{"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":"markdown","source":"**Implementation of the original Grad-CAM paper https://arxiv.org/abs/1610.02391**\n\nGradient-weighted Class Activation Mapping (Grad-CAM), uses the gradients of any target class flowing into the final convolutional\nlayer to produce a coarse localization map highlighting the\nimportant regions in the image for predicting the class. \n\nThis results in a coarse saliency map, which can then be resized to the\noriginal input image size to render a class-discriminative saliency map.\n\n**Goal: To see where our trained model is paying attention to and see if the predictions trained on class labels have anything to do with the real bounding boxes.**\n\n**This notebook uses two network trained in my two seperate notebooks, one trained to classify opacity vs None (Image level classes), the other, for 4 classes of 'Negative for Pneumonia' 'Typical Appearance' 'Indeterminate Appearance' 'Atypical Appearance' (Study level classes)**\n\n[First Notebook](https://www.kaggle.com/sinamhd9/image-label-classification-opacity-vs-none)\n\n[Second Notebook](https://www.kaggle.com/sinamhd9/classification-model)\n\n**Note that the reference notebooks may get updated to achieve better accuracy based on the insights from this notebook**","metadata":{}},{"cell_type":"markdown","source":"![img_xray.jpg](attachment:c0a9e35c-4489-470a-adc2-7bd6be76ee30.jpg)","metadata":{},"attachments":{"c0a9e35c-4489-470a-adc2-7bd6be76ee30.jpg":{"image/jpeg":"/9j/4AAQSkZJRgABAQEAkACQAAD/4RD0RXhpZgAATU0AKgAAAAgABAE7AAIAAAAOAAAISodpAAQAAAABAAAIWJydAAEAAAAcAAAQ0OocAAcAAAgMAAAAPgAAAAAc6gAAAAgAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAFNpbmEgTWVoZGluaWEAAAWQAwACAAAAFAAAEKaQBAACAAAAFAAAELqSkQACAAAAAzIzAACSkgACAAAAAzIzAADqHAAHAAAIDAAACJoAAAAAHOoAAAAIAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAyMDIxOjA2OjAxIDE0OjAzOjA5ADIwMjE6MDY6MDEgMTQ6MDM6MDkAAABTAGkAbgBhACAATQBlAGgAZABpAG4AaQBhAAAA/+ELIGh0dHA6Ly9ucy5hZG9iZS5jb20veGFwLzEuMC8APD94cGFja2V0IGJlZ2luPSfvu78nIGlkPSdXNU0wTXBDZWhpSHpyZVN6TlRjemtjOWQnPz4NCjx4OnhtcG1ldGEgeG1sbnM6eD0iYWRvYmU6bnM6bWV0YS8iPjxyZGY6UkRGIHhtbG5zOnJkZj0iaHR0cDovL3d3dy53My5vcmcvMTk5OS8wMi8yMi1yZGYtc3ludGF4LW5zIyI+PHJkZjpEZXNjcmlwdGlvbiByZGY6YWJvdXQ9InV1aWQ6ZmFmNWJkZDUtYmEzZC0xMWRhLWFkMzEtZDMzZDc1MTgyZjFiIiB4bWxuczpkYz0iaHR0cDovL3B1cmwub3JnL2RjL2VsZW1lbnRzLzEuMS8iLz48cmRmOkRlc2NyaXB0aW9uIHJkZjphYm91dD0idXVpZDpmYWY1YmRkNS1iYTNkLTExZGEtYWQzMS1kMzNkNzUxODJmMWIiIHhtbG5zOnhtcD0iaHR0cDovL25zLmFkb2JlLmNvbS94YXAvMS4wLyI+PHhtcDpDcmVhdGVEYXRlPjIwMjEtMDYtMDFUMTQ6MDM6MDkuMjI3PC94bXA6Q3JlYXRlRGF0ZT48L3JkZjpEZXNjcmlwdGlvbj48cmRmOkRlc2NyaXB0aW9uIHJkZjphYm91dD0idXVpZDpmYWY1YmRkNS1iYTNkLTExZGEtYWQzMS1kMzNkNzUxODJmMWIiIHhtbG5zOmRjPSJodHRwOi8vcHVybC5vcmcvZGMvZWxlbWVudHMvMS4xLyI+PGRjOmNyZWF0b3I+PHJkZjpTZXEgeG1sbnM6cmRmPSJodHRwOi8vd3d3LnczLm9yZy8xOTk5LzAyLzIyLXJkZi1zeW50YXgtbnMjIj48cmRmOmxpPlNpbmEgTWVoZGluaWE8L3JkZjpsaT48L3JkZjpTZXE+DQoJCQk8L2RjOmNyZWF0b3I+PC9yZGY6RGVzY3JpcHRpb24+PC9yZGY6UkRGPjwveDp4bXBtZXRhPg0KICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICA8P3hwYWNrZXQgZW5kPSd3Jz8+/9sAQwAHBQUGBQQHBgUGCAcHCAoRCwoJCQoVDxAMERgVGhkYFRgXGx4nIRsdJR0XGCIuIiUoKSssKxogLzMvKjInKisq/9sAQwEHCAgKCQoUCwsUKhwYHCoqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioq/8AAEQgC8wLzAwEiAAIRAQMRAf/EAB8AAAEFAQEBAQEBAAAAAAAAAAABAgMEBQYHCAkKC//EALUQAAIBAwMCBAMFBQQEAAABfQECAwAEEQUSITFBBhNRYQcicRQygZGhCCNCscEVUtHwJDNicoIJChYXGBkaJSYnKCkqNDU2Nzg5OkNERUZHSElKU1RVVldYWVpjZGVmZ2hpanN0dXZ3eHl6g4SFhoeIiYqSk5SVlpeYmZqio6Slpqeoqaqys7S1tre4ubrCw8TFxsfIycrS09TV1tfY2drh4uPk5ebn6Onq8fLz9PX29/j5+v/EAB8BAAMBAQEBAQEBAQEAAAAAAAABAgMEBQYHCAkKC//EALURAAIBAgQEAwQHBQQEAAECdwABAgMRBAUhMQYSQVEHYXETIjKBCBRCkaGxwQkjM1LwFWJy0QoWJDThJfEXGBkaJicoKSo1Njc4OTpDREVGR0hJSlNUVVZXWFlaY2RlZmdoaWpzdHV2d3h5eoKDhIWGh4iJipKTlJWWl5iZmqKjpKWmp6ipqrKztLW2t7i5usLDxMXGx8jJytLT1NXW19jZ2uLj5OXm5+jp6vLz9PX29/j5+v/aAAwDAQACEQMRAD8A8U1G2ji1zUY4kCol3Kqj0Ac4FEduMDgVa1JM6/qR/wCn2b/0YadDEX6V9HS/hr0MXuQLBluFqZLViOQPyrQhtdoy1WobXcelaCM5bDpwPyq1HZKF5A/KtVbUKvTmpEtR1xzQBlpaBcEoPyrdS3ie3TaqhselR+SdvQGp4VJXaeD2oArSWKoclQxPtUL2StxsA/CtYQv/ABU5YSfvrkeoqQMVIfL42A/hVyBIzjcgB7ZFX2gjOcArQbLO0g5470ASLb21xF5dzErehUVUufDLqpksirr/AHSKvQ2jggqwA9BVmOeW2ychgTjFIDnYbDyVYz2+GHqKdb2MUiuNoAznpXYu1vcQqLhQMjnis9tJQbnt3yp7UAYUumqx3IFwRxxWcdOKkgBevpXVNCY44w6cAGqHlI5IPHNAGKdKLcoiqf50CwaP78YA7cVvfZ0wCDn8KsfYvOUA8g9M0Ac7HYHrhMfSrK2KkchfyrQm0i4iz5K719qqru3bTwR1BoAZ/Z8IHIGfpT1soR/APyqcIM5Jx7VMu3HXNAFX7MnaJfyoa1UjlVA+lWXm2Diqssjv3oAfaLCpKlVP4VdG1h8qLx7VkxZSZT6mtRDjp3oAfGitglFx9KtrBGSAUXB9qgiUHpV+FemeaAJbe0hLbfLXH0q4LWIf8s1/75FJAgVSasxoeSTQBCbaMjAiU5/2RU5tolUARLn/AHRUkYyQamI5y1AFdbaJhny0/wC+RU0dlEw4iT/vkU7GAcVLCzAfLQAsdnCW/wBUn/fIp5soUG4on/fIqeHB6jFSPGWXOBQBlXMahOYUx67RWLcrEM/ul/75rqriLdbntXOXMLA9M80ADQxJbRBFQZ7bRV7TY4mlIES9D/CKrRorqA5Awe9W7R44JgEbcx9KAK8iRvMQIlypP8IoS1hDbhEpP+6Kkl2IxPIYk0Iy+YCM0AIYV2/LDHj/AHRVWayDf8sk/wC+RWhvPt1oIz6UAYzWqqm0wrj/AHaqSQRBv9Wv5VvyruTanJNVRZqn+sOWpgZcdsmCfLVvbbTvsiFP9Wv021rCJVPAA47VItsJMYoAxBYx5/1a/lStYxbeI1/KtaaAxyYaoQjbjxxTAyXsoz/yzX8qqy2MXP7sflW7LGIwAOc1VmQdqYHNT2iL0QflVKe0jIGEX8q6G5iODxmsyaPBGKAMn7Mn90flR9nQfwj8q0Wh3e1IsNMDPFvG/wDCPyqX7Gm37o/Krv2bHI61IkDZ6UAUFsUxxGPyqVbFT/APyrSWEnrUixYPTigDNXT1bOEH5UpsFC/cX8q0wMD5RTQhLe1AGS1mo/gX8qDaIeqD8q10spp/9VEz/QVft/DUrjddOI09O9AHKm0VuFjB+gqaPRJ5hzGqJ6sK65LGzszhFy3qagnlV3KueOwoA5z+x7WDk4d/pTJ7XI/1aj6CteTyQ3FVZWUcgUgMg20ecbFB9cUJaKrchT+FW3Zf7uT9Ks2ulzXUil8InWgCvBYvcfu7eJcjqcV0ljosFlZ+ZcRo8mM4IqS3WG0XEY5HUipZWa4t/kBYUwKl4RNYqViRSOwFYTwKScxqfwrqTYYtgxI/E1SMFurEnBagDFg0zz/uxgfUVcXTLSAgzYZvQDirpWQr8o2r6CkitC2d1AFSWCHjbEoHrjpWfc2EZHCjGfSt97cKuD0+lVvKV354H0pgYcemq3WPj6VJ/YUbchQPwreSFFJ5JxTmQGQeWvHfNAHOP4fUjt+VM/sBB1A/KupCDnjn2pjR5HSmI5Gfw+ueB+lVDoI3dAPwrsnibbVaW3+XJHNAHLnQIz1UH8KafD8Y4AH5V0ezC4IxTSnPNMDmH0Db1AI+lZmoeGjIpMaHPsK7hhjpzSAjvQB5DdWctrIUlTFV8V6ve6NZaohWVAW7FeDXF6x4TuLBi0H72Prx1FAHOYopzIVJBBBHUGm0Abum/wDIPj/4F/M0Uab/AMg+P/gX8zRXz1T436my2NK7j3a5qXH