{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import os\nfrom os import listdir\nimport pandas as pd\nimport numpy as np\nimport matplotlib.pyplot as plt\n%matplotlib inline\nimport pydicom\nfrom tqdm import tqdm\n\nfrom PIL import Image","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:56:37.409458Z","iopub.execute_input":"2022-01-19T02:56:37.409777Z","iopub.status.idle":"2022-01-19T02:56:37.470297Z","shell.execute_reply.started":"2022-01-19T02:56:37.409717Z","shell.execute_reply":"2022-01-19T02:56:37.469527Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"Path = \"../input/osic-pulmonary-fibrosis-progression\"\ndf_train= pd.read_csv(f\"{Path}/train.csv\")\ndf_train.drop_duplicates(keep=False, inplace=True,subset=['Patient','Weeks'])\ndf_test= pd.read_csv(f\"{Path}/test.csv\")\nPatient_list= df_train[\"Patient\"].unique()","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:56:37.47478Z","iopub.execute_input":"2022-01-19T02:56:37.475398Z","iopub.status.idle":"2022-01-19T02:56:37.493671Z","shell.execute_reply.started":"2022-01-19T02:56:37.475358Z","shell.execute_reply":"2022-01-19T02:56:37.492958Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def getDcms(patient,type):\n    image_dir = f\"{Path}/{type}/{patient}\"\n    fig=plt.figure(figsize=(10,10))\n    image_list = os.listdir(image_dir)\n    columns = int(np.sqrt(len(image_list)))\n    rows = columns+1\n    dcms=[]\n    for i in tqdm(range(1, len(image_list) +1)):\n        ds = pydicom.dcmread(image_dir + \"/\" + str(i) + \".dcm\")\n        dcms.append(ds.pixel_array)\n        fig.add_subplot(rows,columns,i)\n        plt.imshow(ds.pixel_array, cmap=plt.cm.bone)\n    return dcms\n\nlist=getDcms(Patient_list[50],\"train\" )","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:56:39.00897Z","iopub.execute_input":"2022-01-19T02:56:39.009551Z","iopub.status.idle":"2022-01-19T02:56:50.546979Z","shell.execute_reply.started":"2022-01-19T02:56:39.009508Z","shell.execute_reply":"2022-01-19T02:56:50.546312Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df_train","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:56:50.57102Z","iopub.execute_input":"2022-01-19T02:56:50.57128Z","iopub.status.idle":"2022-01-19T02:56:50.588255Z","shell.execute_reply.started":"2022-01-19T02:56:50.571247Z","shell.execute_reply":"2022-01-19T02:56:50.587607Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def get_Weeks_Dcms(type):\n    fig=plt.figure(figsize=(10,10))\n    dcms=[]\n    plist=[]\n    wlist=[]\n    for p in tqdm(Patient_list):\n        image_dir = f\"{Path}/{type}/{p}\"\n        week=df_train[df_train[\"Patient\"]==p][\"Weeks\"]\n        for w in week:\n            try:\n                ds = pydicom.dcmread(image_dir + \"/\" + str(w) + \".dcm\")\n                dcms.append(ds.pixel_array)\n                plist.append(p)\n                wlist.append(w)\n            except:\n                pass\n    data = pd.DataFrame({\"Patient\":plist,\"Weeks\":wlist})        \n    return dcms,data","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:56:53.377192Z","iopub.execute_input":"2022-01-19T02:56:53.377835Z","iopub.status.idle":"2022-01-19T02:56:53.38429Z","shell.execute_reply.started":"2022-01-19T02:56:53.377796Z","shell.execute_reply":"2022-01-19T02:56:53.38361Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"x_train,train_data=get_Weeks_Dcms(\"train\")","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:56:55.589331Z","iopub.execute_input":"2022-01-19T02:56:55.589835Z","iopub.status.idle":"2022-01-19T02:56:59.054651Z","shell.execute_reply.started":"2022-01-19T02:56:55.589794Z","shell.execute_reply":"2022-01-19T02:56:59.053935Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for i in range(0,len(x_train)):\n    im = Image.fromarray(x_train[i])\n    im = im.resize((128,128),resample=Image.NEAREST) \n    x_train[i] = np.array(im).reshape((128,128,1))\n    #x_train[i] = np.array(im)\n    \nx_train=np.array(x_train)\ntrain_data = pd.merge(train_data,df_train, how=\"left\", on=['Patient',\"Weeks\"])","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:56:59.056255Z","iopub.execute_input":"2022-01-19T02:56:59.056663Z","iopub.status.idle":"2022-01-19T02:57:00.411191Z","shell.execute_reply.started":"2022-01-19T02:56:59.056623Z","shell.execute_reply":"2022-01-19T02:57:00.410446Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_data[\"Sex\"]=train_data[\"Sex\"].astype(\"category\").cat.codes\ntrain_data[\"SmokingStatus\"]=train_data[\"SmokingStatus\"].astype(\"category\").cat.codes","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:57:02.870586Z","iopub.execute_input":"2022-01-19T02:57:02.871194Z","iopub.status.idle":"2022-01-19T02:57:02.880375Z","shell.execute_reply.started":"2022-01-19T02:57:02.871152Z","shell.execute_reply":"2022-01-19T02:57:02.878934Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"<h3><font color=red>Min-Max Normalization </font> <h3>","metadata":{}},{"cell_type":"code","source":"train_data[\"Weeks\"]=(train_data['Weeks'] - train_data['Weeks'].min() ) / ( train_data['Weeks'].max() - train_data['Weeks'].min() )\ntrain_data[\"Percent\"]=(train_data['Percent'] - train_data['Percent'].min() ) / ( train_data['Age'].max() - train_data['Percent'].min() )\ntrain_data[\"Age\"]=(train_data['Age'] - train_data['Age'].min() ) / ( train_data['Age'].max() - train_data['Age'].min() )","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:57:04.946916Z","iopub.execute_input":"2022-01-19T02:57:04.94743Z","iopub.status.idle":"2022-01-19T02:57:04.95623Z","shell.execute_reply.started":"2022-01-19T02:57:04.947386Z","shell.execute_reply":"2022-01-19T02:57:04.955365Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_data","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:57:09.622882Z","iopub.execute_input":"2022-01-19T02:57:09.62349Z","iopub.status.idle":"2022-01-19T02:57:09.642307Z","shell.execute_reply.started":"2022-01-19T02:57:09.623452Z","shell.execute_reply":"2022-01-19T02:57:09.641573Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"y=train_data[\"FVC\"].values\nfeature=train_data[['Weeks','Percent','Age','Sex',\"SmokingStatus\"]].values","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:57:11.725096Z","iopub.execute_input":"2022-01-19T02:57:11.725365Z","iopub.status.idle":"2022-01-19T02:57:11.732091Z","shell.execute_reply.started":"2022-01-19T02:57:11.725335Z","shell.execute_reply":"2022-01-19T02:57:11.731155Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":" <h3><font