsyTFJ7k3ymXZ8apKdSXYn+VTbcZe2K++nWnxnkuU997iixb+W5Nye+OoW251kw6D7Jmm4JdmTZFeS+5Lc3WInJtme5KH2/YQWT5Jr2/PmK0lO77nP2tb+oSRre+JntPvvbtdm8L2UNJ8sNEmSJA3ee4AHe44/BHykql4FPAFc1uKXAU+0+EdaO5KcRreD788Cq4E/asWrY4CPAucBpwFva20l6Uj8YlW9rqpWteMNwB1VtQK4ox1D96xZ0b7WAddDV5gCrgTOpNso6srJ4lRr846e61bPf3ckDZKFJkmSpAFKsgx4C/DxdhzgTcCtrclm4IL2ek07pp0/u7VfA2ypqu9X1TeA3XRv5l4P7K6qh6vqB8CW1laS5qL3WTT1GXVTdXYAxyc5CTgX2F5V+6vqCWA7sLqdO66qdlRVATf13EvSiLDQJEmSNFh/APw28Hft+JXAk1V1sB3vBU5ur08GHgFo559q7X8Un3LNoeKSNFMF/FmSe5Ksa7Gxqnq0vX4MGGuvj/RZdHJ7PTUuaYQsWegEJEmSjhZJfhl4vKruSTK+wLmso5vqwtjYGBMTE4e9ZmwprF958LDtpprJvfvpwIEDA/+ZR2oYcoThyHMYcoThyRP4+aral+SngO1Jvtp7sqoqSc1nArN5PsHsnlE+n55vGHKE4chzGHKE/udpoUmSJGlw3gj8SpLzgZcBxwHX0E03WdJGLS0D9rX2+4BTgL1JlgCvAL7TE5/Ue82h4s9RVRuBjQCrVq2q8fHxwyZ/3c238eFdR/7n456LD3/vfpqYmGAm/VlIw5AjDEeew5AjDE+eVbWvfX88yafppuR+K8lJVfVom/72eGt+qGfRPmB8SnyixZdN035qDkf8fILZPaN8Pj3fMOQIw5HnMOQI/c/TqXOSJEkDUlVXVNWyqlpOt5j3F6rqYuBO4K2t2VrgtvZ6azumnf9CW9dkK3BR25XuVLoFdb8E3AWsaLvYvaT9jK0D6JqkEZDk2CQvn3wNnAP8Jc99Fk19Rl3Sdp87C3iqTbHbBpyT5IS2CPg5wLZ27ukkZ7X15i7puZekEeGIJkmSpIX3XmBLkquAe4EbWvwG4BNJdgP76QpHVNX9SW4BHgAOApdX1Q8BkryL7k3eMcCmqrp/oD2RNMzGgE93NSCWAJ+sqs8nuQu4JcllwDeBC1v724Hz6TYkeAZ4O0BV7U/yfrriN8D7qmp/e/1O4EZgKfC59iVphFhokiRJWgBVNUE3lYSkx6v+AAAgAElEQVSqephuesrUNt8DfvUQ138A+MA08dvp3vxJ0hFpz6LXThP/DnD2NPECLj/EvTYBm6aJ3w28Zs7JSlq0nDonaWQlOSbJvUk+045PTbIzye4kn2rTSmhTTz7V4juTLO+5xxUt/rUk5/bEV7fY7iQbBt03SZIkSVqMLDRJGmXvAR7sOf4Q8JGqehXwBHBZi18GPNHiH2ntSHIa3TSVnwVWA3/UilfHAB8FzgNOA97W2kqSJEnSUc1Ck6SRlGQZ8Bbg4+04wJuAW1uTzcAF7fWadkw7f3ZrvwbYUlXfr6pv0K0/8Pr2tbuqHq6qHwBbWltJkiRJOqpZaJI0qv4A+G3g79rxK4En29bhAHuBk9vrk4FHANr5p1r7H8WnXHOouCRJkiQd1VwMXNLISfLLwONVdU+S8QXOZR2wDmBsbIyJiYnDXnPgwIEZtVto61cePGybsaUza7fYzbQfi/3fbVh+t2ZilPoiSZI0Siw0SRpFbwR+Jcn5wMuA44BrgOOTLGmjlpYB+1r7fcApwN4kS4BXAN/piU/qveZQ8eeoqo3ARoBVq1bV+Pj4YZOfmJhgJu0W2qUbPnvYNutXHuTDu4b/PzUz7ceei8fnP5k5GJbfrZkYpb5IkiSNEqfOSRo5VXVFVS2rquV0i3l/oaouBu4E3tqarQVua6+3tmPa+S+07Xq3Ahe1XelOBVYAXwLuAla0Xexe0n7G1gF0TZIkSZIWteH/mFmSZu69wJYkVwH3Aje0+A3AJ5LsBvbTFY6oqvuT3AI8ABwELq+qHwIkeRewDTgG2FRV9w+0J320fAYjkyRJkiRpJiw0HSHfkEnDpaomgIn2+mG6HeOmtvke8KuHuP4DwAemid8O3N7HVCVJkiRp6Dl1TpIkSZIkSX1hoUmSJEmSJEl9YaFJkiRJkiRJfWGhSZIkSZIkSX1hoUmSJEmSJEl9YaFJkiRJkiRJfWGhSZIkSZIkSX1hoUmSJEmSJEl9MetCU5KXJflSkv+c5P4kv9fipybZmWR3kk8leUmLv7Qd727nl/fc64oW/1qSc3viq1tsd5INs++mJEmSJEmS5ttcRjR9H3hTVb0WeB2wOslZwIeAj1TVq4AngMta+8uAJ1r8I60dSU4DLgJ+FlgN/FGSY5IcA3wUOA84DXhbaytJkiRJkqRFaNaFpuocaIcvbl8FvAm4tcU3Axe012vaMe382UnS4luq6vtV9Q1gN/D69rW7qh6uqh8AW1pbSZIkSdI8aB/635vkM+3YGSuSjsic1mhqD6H7gMeB7cDXgSer6mBrshc4ub0+GXgEoJ1/Cnhlb3zKNYeKS5IkSZLmx3uAB3uOnbEi6YgsmcvFVfVD4HVJjgc+Dby6L1kdoSTrgHUAY2NjTExMHPaaAwcOzKjdVOtXHjx8owEaW7r4cpqtmfZlNv9ugzTb363FaJT6IkmSpBeWZBnwFuADwG+1GShvAn69NdkM/C5wPd1sk99t8VuBP5w6YwX4RpLJGSvQZqy0nzU5Y+WBee6WpAGbU6FpUlU9meRO4A3A8UmWtFFLy4B9rdk+4BRgb5IlwCuA7/TEJ/Vec6j41J+/EdgIsGrVqhofHz9szhMTE8yk3VSXbvjsEV8zn9avPMiHd/Xln3HBzbQvey4en/9k5mC2v1uL0Sj1RZIkSYf1B8BvAy9vx69khjNWkvTOWNnRc8/ea6bOWDmz3x2QtPBmXaFI8pPA/9eKTEuBN9MNl7wTeCvdmkprgdvaJVvb8f/bzn+hqirJVuCTSX4f+GlgBfAlIMCKJKfSFZgu4tlKuiRJkiSpT5L8MvB4Vd2TZHyBczniGSswu9kegx69PwwzBoYhRxiOPIchR+h/nnMZCnMSsLnNtX0RcEtVfSbJA8CWJFcB9wI3tPY3AJ9oQyf30xWOqKr7k9xCN2TyIHB5m5JHkncB24BjgE1Vdf8c8pUkSZIkTe+NwK8kOR94GXAccA1DMmMF4Lqbbzvi2R6Dni0xDDMGhiFHGI48hyFH6H+esy40VdVXgJ+bJv4wz87B7Y1/D/jVQ9zrA3TzgKfGbwdun22OkiRJkqTDq6orgCsA2oimf15VFyf59zhjRdIRmNOuc5IkSToySV6W5EtJ/nOS+5P8Xou7hbikxei9dAuD76Zbg6l3xsorW/y3gA3QzVgBJmesfJ42Y6WNiJqcsfIg3YwYZ6xII2g0VpGWJEkaHt8H3lRVB5K8GPhiks/RvVH7SFVtSfLHdFuHX0/PFuJJLqJbE/PXpmwh/tPAnyf5++1nfJRu/cy9wF1JtlaVOztJmpGqmgAm2mtnrEg6Io5okiRJGqDqHGiHL25fRbeF+K0tvhm4oL1e045p58+euoV4VX0DmNxC/PW0LcSr6gd0013WzHO3JEmSAAtNkiRJA5fkmCT3AY8D24GvM8MtxIHeLcSnbhV+8gvEJUmS5p1T5yRJkgas7bD7uiTHA58GXj3oHGazffhstg4Htw+fzjDkCMOR5zDkCMOTpyTNlYUmSZKkBVJVTya5E3gDA95CfDbbh89m63Bw+/DpDEOOMBx5DkOOMDx5StJcOXVOkiRpgJL8ZBvJRJKldIt2PwjcSbdFOEy/hTj0bCHe4he1XelO5dktxO+ibSHedq67qLWVJEmad45okjSSkrwM+AvgpXTPulur6sr2ZmwL3fom9wC/UVU/SPJS4CbgDLqRAr9WVXvava6g2/Xph8BvVtW2Fl8NXAMcA3y8qq4eYBclDa+TgM1JjqH70O+WqvpMkgeALUmuAu7luVuIf6JtIb6frnBEVd2fZHIL8YO0LcQBkkxuIX4MsOlo20J8176nuHTDZ4/4uj1Xv2UespEk6ehioUnSqHL7cEmLUlV9Bfi5aeJuIS5JkoaehSZJI6lNKznU9uG/3uKbgd+lKzStaa+h2z78D6duHw58o40omHwjuLu9MSTJ5PbhFpokaYrlsxhdBMMxwmi2fbtx9bF9zkSSpMXBQpOkkdWmpdwDvIpu9NGMtw9P0rt9+I6e2/ZeM3X78DPnoRuSdNSabRFn/crB/jxJkvQsC02SRtawbh8+6O2PZ7NV+UzNdiv0xWam/Vjs21aP0tbao9QXSZKkUWKhSdLIG7btwwe9/fFsFsydqfUrD85qK/TFZqb9GPQW7kdqlLbWHqW+SJIkjZIXLXQCkjQf3D5ckiRJkgZv+D9mlqTpuX24JEmSJA2YhSZJI8ntwyVJi9mufU/NauryMOzEJ0k