color=red>Min-Max Normalization </font> <h3>","metadata":{}},{"cell_type":"code","source":"x_min = np.min(x_train)\nx_max = np.max(x_train)\nxs = x_train - x_min / (x_max - x_min)","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:57:13.475929Z","iopub.execute_input":"2022-01-19T02:57:13.476203Z","iopub.status.idle":"2022-01-19T02:57:13.541173Z","shell.execute_reply.started":"2022-01-19T02:57:13.476172Z","shell.execute_reply":"2022-01-19T02:57:13.540395Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"y.shape,feature.shape,xs.shape,x_train.shape","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:57:15.124132Z","iopub.execute_input":"2022-01-19T02:57:15.124755Z","iopub.status.idle":"2022-01-19T02:57:15.131524Z","shell.execute_reply.started":"2022-01-19T02:57:15.124694Z","shell.execute_reply":"2022-01-19T02:57:15.13076Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import tensorflow as tf\nimport tensorflow.keras.backend as K\nimport tensorflow.keras.layers as L\nimport tensorflow.keras.models as M","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:57:16.429344Z","iopub.execute_input":"2022-01-19T02:57:16.429881Z","iopub.status.idle":"2022-01-19T02:57:18.903067Z","shell.execute_reply.started":"2022-01-19T02:57:16.429842Z","shell.execute_reply":"2022-01-19T02:57:18.897621Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"![Laplace.png](attachment:d0b8a979-fffc-40a4-ba41-24888592cd56.png)","metadata":{},"attachments":{"d0b8a979-fffc-40a4-ba41-24888592cd56.png":{"image/png":"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"}}},{"cell_type":"markdown","source":"![metric.png](attachment:e1028f31-3fec-45bd-a6bf-a3378e086d2c.png)","metadata":{},"attachments":{"e1028f31-3fec-45bd-a6bf-a3378e086d2c.png":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAoEAAADUCAYAAADjj/nVAAAgAElEQVR4Ae3dBbh1Rdk3cBVUEBVEKQFpBJUO6e4SQWkQDFpCSgkVRJCQBkFKQlQQXkLpkBQUpEFCQlQQQSVExZjv+s37zvnWWWfXqefss899X9d+9nn2mjVr5r9mzfrPXfOWFBIIBAKBQCAQCAQCgUAgMOEQeMuE63F0OBAIBAKBQCAQCAQCgUAgBQmMQRAIBAKBQCAQCAQCgcAERCBI4AS86dHlQCAQCAQCgUAgEAgEggTGGAgEAoFAIBAIBAKBQGACIhAkcALe9OhyIBAIBAKBQCAQCAQCQQJjDAQCgUAgEAgEAoFAIDABEQgSOAFvenQ5EAgEAoFAIBAIBAKBIIExBgKBQCAQCAQCgUAgEJiACAQJnIA3PbocCAQCgUAgEAgEAoFAkMAYA4FAIBAIBAKBQCAQCExABIIETsCbHl0OBAKBQCAQCAQCgUAgSGCMgUAgEAgEAoFAIBAIBCYgAkECJ+BNjy4HAoFAIBAIBAKBQCAQJDDGQCAQCAQCgUAgEAgEAhMQgSCBE/CmR5cDgUAgEAgEAoFAIBAIEhhjIBAIBAKBQCAQCAQCgQmIQJDACXjTo8uBQCAQCAQCgUAgEAgECYwxEAgEAoFAIBAIBAKBwAREIEjgBLzp0eVAIBAIBAKBQCAQCASCBMYYCAQCgUAgEAgEAoFAYAIiECRwAt706HIgEAiMfwQOOOCAtOeee6b//ve/478zg+jBqaeemrbccst01VVXTbi+DwKmKBoIdIRAkMCOYIpCgUAgEAh0DwIHH3xwmnvuudPf//73MW/UJptsklZcccV07LHHpltuuSU9+OCD6bHHHhvwefzxx9NTTz01oL2vvvpqOvHEE9NXvvKVtP/++6cLLrhgQJn6DzvssENaeuml0/e+9736ofh/IBAIDAKBIIGDACuKBgKBQCAw1gjcdddd6e1vf3v69a9/PdZNyddHAKeccso01VRTpfe+971pmmmmafh53/vel6affvp+2rtLL700zTbbbGnWWWdNX/va19KXv/zlNO2006aFFloo3XfffU37949//CN97nOfS2ussUa68sorm5aLA4FAINAagSCBrfGJo4FAIBAIdBUCCOBWW23VNW1adNFF01vf+tb0lre8peXnXe96VzbjFvP1jTfemEnhggsumB555JH0l7/8Jf35z39Ot912W0IYP/zhD6ff/e53TfuJ/K2++uppm222Sa+88krTcnEgEAgEmiMQJLA5NnEkEAgEAoGuQoC2DAnsFvn973+f5p9//vTpT386HXbYYdkkzCxc/ey1115p9tlnTx//+MfTs88+29f05ZZbLpPG73//++k///lP3+///Oc/05e+9KU02WSTpZ133rmf5rCv0P/98fWvfz0Tyd13371+KP4fCAQCHSAQJLADkKJIIBAIBALdgABt20c+8pFuaEpuw2WXXZZJ4CWXXJJefPHF9PLLL/f7/OY3v0nHH398+tjHPpZ+9rOf9RG6O+64I5uQ9Yf2ry5XXHFFJogf+MAH0l//+tf64b7///CHP0yLLbZYWm211UIb2IdK/BEIdI5AkMDOsYqSgUAgEAiMGQI333xzJka+u0W+9a1vpQ022CDde++9A5r05ptv5gjeNddcMynHj6+I/yOA73znO/uIYTnmWxCJ48zMCGYzUaegkg996ENZE9msXPweCAQCjREIEtgYl/g1EAgExhECr732WkI2EIdZZpklXX/99bn1Tz75ZPYZ44/m2NRTT52DD0rXBCYsvvji2fSIcHz0ox9NfNXaCR82vmh816aYYoq0/fbbJwEbp5xySjZ5Fr+3Us9zzz2XmCxnnHHGNPnkk6ellloqXXjhhVkLduaZZ6aZZpopa8YEWTz00EPltH7fNID60E3yq1/9KvvwNdLWiRTeeOON06abbpr0vyrrrbde7sv73//+6s99f9MO6qsPM3MzYUY+7bTTclDK+uuv35BQNjs3fg8EAoGUumtGiTsSCAQCgcAQENhvv/0ygbvoootydClSdfXVV6e55porSWFy/vnnp2uvvTYJYnjPe96TzjvvvHTooYemOeaYI0m34thZZ52V3vGOd6QPfvCDia9bI/nXv/6VbrjhhmyCRNiYIy+//PK0yiqr5LqQGlou5YogoggKoohg0uRtscUWub3qWH755dN3vvOddPLJJ2dSiczWSaS6RN92GwnUz3//+98D2vvb3/42HXTQQTl69/bbby9Q9H3PM888uS8IeyNRbyGBooAb4VHOO+ecczI2CHw3pMwp7YrvQGA8IBAkcDzcpWjjJEHguOOOyy9mDunlBdToe5111pkk7YmLdIYALdR0002Xfc9ef/31TMZo9QQjCEp45plnklx0CIJcdO4pEuIj4TBfNsdEmEpP4txG+ecQEVo6+en4oV188cW53r/97W/p29/+diaPzkUmEaMiSIzr3nPPPfk6TJjy4QnwkB5l7733TnzklllmmayR1IZGpEfdCGyn8pOf/CRttNFGGZu3ve1tTcc0k+xIj2nkeOWVV04HHnhgv6CP0nZ9dB/co2ZSnj0RwI3wKOch1iuttFLG0v0MCQQCgc4RCBLYOVZRskcReOKJJzIhKC+ddt9e4CHdgwAyhlAV7Z1cde4hDVs1GlWLv/nNb+ZjiL6AhTfeeKNfRwQiOFd0a12QTdGqTMA77rhjTmlSygiQQCqRLSbSQlpoI0XF+q76xIlqdR0Ehz+dtChy7dFEIpnl/FI/kqg8stNOaOFE67773e/O5ziv1YeGcSTHtP5vttlmae21185YNGqvvmpTJyRw2WWXHYBHtU7BJ9ovB+E3vvGN6qH4OxAIBNogECSwDUBxuLcR8ALxUvdCorF5+OGHczSibyY+vzMzIQA+VTNfbyMzfnp39tlnZzLFP0wC5aL1Yhquyxe+8IV8TxdYYIH0xz/+sd9h2kBaMff8Rz/6Ub9jSBkixqcPcZHWpCo0X0zLkh5X/d/4qyF8VTKKeDJfa+fWW2+dyeHTTz+d9t1336R9xZ+xWn8hr1LEtBIEEGlCilddddXcD7n2tOmkk05K8803X07o/JnPfCaPZxrSkR7TRSuKMDeTgnMnJJCGtE6Kq/W+9NJL2ReTKX4ibqNXxSL+DgQGi0CQwMEiFuV7CgFRhV76++yzz4B+/fznP+9LgvunP/1pwPFOfkA0bIOFCNBSeEHWhSaI1uaLX/xi/VBX/J+DP4x82mllNBipQKI4/zMzDuWz7rrr5sTCnQAgkrREzDLjaic8JR+uC/Ln+C677FI/lOtgckXOkLKqiHRl2nQuzR4/v6rwL2TiFCmLlBQRQKJs1VfNuFIOEaJRJLSErsnc7Fp1+fznP5+v3YjYVsvSFNIm8oN84IEHqofSH/7wh2S/YX3gi9gIn34nDOE/tLEWUzR9CFkzKYE6nZBAfWpFAqWl4VNppxLktlXZZu2J3wOBiYpAkMCJeuej35mQeSHyJ6v6cBVovLyRAmXuv//+8vOgvr14aaqYxlxHVGhdfvzjH+eI0TPOOKN+qCv+L3mvYAhY8LFr95KlkUOabfuF4A71g0ANVoqmD1GrC3JWzJDnnntu/XAm6u41s259PCBxCy+8cL5PG264YT/TLtL72c9+NmvfbHtGu9ZKmKEtPgSvHHLIIa2K9h0TZattrXzeECFjjGb7pptuGnCfaLJLahbjkBvESIsAnEUWWSSbZgXcNBOmW/2hOW0kxpjjPvBuNeZEEtO48pdsV7bRteK3QGAiIxAkcCLf/Qned1GgXjK0dI2En1chgcjcUEVggh0QXI+moi40MrQ2jZLm1suO1f9Fz8JCSpV24oWNCCKPw/k00oi1u3ZJo7LHHnsMKIrIut/6Ud93F+ljdnR8p512GnAuAiUVDKJBa1wlJcYGbad6kSB9biXMvrR1tHHXXHNNq6J9x9SvbbSyzcR+u9LPiELm5lAXv9E2q4e/5FC12/V6q//fbbfdsoaTC0WrRY0obe1AWhsJDB33obWt4l0vHySwjkj8PxDoHIEggZ1jFSV7CAFmq/KS4evVSIrzfrNcZo3OafSbFzcNhb1Qaf3Go9Cw8TPrhKiOFQmk6UOu3Ncf/OAHA2CWb84xWqi6H9yjjz6az0XkirZNNO9jjz2W63n++efzuTPMMMMAcmP80DwyQf/yl7/MwSbXXXddPg8ZZuZFIv1NEFQmZybeqkkW8VG27qvonM033zxfv5k52HhGqPRP3kHXqwsz9Cc+8YncT1HDIy1M2kzBMESoWy0YLIa0FWaNpPp88mVsJcUcLAek1DutCGOreuJYIDAREQgSOBHvevQ5R2N6CXlhNZO55547v6iGGznJT01KEf50TKNFEBFpPOyIINlxNwstJjNpJy/YsTIHI17uKW1Y3acPtgiQ4/wN63L00UfnY0heMedutdVWfQEgiJmxwnxpS7Oq0P7NOeeceWs0/okPPvhgzo+nDF9F4+f000/P0cvIjWAQdVW1lXClURT48Ytf/KJaff6bf522NwsMEV1c0ts0Ol8lUtcgwIJDEMWRFoS5YMz9QZuaibYgwnCoR2g7R9oX/bXwQMZbCRIoSXcEhrRCKY4FAo0RCBLYGJf4tccR4G/mJcNBvZEgZl5QnM3lnmsmNESS/HL0F00sGIJWiOaPIHd8oxZccMG0ww479KvrqKOOSkcccUQu63u0BUERZaqtNEF2vNBeJrw6uePbtuuuu6bDDz88m7JhwS+tXq5Rmyd1YEhpA4LknvK3qwstW0n/cuSRR9YP52TOzv3Upz6V+yiiVvmiJZYL0FgQyFCN3kVA4CfNimAM59nGrKRy8S0ARD5A5mjkyz66UtRwESiCCEkjw5ztPtVFGhzt++QnP1k/lP9vPCNBfB7vvvvuAWVoOrfbbrusBaQtq6arGVB4iD9IwM0krZ3GF5/QZiI3Y/ELbJRM2rhTj/rajblqiphmrh3N2hG/BwITHYEggRN9BEzg/tutwIumbuKkreFcTwtRzHqNYJKOQxJgm9eL/PXSo9njdF9MyLQjyIGXO9JXhEbQ716AiMVyyy1XDuWgBCk9mGD5gtGqIBp81bzA+ZKplx8b8xsfMFog9auT1ohGhuYRSSV8DgUvqE8AwVNPPZX9xgRI1CN+RZDSXB5zzDF5Zw2kFk60Ld0sNHzaueWWWw5oJmLlGDIraKIu8847bz7OZMxES3vHb64I0oQwC+ag2fV/BFDQD6JirMAXrnL0Cf4gNIc0k3YMcV1EzMJDeffJPVQXM7L7i3g3Ij1IovaXcVXaVf0ugSvIZtWfkmnYPWV6hVG7ICfmcO02rquRztVrNfpb0ImxrJ3M18WU3qis3+wogiAzDcPc4kG7EcQll1wyt7eZ+btaJ22rfIva3Mzdoton9y0kEAgE/heBIIExEiYsArb7YpJCjAQKeDnwJeNbREvhZdpMvLile6Hh8/JGsryomeIQNfulEo7/yIQXalXj4WXFRChoxMuZ6bGIF+GVV16ZNYi0SsgqciLXHOKJUH71q1/N5rZtt90254BDNOyBa/9aL33mUIEMNH9ILq2T82j45EBEDrXVi7iq4ZPvDkE54YQTSnNy4l+aq0bBBn2FuuAPOfy0E8mtC7KLnNDmVQlSKYec6TeyjewyfxctoDJM3MzASCC8EA87zLiva621VuI6QMNK0ycNTckLKPGzdgmSQNJpi5nVjS/kh4bu1ltvzUmoaY9bEZQS2VzaXP821owVpN7YZl7mj0orhzwaK67XTrQPVj6IaiNS2qgOiwY4FH9Hi6lWQju7xhprZFIs2MbCiN8ifD0T8gx2cm05Gy2CWm0bV+9Tq3bFsUBgIiEQJHAi3e3o6wAEED2aHOSBlogpUT44W4i1EvnckDsvHsSRudALC+FijrNrAqEhRL6YXmmJitDGIGIIi+t6aRdBUvwf6aDBWWGFFTLpoCmhmaG1QjxEL/vogxevdCPIqXao22+0V3feeWcmJoiNvHTIC3MbQsiEWrSdSOz888+fX8pVH0WaTUSmkxdy6cNYfAu80cei/ay2gTkVziUvX/WYv6UDktoE8eEXiITXRZoYZJK2C+m2AEBA3BPYI3bMwjSFJcUM7R8NHYLJb5BmC7aIZqnHN81juzHHpKx9rQSJcs8tZNx/5A+pogmr5ipsVQdiW0ggQtvpfec3SXvtGapqvVtdyzEaRH1zfzyHzOVya3YiyLn0MLSrnolmba33qZO6o0wgMBEQaD2jTAQEoo+BwBAQ4NiOcNAi0qg1evkwcTERIgFeijQfVfF/miEkADFA8ogXmxeqAAMaRRocWhW/e8l7YfJdU4bWj88hksGkWNrx05/+NJMAZjkvScSGhggpRVIQV2WQ0UIOkFcvf3u+FtGm8lupuxyL70mLAM2xe2EBMppCG0mDTFNpbHbzfbfoKdryVv6A9T6NJn5RdyAwnhAIEjie7la0tWsQsO0XE9Tuu+8+wEyKrCF1SBUtGpMgp3kvIubcIpz5pTRhCvMyYxYsQpMkKtK5tHWImt9okJgkRbO6jg+TGn+ooo30m6TO6ha04oUuyEE0KlOw6NeiMSxtdV0mYCSjmkiZZtNvNF3yylW3RCttje9Jh4B7YUxNCjF2+Kp2MwmUv3LNNdfMCxe+hO2k9KlduTgeCEwUBIIETpQ7Hf0cUQRoHeRl41NXfJ+8LJE1/mBypJ1zzjnZ/4smj9kWgavuosAMxmTHj8+HVrEIDRwTHvPsqaeemgkgzR+zMqd5/lOuR9tIy8fc5f8+/AyZNpdYYom8KwTN38wzz5z9CBu1tZiitZdJrqT2YJbWD7/pE02UMiFjhwBTr2CKSSGbbbZZXoAYU90qnkMm70YJvhu1mWa9WZqdRuXjt0Cg1xEIEtjrdzj6NyoI8PUSiCEyl8/dCy+8kMkgMicFiShPWgc52TjXI3Ic3as5z+zewIcMuZM7TooQ4qUrLQ0zF3McrR8NHv8zUaYc8GkalUPY+KcJShDcQrvI9IwAClJA5JgPV1111ZycWLBDva0lRQxNihfqhRdemP3cnMe/i6nZMQQE6QwZWwRoA/fee+9RbYQgIPe9vkfyqF50kJV7lhBVgVidaAH1ybPS7QFOg4QhigcCw0IgSOCw4IuTJyoCNH40ezRwCBgyiMj5HHrooTkIRHQmR3mRicrVnd2RR6Srnj4GuWN6lVfQMYEqxx57bE5HI/UITR/TsHK0iV5sttZSRsQyAioalflZGVpFxA6B9NGmelvLfXQeM7e0NBztmY2Zn6WJEUwSMvYISEtj8VBPbTSSLeMXSuts/HSrCPLhLmFx1ImUPnVSNsoEAhMFgSCBE+VORz/HDQL89Jhf5QOU805qGVpC34glAqiMD7KGpAnwUMaHhiQ0duPmdg+6oYiZXITcAkaDpBlD9vYtO6cMuoGT4AT5OS1MGqUDanT50ica9JBAIBD4/wgECfz/WMRfgUBXIIDkIXVyrtlJhNZR0IdoYsSPePl7oQn4kACYtk8ZaV6cH9LbCBgLiKA0P6NBBLsZPZpQix/pdsrz0M3tjbYFAt2MQJDAbr470bYJiYC8c1508soJ6ihRv9WXvZxy8vvJqya6WKqYkEAgEAgEAoFAYDAIBAkcDFpRNhCYBAiIXpSE2u4WIobl86uLfH+SXIsu9i1lTUggEAgEAoFAIDAYBIIEDgatKBsIBAKBQCAQCAQCgUCPIBAksEduZHQjEAgEAoFAIBAIBAKBwSAQJHAwaEXZQCAQCAQCgUAgEAgEegSBIIE9ciOjG4FAIBAIBAKBQCAQCAwGgSCBg0ErygYCgUAgEAgEAoFAINAjCAQJ7JEbGd0IBAKBQCAQCAQCgUBgMAgECRwMWlE2EAgEAoFAIBAIBAKBHkEgSGCP3MjoRiAQCAQCgUAgEAgEAoNBIEjgYNCKsoFAIBAIBAKBQCAQCPQIAkECe+RGRjcCgUAgEAgEAoFAIBAYDAJBAgeDVpQNBAKBQCAQCAQCgUCgRxAIEtgjNzK6EQgEAoFAIBAIBAKBwGAQCBI4GLSibCAQCAQCgUAgEAgEAj2CQJDAHrmR0Y1AIBAIBAKBQCAQCAQGg0CQwMGgFWUDgUAgEAgEAoFAIBDoEQSCBPbIjYxuBAKBQCAQCAQCgUAgMBgEggQOBq0oGwgEAoFAIBAIBAKBQI8gECSwR25kdCMQCAQCgUAgEAgEAoHBIBAkcDBoRdlAIBAIBAKBQCAQCAR6BIEggT1yI6MbgUAgEAgEAoFAIBAIDAaBIIGDQSvKBgKBQCAQCAQCgUAg0CMIBAnskRsZ3QgEAoFAIBAIBAKBQGAwCAQJHAxaUTYQCAQCgUAgEAgEAoEeQSBIYI/cyOhGIBAIBAKBQCAQCAQCg0EgSOBg0IqygUAgEAgEAoFAIBAI9AgCQQJ75EZGNwYicMABB6Q999yz34F///vf6U9/+lNaZZVV+v3eS//585//nM4///y02mqrpRlmmCFNPvnk6W1ve1v+ft/73pc22mij9MwzzyRYHH300WmmmWZK//3vf3sJguhLEwRibDQBZhL+XOaglVdeeRJedfxf6tRTT01bbrlluuqqq2K+GsHbGSRwBMGMqroHgcMPPzzNN9986ZVXXunXKBPw73//+zTzzDP3+71X/mOCXGmlldK73/3uNP3006f11lsvnXLKKemKK65IP/nJT9K3vvWtNPfcc6fFF188ffWrX03vec970gYbbBCTaq8MgBb9iLHRApxJeKjMQR/84Acn4VXH/6XefPPNtMMOO+T57fvf//7471CX9CBI4CBuxNe//vX0lre8JR155JGDOCuKmvRuvvnmTDZopr70pS+NKii/+tWv0jTTTJPuuuuuAeRGW373u9+lGWecsW0bVlhhhVzPO9/5znzf3XsatXe84x0DPm9/+9vTZJNNlt761rfmsrvttlsmoDSOjjnXx3H1eQFY1d52223J5FbkpZdeyqRtnnnmydcq56l74YUXLsUGfP/973/PBG+WWWZJU0wxRSZ/1113XdZ6vv766+kf//hH/rz66qvp1ltvTdNNN10mgOpHmEMGj8B4GR8xNgZ/b0fzjDIHmQu7QcwJJ554YvrKV76S9t9//3TBBRd03Kw33ngjXXPNNemb3/xmPveoo47K80snFQzlXFacz3zmM/ldcv3113dymSjTBoEggW0AKof/+te/pqmnnjq/yH37f0hnCDz33HPpoIMOysSJhmqNNdbo7MQhlpp11lmzybNKrkpVZQLuhAQikz/72c/Sfvvt10fiVl111XTllVcO+Fx22WXpy1/+cpptttly2YMPPjibW2+//fZk1brhhhtm7Rxz7HbbbZd++tOfpt/85jfpb3/7W2la/v7Xv/6VSeo555yTtXXvf//70zrrrJO++93vNp1cETwT+Ac+8IFMOPfaa6/01FNPZdLXr/L/+88///nPtOuuu/b16Y477mhULH5rg8B4GB8xNtrcxDE4XOagkSKB6jNvmO8eeeSRdOihh6Zpp502nXDCCW17d+GFF6b3vve9yTzzta99Lc9hFrkf+tCH0t13393y/Jtuuilr5cxp66+/fjrwwAPTuuuum+szZz388MNNz2937kMPPdTwXG4rP/7xj9Pqq6+edtxxxwGWnoYnxY8tEQgS2BKe/z1o4Fkh8a0qmhmmtPCj6gC8lJKV5tlnn50WWGCBtNxyyyXEaLTEipbm7Q9/+EPDS5QJuBMSaGJFmJC7ct+PP/74Pq1a0a6VbwuD008/PWv7TjrppHx9E7PfzzjjjKz9Q9T22WefTAC1pdEYorm7/PLLE00T88evf/3rTBYbkVoXoZlWLy3jzjvvnPveqN4qINdee23uE1JeJ6LVcvF3cwSGOj6MpUk1PmJsNL9/Y3WkzEEjRQLdYwvLjTfeOC2yyCJpyimnzHPBcccd17KLFqLmjMUWW6wfmXr66aezwoOy49lnn21Yxy233JIWWmihNNdcc6Uf/OAHeZ5U0FxvrnPuEksskRe09Qqcu+CCCzY9lxWn2bnqMj/uu+++aY455kjf+MY36tXH/weJQJDADgBDKKx2ONpbJZnEDdTnn3++g7OjyKREALkzMTWTMgF3QgJLHXzs3HOm4FarW+WNFRMrH7wiiOSll16aZp999jyOELsnn3yyHO73jbzRECKKVtM0kc3InxOROZpP7TNxOrcTefnll/M5+taOMHZS30Qus+KKKw5qfLhXIzU+aFSajY8YG905KsscNFIkkAZvs802S9tvv312CTG+fNqRQCRKuRtuuGEAUFxEHPvkJz854JgfzE3cTlybj3VVHnvssfSFL3whE8HPfe5zA+aXtddeO5/LvardudV6q39/73vfy0oFPs+vvfZa9VD8PUgEggR2ABhzoAjKP/7xj1mTVR4yGqKQ7kHg3nvvzRMXrWMzMQEzT3dKAq06EX733DkIXVX4tVj9FvF/JPDxxx8vP+WXNL+ZD3/4w3li3HrrrdOjjz7ad7z6B63hD3/4w+zTZyLmA9OMpLkuEud6fAatyJuVrV7D37RY+mQSDxk6AsZHcRPpdHzAfbTHR4yNod/T0T6zzEEjRQK9l/g5W4Ded999+bk2xlqRQFpAZXy0py4Wn45Z+FowVgVpnHPOOfNx5uT6nGPOcm3nC8Dj51zEuYV8XnTRRU3PNac5t37tUo++ssaZU88888zyc3wPAYEggW1A++1vf5v9JYraWYQddbsBzo8CoQjpDgSkPnFf/vKXvzRtkAnPPe2UBPKXMxGqt9GqmHOyMVHkxRdfzGlZkMEiCBcTuIhc/jfMNgjrf/7zn1Ikfyv3wAMPpC9+8Ytpm222Sb/4xS8yWetXqPIfq+EStMKk06rfldPyn8iL6GmTfcjQETA+vLAGMz5EbY/2+IixMfR7OtpnljlopEhgtb1PPPFEHovGYysSWPyc3/Wud1VP7/sbqVSHz3e+852+3/0hCERgGU0gC0ddjG0Ez7kW0N2afmgAACAASURBVLITFKmey4+6LuVcz5Rz69cu5WHomDJbbbXVADJZysV3ewSCBLbBaPfdd89mPKutIoIDygPCCX88ChX6mmuumfshorREWjFTIiAmB32k5ahqPD30yEyJhP3oRz+abrzxxgEQMFHRTC2//PK5Lj4gJqVmqvuddtop1yllCb87YhW4yy679GlatIkzcF0bVy5uYtLmVlIm4E5JoJQq5V7XJ1W+gFtssUW/ydaqd++99+43KSF7Dz74YDJukEBOzchXvR9W0KeddlommwJB5HRrJvBdZpll+tp22GGHNSva8HfXZpKMAKeG8HT8YzeOjxgbHd++MSlY5qCxJIFLLrlknjuatcE8XeY9ZtuqCP4wF3ORknaqLhazguecb76rWhuq59JG1qWciwQ6t37tUp72kcWHTzM3mPpcWsrFd3sEWr8x25/f0yUQItq+I444ol8/r7766n7awGb+Xf1O6rL/IFM0QXyKPMx8HRE8JOzTn/50foiZFspkIVEnPzUrL47IjinvYXUu7VpVlBcEImqV2l5QyPzzz5/OO++8arH8tyg2x2i+tMXk4QEXoWaVhzBZ4S699NL5mLY3EgEhtLStpEzAnZBAEw3fF+3RT2SXmc0HQdMuk2h1NewFLDK3KurhYE1TaWJbaqmlMgGraoMQSlolfZNehkZQW5uJ41bi2iZgqZ2vYrN64vehI9Ct4yPGxtDv6aQ4s8xBzQjYcNrQqSaQe5O5QzaDZuK4j3mrKoJPzLXaj+w1Eu8V55qPne9ZIdVzqxaUah3ONd+Wc6vHqn8jkdJm8QFvF8lcPS/+7o9AkMD+ePT7H7OcAcb/oCoe4kIODHTlxpPQsNHk8T0jkgeXB75OeIumw4pLBGqd8DJtObdo79RH60VLJWcUcijvlEgyufGYIepCg8d8QKptQTarUtoi/UFdTH4mDhrHVlIm4E5IoETTyukfkzBiK92LjyjgT3ziE3klWndubnR9QUR8AWlWEWLYF62oCZJbwTHHHJM1iz/60Y/6Res1qs/1yz0TcGIFHTJpEejW8RFjY9KOg8FercxBY0kCp5pqqjx/mDuaSZlfll122X5FSg5Tc2MrIud8CgLnFxJYPZcypZEUEljObVTGbxY7AlAQ2pNPPrlZsfi9DQJBApsAJN8SLaAXcyMxUA1SA1055ceL8KXQ9qJCL/3g81Ye1tIXOafKZCD5cF303fEqTsyha621VvaD4yMnZx4C+bGPfSxdfPHF/apghnb9Iq3awi/TtWjT6kJziASKkGslZQLuhAT+/Oc/70sLhMAyk/sgc0y7iDHSqc52Age5+Wg655133qzpLKZYGkGRnCY05g9Eu+4vWK9fwu1yXzbZZJP64fj/JECgW8dHjI1JcPOHcYkyB40lCSzzbCck0IK+Ks4x13ZKAp1f3ivVczshgfVrV9tBOVNSIFVdlqpl4u/2CAQJbIIRHzWrlqovYLUosx//hvIiVn4o4mXPQVayTdrFoX60xQ4Unci5556bPvvZz+aiTJClD42ILILheKO9dmmyTAaOV80CtH182mjImHiZl6n2tZHGrio0Yvz+SKdtoUmrC99M7aCtayVlAu6EBNpXt2Aj1QHcfBBORNexTu878zHfGJpTEyHNKWLo/kvrwqfPNaysO8nbx0xe2nbIIYe06vKAY/wWGxH6AQXjh5YIdOv46MWx4YUvqMCY5wZR104JtLIoY5UphKPlzRvlg63aW+agIIHtNYGtSCCLlqATC2tpt0KGhkCQwAa4MWfScFVNnA2KJUkvy5ZgyjtvsIIEIAj3339/jhg1mQ3104jEtWtPMbFqfyMpmj4TcF3saoEEMi0XrWK9jKS4fA/5ntAItpJqWxpN5MVfUKLQupQXX6OIs2rZMgF3QgKZe7109NGqlfnPhwZPAme/C37pRBBmQR/8HJkv+EkyAauPdhQBLObzRn2vX0Ny2EICJcgejMDIDiYhw0OgW8dHL44Nzy1ixX+YX3I9IM8uEnxjmcK7QVq1t8xBY0kCEWbzRyeaQAGEVWHJ4B7TiSYQYXd+mdOq57bTBJZzq9eu/o0EWvQbD5tuumn1UPw9CASCBDYAi1nuIx/5SNbUNDjc9xPiw4RaXsadaoX6KkgpPxyIoLqG+6GdHKzYwk37G5kUmSVL3xppjj7/+c/n4420hNohGMILyYTjOvbJbSUlWllbyqRRyiPcpS2NJg+kxvFGiU9LHb7LBNyOBDLRylOlTn589UhdE5DjzzzzTLX6pn/bv5XfI59HpnEaC0TSwsFuNEzF+ihApBNxTsFjMP4wxgiNcfFjhbPfJHiVlxDBrge2dNKeSV2Glhnx59tU3cmnYNLuW3SjMSDP4lCkm8fHSI2NoeAymuewyojQd9/qiz3pkfiI1Z/TVu0Z7bHfrL1lDhpLEsjX3TPSbB70TirPkL16q0I7x5zMN7tRdDBcBdE533Pm/DKfV89tFB1czrXALudWr1392722sOYeFIvaKjKD+ztIYA0vEwmNU7P8RLXi6Ze//GXfS4jWzPmDEYN+LEmgB83D+u1vf3tAsz1gjjXS9CGs5dyyBVa9AlGzInoFfogsbpfHTmRys7bsueee+Zh700hKjsBmjsrlnDIBN5v8Srl77rmn777qQ5nEynHpXJh3Ow3IQO6kZBEFbeW67bbbZlIsn5tJ0oq2metBuWb1W5S1iRJegzEHi6Kz+i/RdPCQM5BvJlKFHEbuyyrSjf/u5vExUmOjcc/H7ld7NdO+ctNpN5d00srRHvvN2lvmoLEkgdyPzB3mokbC2uG4T90KZO5CvCzu62RcXRaVyKFzleNXXqR6rq0x61LONbc5t37tavmiCbRI53IUMjQEggTWcKO5ogXsdJJBAgzA8sA4fzAylubganZ5f9dFuhL9koi4LnzjPKgIYskIL6jhrLPOykU9zIgbsrXooosmZAfZoA0UWFGXdm0ppmB1EmR75ZVX7qumaCVHyieQxqGQLGlu6qJ/gyFtylsdy2ll4kVakWyki2kLqeiUUGoLTRiNovtj66RONIjaK68hrIpId6MuZjTBPAilvUNDWiPQzeNjpMZGawQm/VF+x3KaCsoq4plBOC655JK+aPtyrN33aI997bWlY7W92tQNJFAWCHOHOa6R3HXXXX3vNEGQVWF54NLi3Hqgn3I0dJQo6jdHVfPIVs91z+pSzlW3c+vXrpZHAl3Hu6E6p1XLxN/tEQgSWMHIVl60UZ1qAcupXuBlVwnnN9sSrJSvfiOBVlOTMjCkXJ/2z4OqzY1ERKzjBx100IDDiJxjTMFFS4bclD1RH3rooUzS+HUIQoEJcsEXrpGzb6u2VE3B/iaIGdJSpExqxcxZfq9/lwm4nSawmPlNRo3MFvV62/3fdY0TiaKtXBFrfiz2/KS5GYwZq1yL+0HJjYhcI5qNxBhD1K3IOdTbDaCIYBz+VByr5S9kSix5D/lg+b8dTrSXmV7ydP/Xj7I4QByZY/h97rzzzqXq/LIzhurmJFumWeFvsMEGmQyLuEZk5UccL1I0z8ZHNShqqO0f6fExEmOj0750Ok7UN5h777mw6Cu5Ro0XxKA6xvjRevaNv3p6K9eTmgnxcK65yjj7+Mc/nsu3Gvul7yPdXvWWOWgsNYHumQW8OZxbSl0KiWvURotJwXnGPh/NkuWg1MGdxH3jpiGPX3k/OO5cWSKcy5e53bmlzkbf+uD+GxONUo81Oid+G4hAkMAKJl5WHowDDzwwR2uK2OzkI3VJIUweqvpLr3KJrvqzRP428qcweeqLTyO1vRW5Y4I5iAe66mTMTE7rZcVocvbwe1nS3p1//vkDcCikstHWbNLPuBbzM7ECRDirL98777wzTyz1qMH6hcoE3IgE0qb5COIo+Q9NZKJ4EazBaOrq1zURioxGoJg5EEGmYZOXSRhRG6wgc5KvaiNtKxMv7YY+aC+TvUhj0ceIGvIt6KgqorMR8xVWWCETRC/cYt4XBS04Sl5Eib/tDmOs2y8UAXQP1Gf8mPjlgrSXZxHaXSS1GhBEg2tRIArei8Z52kez3O7elXrH6rs6PrwcjUnYcw3otvExEmPDs8I1gI+0+7X22mtnoo5g0iivuOKKmdR1Mk7cs8Hce1py13Qt2n3PiaACe9Yaj8TYM4fZktHcY4xWCYccpcbzaqutlset8UhraLHDdUdQV7OxPxrtzY3ugARyaaFB1G5zXTtxn4w/H4tv49JH9Hr5vdHchUwrZ+Fcxc31zAeOseA0Eos4PtHuEfyrcvPNN+e5QB/MFXVpdy7i2Ozcal38qS1ezeXHHnts9VDf31Usi8Wq72D8kREIEvh/A8ELuuzDKp3JYD9FE+jBUU89FUo3jrdCdBr5A4r81RcOwMhEXYom5IADDsiTsj5XVfceUE73tEPKMDc6RwBEdaeMUm+rthQnYxM3cmmC4htUFW30Qm6m1Sxlm5FA55swEEurSytV/WdqcH054ZhIhiNeSrReSCC8aCYQbGbyoYqocgSQxlXQieAOgTO0HO6HFzRzlMla8EddaGjda4E7TCom8KIZYN6Xv9C9QzaRPThI+6MfFjvOlQ7IqhxWn/rUp/ougUxWNWW0nXLYcbeQ18suJxYb6qNtRpC7VVqNDzh34/gY7thAIDwPSLxFnoWfRbGAL2PZfUSiOhkng7n3iJq6kc/iYy0wzRaUFpauTWiwjU3BWfzT+LQWMoOwm9fk8UQk1WPeMdYQWFp42RSajf3RaG8Z22UOaqRlU4bPo7nHp7i/lHMbfVtIIWK33npr1nqWc2Uc8Jtj9cWfemBEK6o8xQcczQdFOeAZbyaeB+5CgkOk/jKPCSS0wLeYQ8zM+43EucjbUM6t1meeMx4sDOpEtJSrY1nGRzke3ykFCfy/UVD838oDNNxvg7zbBxytDQ2mF3Fd+Hp4gcOlkXjgEQM4mcyqWrlSnlkRATFBm9QRlGa+lq3aoj6aAITUx+RmIqlL8Rus/179f5mA65pAkx8iRXOFTFbvP42XPnoJNrputf5Wf/PVQoKZL6SK8fLsNLq4Vb20NV4WMNQHH6TaatqEzE9TKppGQtNAC0fLQ6trLHj5E9pVGlxaRFo+gTBW00w4XiperHDTL5O/hRATXpHNN988368XXnihrz5mJGPOuOJXZ0wghOpuFIFe6hrr7/E6PoYzNowDL3cuHl7oFl/cMWiVjAOaOS4ZnYwTZTq594gaVwPPMi1SGYvmI4SOObcIkmu80iibq6oWCxoxpJH2msaPRtx8jNzRmAvaaDf2R7q9pd1lDmpGAmkAy/yjDe1E0J3gO4vkoshwvnnMb47Z/rOZeO4t1mHIEsZca5enToRfnzHAsuH5dz2LubIXfas6hnOue4nAGyf2sm82L9ex7PZ3ciu8RutYkMDRQjbqneQI0Dya/Fpp1soEXCeBk6KxzIa0KtIz8AG197DV+FgKgobkaQ8ih+DBiJSXvYg+LxEvhkZmJWVpeWlhi7+hOvSTydvE6zwmPi8+xJRZmanJC5rPpRe6NDoTWbptfHBRsAig1ac5Mw5ofP1O84k40Px2Mk46vfcImsUW7TYtFnE9WmVkxiKqKl7+XEyYg7lxlPFp4WOxQQNO+9fo5d9q7I9We7W9zEHNSCByymRK0151san2e6L/DUPk3wKd5rLR/YVRHctm5SYynkECJ/Ld77G+C7BBAhulLShdLRPwWJBA/nq0bl6cXjJ8DcdaaHRpVuUOZPamqS1ETtto+/haWe1Xzf31dnuhMc0UYbKjVRUtT7vjXC82WtA99tgjm4KL/6KJ2YveS3wiSzeOD8+LgCZpkmhrLVr8ZqcG/r78zohxQvNmnFjc1KXTey/QwBjhHlDMvhYHrmNBUZ6bksaISdc4ow3TtkIcaZeZGwUy8TmtShlrrcb+aLVXO8oc1IwElrYKaPHshQxEwNxp8crciwy2k4JlkMCBSAUJHIhJ/DKOEWCWYLpsJmUCHgsS2KxNY/m7qGI+PTQstDDMaPysijDDMeXbk1mkZDNhRqIpIl6yTDXMS3zJmN5E/tL8cSbnR8aETEzKXvL8zKrmvHww/hlzBGjW+MoiYAI0PD+0g/zFaIeLr1m7cdLpvTdukCPXYnJ2Pb6j3Ackk6cxPuOMM/L4Ag5TJjMkf10fmj9jiqsFzSEy2WiscVVpNfZHqr2CUart1eYyB7UjgbSviHfIQASYki06uLHQ9rWTgmWQwIFIBQkciEn8Mo4R8LLgN+hF1UjKBBwk8H/R8YLnt8nnkh+PQBIvrSIIG19DG7W3iq5jeuMLxMRGwyLam4bGS10dgkcQTNcSgCJwRVkvaC9v9Y+nFDEFn17+9sLkWlEi/2njEDLRudwHRO0X7S2y0mqcdHrvaXj4GiN3/DAFPYjwpR2U5sW4EV1eUotYvNg1xvNuzJXdhGie+YpZvPBHrI8157Ua+yPVXgEL1fYaL2UOakUCaS/59NW1mL083jrtm0WjMcGv2tzSjtgVLM1LIQMRCBI4EJP4ZZwjwDlZYA6NVF3KBBwk8H+RoalAykT+ShXhhV8NVhFZKSCA315x0q9j6v98HRFBaSf46CAF/JmYB/lqMjPTKnH2RyB8RH96cfvIYeiFH9I9CHi58lP0LNEGiyCVisMLWOog48TzRNx340Rkb6NxMph7b4wgglK5GD+0cmeeeWb2MaX5qZr/EDxmX+0T1V8EUWCyFhiAPCKD9bHWauyPVnu1r8xBrUggP0fa1pCBCMhgwKWGG0urhWk5s2DZjiyW8hPtO0jgRLvjE6C/gg1EjTUiFWUCDhI4AQZCdHFYCFhEMdMj8XxGabVo2nwLEPEsxYt18BCXOagZCYSxwJiiZR38FXr3DJhYtCJ2XBDaSWDZDqFIEdMeoSgx7hDwYrLqp0moR9+WCZhTe0ggEAg0R8CzwgeP872E77RrUq2IyG2kZW9eUxypIlDmoFiIVlFp/zcN87777pt9AasuK+3PjBKtEAhNYCt04ti4RcCLyp7OogSrYgKmxZC+JCQQCASaI8DUxkdUMl4m2ZJrL7R/zTHr5EiZg+Aa0jkCxiJ3gIsuuigWIZ3D1rZkkMC2EEWBQCAQCAQmHgL8OiVblotN1O1I7KE98VCMHgcC3Y1AkMDuvj/RukAgEAgEAoFAIBAIBEYFgSCBowJrVBoIBAKBQCAQCAQCgUB3IxAksLvvT7QuEAgEAoFAIBAIBAKBUUEgSOCowBqVBgKBQCAQCAQCgUAg0N0IBAns7vsTrQsEAoFAIBAIBAKBQGBUEAgSOCqwRqWBQCAQCAQCgUAgEAh0NwJBArv7/kTrAoFAIBAIBAKBQCAQGBUEggSOCqxRaSAQCAQCgUAgEAgEAt2NQJDA7r4/0bpAIBAIBAKBQCAQCARGBYEggaMCa1QaCAQCgUAgEAgEAoFAdyMQJLC770+0LhAIBAKBQCAQCAQCgVFBIEjgqMAalQYCgUAgEAgEAoFAINDdCAQJ7O77E60LBAKBQCAQCAQCgUBgVBAIEjgqsEalgUAgEAgEAoFAIBAIdDcCQQK7+/5E6wKBQCAQCAQCgUAgEBgVBIIEjgqsUWkgEAgEAoFAIBAIBALdjUCQwO6+P9G6QCAQCAQCgUAgEAgERgWBIIGjAmtUGggEAoFAIBAIBAKBQHcjECSwu+9PtC4QCAQCgUAgEAgEAoFRQSBI4KjAGpUGAoFAIBAIBAKBQCDQ3QgECezu+xOtCwQCgUAgEAgEAoFAYFQQCBI4KrBGpYFAIBAIBAKBQCAQCHQ3AkECu/v+ROsCgUAgEAgEAoFAIBAYFQSCBI4KrFFpIBAIXH311emyyy4LIAKBQCAQCAS6FIEggV16Y6JZgcB4R2D22WdPZ5xxxnjvRrQ/EAgEAoGeRSBIYM/e2uhYIDA2CDzzzDNp/vnnT295y1uG/JlsssnSjDPOmFZaaaWx6URcNRAIBAKBCYBAkMAJcJOji4HApEbgIx/5SDr44IPTf//730l96bheIBAIBAKBQIcIBAnsEKgoFggEAp0hcNFFF6UZZpgh/fGPf+zshCgVCAQCgUAgMCYIBAkcE9jjooFA7yKwxBJLpL322iv95z//mSSd/PrXv57NzkcdddQkuV5cJBAIBAKBXkEgSGCv3MnoRyDQBQhcccUVabrppktPPfVUbg1z8L/+9a/061//On3lK19JCy+8cJp22mnTVFNNlT7wgQ+k5ZdfPp100klZazgU0vjXv/41TT311JkE+n7llVe6AIVoQiAQCAQC4wOBIIHj4z5FKwOBcYHAKqusknbYYYf073//O7cXsXviiSfSjjvumDbccMN05JFHpksuuSR95zvfSRtttFGaaaaZ0qyzzpr23HPPQZuPEcz9998/TT755H0BKF/96lfDD3FcjJRoZCAQCHQDAkECu+EuRBsCgR5A4LrrrsvavUceeaSvN//4xz/SEUcckTV+J554YhI5/MYbb6RXX3013X777WmLLbZI00wzTZplllnSCSec0HdeJ3/84Q9/SO973/vSaqutlt7xjndkIqiu559/vpPTo0wgEAgEAhMegSCBE34IBACBwMggsP7666ctt9yyny/gyy+/nD75yU9mjd+nP/3p9LOf/SyTQFo8x84888w055xzpimnnDKtuOKKg9LiffnLX871CkBZdtll+7SBfg8JBAKBQCAQaI9AkMD2GEWJQCAQaIPAnXfembWA99xzT7+StHWrr756eve7352kjZE8GvkjtIRI4TLLLJM1eUzDnaaU+e1vf5uv941vfCPXddVVV6UpppgiE8H3v//96bnnnuvXjvhPIBAIBAKBwEAEggQOxCR+CQQCgUEiQANI41cP7njttdfSAQcckJZccsnsE/jTn/40+Y0IGPnFL36RVl555fTOd74zm3Y7JYF77LFHsiNJNQ0Nf8SSoFp0ckggEAgEAoFAawSCBLbGJ44GAoFAGwTuu+++HBH885//fEBJASIPP/xwDga5/vrr0wsvvNAXNPL3v/893XjjjZkg0uLNO++8HWkCn3zyyawF5GtYFXsVV7WByoUEAoFAIBAINEcgSGBzbOJIIBAIdIDALrvsktZcc82mBI7G7/XXX8++gEVTSOP30ksvpdNOOy3NMcccWQv4hS98oWkd1Wbstttuaa655kq///3vqz9ncrn22mv3aQO/+MUvdlRfv0riP4FAIBAITCAEggROoJsdXQ0EOkFAdK8EzPz12on8fzPPPHNHZat1/fOf/0x333132m677XJwB5PwtddeWy3S8G9tk1/wmGOOaXj8pptu6osU5htICxkSCAQCgUAg0BiBIIGNcYlfA4EJhwAt3WWXXZa23nrrrE1rpd0r4IjEFdU7GGEiFthx/PHHp49//ONppZVWylHCEj+3k5122inNM8882azcqOybb76Z1l133T5toPKd+hk2qi9+CwQCgUCglxEIEtjLdzf6FggMAoGzzz47rbPOOpmUCbCYbLLJ0l133dW0hqeffjqbcn/yk580LVM/gGj+6U9/ShdeeGHaZJNN0gYbbJBOP/30bNptR9YefPDBRLt33HHH1avt9/9bbrklvf3tb89EUPkHHnig3/H4TyAQCAQCgcD/IhAkMEZCIBAIZARs63b44Ydn025Jvrz55ps31aQddthhabHFFsupXjqBEMmzrds111yT+BFuu+226fzzz88EkN+gQJFWwmdQmplqRHCj8kzNIpVLpHBoAxuhFL8FAoFAIJBSkMAYBYFAIJARQJ6KlOTL73nPe9Ljjz9efu77fvHFF9P888+fLrjggr7f2v1hpxC7hNjqTXCH7eMQOgRQ7sCbb765aRW0eXYHsd1cJ/LLX/6ybzs52sD777+/k9OiTCAQCAQCEwqBIIET6nZHZwOBzhC49NJLszmYNq3RDhyieuebb76OtYASQ997773pkEMOSfvuu2+68sorc3Qw87Bj0svYUaSZSVgACS3gX/7yl446gFiqr2gDnR8SCAQCgUAg0B+BIIH98Yj/BQKBwP8h8NGPfjSTKPv6VvfjRcQWXnjhdNJJJ3WElUCQxx57LB188MFJ2hYawGeffTYTOsEgdhXhj7jUUks1JIGPPvpoopHsVAtYGmX3En6NiKDz1RMyPhCQ81GQUkggEAiMLgJBAkcX36g9EBi3CAjYeOtb35pJlEjeIj/84Q/zbh38+9oJzZ58foceemiOIqaR43d4wgknpBNPPDF/+Baut956abXVVmtIAj/zmc9kMnfggQcmZTv92FLugx/8YJ82UD0hw0PA7i977rnn8CppcXZZeBQN7lC+7T4Tmt8WIHdw6NRTT837gNuOsZl2voNqosg4QCBI4Di4SROxibRHfMREj84wwwzpS1/60kSEYUz7jOTRAnoRL7TQQjmoQ9LnpZdeOhOxThon2IPGUHJnL+cpp5wyTTXVVHkvYfsJ+/j/e9/73hwtXH/hPPHEE/k8bXDuYD9ve9vb+kig66uvF8R2e9/61reSYJ7tt98+feITn0hf+9rXRrVrNLlzzz13xy4AQ2lMkMChoDY65+ywww75Wf/e9743OheIWrsCgSCBXXEbohF1BGiPvHQQBKY8O0GETHoEEIuijWHGtfcvUi7NSyfC7Lvccsv1bedW6qJhrH7cY1qmOgn0IirnjMS3+npBvJj5ZL7rXe/K2Aqa+fznPz9qXZMqSNodycFHW2hvjz322NG+TNTfBgG+up/73OfSGmuskX142xSPw+MUgSCB4/TGjXWzaXhuuOGGtPHGG6cNN9wwyeFWf4EPp418xZgjveiWXHLJTD6GU1+cOzQE5AKceuqpMxFbddVVcw7BRoEiQ6s9zhoqAq+99lp67rnn0re//e20wAILZI3qZz/72aFW1/Y8BHCrrbZqW264BTzz0003XZL0W3AP0knbyQd12mmnzYtCO8Ysv/zyWcMsurxsRTjca8f5AxEQwLX66qunbbbZJlsCBpaIIMFSTAAAHgRJREFUX8Y7AkECx/sdHKP2c+Y/4ogjEvPNWmutlbfnGkkSaGL/85//nNOT/OY3v0l/+9vfxqincVnagKKFo7H73e9+F6B0CQK29rOzC7P6aJFA2mAkcFLIrLPOmucVzz/T/Y477pgXmUceeWQOKBIctNFGG+WtBpXln9gub+SkaHcvX8MWktNPP33afffde7mbE7ZvQQIn7K0fXset0CXv5SPECbsaPTq8muPsbkPgvvvu69uPV5LnkST73dbX8dYeqXUE1YwmCWS2l55ntOXMM8/M2j5mSB+LTBo/AUTPPPNMkmfy1Vdfzbkmt9hiizTNNNNkn1XHQ0YPAYFgksIL3OokGGz0WhI1jwYCQQJHA9UerdPL34ep5tZbb817tHo5WKXT2gU56NEbn1LWNvHP9DIO6R4ERpsECs5CAlsl8h4pNOacc84cRa4+ycPt+jLTTDPlfI80nkigOcYxhFF5gUKD3bt6pNo7UepByBHtD33oQx0HhE0UbHqhn0ECu/Qu8vlh5mGGE6F5/fXX55Y++eST2T+DQ7hj/LWqPlqS/C6++OI5pYbJm7n2xhtvbNtLvh80CqVeZr+vfvWr/c5johG1azVudbjoooumRRZZJF100UVJ1KhjVWGqkadNXYI8yLnnnpsnE23nzH7TTTf1neJ85FIqECt9EcH2pW23nRh/RCbL4rumD3xYvCxCRgYBgQFy+YV0FwKjTQJLtO5o99rYEiGOcBDuJnzRaDgtNM8444y+51kZpFCUuu0NEcWQ0UPAvC85vEXg+uuvP3oXiprHBIEggWMCe/uL7rfffpnAIVgcok10Eqha/W6yySZ5z9Vrr702EzEk67zzzsur6DnmmCMTLsfOOuus7Msj2q6ZH5ftv7bccstkay0rbxGgv/rVr9JRRx2VJ2X1FpEzitYPSRMMIkrU9ZiD5WQzURS5+OKL8wrdlmP8SUzwdnDwUnGNK664IptzEMkiAk1EOPJt4vez4IIL5tQXt9xySyky4Ft7tEO9Rx99dN6XFnk1YdmbNiQQ6GUERpsEImEWbKMt88wzT99