6ujl1TpIkSZIkSX1hoUmSJEmSJEl94dQ5SZIkaUgsn+VOoU65kyQNioUmSZIkacTNtkB14+pj+5yJJGnUWWiSpBEx2zcRkiRJGgw3AtDRwDWaJEmSJEmS1BeOaJIkSZIkkeRlwF8AL6V7r3hrVV2Z5FRgC/BK4B7gN6rqB0leCtwEnAF8B/i1qtrT7nUFcBnwQ+A3q2pbi68GrgGOAT5eVVcPsIt9M9uR5OtX9jkRaRGy0CRJkiRJAvg+8KaqOpDkxcAXk3wO+C3gI1W1Jckf0xWQrm/fn6iqVyW5CPgQ8GtJTgMuAn4W+Gngz5P8/fYzPgq8GdgL3JVka1U9MMhO9nLpAan/nDonSZIkSaI6B9rhi9tXAW8Cbm3xzcAF7fWadkw7f3aStPiWqvp+VX0D2A28vn3trqqHq+oHdKOk1sxztyQNmCOaJEmSJE1rNgsXu2jxcEtyDN30uFfRjT76OvBkVR1sTfYCJ7fXJwOPAFTVwSRP0U2vOxnY0XPb3msemRI/cx66MXJmO/LK/z9qIVhokiRJkiQBUFU/BF6X5Hjg08CrB51DknXAOoCxsTEmJiZmdN3YUli/8uDhGy6gQec40//teh04cGBW1w3aMOQ5DDlC//O00CRJkiRJeo6qejLJncAbgOOTLGmjmpYB+1qzfcApwN4kS4BX0C0KPhmf1HvNoeK9P3sjsBFg1apVNT4+PqOcr7v5Nj68a3G/xV2/8uBAc9xz8fgRXzMxMcFM/zdfSMOQ5zDkCP3P0zWaJEmSJEkk+ck2kokkS+kW7X4QuBN4a2u2Fritvd7ajmnnv1BV1eIXJXlp27FuBfAl4C5gRZJTk7yEbsHwrfPfM0mDtLjLvZIkSZKkQTkJ2NzWaXoRcEtVfSbJA8CWJFcB9wI3tPY3AJ9IshvYT1c4oqruT3IL8ABwELi8TckjybuAbcAxwKaqun9w3ZM0CLMuNCU5BbgJGKPbiWBjVV2T5ETgU8ByYA9wYVU90XYfuAY4H3gGuLSqvtzutRb4nXbrq6pqc4ufAdwILAVuB97TKuSSJEmSpD6qqq8APzdN/GG6HeOmxr8H/Ooh7vUB4APTxG+ne2+nAZjNIuI3rj52HjLR0WQuU+cOAuur6jTgLODyJKcBG4A7qmoFcEc7BjiPbsjkCrqF3a4HaIWpK+l2G3g9cGWSE9o11wPv6Llu9RzylSRJkiRJ0jyadaGpqh6dHJFUVX9DN3f3ZGANsLk12wxc0F6vAW6qzg66BeVOAs4FtlfV/qp6AtgOrG7njquqHW0U000995IkSZIkSdIi05fFwJMspxtiuRMYq6pH26nH6KbWQVeEeqTnsr0t9kLxvdPEJUmSJEmStAjNeTHwJD8O/Afgn1bV091STJ2qqiTzvqZSknV00/EYGxtjYmLisNccOHBgRu2mWr/y4BFfM5/Gli6+nGZrpn2Zzb/bIM32d2sxGqW+SJKkwZjNmjAAe65+S58zkSQthDkVmpK8mK7IdHNV/WkLfyvJSVX1aJv+9niL7wNO6bl8WYvtA8anxCdafNk07Z+nqjYCGwFWrVpV4+Pj0zV7jomJCWbSbqpLZ/kfzvmyfuVBPrxrNDYPnGlf9lw8Pv/JzMFsf7cWo1HqiyRJkiRp/s166lzbRe4G4MGq+v2eU1uBte31WuC2nvgl6ZwFPNWm2G0DzklyQlsE/BxgWzv3dJKz2s+6pOdekiRJQyfJKUnuTPJAkvuTvKfFT0yyPclD7fsJLZ4k1ybZneQrSU7vudfa1v6htoPvZPyMJLvaNdemd7i5JEnSPJvLGk1vBH4DeFOS+9rX+cDVwJuTPAT8UjuGbgvLh4HdwMeAdwJU1X7g/cBd7et9LUZr8/F2zdeBz80hX0mSpIXmrr2SJGmkzXrOVVV9ETjUJ2RnT9O+gMsPca9NwKZp4ncDr5ltjpIkSYtJG7H9aHv9N0l6d+0db8020y0j8F56du0FdiSZ3LV3nLZrL0CSyV17J2i79rb45K69flgnSZIGYjQW95EkSRoy7torSVqMdu17alZrE7ugvyZZaJI0cpKcAtxE90atgI1VdU2bavIpYDmwB7iwqp5o65dcA5wPPANcWlVfbvdaC/xOu/VVVbW5xc8AbgSW0k0Nfk8bcSBJhzWsu/YOy263w5DnMOQIg81ztjvdDssuucOSpyTNlYUmSaNocg2ULyd5OXBPm1ZyKd0aKFcn2UC3Bsp7ee4aKGfSrW9yZs8aKKvoClb3JNlaVU/w7BooO+kKTatxaoqkGRjmXXuvu/m2odjtdhh25R2GHGGwec52Z+Fh2SV3WPKUpLmay2LgkrQoVdWjkyOSqupvgN41UDa3Zpvp1i2BnjVQ2romk2ugnEtbA6UVlybXQDmJtgZKG8V0U8+9JOmQ3LVXkiSNusX/MYokzYFroEhaZCZ37d2V5L4W+xd0u/TekuQy4JvAhe3c7XTTenfTTe19O3S79iaZ3LUXnr9r7410U3s/h6MtJUnSAFlokjSyhnUNlNmu4bAY1/oYljVIDmem/Vjsa2+M0vogw9oXd+2VJEmjzkKTpJE0zGugzHYNh9nsDjLfhmUNksOZaT9mu77IoIzS+iCj1BdJkqRRMvx//UvSFDNYA+Vqnr8GyruSbKFbDPypVozaBnywrX8C3RooV7QpK0+39VJ20q2Bct28d0ySJElapJbP8kPPPVe/pc+ZaKFZaJI0ilwDRZIkSZIWgIUmSSPHNVAkSRo+sx0NcePqY/uciSRpLl600AlIkiRJkiRpNDiiSZIkSZJEklOAm4AxoICNVXVNkhOBTwHLgT3AhVX1RFsX8xq6JQieAS6tqi+3e60Ffqfd+qqq2tziZ/Ds8gO3A+9po8t1lHI04+hxRJMkSZIkCeAgsL6qTgPOAi5PchqwAbijqlYAd7RjgPOAFe1rHXA9QCtMXUm3ycrrgSt7Nle5HnhHz3WrB9AvSQNkoUmSJEmSRFU9Ojkiqar+BngQOBlYA2xuzTYDF7TXa4CbqrMDOD7JScC5wPaq2l9VTwDbgdXt3HFVtaONYrqp516SRoSFJkmSJEnScyRZDvwcsBMYq6pH26nH6KbWQVeEeqTnsr0t9kLxvdPEJY0Q12iSJEmSJP1Ikh8H/gPwT6vq6W4ppk5VVZJ5XVMpyTq6qXiMjY0xMTExo+vGlsL6lQfnMbO5M8f+OXDgwIx/NxbKMOQI/c/TQpMkSZIkCYAkL6YrMt1cVX/awt9KclJVPdqmvz3e4vuAU3ouX9Zi+4DxKfGJFl82TfvnqKqNwEaAVatW1fj4+NQm07ru5tv48K7F/RZ3/cqD5tgnN64+lpn+biyUiYmJRZ8j9D/Pxf/bI0mSJEmad20XuRuAB6vq93tObQXWAle377f1xN+VZAvdwt9PtWLUNuCDPQuAnwNcUVX7kzyd5Cy6KXmXANfNe8c0knbte4pLZ7Fj3Z6r3zIP2aiXhSZJkiRJEsAbgd8AdiW5r8X+BV2B6ZYklwHfBC5s524Hzgd2A88AbwdoBaX3A3e1du+rqv3t9TuBG4GlwOfal6QRYqFJkiRJkkRVfRHIIU6fPU37Ai4/xL02AZumid8NvGYOaUpa5I7aQtNsh9lJkiRJkiRpei9a6AQkSZIkSZI0Go7aEU2SJEmShp8LAks6Est9Xsw7RzRJkiRJkiSpLyw0SZIkSZIkqS8sNEmSJEmSJKkvLDRJkiRJkiSpLyw0SZIkSZIkqS8sNEmSJEmSJKkv5lRoSrIpyeNJ/rIndmKS7Ukeat9PaPEkuTbJ7iRfSXJ6zzVrW/uHkqztiZ+RZFe75tokmUu+kiRJkiRJmj9L5nj9jcAfAjf1xDYAd1TV1Uk2tOP3AucBK9rXmcD1wJlJTgSuBFYBBdyTZGtVPdHavAPYCdwOrAY+N8ecJUmSJEmSZmT5hs/O6robVx/b50yGw5xGNFXVXwD7p4TXAJvb683ABT3xm6qzAzg+yUnAucD2qtrfikvbgdXt3HFVtaOqiq6YdQGSJEmSJElalOZjjaaxqnq0vX4MGGuvTwYe6Wm3t8VeKL53mrgkSdLQcukBSZI0yuY6de4FVVUlqfn8GQBJ1gHrAMbGxpiYmDjsNWNLYf3Kg/Oc2fwblX7AzPsyk3/fhXTgwIFFn+NMjVJfJGkRuRGXHpAkSSNqPgpN30pyUlU92qa/Pd7i+4BTetota7F9wPiU+ESLL5um/fNU1UZgI8CqVatqfHx8umbPcd3Nt/HhXfNaZxuI9SsPjkQ/YOZ92XPx+PwnMwcTExPM5HdwGAxzX5JsAn4ZeLyqXtNiJwKfApYDe4ALq+qJ9mn/NcD5wDPApVX15XbNWuB32m2vqqrNLX4G3ZvFpXRv5N7TpvlK0guqqr9IsnxKeA3P/j20me5voffSs/QAsCPJ5NID47SlBwCSTC49MEFbeqDFJ5cesNAkSZIGYj4qFFuBtcDV7fttPfF3JdlC94ncU60YtQ344OQQceAc4Iqq2p/k6SRn0X0idwlw3TzkK2k03ciQjhjYte8pLp3lgoOShpZLD0iSpJEwp0JTkn9H94naTyTZS/eG7GrgliSXAd8ELmzNb6cbLbCbbsTA2wFaQen9wF2t3fsmP50D3smzIwY+h5/GSZohRwxIGlYuPTB3w5DnMOQIw5HnbHMc9PIALkkg6Wgxp0JTVb3tEKfOnqZtAZcf4j6bgE3TxO8GXjOXHCWphyMGJC1WLj3QR8OwtMAw5AjDkedscxz0cgzDvCSBJB2Jxf1fDUmaJ44YGIxR6YubFSw+o9SXxqUHJEnSSLDQJOlo4oiBARuGT8Jnws0KFp9h7otLD0iLw/JZroe45+q39DmTxcPNVKT+mu3aq8P+nHnRQicgSQM0OWIAnj9i4JL/n717D7Krug98//1F4uUHL5P0JRKxNGPFuRjFMVYBKd9k2sgGAb6IqsGOCAmSo7HuxOCQRDdGJFNDxoYU3NjG4FdGNgqCcSwwiQeNEVEUoMvjqUg8jAMWhNAWspEGGwcJYQUb0uR3/9irzaE5p9Xq3n1e/f1Uneqzf3vtfX67T2vp9K/XXisqp1FGDACbgTMi4pgyauAMYHPZ91xEnFY+ZF3UcC5JGldmXpCZx2fmIZk5NzNvyMxnMnNxZi7IzHeNFo2ycnFm/tvMXFimFRg9z7rMfFN5/HlD/P7MPKkcc4m/xEk6CDdSLXDSaHQxlQXAXWUbXrmYyiqqhVJGC1NXUI3CPAW4omH05ehiKqPHjX0tSX3AQpOkvlRGDPwd8OaI2FVGCVwNvDsiHgfeVbah+ovaDqoRA5+nGg1A+UVvdMTAfbx6xMAXyjHfxhEDkiSpx2Xm14A9Y8JLqRZRoXw9ryF+UymIbwVGF1M5k7KYSlmpd3QxleMpi6mUAvhNDeeS1Ed6/34GSWrCxQokSZJq4WIqkg6KhSZJkiRJ0gF182Iq0BuLkJhjfXohz8nm2O4FT+peZMVCkyRJkiSplZ5YTAV6Y0GVXlgopRdyhN7Ic7I5tnuBmboXWXGOJkmSJElSKy6mIumgdHf5T5IkSZLUFmUxlUHguIjYRbV63NXArWVhle8A7yvNNwFnUy2M8jzwfqgWU4mI0cVU4NWLqdwIHEG1kIqLqUh9yEKTJEmSJMnFVCTVwlvnJEmSJEmSVAsLTZIkSZIkSaqFt85JkiRJkiR1iXlr7pjUcTuvPqfmTCbHEU2SJEmSJEmqhYUmSZIkSZIk1cJCkyRJkiRJkmphoUmSJEmSJEm1sNAkSZIkSZKkWlhokiRJkiRJUi0sNEmSJEmSJKkWFpokSZIkSZJUCwtNkiRJkiRJqsXsTicgSZIkSb1i3po7JnXcjUteW3MmktSdHNEkSZIkSZKkWlhokiRJkiRJUi0sNEmSJEmSJKkWztEkSZIkSZLU47plDjlHNEmSJEmSJKkWFpokSZIkSZJUCwtNkiRJkiRJqkXXF5oiYklEPBYRwxGxptP5SNIo+ydJ3cr+SVI3s4+S+ltXF5oiYhbwGeAs4ETggog4sbNZSZL9k6TuZf8kqZvZR0n9r6sLTcApwHBm7sjMF4ENwNIO5yRJYP8kqXvZP0nqZvZRUp/r9kLTHODJhu1dJSZJnWb/JKlb2T9J6mb2UVKfm93pBOoQEauAVWVzf0Q8NoHDjgP+afqyao/f6ZPrgIlfS1zThmSmpm/eE6Z+LW+sK5FeNZP7J+ifPsr+qSvZP01RP/dPvdD39EKO0Bt59kKOAO+8ZsJ52j9Nrn+CHvhZ6IWf117IEXojz17IEervn7q90LQbOKFhe26JvUJmrgXWHsyJI+L+zFw0tfQ6r1+uA/rnWvrlOqC/rmUa2D9NQL9ci9fRffrpWqbBjO+feiHPXsgReiPPXsgReifPNjhgHzWZ/gl643tsjvXphTx7IUeoP89uv3XuPmBBRMyPiEOBZcDGDuckSWD/JKl72T9J6mb2UVKf6+oRTZk5EhGXAJuBWcC6zNze4bQkyf5JUteyf5LUzeyjpP7X1YUmgMzcBGyahlMf9FDMLtUv1wH9cy39ch3QX9dSO/unCemXa/E6uk8/XUvt7J96Is9eyBF6I89eyBF6J89pN8P7KHOsTy/k2Qs5Qs15RmbWeT5JkiRJkiTNUN0+R5MkSZIkSZJ6RN8XmiJiSUQ8FhHDEbGmyf7DIuKWsn9bRMxrf5YHNoHrWBERP4iIb5bHf+hEngcSEesi4umI+FaL/RER15frfCgiTm53jhMxgesYjIh9De/Hf253jhMVESdExD0R8UhEbI+IS5u06Yn3pdfYP3WXfumfoH/6KPunzumV/mkCef5++fl5KCLuioi2Lxt/oBwb2v37iMiIaPvqRBPJMSLe1/Bv8S/anWPJ4UDv98+VPuPB8p6f3YEc++b/km5l/9S+HBvadax/Kq/f9X2U/dMYmdm3D6rJ5b4N/BvgUODvgRPHtPkg8Gfl+TLglk7nPcnrWAF8utO5TuBafhU4GfhWi/1nA3cCAZwGbOt0zpO8jkHgq53Oc4LXcjxwcnn+euAfm/x89cT70ksP+6fue/RL/zTBa+mJPsr+qWPf957onyaY5zuB15Tnv93uPCeSY2n3euBrwFZgUbflCCwAHgSOKds/06Xv91rgt8vzE4GdHcizb/4v6caH/VN7cyztOtY/HcT3sqN9lP3Tqx/9PqLpFGA4M3dk5ovABmDpmDZLgfXl+W3A4oiINuY4ERO5jp6QmV8D9ozTZClwU1a2AkdHxPHtyW7iJnAdPSMzn8rMb5TnPwQeBeaMadYT70uPsX/qMv3SP0H/9FH2Tx3TK/3TAfPMzHsy8/myuRWY2205Fh8FrgF+3M7kionk+AHgM5m5FyAzn25zjjCxPBM4sjw/CvjfbcyvSqCP/i/pUvZPbcyx6GT/BL3RR9k/jdHvhaY5wJMN27t49QfUn7TJzBFgH/CGtmQ3cRO5DoB/X4a43RYRJ7QntdpN9Fp7wS9HxN9HxJ0R8ZZOJzMRZWjx24BtY3b10/vSLeyfek+//TvoqT7K/qmteqV/Otj3fiXVX2rb6YA5llsTTsjMO9qZWIOJfB9/Hvj5iPhfEbE1Ipa0LbuXTSTPPwZ+IyJ2Ua1o9qH2pHZQ7LOmxv6pPr3QP0Fv9FH2T2P0e6FpJvkfwLzM/EVgCy9X8dUZ3wDemJlvBT4F/PcO53NAEfE64C+B383M5zqdj/qK/VP36ak+yv5JUxURvwEsAv6007k0ioifAoMQ+Z8AACAASURBVD4BrO50Lgcwm+rWlEHgAuDzEXF0RzNq7gLgxsycS3ULyM3leyx1LfunWvRCHzWj+qe+vbBiN9D4l/O5Jda0TUTMphrG9kxbspu4A15HZj6TmS+UzS8Ab29TbnWbyHvW9TLzuczcX55vAg6JiOM6nFZLEXEI1S9xX8zMv2rSpC/ely5j/9R7+ubfQS/1UfZPHdEr/dOE3vuIeBfwR8C5DX1Ruxwox9cDJwFDEbGTak6MjW2ecHci38ddwMbM/JfMfIJqvrQFbcpv1ETyXAncCpCZfwccDnRb32afNTX2T/Xphf4JeqOPsn8ao98LTfcBCyJifkQcSjUZ3MYxbTYCy8vz84G7M6uZsLrIAa9jzL2T51LNY9GLNgIXlRnvTwP2ZeZTnU7qYEXE/zF6L3hEnEL1b63bCgRAtboAcAPwaGZ+okWzvnhfuoz9U+/pm38HvdJH2T91TK/0TxPpf94G/FeqX+I6Ma/QuDlm5r7MPC4z52XmPKp5Ws7NzPu7Jcfiv1ONFKAUpX8e2NHGHGFieX4XWAwQEf8n1S9yP2hrlgdmnzU19k9tyrFL+qcD5ll0uo+yfxpjdr15dZfMHImIS4DNVDPBr8vM7RHxEeD+zNxI9QH25ogYppoYa1nnMm5ugtfxOxFxLjBCdR0rOpbwOCLiS1SdwHHl/tQrgEMAMvPPqO5XPRsYBp4H3t+ZTMc3ges4H/jtiBgBfgQs68ICwah3AL8JPBwR3yyxPwR+Dnrrfekl9k/dp1/6J+irPsr+qQN6pX+aYJ5/CrwO+HKprX43M8/tshw7aoI5bgbOiIhHgJeAP8jMthanJ5jnaqpbZn6PauLdFe3u2/rp/5JuZP/U9hw7rhf6KPunJq/VnZ8rJUmSJEmS1Gv6/dY5SZIkSZIktYmFJkmSJEmSJNXCQpMkSZIkSZJqYaFJkiRJkiRJtbDQJEmSJEmSpFpYaJIkSZIkSVItLDRJkiRJkiSpFhaaJEmSJEmSVAsLTZIkSZIkSaqFhSZJkiRJkiTVwkKTJEmSJEmSamGhSZIkSZIkSbWw0CRJkiRJkqRaWGiSJEmSJElSLSw0SZIkSZIkqRYWmiRJkiRJklQLC02SJEmSJEmqhYUmSZIkSZIk1cJCkyRJkiRJkmphoUmSJEmSJEm1sNAkSZIkSZKkWlhokiRJkiRJUi0sNEmSJEmSJKkWFpokSZIkSZJUCwtNkiRJkiRJqoWFJkmSJEmSJNXCQpMkSZIkSZJqYaFJkiRJkiRJtbDQJEmSJEmSpFpYaJIkSZIkSVItLDRJkiRJkiSpFhaaJEmSJEmSVAsLTZIkSZIkSaqFhSZJkiRJkiTVwkKTJEmSJEmSamGhSZIkSZIkSbWw0CRJkiRJkqRaWGiSJEmSJElSLSw0SZIkST0uIn4lIh7rdB6SNJb908xjoUm1iIj/FhF/Pib27yLimYj4rxGREXHpmP2Xlvgf1/D674yIeyJiX0TsnOr5JPWPLuif/iAivhURP4yIJyLiD6Z6Tkn9pY5+KjP/Z2a+eYKvd2hE3BYRO8s5Buu6Fkn9pQP902kRsSUi9kTEDyLiyxFxfG0XpLaw0KS6XAqcFRHvBoiIw4HPA6uBp4B/BC4ac8zyEq/DPwPrAH+BkzRWp/unKOc/BlgCXBIRy2o6t6T+0Il+6uvAbwDfm8I5JPW/dvdPxwBrgXnAG4EfAn8+3gHqPhaaVIvMfAb4ELA2Il4LXAF8OzNvLE3uA14TEW8BKF8PL3FK7JiI+GqpXO8tz+eWfcdGxK6I+L/L9usiYjgiLiqvf29m3gzsaM8VS+oVXdA//X+Z+Y3MHMnMx4DbgXe05eIl9YSa+qnBiNjVsL0zIv7fiHiojPi+pfyCSGa+mJmfzMyvAy+14xol9aYO9E93ZuaXM/O5zHwe+DR+buo5FppUm8z8MvAN4EvAqvJodDMvV7uXl+1GP0VVrX4j8HPAj6g6FjJzD/BbwOcj4meAa4FvZuZN9V+JpH7TLf1TRATwK8D2qV+VpH5SQz/VzPuoRlLOB34RWFFHrpJmlg73T7+Kn5t6joUm1e2DwOnARzLzyTH7/htwQUQcAiwr2z+Rmc9k5l9m5vOZ+UPgKuDfNez/G+DLwF3A2cD/M32XIakPdUP/9Me8XLSSpLEm3U+1cH1m/u9SEP8fwC/Vmq2kmaTt/VNE/CLwn3F6lJ5joUm1yszvA/9Ek6pzZn4XGAb+BHh8bAcVEa8pE8p9JyKeA74GHB0RsxqarQVOAm4swzglaUI63T9FxCVUf+07JzNfqOu6JPWPqfRTLTTOv/Q88Lo68pQ087S7f4qINwF3Apdm5v+cbN7qDAtNarebqCaOa3bL22rgzcCpmXkk1TBJqCbSpfxCt7Yc+8HS+UhSXaatf4qI3wLWAIszcxeSNDnj9VOS1Em19U8R8Ubgb4GPlnl41WNmdzoBzTi3ALuA/9Vk3+up5j15NiKOpZportEfAkk1F8plwE0R8SuZ+VJE/BRwKHAI1TQohwP/mpkvTtN1SOo/09U/XUj1F753ZqYLFkiaivH6qYMSEYdRiuXAoeWz0wuZmVM9t6QZqZb+KSLmAHcDn87MP6sjMbWfI5rUVpn5o8z828z8UZPdnwSOoBqSuRX469EdEfF24PeBizLzJeAaql/q1pQmv0r1S+AmXp6o92+m6zok9Z9p7J+uBN4A3BcR+8vDD06SDtoB+qmD9RjV56U5wOby/I01nFfSDFRj//QfgH8D/HHD56b9NaSoNgr/aCFJkiRJkqQ6OKJJkiRJkiRJtbDQJEmSJEmSpFocsNAUEesi4umI+FZD7E8j4h8i4qGI+EpEHN2w7/KIGI6IxyLizIb4khIbjog1DfH5EbGtxG+JiENL/LCyPVz2z6vroiVJkiRJklS/iYxouhFYMia2BTgpM38R+EfgcoCIOBFYBrylHPPZiJhVln3+DHAWcCJwQWkL1aSp12bmm4C9wMoSXwnsLfFrSztJkiRJkiR1qQMWmjLza8CeMbG/ycyRsrkVmFueLwU2ZOYLmfkEMAycUh7DmbmjLDe/AVgaEQGcDtxWjl8PnNdwrvXl+W3A4tJekiRJkiRJXWh2Def4LeCW8nwOVeFp1K4SA3hyTPxUquWen20oWjW2nzN6TGaORMS+0v6fxkvmuOOOy3nz5rXc/8///M+89rWvHf+K2sRcmjOX5noxlwceeOCfMvOn25BST7B/mhxzac5cmrN/mpwD9U+juum9Ho951qcXcoTeyNP+aXIm2j9Bf/0cdFov5NkLOUJv5Fl3/zSlQlNE/BEwAnxxKueZqohYBawCGBgY4GMf+1jLtvv37+d1r3tdu1Ibl7k0Zy7N9WIu73znO78z3blExDrgPcDTmXnSmH2rgY8BP52Z/1RGRV4HnA08D6zIzG+UtsuB/1QOvTIz15f426luIT4C2ARcmpkZEcdSFdnnATuB92Xm3vFynTdvHvfff3/L/UNDQwwODk742qeTuTRnLs31Yi4RMe39Uy85UP80qpve6/GYZ316IUfojTztnyZnov0T9NfPQaf1Qp69kCP0Rp5190+TLjRFxAqqX+4WZ2aW8G7ghIZmc0uMFvFngKMjYnYZ1dTYfvRcuyJiNnBUaf8qmbkWWAuwaNGiHO8b1E1vsrk0Zy7NmUtLNwKfBm5qDEbECcAZwHcbwmcBC8rjVOBzwKmlaHQFsAhI4IGI2FgKR58DPgBsoyo0LQHuBNYAd2Xm1WWBgzXAZdN0jZIkSZLUEyYyGfirRMQS4MPAuZn5fMOujcCysmLcfKpf5u4F7gMWlBXmDqWaMHxjKVDdA5xfjl8O3N5wruXl+fnA3Q0FLUkCms8jV1xL1U819htLgZuyspWq0H08cCawJTP3lOLSFmBJ2XdkZm4t/c9NNJ9HrnF+OUmSJEmasQ44oikivgQMAsdFxC6qv/pfDhwGbCnzc2/NzP+Ymdsj4lbgEapb6i7OzJfKeS4BNgOzgHWZub28xGXAhoi4EngQuKHEbwBujohhql8il9VwvZJmgIhYCuzOzL8fs4bAT+Z+K0bnhRsvvqtJHGAgM58qz78HDNR2AZIkSZLUow5YaMrMC5qEb2gSG21/FXBVk/gmqttOxsZ3UK1KNzb+Y+C9B8pPkhpFxGuAP6S6ba4typxNTUdcjp1DbmhoqOV59u/fP+7+djKX5sylOXORJEnSqDpWnZOkbvJvgfnA6GimucA3IuIUWs8jt5tq5GZjfKjE5zZpD/D9iDg+M58qt9g93SwZ55CbOnNpzlya66ZcJEmSZqJJzdEkSd0qMx/OzJ/JzHmZOY/qdreTM/N7VHO/XRSV04B95fa3zcAZEXFMRBxDNRpqc9n3XEScVlasu4jm88g1zi8nSZIkSTOWhSZJPa3MI/d3wJsjYldErByn+SZgBzAMfB74IEBm7gE+SrVwwX3AR0qM0uYL5ZhvU604B3A18O6IeBx4V9mWJEmSpBnNW+ck9bQW88g17p/X8DyBi1u0WwesaxK/HzipSfwZYPFBpitJkiRJfc0RTZIkSZIkSaqFhSZJkiRJkiTVwlvnpC4yb80dLfetXjjCihb7d159znSlJEmA/ZPULcb7tzge/y1K6laT7dduXPLamjNRXRzRJEmSJEmSpFpYaJIkSZIkSVItLDRJkiRJkiSpFhaaJEmSJEmSVAsnA5ckSZL6nJPtSpLaxRFNkiRJkiRJqoWFJkmSJEmSJNXCQpMkSZIkSZJqYaFJkiRJkiRJtbDQJEmSJEmSpFq46pwkSZIkSZqSya5uqf7jiCZJkqQ2i4hZEfFgRHy1bM+PiG0RMRwRt0TEoSV+WNkeLvvnNZzj8hJ/LCLObIgvKbHhiFjT7muTJEkzm4UmSZKk9rsUeLRh+xrg2sx8E7AXWFniK4G9JX5taUdEnAgsA94CLAE+W4pXs4DPAGcBJwIXlLaSJEltYaFJkiSpjSJiLnAO8IWyHcDpwG2lyXrgvPJ8admm7F9c2i8FNmTmC5n5BDAMnFIew5m5IzNfBDaUtpIkSW1hoUmSJKm9Pgl8GPjXsv0G4NnMHCnbu4A55fkc4EmAsn9faf+T+JhjWsUlSZLawsnAJUmS2iQi3gM8nZkPRMRgh3NZBawCGBgYYGho6IDH7N+/f0LtOq2f81y9cOTAjWrUz9/LduuFHCWpDhaaJEmS2ucdwLkRcTZwOHAkcB1wdETMLqOW5gK7S/vdwAnAroiYDRwFPNMQH9V4TKv4K2TmWmAtwKJFi3JwcPCAyQ8NDTGRdp3Wz3muaPOqTjcueW3ffi/brRdylKQ6eOucJElSm2Tm5Zk5NzPnUU3mfXdmXgjcA5xfmi0Hbi/PN5Ztyv67MzNLfFlZlW4+sAC4F7gPWFBWsTu0vMbGNlyaJEkS4IgmSZKkbnAZsCEirgQeBG4o8RuAmyNiGNhDVTgiM7dHxK3AI8AIcHFmvgQQEZcAm4FZwLrM3N7WK9GEzFtzB6sXjrR9hJIkSdPNQpMkSVIHZOYQMFSe76BaMW5smx8D721x/FXAVU3im4BNNaYqSZI0Yd46J0mSJEmSpFpYaJLUsyJiXUQ8HRHfaoj9aUT8Q0Q8FBFfiYijG/ZdHhHDEfFYRJzZEF9SYsMRsaYhPj8itpX4LWW+E8qcKLeU+LaImNeeK5YkSZpeEbEzIh6OiG9GxP0ldmxEbImIx8vXY0o8IuL68pnooYg4ueE8y0v7xyNieUP87eX8w+XYaP9VSppOFpok9bIbgSVjYluAkzLzF4F/BC4HiIgTqeY2eUs55rMRMSsiZgGfAc4CTgQuKG0BrgGuzcw3AXuBlSW+Ethb4teWdpIkSf3inZn5S5m5qGyvAe7KzAXAXWUbqs9PC8pjFfA5qApTwBXAqVS3BV8xWpwqbT7QcNzYz3KSepyFJkk9KzO/RjU5bmPsb8ry4ABbqZb2BlgKbMjMFzLzCWCY6oPPKcBwZu7IzBeBDcDS8te104HbyvHrgfMazrW+PL8NWOxf4yRJUh9r/Owz9jPRTVnZChwdEccDZwJbMnNPZu6l+kPgkrLvyMzcWlbQvKnhXJL6xAELTS1uTZn2oZOtXkOSDsJvAXeW53OAJxv27SqxVvE3AM82FK1G4684V9m/r7SXJEnqdQn8TUQ8EBGrSmwgM58qz78HDJTnB/v5ak55PjYuqY9MZNW5G4FPU1WbR40Onby6zGeyhmpZ3sahk6dSDYs8tWHo5CKqjuuBiNhYqtujQye3Ua2QsoTqF8NWryFJBxQRf0S15PcXO5zHKqqh5AwMDDA0NNSy7f79+8fd307m0txMzmX1wpGW+waOaL2/3d+vbnqPJKlH/V+ZuTsifgbYEhH/0LgzMzMicjoTOJjPT4164f+AXsgRJpfneJ8VpkM/fy/bre4cD1hoysyvNZnodikwWJ6vp1qa9zIahk4CWyNidOjkIGXoJEBEjA6dHKIMnSzx0aGTd47zGpI0rohYAbwHWFz6I4DdwAkNzeaWGC3iz1AN/55dRi01th89166ImA0cVdq/SmauBdYCLFq0KAcHB1vmPTQ0xHj728lcmpvJuaxYc0fLfasXjvDxh5t/pNh54eA0ZdRcN71HktSLMnN3+fp0RHyFapqB70fE8Zn5VPn97unSvNXnq928/LvcaHyoxOc2aT82hwl/fmrUC/8H9EKOMLk8x/usMB1uXPLavv1etlvdOU52jqZ2DJ1s9RqS1FJELAE+DJybmc837NoILCsrxs2nGnl5L3AfsKCsMHco1YThG0uB6h7g/HL8cuD2hnON3gJ8PnB3Q0FLkiSpJ0XEayPi9aPPgTOAb/HKzz5jPxNdVKZQOQ3YV36H2wycERHHlClQzgA2l33PRcRpZcqUixrOJalPTOTWuXG1Y+jkgV7DW1Omzlya89aU5rrlPYqIL1H9tey4iNhFdYvu5cBhVEO9AbZm5n/MzO0RcSvwCNUtdRdn5kvlPJdQfSCaBazLzO3lJS4DNkTElcCDwA0lfgNwc0QMU01GvmzaL1aSJGn6DQBfKZ+hZgN/kZl/HRH3AbdGxErgO8D7SvtNwNlUi6w8D7wfIDP3RMRHqf6gB/CR0btbgA9STc9yBNWdLKPzaUrqE5MtNLVj6GSr13gVb02ZOnNpzltTmuuW9ygzL2gSvqFJbLT9VcBVTeKbqD4ojY3voBouPjb+Y+C9B5WsJElSlyuffd7aJP4MsLhJPIGLW5xrHbCuSfx+4KQpJyupa0321rl2DJ1s9RqSJEmSJEnqQgcc0dTi1pSrmf6hk61eQ5IkSZIkSV1oIqvONbs1BaZ56GSr4ZmSJEmSJEnqTlOeDFySJEmSJKmdHt69b9w5blvZefU505CNGk12jiZJkiRJkiTpFSw0SZIkSZIkqRYWmiRJkiRJklQLC02SJEmSJEmqhZOBS5IkSZM0bxIT0UqS1M8c0SRJkiRJkqRaWGiSJEmSJElSLSw0SZIkSZIkqRYWmiRJkiRJklQLC02SJEmSJEmqhYUmSZIkSZIk1WJ2pxOQJEmS1J0e3r2PFWvuOKhjdl59zjRlI0nqBY5okiRJkiRJUi0sNEmSJEmSJKkWFpokSZIkSZJUCwtNkiRJkiRJqoWFJkmSJEmSJNXCQpMkSZIkSZJqYaFJkiRJkiRJtbDQJEmSJEmSpFpYaJLU0yJiXUQ8HRHfaogdGxFbIuLx8vWYEo+IuD4ihiPioYg4ueGY5aX94xGxvCH+9oh4uBxzfUTEeK8hSZIkSTPZ7E4nIElTdCPwaeCmhtga4K7MvDoi1pTty4CzgAXlcSrwOeDUiDgWuAJYBCTwQERszMy9pc0HgG3AJmAJcOc4ryFJkiT1rHlr7mD1whFWrLmj06moRzmiSVJPy8yvAXvGhJcC68vz9cB5DfGbsrIVODoijgfOBLZk5p5SXNoCLCn7jszMrZmZVMWs8w7wGpIkSZI0Y1loktSPBjLzqfL8e8BAeT4HeLKh3a4SGy++q0l8vNeQJEmSpBnLW+ck9bXMzIjITr1GRKwCVgEMDAwwNDTU8jz79+8fd387mUtzMzmX1QtHWu4bOKL1/nZ/v7rpPZIkSZqJLDRJ6kffj4jjM/Opcvvb0yW+Gzihod3cEtsNDI6JD5X43Cbtx3uNV8jMtcBagEWLFuXg4GCzZkD1C/l4+9vJXJqbybmMN0/D6oUjfPzh5h8pdl44OE0ZNddN75EkSdJM5K1zkvrRRmB05bjlwO0N8YvK6nOnAfvK7W+bgTMi4piyetwZwOay77mIOK2sNnfRmHM1ew1JkqSeFRGzIuLBiPhq2Z4fEdvKCry3RMShJX5Y2R4u++c1nOPyEn8sIs5siC8pseGymIqkPuSIJkk9LSK+RDUa6biI2EW1etzVwK0RsRL4DvC+0nwTcDYwDDwPvB8gM/dExEeB+0q7j2Tm6ATjH6Ra2e4IqtXm7izxVq8hSepR81xhSQK4FHgUOLJsXwNcm5kbIuLPgJVUq/KuBPZm5psiYllp92sRcSKwDHgL8LPA30bEz5dzfQZ4N9W8l/eVVX4fadeFSWoPC02SelpmXtBi1+ImbRO4uMV51gHrmsTvB05qEn+m2WtIkiT1qoiYC5wDXAX8fhnRfTrw66XJeuCPqQpNS8tzgNuAT5f2S4ENmfkC8EREDAOnlHbDmbmjvNaG0tZCk9RnpnTrXET8XkRsj4hvRcSXIuJwh1ZKkiS1Vj4v3RsRf18+R/2XEvczlKRO+yTwYeBfy/YbgGczc3TFh8YVeH+yam/Zv6+0P9hVfiX1mUmPaIqIOcDvACdm5o8i4laqIZJn49BKSZKkVl4ATs/M/RFxCPD1iLgT+H38DCWpQyLiPcDTmflARAx2OJcJr9rbqBdWHu2FHFcvHBl3RdluMdkcXRH31erOcaq3zs0GjoiIfwFeAzyFQyslSZJaKrfx7i+bh5RH4mcoSZ31DuDciDgbOJxqjqbrgKMjYnYZtdS4Au/oar67ImI2cBTwDK1X+WWc+CsczKq9jXph5dFeyHHFmjvGXVG2W0w2R1fEfbW6c5z0rXOZuRv4GPBdqgLTPuABHFopSZI0rrKq0zeBp4EtwLfxM5SkDsrMyzNzbmbOoxoteXdmXgjcA5xfmo1dzXd0Bd7zS/ss8WXltt/5wALgXqpFVxaU24QPLa+xsQ2XJqnNpnLr3DFUfx2bDzwLfBlYUlNeB5vLhIdWdtOwNXNpbibnMt7Qz/GGhjr8U5J6S2a+BPxSRBwNfAX4hXbnMJlbU3ql/59snu2+TaRfb03pxM9IL/xs9kKOLVwGbIiIK4EHgRtK/Abg5jKacg9V4YjM3F6mVXkEGAEuLn0eEXEJsBmYBazLzO1tvRJJbTGVsXDvAp7IzB8ARMRfUQ237Oqhld00bM1cmpvJuawYZ1nl8YaGOvxTknpTZj4bEfcAv0ybP0NN5taUXun/J5vneP8PT4d+vTWl3Z9LoDd+Nnshx1GZOQQMlec7ePm23MY2Pwbe2+L4q6hWrhsb3wRsqjFVSV1oKqvOfRc4LSJeU+YJWExVtXZopSRJUgsR8dNlJBMRcQTVpN2P4mcoSZLUByb9J5TM3BYRtwHfoBoS+SDVX8XuwKGVkiRJrRwPrI+IWVR/9Ls1M78aEY/gZyhJktTjpjRWNzOvAK4YE3ZopSRJUguZ+RDwtiZxP0NJkqSeN5Vb5yRJkiRJkqSfsNAkSZIkSZKkWlhokiRJkiRJUi0sNEmSJEmSJKkWFpokSZIkSZJUCwtNkiRJkiRJqoWFJkmSJEmSJNXCQpMkSZIkSZJqYaFJkiRJkiRJtbDQJEmSJEmSpFpYaJIkSZIkSVItLDRJkiRJkiSpFrM7nYAkSZJUp4d372PFmjs6nYYkqQvNm8T/DzuvPmcaMulfjmiSJEmSJElSLSw0SZIkSZIkqRYWmiT1pYj4vYjYHhHfiogvRcThETE/IrZFxHBE3BIRh5a2h5Xt4bJ/XsN5Li/xxyLizIb4khIbjog17b9CSZIkSeo+Fpok9Z2ImAP8DrAoM08CZgHLgGuAazPzTcBeYGU5ZCWwt8SvLe2IiBPLcW8BlgCfjYhZETEL+AxwFnAicEFpK0mSJEkzmoUmSf1qNnBERMwGXgM8BZwO3Fb2rwfOK8+Xlm3K/sURESW+ITNfyMwngGHglPIYzswdmfkisKG0lSRJkqQZzUKTpL6TmbuBjwHfpSow7QMeAJ7NzJHSbBcwpzyfAzxZjh0p7d/QGB9zTKu4JEmSJM1oszudgCTVLSKOoRphNB94Fvgy1a1vnchlFbAKYGBggKGhoZZt9+/fP+7+djKX5mZyLqsXjrTcN3BE6/3t/n5103skSZI0E1loktSP3gU8kZk/AIiIvwLeARwdEbPLqKW5wO7SfjdwArCr3Gp3FPBMQ3xU4zGt4q+QmWuBtQCLFi3KwcHBlkkPDQ0x3v52MpfmZnIuK9bc0XLf6oUjfPzh5h8pdl44OE0ZNddN71E/enj3vnF/FlrZefU505CNJEnqRhaaJPWj7wKnRcRrgB8Bi4H7gXuA86nmVFoO3F7abyzbf1f2352ZGREbgb+IiE8APwssAO4FAlgQEfOpCkzLgF9v07VJktTV5k2iGAkWJCWpX1hoktR3MnNbRNwGfAMYAR6kGlV0B7AhIq4ssRvKITcAN0fEMLCHqnBEZm6PiFuBR8p5Ls7MlwAi4hJgM9WKdusyc3u7rk+SJEmSupWFJkl9KTOvAK4YE95BtWLc2LY/Bt7b4jxXAVc1iW8CNk09U0mSJEnqH646J0mSJEmSpFpYaJIkSZIkSVItLDRJkiRJkiSpFs7RJEmSJElSH5rsKpDSVDiiSZIkSZJERBweEfdGxN9HxPaI+C8lPj8itkXEcETcEhGHlvhhZXu47J/XcK7LS/yxiDizIb6kxIYjYk27r1HS9LPQJEmSJEkCeAE4PTPfCvwSsCQiTgOuAa7NzDcBe4GVpf1KjaNN+gAAHvhJREFUYG+JX1vaEREnAsuAtwBLgM9GxKyImAV8BjgLOBG4oLSV1EemVGiKiKMj4raI+IeIeDQifjkijo2ILRHxePl6TGkbEXF9qVw/FBEnN5xneWn/eEQsb4i/PSIeLsdcHxExlXwlSZIkSc1lZX/ZPKQ8EjgduK3E1wPnledLyzZl/+LyO9tSYENmvpCZTwDDwCnlMZyZOzLzRWBDaSupj0x1RNN1wF9n5i8AbwUeBdYAd2XmAuCusg1V1XpBeawCPgcQEccCVwCnUnU8V4wWp0qbDzQct2SK+UqSJEmSWigjj74JPA1sAb4NPJuZI6XJLmBOeT4HeBKg7N8HvKExPuaYVnFJfWTSk4FHxFHArwIrAEpF+sWIWAoMlmbrgSHgMqpK9U2ZmcDWMhrq+NJ2S2buKefdQjVEcwg4MjO3lvhNVJXzOyebsyRJknrHZCexXb2w5kSkGSQzXwJ+KSKOBr4C/EK7c4iIVVSDExgYGGBoaGhCx+3fv3/CbTul3TmuXjhy4EZNDBwx+WPbpZ05TuU9m4k/l1NZdW4+8APgzyPircADwKXAQGY+Vdp8Dxgozw+2qj2nPB8blyRJkiRNo8x8NiLuAX4ZODoiZpdRS3OB3aXZbuAEYFdEzAaOAp5piI9qPKZVvPG11wJrARYtWpSDg4MTynloaIiJtu2Udue4YtIF+xE+/nB3L1Lfzhx3Xjg46WNn4s/lVN6V2cDJwIcyc1tEXMfLt8kB1T2+EZFTSXAiDqbi3U3VRHNpbibnMl5FfryKfbu/X930HkmSJKkeEfHTwL+UItMRwLupJvi+Bzifak6l5cDt5ZCNZfvvyv67y++AG4G/iIhPAD9LNQ3KvUAACyJiPlWBaRnw6+26PkntMZVC0y5gV2ZuK9u3URWavh8Rx2fmU+XWuKfL/lZV7d28fKvdaHyoxOc2af8qB1Px7qZqork0N5NzGe8vDuNV7KdSYZ+MbnqPJEmSVJvjgfVldbifAm7NzK9GxCPAhoi4EngQuKG0vwG4OSKGgT1UhSMyc3tE3Ao8AowAF5db8oiIS4DNwCxgXWZub9/lSWqHSReaMvN7EfFkRLw5Mx8DFlN1JI9QVbWv5tXV7ksiYgPVxN/7SjFqM/AnDROAnwFcnpl7IuK5spzmNuAi4FOTzVeSJEmS1FpmPgS8rUl8B9XCTWPjPwbe2+JcVwFXNYlvAjZNOVlJXWuqNzR+CPhiRBwK7ADeT6l8R8RK4DvA+0rbTcDZVEtbPl/aUgpKHwXuK+0+MjoxOPBB4EbgCKpJwJ0IXJIkSZIkqUtNqdCUmd8EFjXZtbhJ2wQubnGedcC6JvH7gZOmkqMkSZIkSZLa46c6nYAkSZIkSZL6g4UmSZIkSZIk1cJCkyRJkiRJkmphoUmSJEmSJEm1sNAkSZLUJhFxQkTcExGPRMT2iLi0xI+NiC0R8Xj5ekyJR0RcHxHDEfFQRJzccK7lpf3jEbG8If72iHi4HHN9RET7r1SSJM1UFpokSZLaZwRYnZknAqcBF0fEicAa4K7MXADcVbYBzgIWlMcq4HNQFaaAK4BTgVOAK0aLU6XNBxqOW9KG65IkSQIsNEmSJLVNZj6Vmd8oz38IPArMAZYC60uz9cB55flS4KasbAWOjojjgTOBLZm5JzP3AluAJWXfkZm5NTMTuKnhXJIkSdNudqcTkCRJmokiYh7wNmAbMJCZT5Vd3wMGyvM5wJMNh+0qsfHiu5rEO2remjs6nYIkSWoTC02SJEltFhGvA/4S+N3MfK5xGqXMzIjINuSwiup2PAYGBhgaGjrgMQNHwOqFI9Oc2dSZZ33ameNEfgZb2b9//5SOb4deyFGS6mChSVJfioijgS8AJwEJ/BbwGHALMA/YCbwvM/eWiXKvA84GngdWjN7aUibY/U/ltFdm5voSfztwI3AEsAm4tNymIknjiohDqIpMX8zMvyrh70fE8Zn5VLn97ekS3w2c0HD43BLbDQyOiQ+V+Nwm7V8lM9cCawEWLVqUg4ODzZq9wqe+eDsff7j7Pz6uXjhinjVpZ447Lxyc9LFDQ0NM5Ge4k3ohR0mqg3M0SepX1wF/nZm/ALyVah4UJ9uV1FGlsH0D8GhmfqJh10ZgdOW45cDtDfGLyupzpwH7yi12m4EzIuKY0i+dAWwu+56LiNPKa13UcC5JkqRpZ6FJUt+JiKOAX6X6ZY7MfDEzn8XJdiV13juA3wROj4hvlsfZwNXAuyPiceBdZRuqEZM7gGHg88AHATJzD/BR4L7y+EiJUdp8oRzzbeDOdlyYJEkSeOucpP40H/gB8OcR8VbgAeBS+nyyXUndLzO/DkSL3YubtE/g4hbnWgesaxK/n+q2YUmSpLaz0CSpH80GTgY+lJnbIuI6Xr5NDujOyXa7aZJQc2luJucy3mTA400W3O7vVze9R5IkSTORhSZJ/WgXsCszt5Xt26gKTV092W43TRJqLs3N5FxWjLM8/XiTBU9lct/J6Kb3SJIkaSZyjiZJfSczvwc8GRFvLqHFwCM42a4kSZIkTStHNEnqVx8CvhgRh1JNpPt+quL6rRGxEvgO8L7SdhNwNtXEuc+XtmTmnogYnWwXXj3Z7o3AEVQT7TrZriRJkqQZz0KTpL6Umd8EFjXZ5WS7kiRJkjRNvHVOkiRJkiRJtbDQJEmSJEmSpFpYaJIkSZIkSVItLDRJkiRJkiSpFk4GLk2DeWvu6HQKkiRJkiS1nSOaJEmSJEmSVAsLTZIkSZIkSaqFhSZJkiRJkiTVwjmaJEmSJEmSWpjsHLw7rz6n5kx6gyOaJEmSJEmSVAsLTZIkSZIkSaqFhSZJkiRJEhFxQkTcExGPRMT2iLi0xI+NiC0R8Xj5ekyJR0RcHxHDEfFQRJzccK7lpf3jEbG8If72iHi4HHN9RET7r1TSdJryHE0RMQu4H9idme+JiPnABuANwAPAb2bmixFxGHAT8HbgGeDXMnNnOcflwErgJeB3MnNziS8BrgNmAV/IzKunmq8kSZKk7uMcKF1hBFidmd+IiNcDD0TEFmAFcFdmXh0Ra4A1wGXAWcCC8jgV+BxwakQcC1wBLAKynGdjZu4tbT4AbAM2AUuAO9t4jT1psv8+pE6oYzLwS4FHgSPL9jXAtZm5ISL+jKqA9LnydW9mvikilpV2vxYRJwLLgLcAPwv8bUT8fDnXZ4B3A7uA+0rn9EgNOUsT8vDufaywU5ckSdIMkJlPAU+V5z+MiEeBOcBSYLA0Ww8MURWalgI3ZWYCWyPi6Ig4vrTdkpl7AEqxaklEDAFHZubWEr8JOA8LTVJfmdKtcxExFzgH+ELZDuB04LbSZD1VxwFVJ7S+PL8NWFzaLwU2ZOYLmfkEMAycUh7DmbkjM1+kGiW1dCr5SpIkSZIOLCLmAW+jGnk0UIpQAN8DBsrzOcCTDYftKrHx4ruaxCX1kamOaPok8GHg9WX7DcCzmTlSths7jp90Npk5EhH7Svs5wNaGczYeM7ZzOnWK+UqSJEmSxhERrwP+EvjdzHyucRqlzMyIyGl+/VXAKoCBgQGGhoYmdNz+/fsn3LZTJpvj6oUjB25Uo4Ej2v+aB6sXchwaGurrn8tWJl1oioj3AE9n5gMRMVhbRpPLZcIdUTe9yebSXDfl0k2d13i5tPv71U3vkSRJkuoTEYdQFZm+mJl/VcLfj4jjM/Opcmvc0yW+Gzih4fC5Jbabl2+1G40PlfjcJu1fITPXAmsBFi1alIODg2ObNDU0NMRE23bKZHNs93QeqxeO8PGH65hpZ/r0Qo47Lxzs65/LVqbyrrwDODcizgYOp5qj6Trg6IiYXUY1NXYco53QroiYDRxFNSl4q86JceKvcDAdUTe9yebS3HTlMpkJ9FYvpGs6r/E60p0XDrY1l276eZF0cJxMVJLUSpna5Abg0cz8RMOujcBy4Ory9faG+CURsYHq7pN9pRi1GfiT0dXpgDOAyzNzT0Q8FxGnUd2SdxHwqWm/MEltNek5mjLz8sycm5nzqCbzvjszLwTuAc4vzcZ2QqPLWp5f2meJL4uIw8qKdQuAe4H7gAURMT8iDi2vsXGy+UqSJEmSxvUO4DeB0yPim+VxNlWB6d0R8TjwrrIN1apxO6jm2f088EGAMgn4R6l+p7sP+MjoxOClzRfKMd/GicClvjMdQzUuAzZExJXAg1QVccrXmyNiGNhDVTgiM7dHxK3AI1TLaV6cmS8BRMQlwGZgFrAuM7dPQ76SJEmSNONl5teBaLF7cZP2CVzc4lzrgHVN4vcDJ00hTUldrpZCU2YOUd1zS2buoFoxbmybHwPvbXH8VcBVTeKbqKrkkiRJkiRJ6nLdMfmMdBCcX0SSJEmSpO406TmaJEmSJEmSpEYWmiRJkiRJklQLC02SJEmSJEmqhYUmSX0rImZFxIMR8dWyPT8itkXEcETcEhGHlvhhZXu47J/XcI7LS/yxiDizIb6kxIYjYk27r02SJEmSupGFJkn97FLg0Ybta4BrM/NNwF5gZYmvBPaW+LWlHRFxIrAMeAuwBPhsKV7NAj4DnAWcCFxQ2kqSJEnSjGahSVJfioi5wDnAF8p2AKcDt5Um64HzyvOlZZuyf3FpvxTYkJkvZOYTwDBwSnkMZ+aOzHwR2FDaSpIkSdKMZqFJUr/6JPBh4F/L9huAZzNzpGzvAuaU53OAJwHK/n2l/U/iY45pFZckSZKkGW12pxPQzDVvzR1N46sXjrCixT5pIiLiPcDTmflARAx2OJdVwCqAgYEBhoaGWrbdv3//uPvbyVya64dcVi8cOXCjgzRwROvztvv71U3vkSRJ0kxkoUlSP3oHcG5EnA0cDhwJXAccHRGzy6ilucDu0n43cAKwKyJmA0cBzzTERzUe0yr+Cpm5FlgLsGjRohwcHGyZ9NDQEOPtbydzaa6bcvnUF2/n41//50kcWf9//asXjvDxh5ufd+eFg7W/3ni66T2SJEmaibx1TlLfyczLM3NuZs6jmsz77sy8ELgHOL80Ww7cXp5vLNuU/XdnZpb4srIq3XxgAXAvcB+woKxid2h5jY1tuDRJkiRJ6mqOaJI0k1wGbIiIK4EHgRtK/Abg5ogYBvZQFY7IzO0RcSvwCDACXJyZLwFExCXAZmAWsC4zt7f1SiRJkiSpC1loktTXMnMIGCrPd1CtGDe2zY+B97Y4/v9v7/5j7a7vOo4/3w6nqEzY0Kah3bol3SJSzVhD6zTahY0VMHaJhkDYaBek2QAzY/+wOhMWcKb+wRIhBFOxoSzID38kNKGkQcINcVsZyJACC1KwC60djRZhDVHX+PaP87nj9Pbc23PP+X7P9/s9fT6SE875nM8599XvPX3R+7nfH18FvjpgfDewu8KokiRJktR5HjonSZIkSZKkSrhHkyRJDZvvKpwL2bKqhiCSJEnSmNyjSZIkSZIkSZVwoUmSJEmSJEmV8NA5SZIkSVLn7Tv0JpsWeTj6gW2X15RmsFEySl3jQpMkSRUZ5VxLOv1ExA7gN4EjmXlBGXsv8ACwAjgAXJGZb0REAH8BXAa8DWzKzGfKazYCf1Le9k8zc2cZ/xhwN3AmvatjfikzcyJ/OEmSdNrz0DlJkqTJuhtYP2dsK/BYZq4EHiuPAS4FVpbbZuBO+NHC1E3AGuAi4KaIOKe85k7gur7Xzf1akiRJtXGhSZIkaYIy8wng6JzhDcDOcn8n8Jm+8XuyZy9wdkQsBT4NPJqZRzPzDeBRYH157j2ZubfsxXRP33tJkiTVzoUmSZKk5i3JzMPl/veBJeX+ecBrffMOlrGFxg8OGJckSZoIz9EkSZLUIpmZEVH7OZUiYjO9w/FYsmQJMzMzp3zNkjNhy6rjNScbnzmr04WMMzMzHDt2bKjPcJO6kFGSquBCkyRJUvNej4ilmXm4HP52pIwfApb3zVtWxg4B6+aMz5TxZQPmnyQztwPbAVavXp3r1q0bNO0Et9/7ELfua/8/H7esOm7OinQh44Gr1zEzM8Mwn+EmdSGjJFXBQ+ckSZKatwvYWO5vBB7qG78metYCb5ZD7PYAl0TEOeUk4JcAe8pzb0XE2nLFumv63kuSJKl27f71hDrBy3lLmjbz9dqWVcfZZOdpTBFxH729kc6NiIP0rh63DXgwIq4FvgdcUabvBi4D9gNvA58HyMyjEXEL8FSZd3Nmzp5g/Hp6V7Y7E3ik3CRJkibChSZJkqQJysyr5nnq4gFzE7hhnvfZAewYMP40cME4GSVJ0vhWbH14pF9UHth2eU2JJsND5yRJkiRJklQJ92iSJEmS1Fmn6x4DktRW7tEkSZIkSZKkSrjQJEmSJEkiInZExJGIeL5v7L0R8WhEvFz+e04Zj4i4LSL2R8RzEXFh32s2lvkvR8TGvvGPRcS+8prbytUxJU2ZkReaImJ5RDweES9GxAsR8aUybhFJkiRJUvfcDayfM7YVeCwzVwKPlccAlwIry20zcCf0fh6kdzXNNcBFwE2zPxOWOdf1vW7u15I0BcY5R9NxYEtmPhMRZwH/HBGPApvoFdG2iNhKr4j+kBOLaA29klnTV0SrgSzvsysz3+CdInqS3uV91+MleiVJkiSpcpn5RESsmDO8AVhX7u8EZuj9fLcBuKdcHXNvRJwdEUvL3Ecz8yhA+RlxfUTMAO/JzL1l/B7gM3T057sVizwn2KwtqyoOIrXQyHs0ZebhzHym3P8B8F3gPHqFs7NM20mvPKCviEq5zBbRpylFVBaXZotoKaWISnnd0/dekiRJkqT6LcnMw+X+94El5f55wGt98w6WsYXGDw4YlzRlKrnqXFn1/ii9PY8sIkmSJEmaMpmZEZF1f52I2EzvcDyWLFnCzMzMUK9bciZsWXV8UV9r2Peea7FfZ9YoGZvQhZxdyAiT/VyO6tixY5V+zbEXmiLiZ4C/B34/M9/qP41SG4uo6g04jmnJUvVf7jYVRleydL2IJEmS1FqvR8TSzDxcjjo5UsYPAcv75i0rY4d451C72fGZMr5swPyTZOZ2YDvA6tWrc926dYOmneT2ex/i1n2L+xH3wNXDvfdcm0Y+dO74ojM2oQs5u5ARRss56udyVDMzMwz792wYY31XIuLH6S0y3ZuZ/1CGW11EVW/AcUxLllFLdj5tKoyuZOl6EUmSJKm1dgEbgW3lvw/1jd8YEffTOwfvm+VnwD3An/WdAPwS4I8y82hEvBURa+kdCXMNcPsk/yCSJmOcq84F8NfAdzPza31PzRYRnFxE15Srz62lFBGwB7gkIs4pZXQJsKc891ZErC1f65q+95KkeXlVTEmSpMWLiPuAbwEfiYiDEXEtvQWmT0XEy8Any2PoXazpVWA/8FfA9QDlJOC3AE+V282zJwYvc+4qr3mFjp4IXNLCxtlV41eBzwH7IuLZMvbH9IrnwVJK3wOuKM/tBi6jVypvA5+HXhFFxGwRwclFdDdwJr0SsogkDcOrYkqSJC1SZl41z1MXD5ibwA3zvM8OYMeA8aeBC8bJKKn9Rl5oysx/Aub7Db5FJKkxZY/Iw+X+DyKi/6qY68o0L897Ghj10sOSJEmSRjPyoXOS1AVeFVOSJEmSJqcdZzmWpBp4VczRTUsWr4o5GV4VU5LUVe79LFXPhSb9iCWraeJVMcczLVm8KuZkeFVMSZIkzfLQOUlTx6tiSpIkSVIz2vGrUEmqllfFlCRJkqQGuNAkaep4VUxJkiRJaoaHzkmSJEmSJKkSLjRJkiRJkiSpEi40SZIkSZIkqRIuNEmSJEmSJKkSLjRJkiRJkiSpEl51TpLUeiu2Ptx0BEmSJElDcI8mSZIkSZIkVcKFJkmSJEmSJFXCQ+ckSZIknXZGPSz7wLbLK04iSSfqej+50DSlFvvB3LLqOH4cJEmSJEnSODx0TpIkSZIkSZVwoUmSJEmSJEmVcKFJkiRJkiRJlfCkPJKkifIccpIkSdL0co8mSZIkSZIkVcKFJkmSJEmSJFXChSZJkiRJkiRVwpNetNxiz2UiSZIkSZLUFPdokiRJkiRJUiXco0mSNBL3uJQkSZI0l3s0SZIkSZIkqRLu0TQhg37zv2XVcTa5R4AkSZIkSZoS7tEkSZIkSZKkSrhHkySd5tzjUpKk4Y16jsK71/90xUkk6URt6afW79EUEesj4qWI2B8RW5vOI0mz7CdJbWU/SWozO0qabq3eoyki3gXcAXwKOAg8FRG7MvPFpjJ5lSVJ0M5+kiSwnyS1mx0lTb9WLzQBFwH7M/NVgIi4H9gAWEKSmta6fnIhXFLRun6SpD52lDTl2r7QdB7wWt/jg8CaKt7YH8gkjcl+ktRWtfWTJFXAjpKmXGRm0xnmFRG/A6zPzN8tjz8HrMnMG+fM2wxsLg8/Ary0wNueC/xHDXFHYZbBzDJYF7N8IDN/ru4wTbCfJsosg5llMPupnn6a1abv9ULMWZ0uZIRu5Dzt+wmG66gR+wmm63PQtC7k7EJG6EbOSvup7Xs0HQKW9z1eVsZOkJnbge3DvGFEPJ2Zq6uJNx6zDGaWwczSOvbThJhlMLMM1qYsDaq8n2Z1ZfuaszpdyAjdyNmFjBNyyo4apZ+gG9u4CxmhGzm7kBG6kbPqjG2/6txTwMqI+GBEvBu4EtjVcCZJAvtJUnvZT5LazI6Splyr92jKzOMRcSOwB3gXsCMzX2g4liTZT5Jay36S1GZ2lDT9Wr3QBJCZu4HdFb7lonfBrJFZBjPLYGZpGftpYswymFkGa1OWxtTQT7O6sn3NWZ0uZIRu5OxCxok4zTuqCxmhGzm7kBG6kbPSjK0+GbgkSZIkSZK6o+3naJIkSZIkSVJHTO1CU0Ssj4iXImJ/RGwd8PxPRMQD5fknI2JFg1n+ICJejIjnIuKxiPhAU1n65v12RGRE1HZ2/GGyRMQVZdu8EBF/01SWiHh/RDweEd8p36fLasqxIyKORMTz8zwfEXFbyflcRFxYR44hs1xdMuyLiG9GxC/XlWXa2E+jZembZz+d+Lz9dPLz9tOY2tRTC2lTh42asW9e7d22kDb13qgZJ9WHp8jYmq6cVvbT5DL2zbOfhmBH9cnMqbvRO6ncK8CHgHcD/wKcP2fO9cBflvtXAg80mOUTwE+V+19sMkuZdxbwBLAXWN3gdlkJfAc4pzz++QazbAe+WO6fDxyoKcuvAxcCz8/z/GXAI0AAa4En68gxZJaP931vLq0zyzTd7KfRs5R59pP9ZD/VfGtTT1WQcyIdNk7GMq/2bqtgW06k98bMOJE+PEXO1nTlNN7sp8lmLPPsp+pynjYdNa17NF0E7M/MVzPzf4H7gQ1z5mwAdpb7fwdcHBHRRJbMfDwz3y4P9wLLasgxVJbiFuDPgf+uKcewWa4D7sjMNwAy80iDWRJ4T7n/s8C/1xEkM58Aji4wZQNwT/bsBc6OiKVNZMnMb85+b6j3cztt7KcRsxT2k/1kP9WvTT21kDZ12MgZi0l020La1HvjZJxIHy6kTV05peynCWYs7Kfh2FF9pnWh6Tzgtb7HB8vYwDmZeRx4E3hfQ1n6XUtvBbEOp8xSdo1bnpkP15Rh6CzAh4EPR8Q3ImJvRKxvMMtXgM9GxEF6V8j4vZqynMpiP0+TUufndtrYTyNmsZ/mzfIV7KeF2E+L16aeWkibOmw+beq2hbSp9+bTpT5cSFu7sivsp+rYT9Wyo/qcUVkcjS0iPgusBn6joa//Y8DXgE1NfP0BzqC3G+Q6eiv8T0TEqsz8rwayXAXcnZm3RsSvAF+PiAsy8/8ayNIqEfEJev9z/LWms6g+9tNJ7KcOsJ80q+kOm08Lu20hbeq9+diH6hz7qRJd6Cc4jTpqWvdoOgQs73u8rIwNnBMRZ9Dbde0/G8pCRHwS+DLwW5n5PzXkGCbLWcAFwExEHKB3TOaumk76Nsx2OQjsyswfZua/Af9Kr0CayHIt8CBAZn4L+Eng3BqynMpQn6dJiYhfAu4CNmRmHX9/ppH9NFoW+2n+LPbTAPbTWNrUUwtpU4fNp03dtpA29d58utSHC2lVV3aQ/VQd+6ladlS/nPDJpyZxo7ei+SrwQd45EdcvzplzAyeeJO7BBrN8lN6Jw1Y2vV3mzJ+hvpPtDrNd1gM7y/1z6e3C976GsjwCbCr3f4He8bRR07ZZwfwnZ7ucE0/O9u2aPzMLZXk/sB/4eJ0Zpu1mP42eZc58++mdOfbTyc/ZT+Nt29b0VAU5J9Jh42ScM7+2bqtgW06k98bMOLE+PEXW1nTltN3sp8lmnDPffho/52nTURP9A014411GbyXzFeDLZexmeqvF0Fs9/Ft6/xD9NvChBrP8I/A68Gy57Woqy5y5tZbJENsl6O2u+SKwD7iywSznA98ohfEscElNOe4DDgM/pLcyfy3wBeALfdvkjpJzX83fn1NluQt4o+9z+3RdWabtZj+NlmXOXPvJfrKfary1qafGzDmxDhs145y5tXbbmNtyYr03RsaJ9OEpMramK6f1Zj9NLuOcufbT+DlPm46K8maSJEmSJEnSWKb1HE2SJEmSJEmaMBeaJEmSJEmSVAkXmiRJkiRJklQJF5okSZIkSZJUCReaJEmSJEmSVAkXmiRJkiRJklQJF5okSZIkSZJUCReaJEmSJEmSVIn/Bx8oFuBG1LrTAAAAAElFTkSuQmCC\n"},"metadata":{}}]},{"metadata":{},"cell_type":"markdown","source":"# Test Dataset"},{"metadata":{"trusted":true},"cell_type":"code","source":"df_sample.head()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"'PredictionString' looks odd for me :(\n\nLet' have a look!"},{"metadata":{"trusted":true},"cell_type":"code","source":"values_what = df_sample[df_sample['ImageId']=='b4c3b52a8723d431']['PredictionString'].values\nvalues = str(values_what)[2:-2].split(' ')\n\nprint('confidence: ', values[0])\n\nprint('label 1: ', values[1])\nprint('XMin1: ', values[2])\nprint('YMin1: ', values[3])\nprint('XMax1: ', values[4])\nprint('YMax1: ', values[5])\nprint('Label 2: ', values[6])\nprint('XMin2: ', values[7])\nprint('YMin2: ', values[8])\nprint('XMax2: ', values[9])\nprint('YMax2: ', values[10])\nprint('Relation Label: ', values[11])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"print('confidence: ', values[12])\nprint('label 1: ', values[13])\nprint('XMin1: ', values[14])\nprint('YMin1: ', values[15])\nprint('XMax1: ', values[16])\nprint('YMax1: ', values[17])\nprint('Label 2: ', values[18])\nprint('XMin2: ', values[19])\nprint('YMin2: ', values[20])\nprint('XMax2: ', values[21])\nprint('YMax2: ', values[22])\nprint('Relation Label: ', values[23])","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Let's have a look 10 random sample images"},{"metadata":{"trusted":true},"cell_type":"code","source":"image_filenames = os.listdir(\"../input/open-images-2019-visual-relationship/test\")\n\nimport random\nfor i in range(10):\n    index = random.randrange(len(image_filenames))\n    path = \"../input/open-images-2019-visual-relationship/test\" + \"/\" + image_filenames[index]\n    src_img = cv2.imread(path)\n    fig=plt.figure(figsize=(18, 16), dpi= 80, facecolor='w', edgecolor='k')\n    plt.imshow(cv2.cvtColor(src_img, cv2.COLOR_BGR2RGB))\n    plt.show()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"display_name":"Python 3","language":"python","name":"python3"},"language_info":{"name":"python","version":"3.6.4","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat":4,"nbformat_minor":1}