C0bUsgkWLCwoz14hK9xsRMCJFzsorr5zTB1lMhowuAuecc06eU82zhaiP7hWj9kmFwOg/3ZOqJz10HZGxIuQk6KVhQ7Ro9Wafffa01157ZZMcbRwNmcg5k7RJ1AdRQ+wc479hgnRuIy2ONB/C/5FIecaY+lyv7OUqMrfqbE4zx/QiUg+xm2KKKTI5pQ1cYYUV8m4Q5TYss8wySRoLq8jZZpstt9E511xzTd4pAsGjJdRPYuXvWjvttFOe8E8++eRMAvUZIW0kBx10UG6H6Ei7Q9D8MRl9/OMfz9dDDkMCgV5GoBkJpIUvBI7JlAbnxz/+cTbneV79RmPuOfWC99w3EnkWzQ+jKawK2lLdu5pV4OGHH84LRlaQF154oc/SYG5j3UAQzUHzzjtv2+axKFhYmm+quSPNndWPXJLrrLNO2/omWgF4r7TSSnkup4wI6R0EggR24b2kRRONJ1ceoUUzUXGS/u1vf9uvxd/85jfzMYQKaUSCqkLD59xjjjmm+nP+e9ddd80vA2SJSbUufEBs81UEOVQ//yCmmg9/+MN59wdBISWlg7JM1q6LlDmnmJgFFYjy9dE/7dpss81y9V5EVvYXXnhhDjLhg4L8SjLcaHsyL7QZZ5wx18OUXUzR2ucFp27ENCQQ6GUEmpFAWnWLpJI+hgXB88XCYMEkf6P0Mp4viz0kyTNclTvuuCMvIL38R1NcXwRqXcwdyKln2mKS8Al86aWXstXB/GCRK0CtmZgvWSAKITYvtPqwIOy///7Nqpuwv8vQABcLeVafkN5BIEhgF95LWjsky8QnCresXJmG62ICNKmZ7OupEqyY+cw4brVdlXvvvTc/0I4hg2WSVcbesMUcXF31mYAROASMNsEuEt///vezmcb5JTCE/9jOO++cL+c6NJGuc+edd+bfTO7HHXdc3uHg/vvvz7/tu+++OQXBQw89lFNDaBMfv2WXXTbddtttuUz5R/v4Papz3XXX7Zc+Rt1SGWy99db9/A3LufEdCPQSAs1IoD4iTxaGXEmQJelWHnnkkfycel49m7Rj3EX22GOP7OtbxaYsMEdzJxJziTmqLOKq12/0N23h3XffnV1Q9Aux5frSSBBA84cFpxyXP/rRj7JbjPyKFpbIpwW27QRZX1hXzB8hAxFAvE855ZS8uKdlLnP9wJLxy3hDIEhgF94xfnQlGo9JFdmxkkV+6oL8OU7LVhd1OIZEPvXUU/0OC9RwzGe33XZLP/jBD/Lqmh+OpLC0jiZ/k2NVCkGcf/75c+CKwI66BoHpxsuGfPe7383XkM6hmuuPlrDqp2iz+Ntvvz1fj7/gKquskiccJuL67hS0hZNPPnmut5GZW34xaU3q7ar2o/xt0vdyEBTDDDSUDyLKrzImxoJqfE8qBFqRQItAL+6ZZ545+8x6xhGdIo8++mjafPPNs0bdM19/1ssC0/M2WsInuJEWsNH1EEXEDrHl8kFDKUq43u5yruMIJh/qBx54oPycv7mfmOvMf1xcGs2t/U5o8h8L7wsuuCD3gYaMdrUuyLb5u2pVqZcZq//DoZB9pnWkudE8Zr6Wo9E8jjQ3KjNWfYjrDg+BIIHDw2/Uzy4TsUmvLlZnxfQp6rYuJleTHF/B+grXw+6Y82ndmImQQUEetAIm10a5//wmEpcZybcXSSspiYblmOpk4mD+QVCZovkNIpF1ob3Qdv47dfN4vWy7/9OIIJlIo8l6qJ9CettdrxwX+W3S1Y+hfPTdvWomfMAQ2/gMH4NmGDf6vUTTDuWeOsc97eQ5KdfuhAR6ljzv/IURwyKFBHIHEY1fJ1Of+tSn8tjkbtGpcGXRj06InUwG3Fg6kfKcIqRInT3F7S7CZaYRXggL/z8aUAvVehl9tfeyti611FJD3lMaibIQta2l66mrfi3aTotWEc7dJoi1PpxwwgmZ4PE5r7dfm6UAE/jHP1SgTqMy3da3aE9nCAQJ7AynMSslPYKJCjGrC78fx5hb63t6erilUHBcsEX9obX1kmOI1nXXXZc/6uMHZFJoJOrw4qAtK348TCtF6tfwe2k//6RGx8u55Zt52MuHKZgvT30FrxzzlbYjt8MVbfKCYWYazqcTrWO1raNNAvlabrrppvEZAQyq963d391KAmn2mU2rkZ3tSCANt+esWCVa9d0ik1ZReZ9OTLwI06GHHtqq2nzMMyrITVAZi4eo//PPPz8TQNetEttSmeA1xGv77bdP/Nnq4jeaOW2FTd3aUC/f6v8C6gTNmZcaacloGc1jiFQ3Cm0m9xw+1oh5IwkS2AiV3vgtSGAX30eavuLTx5RTl8MOOyxPYpx165o+EzxfGASRBoBwCH/sscfy3yVYg6auU3ENPoIf+9jHsj+g9DElg3whU9W6TC4lAKTZ5FItrw6rahO4LaFoE9RvEi2BKwhbqRNZHK6Udg+HADp3sCRwuO2O8wMBCHSqCRwKCaQdRJI6MQcLEmOtYH4uRND81EykfPEcN4tKrp7Ht5GriMAE1goppswt5iNmyjpJpR00J2pHM3Mx3GREML+yLAxHpLCiHYMBrd94EynBYGFR3YyoFnOwxblxYd4M6Q0EggR28X2koTORWdE+/fTTA1pq4nHcir0ucuY5xtRT/IAESwjkIPyEHG9ltkHAqhpGWgRmlhJpiBDy85O/S/LWuklaxKFrmOybaRer7XY9K3dayuWWWy4xLTHbqIdWkJj4iz9gNUl2tR5/m8waaRHr5YqZaVKbg+vtiP8HAkNBYDRJIO2W57eTwJCqv2/JZsA82izggx+e9FbtxJzDReOQQw5JgseYpi2OPbeO6X+ZG0pdAslkJ7AAlk+wkcihiiiyaCCKwxF+27ZVY3LX1iLmKnMX0lpyHJZj3fQNU4tuwTPNyB0SGIEh3XTXRq4tQQJHDssRr8nkaxLm01MX2ieTrOMSJteFT5hjtGUebEEYyjP3EhOn4yK9GgnyRdNXdvpQxqqdj6IdTDiSI1kmYlo6vkzSvFSltF+ewGaTS7W8+rwcpGkQyShABgnde++9+2Wqt2LV9n322ad6et/fJiz+Ql4a7cREHYEh7VCK492KwGiSQIswREoS+cGIPX89nz6NAreQNBq4dlpABJLlwhzEdItMSW3DMmB+Kv54zMpV4Z9rIcnfWSRxXVhJzFfaQKvVyJxczuEDLWcpH0TBanxsaTwtwAlyp33canbYYYd+fZJhARbK+h5NMVcK8NBOms1tttkmt5XmtD73mtfN+0i4xPrOgZeMDvWypc3VFDGtFAelfHyPHwSCBHbxvSo+OSJP6+KhN0H7VLdeK+UkUDUJH3744XnSZEphEiryP//zPzliTcLVYlpBLGnQaB2tbvnzVRM1m/BMmkidSVkbkEs5/UyA9UCSavubTS6lPb5N2CZYvjXarR0ImoTWXgBFaAUERli5aruXBXOsl4NzODAjiohgSCDQiwgY88gLP9611lor5+Kk6Wc6tTAjjtPeWEQOxRysLvMLrdpgpQQ90bTVxXPLx7nVnOAYsy6fQQtDpM2cIIDBfOPD3IyUCTqrixRWLAbwqbpqII+igkXrmp9Kiqr6+f4v6AxRUr/IX9YCmj2ks2CCpApUqc+VNIJ+Z8aW8J5lg7hv/KgtpvlWCyhB1Phtm1v11bzaaTl1WjzLoqA+lhqZIJA2lpl6tK+5UTnXO/XUU3NAoHcF6w4/y2bC5C5tGY2hxUFVaGVlhHBtRNncqx8h4wOBIIFdfJ846oqe83DVxQNughayX53kSjmTJjOsCcaLALEyIVWFNlC0F42ZScsKXf6sjTfeOL846tFsXiomXr4vnLOldfF/q04EzsRVlVbtr5Yrf1vBe6GZkJiiTPjaqH3V1Tqy56VWErvKLcjsY1LTdpNuXStZrhHfgUAvIICgXH755fk58TzSaiEafIe5aXiemTk9D/y4uH8IWrAwMg/wE0ZspGhBiJAmwQHVvKBwKr7Dg8WMhqloA6sLOM+pBVy7QAzPu7lIFgLlafU879paPv7P9Ixs1UUQCYsFEgQnWkO+e7RkCJz5i0awmSDS0r3Q8CFMiBZSrP3ImqAr4jrmIkSxOr/yDYS1oBHtZTkh5mrmV9pDRFYbzaE0nO4pMgm7duUEICln0W6udB73GLusWIxrp2A897+QbQRWOQRUfwiXI+1fYoklchvyjw3+4UZk+9JG28aZ9/WTZUb9zOyxn3MDELv0pyCBXXpjNMvkjrzVNWyOWY3J/8ds2kjs2iHAwkTML5Ajdl1o/mgITYrKIZxWcVa/ja5pxedht4qnDbQqFFgiObS66tKq/fWy5f8CSGgKTJwmlKLJLMfLN5Mv/0EvuNJ2uQvta9yo7eW8+A4EegEBC6Ti11vIlm9k0CKKZr6kjyrHPd8SvEsJRVNTtHXluEVjfXcQCyrHByueT/U5F8koIiq+ldmxlEOKkNp6G4v1o3xbxNLsNZJiKjdHmCvNc0hRPVVOo3MtLLUb6UGs+TwiU0iXBa9gCoJIm6cshKu5WGkYzUMW8NqKiBLkzt9yoWq3XY1o2bil8HWklbMQblQOiS3lzK3KScAvV6z3hHcBIm/ORAi54yB5BIGlyZMyR6CP+on3CKJLo1rIYj5Q+ce8T4NoQeCadUECWYSYvL1zuB3BLWR8IDD4p3t89CtaGQgEAoFAIDBMBGizkBikaLCC7BWCyRcZcUJEy3aYg61vUpYXOIJYMZ3SqjUiSCwSzLCID7eZ+kLY/5lHEWmuNAgcQsWFhR81bSKNJDz8jrRaXPPjblSONrOUU6dyyJlFPo0nQsqNB2m16EdEiwUFcaV1pPm99dZbM5Tqom1F2iwMGvVRQVrRorUMf8BJOQonzbWCBE4anOMqgUAgEAiMSwSQwHrwRScd4S9XtryUgoS26fOf/3xTstFJnZOqjOT5zJ+2oKznGUSekDrECi7ysQq2oP2sJq/n3kIzy0yKSBXyRXMmXZfzaOsQNb9x22FGldmB1Mupo15OMIigDtYZpmCZIJRD6Eo71UVzrBwTcWmj67oWCwoy+eKLL2Z/xTrG8kvS4NqiT5mQ3kIgSGBv3c/oTSAQCAQCI4qATAP88oYiJQsBjSBCyMw7HoTGi4YMaSqaS8QKcdIHbiv8jrnPFP9oJK6aTcGOJPrMXcaHVpHQCDJJI1/8mBE72kE+mwL1SpqZTsrR/HEL4B/YqJ3FDK2tNIaux+zsmkgjEzKCJ+iFD3gjX2pYMKXzjQzpPQSCBPbePY0eBQKBQCAwogggcVI1DVZoy4pJGFkaLyLjwuKLL579jvndiX5FspA5KbmYT6WOEf0s3QpyxfxNw1dEBgXBKwieyG1BN4ik1DjMq0y6NHHImMALZPuYY47JWsZOyzHT86GmYWVyr7ezpIihwRP8g5gKiOG/yPebdlCKGEEtTNtM3FXRH7sPCYYJLWAVmd75O0hg79zL6EkgEAgEAqOCgHx/CI3AiMGKYBO+gNI3jReh8aPZE/WLJAlCQ+R8pK1BovjoibSlDVTuggsu6Nc95JEGTeBISbWF3ImMFrjjd4Fsxx57bCZkxx9/fA68o6XrtBxtoUAPBNJHe+rtLI1SDhFkwhbgQYsoglwEtlQ1jbJQ0BQyW5dNBkpd8d07CAQJ7J17GT0JBAKBQGBUEEBKmHaZKwcrtGeNkkYPtp5eKM9PjzlZdLb8rzRwNIS+kcpCADstN5qY0BTK29qIHI7mdaPuSYtAkMBJi3dcLRAYcwTkC5Mvjc9TcdwvJrvqNz+w4sc05o2OBow5AqJdEcFW2zWOeSO7vAFInqjdueeeO6dUoXGUfga2iF+RTsuV8iP9jbTTDl500UX92jXS14n6xh6BIIFjfw+iBYHAJEFAVKCXuByMVbLX7G/mP3kkQwKBQGBkEBCUgWDJ5cgcW/IP0rRWpdNy1XPi70BgKAgECRwKanFOIDDOEEAA5RKTwJcjue34JHi1hZVcYRzc7b4gQtHWWlJN8DcKCQQCgZFDQEoYiZTt9ETL3iiJv6t1Wm7kWhY1TVQEggRO1Dsf/Z5QCHDOl7PMFltly6gCAF8kuxfQCPJVqkcIlnLtvkVBco6XUkLiWttu1UX6C5pIkZPdJnCQdw0OdqpAmkMCgUAgEOhlBIIE9vLdjb4FAillx27+f7aMkvqibnqi+ZM6AvkROfjEE08MCTckqvgSuZ666teyp+rkk0+ec5IN6SKjeBI/LH2QWJemZq+99hrFq0XVgUAgEAiMPQJBAsf+HkQLAoFRRcB+sYiXNBf13Q9c2G80c0ig/VKlsBiq2Fvahva23GJarpNAWkaayKGkGhlqmwZzHm2m/VdnnHHGdNlllw3m1CgbCAQCgcC4QyBI4Li7ZdHgQKBzBCS4pZVD8OQEo/Wriz1LbevFXCxqeDhik/sNN9ww2eBeZOF4E/vbwkLut6Gaxcdbn6O9gUAgMHERCBI4ce999HwCIHDbbbflPVDt//qLX/yiYY/POuusNP300+fgEERxOCLicbHFFsv+dAhVEUEmUtNccskleUeE8nu3fV955ZVp1llnzcEz3da2aE8gEAgEAiONQJDAkUY06gsEuggBm9jbtWDKKafMG97Xm2YLqe222y5rAW0NZfP5ZvL888/nrbLsoLDKKquk9dZbLyeTpfkjr732Wt47dcEFF8w7Ivh/ETsmHHHEEVlL6Hs05fHHH88BHtpJs7nNNtvkttpCqy7nnXde3vZLYlz7rzrHVlq2AAsJBAKBQKDXEQgS2Ot3OPo34RFYeOGFs0/gDTfckN58880+PJiGRQWL1l133XUHRA33FUwpSTGDKNkeS+SvDedp9gR/IJnksccey9HH1W2y/C4iWFSyXRHsl7rccsvl8vwFBWPY95S/4rTTTpu311LWfqzMsiJ0SyobW2JJqyGJLUJnQ3vEVUQz4kf4G9oDVTk7Hdjei8/jueeeOyDa97TTTsvl1GPv1/322y/NO++8OYdbbJOV4Yx/AoFAoMcRCBLY4zc4uhcIXHPNNWmWWWbJJOjyyy/PEbB892jJELhtt902byrfDCnaQeleaPgQJkTrjTfeyOZlJGzTTTfNp7qOwBJEkRm6iIjg008/Pe+LinButdVW+RASqJzrI5lyGC6++OK5XgTVXrNStRxyyCGJmfbggw9OJ554Yu7LYYcdlrffch7SSYsn2ERQiv/b1eLhhx9OtJfM4AcddFDaeOONS5MygVUOsSwpc6677rrc/iWWWCK3p69w/BEIBAKBQI8iECSwR29sdCsQqCIg+IP5duqpp85bxSF/SNFVV12VbF3VSh566KFMjiS5lQew7HKAdN19992p+P7RECJW22yzTb8oZNtkIWM0c3wTEVGCBIrA9aGlnGyyyTLRQzodc45t7Wj1nHPFFVdkzeUKK6yQbr755pzM2s4KtHdrrbVWuvPOO9MCCyyQI5NtfC/K9/DDD8+EUPJdJI8gsKuvvnpOmXPhhRf2aUcvvvjiTHThFBIIBAKBwERAIEjgRLjL0cdAYBgICByR8gUZu+uuuwakfVG1SFpmWKZn/n/2Qq2K/6+88srZHMxXsOxGIiULrR8tH00grR0CSOymgARuttlmSbkHH3wwax1pDkU922sVuZ1tttnSpz71qcS8O8MMM2SNJ0IqSOUHP/hBrgcRLWQXcaUZXXTRRdOtt96ar6UuO6eIpN51112rTY+/A4FAIBDoWQSCBPbsrY2OBQIjgwBfOf56u+++ez8Nn9qRJ6QOseIfuPTSSyfBFi+//HI/k6oAFSRvn332ycEnVfKljjXXXDNH5b7++uuZBPpNsmZpa04++eTsO3jPPffk+mn1EDr+hKecckqaaaaZ0tFHH52DQQR17LHHHtkUbOu7olUs7dRmyaCVYyJmTibqU8f888+fyeSLL76Yt9TLB+OfQCAQCAR6FIEggT16Y6NbgcBIIWAbOBoypIkGjtDWIU7PPvtsuvTSS9M555yTc+vR0jHvIme0e0XsSEKrR5PoQ6tY6qFFlJyZLyGypm5Jp2kV+efZwYTmkOkaSRPEgQDSDjI9L7nkkjn4hOZv5plnzv6BjdpZzNDaSmOoHuZkdfEfZEJGRgW9nHHGGblPpf3xHQgEAoFALyIQJLAX72r0KRAYQQRsNSdgQwQvv7sXXnghk0Fk7sgjj8zmU9q6+eabL0f1IldSrNDcFWFinWqqqTJx23rrrdPVV1+dDyF8gkNoCaVnkbKG1pBJGQGUt5ApmYZQUIjoYho7pE0E7/rrr5+OOeaYrI3ku7jqqqvmaOM77rhjQDtLihhavkUWWSRvled6IohFPtMOaoMoZqbtRom1S3/iOxAIBAKBXkAgSGAv3MXoQyAwigjQ+NHs0dQhScggIudz6KGHZhJlz12RtnbaUE4ASVWQRyZlgSMIXhEkkJbQtna77LJLOvbYY3MkMtJIG8esTDtoK7sddtghn3/ggQfmcojb8ccfn7WGtHm0hQI9EEMf7am3s1xXOUSQCVtQCS0iwjnXXHPlFDSCWEICgUAgEOh1BIIE9vodjv4FAl2KAAKI4DEN8+tjhqUh9KFFLNu2KcPsLB/glltumTV1ytDYIZ8IoLpCAoFAIBAIBAaHQJDAweEVpQOBQGCEEEDcBJXwN5SsWiJr6WekcEHsivhbdO/cc8+ddx2hmVSOmRhBDAkEAoFAIBAYGgJBAoeGW5wVCAQCw0TglVdeybn7mIJFDRfNX71awRtnn3123smD2bbkKQztXx2p+H8gEAgEAoNDIEjg4PCK0oFAIDBCCMjlt9BCC+WoYd9S0TQidlLCSFQ9zTTTZNOx/IEhgUAgEAgEAsNHIEjg8DGMGgKBQCAQCAQCgUAgEBh3CAQJHHe3LBocCAQCgUAgEAgEAoHA8BEIEjh8DKOGQCAQCAQCgUAgEAgExh0CQQLH3S2LBgcCgUAgEAgEAoFAIDB8BIIEDh/DqCEQCAQCgUAgEAgEAoFxh0CQwHF3y6LBgUAgEAgEAoFAIBAIDB+BIIHDxzBqCAQCgUAgEAgEAoFAYNwhECRw3N2yaHAgEAgEAoFAIBAIBALDRyBI4PAxjBoCgUAgEAgEAoFAIBAYdwgECRx3tywaHAgEAoFAIBAIBAKBwPARCBI4fAyjhkAgEAgEAoFAIBAIBMYdAkECx90tiwYHAoFAIBAIBAKBQCAwfASCBA4fw6ghEAgEAoFAIBAIBAKBcYdAkMBxd8uiwYFAIBAIBAKBQCAQCAwfgf8HqxM58oJvESQAAAAASUVORK5CYII="}}},{"cell_type":"code","source":"C70, C10 = tf.constant(70, dtype='float32'), tf.constant(1000, dtype=\"float32\")\n#=============================#\ndef LaplaceLogLikelihood(y_true, y_pred):\n    tf.dtypes.cast(y_true, tf.float32)\n    tf.dtypes.cast(y_pred, tf.float32)\n    \n    sigma_clip = tf.maximum(y_pred[:, 1], C70)\n    \n    delta = tf.minimum(tf.abs(y_true[:, 0] - y_pred[:, 0]), C10)\n  \n    sq2 = tf.sqrt( tf.dtypes.cast(2, dtype=tf.float32) )\n    metric = (delta / sigma_clip)*sq2 + tf.math.log(sigma_clip* sq2)\n    return K.mean(metric)\n#=============================#\ndef regressionloss (y_true, y_pred):\n    tf.dtypes.cast(y_true, tf.float32)\n    tf.dtypes.cast(y_pred, tf.float32)\n    spread = tf.abs( (y_true[:, 0] -  y_pred[:, 0])  / y_true[:, 0] )\n    #spred = tf.square(y_true, y_pred[:, 0])\n    return K.mean(spread)\n#=============================#\n\ndef OSICloss(_lambda):\n    def loss(y_true, y_pred):\n        return _lambda * LaplaceLogLikelihood(y_true, y_pred) + (1 - _lambda)*regressionloss(y_true, y_pred)\n    return loss\n#=================","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:57:25.124472Z","iopub.execute_input":"2022-01-19T02:57:25.124903Z","iopub.status.idle":"2022-01-19T02:57:25.134205Z","shell.execute_reply.started":"2022-01-19T02:57:25.124864Z","shell.execute_reply":"2022-01-19T02:57:25.133468Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def model_2D():\n    ct = L.Input((128,128,1),name=\"Ctinput\")  \n    Patint = L.Input((5,), name=\"Patient\" ) \n    x =L.Conv2D(64,(6,6),activation=\"relu\",name=\"conv1\")(ct)\n    x =L.MaxPooling2D(pool_size=(3,3), name='pool1')(x)\n    \n    x = L.Conv2D(64,(6,6),activation=\"relu\",name=\"conv2\")(x)\n    x = L.MaxPooling2D(pool_size=(3,3), name='pool2')(x)\n    \n    x = L.Conv2D(128,(6,6),activation=\"relu\",name=\"conv3\")(x)\n    x = L.MaxPooling2D(pool_size=(2,2), name='pool3')(x)\n    \n    x = L.Flatten(name=\"features\")(x)\n    x = L.Dense(64, activation=\"relu\", name=\"d1\")(x)\n    P = L.Dense(32, activation=\"relu\", name=\"d2\")(Patint)\n    x = L.Concatenate(name=\"combine\")([x, P])\n    x = L.Dense(32, activation=\"relu\", name=\"d3\")(x)\n    preds = L.Dense(2, activation=\"relu\", name=\"preds\")(x)\n    \n    model = M.Model([ct, Patint], preds, name=\"CNN\")\n    model.compile(loss=OSICloss(0.5), optimizer=\"adam\", metrics=[LaplaceLogLikelihood])\n    \n    return model","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:57:27.726545Z","iopub.execute_input":"2022-01-19T02:57:27.727239Z","iopub.status.idle":"2022-01-19T02:57:27.738182Z","shell.execute_reply.started":"2022-01-19T02:57:27.727202Z","shell.execute_reply":"2022-01-19T02:57:27.7361Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"net_2D = model_2D()\nprint(net_2D.summary())","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:57:29.943777Z","iopub.execute_input":"2022-01-19T02:57:29.94437Z","iopub.status.idle":"2022-01-19T02:57:30.035536Z","shell.execute_reply.started":"2022-01-19T02:57:29.944332Z","shell.execute_reply":"2022-01-19T02:57:30.034871Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"y.shape,feature.shape,xs.shape,x_train.shape","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:57:32.567532Z","iopub.execute_input":"2022-01-19T02:57:32.568239Z","iopub.status.idle":"2022-01-19T02:57:32.574559Z","shell.execute_reply.started":"2022-01-19T02:57:32.568197Z","shell.execute_reply":"2022-01-19T02:57:32.573662Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"y=y.astype(\"float32\")","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:57:33.985556Z","iopub.execute_input":"2022-01-19T02:57:33.986096Z","iopub.status.idle":"2022-01-19T02:57:33.989794Z","shell.execute_reply.started":"2022-01-19T02:57:33.986054Z","shell.execute_reply":"2022-01-19T02:57:33.98904Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"net_2D.fit([x_train, feature], y, batch_size=50, epochs=100)","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:57:37.808158Z","iopub.execute_input":"2022-01-19T02:57:37.808422Z","iopub.status.idle":"2022-01-19T02:59:00.625471Z","shell.execute_reply.started":"2022-01-19T02:57:37.808391Z","shell.execute_reply":"2022-01-19T02:59:00.624642Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def model_1D():\n    ct = L.Input((128,128), name=\"input\")\n    Patint = L.Input((5,), name=\"Patient\")\n    x = L.Conv1D(50, 4, activation=\"relu\", name=\"conv1\")(ct)\n    x = L.MaxPool1D(2, name='pool1')(x)\n    \n    \n    x = L.Conv1D(50, 4, activation=\"relu\", name=\"conv2\")(x)\n    x = L.MaxPool1D(2, name='pool2')(x)\n    \n  \n    x = L.Conv1D(50, 4, activation=\"relu\", name=\"conv3\")(x)\n    x = L.MaxPool1D(2, name='pool3')(x)\n    \n    x = L.Flatten(name=\"features\")(x)\n    x = L.Dense(50, activation=\"relu\", name=\"d1\")(x)\n    l = L.Dense(10, activation=\"relu\", name=\"d2\")(Patint)\n    x = L.Concatenate(name=\"combine\")([x, l])\n    x = L.Dense(50, activation=\"relu\", name=\"d3\")(x)\n    preds = L.Dense(2, activation=\"relu\", name=\"preds\")(x)\n    \n    model = M.Model([ct, Patint], preds, name=\"CNN\")\n    model.compile(loss=OSICloss(0.5), optimizer=\"adam\", metrics=[LaplaceLogLikelihood])\n   \n    return model","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:59:44.334182Z","iopub.execute_input":"2022-01-19T02:59:44.334481Z","iopub.status.idle":"2022-01-19T02:59:44.34532Z","shell.execute_reply.started":"2022-01-19T02:59:44.334448Z","shell.execute_reply":"2022-01-19T02:59:44.344592Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"net_1D = model_1D()\nprint(net_1D.summary())","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:59:46.04387Z","iopub.execute_input":"2022-01-19T02:59:46.044479Z","iopub.status.idle":"2022-01-19T02:59:46.134508Z","shell.execute_reply.started":"2022-01-19T02:59:46.044437Z","shell.execute_reply":"2022-01-19T02:59:46.133761Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"y.shape,feature.shape,xs.shape,x_train.shape","metadata":{"execution":{"iopub.status.busy":"2022-01-19T03:10:43.673554Z","iopub.execute_input":"2022-01-19T03:10:43.67429Z","iopub.status.idle":"2022-01-19T03:10:43.680245Z","shell.execute_reply.started":"2022-01-19T03:10:43.674253Z","shell.execute_reply":"2022-01-19T03:10:43.679532Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"y=y.astype(\"float32\")","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:59:50.720975Z","iopub.execute_input":"2022-01-19T02:59:50.721525Z","iopub.status.idle":"2022-01-19T02:59:50.725436Z","shell.execute_reply.started":"2022-01-19T02:59:50.721485Z","shell.execute_reply":"2022-01-19T02:59:50.72443Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"net_1D.fit([xs, feature], y, batch_size=50, epochs=100) #, validation_split=0.1","metadata":{"execution":{"iopub.status.busy":"2022-01-19T02:59:52.462621Z","iopub.execute_input":"2022-01-19T02:59:52.463189Z","iopub.status.idle":"2022-01-19T03:00:07.534284Z","shell.execute_reply.started":"2022-01-19T02:59:52.463152Z","shell.execute_reply":"2022-01-19T03:00:07.533567Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"predict_2D=net_2D.predict([xs, feature], batch_size=100, verbose=1)","metadata":{"execution":{"iopub.status.busy":"2022-01-19T03:10:01.777214Z","iopub.execute_input":"2022-01-19T03:10:01.777468Z","iopub.status.idle":"2022-01-19T03:10:02.132428Z","shell.execute_reply.started":"2022-01-19T03:10:01.777438Z","shell.execute_reply":"2022-01-19T03:10:02.131719Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"predict_1D=net_1D.predict([xs, feature], batch_size=100, verbose=1)","metadata":{"execution":{"iopub.status.busy":"2022-01-19T03:10:05.302555Z","iopub.execute_input":"2022-01-19T03:10:05.303114Z","iopub.status.idle":"2022-01-19T03:10:05.652136Z","shell.execute_reply.started":"2022-01-19T03:10:05.303075Z","shell.execute_reply":"2022-01-19T03:10:05.651342Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":" pdata_2D = pd.DataFrame({\"FVC\":predict_2D[:,0],\"CON\":predict_2D[:,1]})    ","metadata":{"execution":{"iopub.status.busy":"2022-01-19T03:12:01.277136Z","iopub.execute_input":"2022-01-19T03:12:01.277398Z","iopub.status.idle":"2022-01-19T03:12:01.282261Z","shell.execute_reply.started":"2022-01-19T03:12:01.277368Z","shell.execute_reply":"2022-01-19T03:12:01.281473Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":" pdata_1D = pd.DataFrame({\"FVC\":predict_1D[:,0],\"CON\":predict_1D[:,1]})   ","metadata":{"execution":{"iopub.status.busy":"2022-01-19T03:12:24.551235Z","iopub.execute_input":"2022-01-19T03:12:24.551504Z","iopub.status.idle":"2022-01-19T03:12:24.555926Z","shell.execute_reply.started":"2022-01-19T03:12:24.551473Z","shell.execute_reply":"2022-01-19T03:12:24.555216Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"pdata_2D","metadata":{"execution":{"iopub.status.busy":"2022-01-19T03:12:05.133747Z","iopub.execute_input":"2022-01-19T03:12:05.134313Z","iopub.status.idle":"2022-01-19T03:12:05.146871Z","shell.execute_reply.started":"2022-01-19T03:12:05.134269Z","shell.execute_reply":"2022-01-19T03:12:05.145989Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"pdata_1D","metadata":{"execution":{"iopub.status.busy":"2022-01-19T03:12:30.113797Z","iopub.execute_input":"2022-01-19T03:12:30.114259Z","iopub.status.idle":"2022-01-19T03:12:30.128668Z","shell.execute_reply.started":"2022-01-19T03:12:30.114223Z","shell.execute_reply":"2022-01-19T03:12:30.126803Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{"trusted":true},"execution_count":null,"outputs":[]}]}