/AC+zf+jDVy3t9qjAp7Q7tY1A463s3/oxq04LXgHFezS/hr0M3uQxQbuCK0IINvap4bfGDir1vbBmOTWgiosBODgVKIR3ABqz5Z6Y49qmgtXlbHRfegDOeHbxTPJfcCFP1xXSm0ijjA2596eqbVAAXH0oAwY0YrzU6J2PAraECbTuRTTjp8Dx5U7W9DSAyBCp4Hp1NDwYXgdBVua3kt2HmDK9iKQGNmI3YGOKQFDy5FIIPHpRjuevpWm9luUNGwJqq1qQx3daAHwziRdsijpjNPt4wpbBP0qAJtYYrRtwjoxPBI4oAib95CFI4FY89mfMJjOM10htkCLtYciqElk4JO7IzQBiqJk+VgfyqVZ2B2scVpC2J/1i1FcafuO9RQA63uSoxkmp3sbXUFOcRydiKpJGyOABV2LcrDAx70AZF5o1xZ9fnX+8KrRrXaQMDHslAdW/Ss/UNGjbMtrwO6igDmnXNJ5Qx1q40W1irDmkWLJwBQBRkiOMjtVi3k3ADuKtfYpJF+VakTR5I2DngGgBkYOa1bZRt680W1lEv3yBWpbxwR4O7NAEESNirqRPs+6alSaFWyOlS/bo0GF5oAgSJwuNpq0luVjyQfxqW2uleQBhgGrbzREEUAZ3k8c5pBFJH9zkd6umaJfvCj7REORQBDHkjJ6+lWkyV5oSeDr3NSoEccNQBUuSRHgd6567D78ZxzXVS2nmxHa1Zd1pobB9ODQBzTBy5G49aksy0dzHjOc1qHT0jbJ5qSGGMMMJ0NAFZkMhLHnBxUkYdRjZVsxBWYY75pQy7hxQBEEO3JWoW3eZ6CrVw53AL0qIgFc0AVixXI/Wo85wetWntiV3Bhz2pv2ZjQBGq1OPlUHOKdHbYGSakDRA7WGcUwIJf37jHao2t3QEjBBqcsNx2UGUgHNAGZMMDDcGqpwWNX7lfMBcdRWc+UUnvTArSgknPSs6WLElartuj24wetV3gyM0AZxhPUCnLAcZ71eW2JHyg1cgsAVy4pgZccBPJqYR4HStB7dV+7TCm1TkZoAqBBj37UhUj61cS3eUjYn41q2eipHiS4bJPagDJtNLmuznG1PWtm10W2tRukIf2rQ8xYl8uMbVqJmDA0AMd44k/wBHVU/Cs25uWLfeyxrQkj3JxUC2W5slenegDGmdsEN19aq+WzvnBNbz20KFi5BPpWdPOqA+XgUAZk0IRsn8qqSfOxVBmrrhp3/rViG0jgwSNzGkBWsLBfvTKCe1ae8BMAAU+OMAZIxQwTZwMmgBVQEN0waFuvs8TIMGoWL5wOPalW3yw3d6YEEhkkXIYgHsTUkNo2FLDIq/DbKnM2MChrotKI7VN2ePpQBCIgPvH6Cpo7R2GT8q+9XoYPKUtKoL0rbmbgcUAUvsKu2C5I9qVNGVpM+YNvoavD5OcVKpHXvTAh/sm029W3d8U19DtxykrD61opFLJHlVx70yVZoV5XcKAMibTJIYS0bBwOuKprGGX5uDW+sgVM9jwRWfeW/lSh0GVamIzmjBPSoXg3ZJ6Vo7AetNMP5UAYcsPfHFQFM+1bc0AA6Vmzx7eRTAqbMDpUbR8VO3HApgyKAKhiwc9PekZwVKyc5HXFWJE3DJqKRMrxjFAHM6z4ct7yMtagJJ1yO9cVd2E1pIVkU8HrXqDDaeKzL60iuIiJEBJ70Acrpv/IPj/wCBfzNFWorbyVMY6KzD9TRXz9T436my2OrgtfM1G9b1vZ//AEY1bENqFTGOc1DYRA3N23f7ZP8A+jGrUjjYnPbNexS/hr0M3uMhtzIwCLzWzZ6Rj55DjNS2Fr5cAmxkk1aYuep6VoISOwgjHzAH8KkFrET8sfHtTFlOamEx4xxQBG1sMPhcD3qDyl6AVe3bs+9M2+o4oAgSEcEdqmEfB45p6J1xxT16j0HWnYCPy1eNkkTcP5Vi6hpzW7l4+YyR+FdBgrISO4pHjEiOhGdw49qOUDlYpJYJMoSRWgjpcD/b70y5tjDMVHFV42MTb+hzRygTSW2zOKdChjtznrViM+cvualeHZGARmnyAV5ObVSODiqUcsqE/Nn61qyWztCEUYAFMTTkVRu5NV7MVyg11Jt7Zp0N6Adswq41jGeCpFRS6aQvyfN/Sq9iwuTfZ45U3xEVEqMrnctVIpZrKT+8AeRW1aT22oDIOxwOlV7ALlcAptx6VYifZ2yDT2tpFbgZqeGDamZDT9gFzPvdG+1fvYRg+lQRaWtv802M1um5VAAlVLy2Nwu5Tg+lP6uK5nvNFEDsXmq8szSAE1J9ncuQw6VKsQPBHSj6swuUAr/hViMscdxU3k9u1Cx7D7UfVmFx8eTmrUUPmScCo4U+b61pwgRqdo+b1o+r2C4RwBQOcHvU5hBX5WqME7Tkc0pfaAAeaXsEO4ySJz1FMMZAqVpT1HNIZSetHsAuRIrE96sRu8Z9qjWXnpVlQJF44qXRsFyQTOAClRT3OM9/WnEbRgVUlHzbs5FL2IXHCeORgCKtxRwKo3YDE1lIAJfxq3Jy689MUvZBcnms8OWU9aqGBo2JJzV1pCI+fpUJlzwRS9mFyqYzJ0pzRFI+RVpdh9jSSRmRSA1L2YzP8wDg0wyFvuU97Z8k9RVdyUICrS5QJwz9BUTFuflpoMmODUZnkTOaVgLEYPBpJGqBLwdDwTUykSKcdaQFeTNV5YQy5FXmjJHzDFRiEjqeKAMlom3YFSx2rfxdK0/JUjpz61G0WeDQAxIlCjGKcy4FLjbxmpo7d5OvApgUzCXHHP0qeDTc48/ir6RJEvy8mleTOM9RQA1Fjh+VABilL7j81RkFm5709FZvlI/GgAK5HPSnpbSNgKOPU1J5sMK/PzVO51F2G2HgeooAts8FsMTMCwrMu9XiJPlk/hVGbc7ZckmoZE+XgZzQBVu7maXOTgHpiqYgll6scfWtRbXON3SpTZFVyeAelAFCMGHFTCQkcVOLAnq2aR7cx4yMCkAKxJHOacFKvnsaVUwMjpUqIWwBzQA3ydzD1qysaxDc9SRQldzOOKqXcgmYKh470wIZpZLqQInHOK2bGyW0t8kfvG71BpVoNzzMOF6e9apbceelAETJvUZ60mMDGKkJxkdBUfBPWgBgGXwelSGJTIAp4NNIwfam+ZtYY7GmBsxL91c/KByPpUzCG4jbZ8uwdPWqauWUSJ6dKljl4Kou3d1JoAyrmHyZGX15xUYHmRbW5qbUJQ1zx2GKjiYemaYiiyBCQetIMrVq9QCRCo696qMNw565oAR1UqdwqjcQDpjNXXT5RzUbEqvrTAxJ4Sh4qtjHU1syoJAcjBrNkgw3TNAFYrupjx4XjmpipU4I5pDz3x60AUZIwFxjiqNxGdprWnGOgyKoT8qaAOWIxI/++38zRT5B++k/32/maK+fqfG/U2Wx3mlwb3uz/wBPc/8A6MatWGIkqg7mqelJ+4uioPN5P/6NatqzjzcRj35r16X8Nehm9zS8sJAsaj7oBphwc1PNkSEDpUO3I3YrQREFX1xTgFY9MmlSMF8nrUoRQ3HWmA5HVF5GTQTvXO3ilEWELClEm1QBVIBBGQB+lOCYbJ6d6dyzZJ6U4kEetWkIOM5x8uKaMKSU5pwbPy44pxUAVaiBQ1CESW+4D5k5NZGwSMNo5PaukKh0KEcY/SsyC0JumwMbe9UogNsoUgmAbkk9K3BDHMfu9qoi3CSZbk9RVi0lZptrfeFaKGgi1LDAEA6NVGWAKeKvzDIBHWqzjmqgBTaPA+brUZ4HHXvV0QhiS3B7VE1uOp61unEhuxkXcCsSV64qraxSRygp61vx2XntgipDBHboRtG6r54rQh1Eie2kBhAcjIHSmyRMwOKpKHEgb3rbtI/OQd65pNQ1Rm6yMv7M3GOtWIrdsjIrZSzB5K082XzAgVk8QnoHtjEm04PyBg1SazdWxjArqjB2NRtZiTqKI4lrcPbHMCzYtQ9oTwBzW/LZmNj8vFNitdzbiK1+saXF7YyYrQqoNWEgY9K1BaAN0qVLTjkc1lLEXH7Yy1tn2nmmyQEYx1rWMQBwRk1A9uWzUqrqHtjL2diaZtIJwc1pfZMjBFM+yAVqqiKVVGcykLmpbdip61O1qR1GaRIsNwMVTkmi1UQEnJqnMdnWtMRZHNZ92NueOlFNq9i+ZFLzMYPvVsyfMuPQVnSSEHKjFOhu3Vx5gyK6HT6oLm0eYTmq5wKhNwzdOlODbmHNY8ltxkhzt4ppd1HBpSxFP27196hxQECXDb8dR3qdliZegzURTYeOTUbxyD5jwKzcRhNDtXg/lWZcJIyHPAq2155TYPNRSX8cqkNgVlKIzNidYSSzbjUkd4QdxPemSRJIx2OOagkhdOAN30rFjNy1vY5MK4yasSIp5K8VzUUsgbAVhWzZSyNHtk/WpAnMQI+Q8077OxUZ61MAijtR53y4PSmBAFSPORk0okPODgUPg5NQ9CaAJQ49aUMD96oQeaecbcnimBNGctzRPL5f3SKpy3WCFU9qrCRuS9ABcSNI3WkAK9PxqQASDIAzUsdsWPNAEBiMg4p6QKqgNzVnYRx2phXnHQUAJFDlt23IFEsBkcccDtVyBflwvNDqxJ29utAFdY1GB3qQW8c+VdeT0NGKlQncpWkBVfTVRsZ3AU6G2VZM7avyqdw2/jTZBsj3L1oAq3UqrEVx1rNa224ZOh61alJmlz2qxHGXZVA4zTAsW8ZjtY0/E/jUijPIFOYbWNGCideGoAY6gio24HAp5bsRjFMY5xjBoAa2QmTUbMQORwaWWTBqFzmmBNFcvC2UOV9KkOosykKMGqQZhwBUbsQvHWgC1kN8znJqSN1HSqQLE89KmQg98HtTAnuV8yxZl6oaoggnPtWlbr5lrcREdV4rIDdiehxQIc7VGzDb0p023bwahAzgUwGsMDpUUkOUzU7oN2KQjjG6gDLljYdOTVU8ZzWtPFu6dqoSoVPSgCoVyDu6dqpTp1rSc/L0qpNgqcCgDkJhi4l/66N/OinXAxdTf9dG/nRXz9T436my2PRtHLeTcen2yf8A9GtW7ZANdoPWsbRiPs9xn/n8n/8ARrVt6eQ1+gAxgGvXpfAvQze5elxuIqIKAP6GpZF/etUJ5Yk/NWghm75uBjFSKc845pgOSf5VLGMU0Am/C8Cmhc05iCcCmE7RjvWiETA4TJ6+lOT5hhRzVeM5Y5NTI+OVPNbRQE+3H1p3XtTAdy5JzSBsHNaqIiRTlvQ1H5XlyEnvzUm5SQe9En3h6VS3EMkGI+Kij+UGToamn+58tMIIjCkdatbAXbRxNGQBn3p/2Yg5PAqGxURcd62UjSSMK2ea5akuSWhnKZlmEZ9aVbTzGAUVq/2eD/q6u2+nsqA8ZxzWMsQkjmlUMOaBbeI7euKy1haRyWGea6a80mS4k+U4FLFpG04I6URxEVHc5J1TFt7HdwV4rXs7DylzirqWQTAxzVpEYcY4Fc9TEOWxzSrEUdurYxUhtfyqaFQrcirGAa45VGmJVmzOa0B6CozalT061rYGelNKA9qFVY/aMzPsm5RkZHrSNYgfdHFaioF7UbelP2rD2rMo2eMEjFMkQLwvJrZZQVwRVGS1I/1fNVGrfcTqtFJYeCWFMa3GeCDmrLrtGDnNMMZ3D5c1spC9uQvbZXiomtm4wKvL70pHIxTU2jWNYy2tTjlaiNqT0HNbe0N1FIYE6kc1SrtG8apjfZ/LXL1mXUWWPGQa6WeAIvPIqnJboy8jrW9Ot1NlUOUltct93ioha7pAAK6GeCPpiqwgBlGzrXfGvoaRqFOKE4KmnLbsG6VouiowAHPemODxtqPaNnRGZTaPZ1p2RxtqWWP+9VXBXOOlUtUaXHbsMaq3U7eWeeKlDgMcisbV75YyVWr5RkMsyDLO35Vm3F2i/cUtmoJbgu3ymoZbg8AYzWE4jJFurhpMINoq9BevH/rZM/hWV577SajeTcMk1yyKOnivI5F+QqW+lRT3csJzs/KufilaNsoTV+3v85SY59M1mBu2+oLJGN2Q1WkmVhwc1ixMrkAHv2q0uUY46UAanXvmkIx71RjnIPJq0sysuAeaYEnyry1Up7gngdKWVn5BPFRLhlwetAEa7gcsPpT0O9sUeRKx9qljjKnpTAljRQwx+VWVBOR0qEJnHGKmUqhy/FAEwhZl4FRvCFXLnkdqsR3gK4QU/wArzlO6gCrFPnhRtPt3qwFL9Bg96elgcgrxVtLRWwAw3fWkBm+UfM5xUqRHcOBVtrBt+OPzoa3KdPvCgBmwE5Haql25XgCrkmPLH96sy4YyP9KAEittzF+3pVi2X/SVFFuGKdadACt0CaYE7KSxzTJORj0qwy5BqrJxjjNADJCSOgFROw2YHNSs/tiq0p5+XpQBGWyaYWG7g0uFPSo2bFMB7ON3BqJmGcd6QsMZHWoi2589KAJWY7Tg4pVkyF4+7+tVwTkk9KcG2rkGgDZs33bM8bvve1Y9wPKu5EA6Gr1nMQy+9QawhjuRIB97imIqMx+tKpzyBUaMd3NOAP8AC1MCQgE5z+FNYDPvSj3pxVuoxigCBx8uKp3Efy5xV/KgEdzUMq8Yf8KAMaZSDjHFUpSec8VqXKccVnzJng9aAORuM/apv+ujfzopbkYvJv8Aro386K+fqfG/U2Wx6Top/wBHuOOl5P8A+jWrZ0/B1BDmsPRsiC5wf+Xuf/0a1bemLi+jzzg161L4F6Gb3NO4A81mzxmqbrsyRzV652+YT056VRdgCQTWghfMJ42DOKUuAmOQ1Rk5bJJpC+TTQDw2RnvTuqVEg5J7GnAk/QVvEQ9TgYpy9800EHgdaXJXrzXRFCJ1ZQB60p7nNQ5DEEUrNuZR71ukIsxncRQcu59KZnj0oBx0PWlYCcKG2imycvgdqfBkKS1OC5bgVnezM5yshbeJmlBroYNqxgAc1mWigc4rUtojJIPSuKvK+5xTnYuwRHdkmpJGO7ap5qYRhI8d6FiUEEjmvMcru5xVKhCkbLkk81KjFfvDJqU49KTbUuV9zklJvYbuyegpw59KUKKeF71yVMRGJUKUpbjAKeo9KXZmnBKz+tRZsqDQu2jYaepxxUoTPSrVaLN1QbK+zNIUIqzs5qKU84odaKE6DSK7UwjFSOAByeKpyXqKdsfJrP60kc7oO5P5at1FQvbIDnd+tRfaSeG4+lNZmJ+U044xEOhoTkxKeRUUjpuGOKcjq/XrUbwbm4bntXVCvFmLjKJIi9wc1K0eccYqBFePAaraNkc1s5X1Q41GtyrMg8s561ReDKg+9a0wG3p1qvMoEfFaQm0dEahiTxBmNQJDh8gVqNGOTimQxAuxAzXYqlkbxqFN0wp461WeM/wj61qzQ/LnFU5UIPPFXCdzqjMpOqnHFU7lRGx2960Zkxz6Vm3WT171103dnTGZk3c3kqQvU1z90d5Jbkmtm+xk5NYVw+SccV6HLobp3KrL15x7VWcheR0qd2G05qncHcF2ZHPSuWpEoFcHIoBxkEcUza5b5UOakVXH3l68VwTKE37cKucVLyXAqNlKuMg/lUyIS2RWIyzbPJGRt6d63YLjdCC3IrnjcJAh5BNTR3jPHgHANMDYnkDD9yaprNIsnJwaqLOyt1NW4ZEmAOPmoAvwXPmDbKMe9WNozyNw9qy2VlbJPFWrW+ET7TwDTA0kc4AJVfQUvmbD8zA/hUDjJEkZyDTlbf1FAD/ODc4qRZVKncoNVxG24gdKlSBj7UASrOF+5GBUsd7IBnAqEWzk/eqZbUtkDpQBaW/MgAPH0oDmNwyk1XS2aPvUyIepOfagCfzWbnJpDcFfvdDTY2bH3Rj3pJsMo46elAE+VeE7euKzTHhjn1qfzQh4/Km7w55oAfGgUZpFbEgLetO3Yzmo1OScimBdf5cOORiq8jKRnFSRvujwKgkPX2oAryygHpUDPnoKWaTH1qo0wJPOKAJWbHOKgJAzzTTIcYNRNnOcUwJBJtamM2ZM1E27OKjdjuGKAJXlO0imo5GMVEX7UiuVfJ7UAacEgDA+lXtSQT2CzLyVrHhm5ORW3ZMJ7VovUd6YjBRtwBP41MoGO4qKZfs9w8Z7HilDEL1zQBYHTGc0pAPOajRgVyaQtTAVlwuehpj/ACr1BJFSZ+XnmoXAxQBRuVHRs1nSrycdK1peVyapzRjHFAHDXeBeT/8AXRv50Ut6u2/nH/TRv50V8/U+N+pstj0TRwDa3OOv2yf/ANGtW5pK7rtfbrWBpTEWdzgn/j8n/wDRrV0OikfaHI/u161L4F6Gb3L16cOfrWc5O7irl4xaQkVnsxPWtBD8jHWgMAOtQlvfFPBG3NOIEu4npxS76jD4HrTDIN3vXTARaBGOBUoGEyarpllzniiSbcuDniuqKJJGkw3FKrtuqsp4zk5qSNizZrpUdBFwNyM1IuGbAquqk1ZjTGCelZSsiWydW2/KakTk8VWB3SE84q7boG61hLRHJUmXbZCQMVqxXMFgqm5fZu6cZqrYwfONo4NLq8QF7YowypfBB78ivmM6xlTDYaVSlbmulrtq0v1OejGNWqoy21/BF0a7YluZv/HTSnXbAn/Xf+Ompxp1p/z7Rf8AfIp4020/59ov++BXlOObL7VP7pf5mPNhJPaX3r/Iq/25p/Xz/wDx007+3dO/57/+OmrY0yz/AOfaL/vgU8aZZf8APrD/AN8CuCtWzOO8ofdL/M6qdHCvZS+9f5FMa9po/wCW/wD46aePEGmf89//ABw1cXSrI/8ALpD/AN8CpBpNj3tIf++BXi1sVj1vKP3P/M9OlhqPRP8AAz/+Eg0z/n4/8cNO/wCEi0v/AJ7/APjhq9/ZNh/z6Q/98Cnf2TYf8+cP/fArnWLx380fuf8AmdH1al2ZQXxFpY/5eP8Axw1IviXSgf8Aj4/8cNXRpOnj/lzg/wC+BQdK0/H/AB5wf98CtFisevtR+5/5lLD0l0ZTbxLpRBxcf+OH/CqcviHTz92bJ/3TVm9s7GEDZZw9f7gqB7WxeAlbaIMP9gVlLG41/aj9z/zIlQpPoyjLrtuzYEvH+6arnUbPORJz67TU8lpbo4/cR8/7IpUsrc5zBH/3zUfW8b/NH7n/AJkfVqXZlcanaZG6T/x01KNXswf9Z/46aeLCA/8ALFMf7tPNpa9reP8A75FVHF42/wAUfuf+YnhqNtmQjVrINkS/+OmlOsWhP+tP/fJp7Wduq8QRkn/ZFR3NjCljK3koGCEghfauxYzMIRcuaGnk/wDM5JYbDt2af4F+3nSeISQtuU9CatKc9ay9FXOlxY9T/OtUDCV9HgcdKrQhUlu0meTiMMoVJRXRiEhutDIGXpUSMC5DVYUZHFezTqqRwuDiym8BPaiGABTxirbDFEI4IYda6Od2HCWtinPCQq1QniIbkfStqVRg56YwKzJUZiSTWtKR1wmZlxGQuay7lPlJPatudCVrJuEPO7mvSoyOyEzlNQyXYVjSwM30rpLm23SMW6Vk3UeFOD+FetBnXCRk+QQcHvSPbiLBOKnZ/XrUV0+YxnOKyqI3RBKxx8vFRLNjhzk0jTYPHI9KauDJnvXm1C0SiYF9uMn1oubxIo9qgZ7mo5M28LMcbj0rJ3MxJPzE1zjLanz2w3er6xlFwv8ACKoWsbBg3T2q20rAnNAEsTMx+aplfy2wDVSOTcfWrKRu7ZApgaEF4jrscVYNsXUGOqscCKN0hqf+0Y7cAAg0wNOzDRgB+R3q+0carv7GsaO/E+McCr8QIUHO5aALQkVcDHWka4w3GRj0qIKN2c5U9KXblSD1oAf9rYf3qFvHCmoMHpTT9c4oAt/aHK9actywHNUwx608PQBb+1MF45qxFcKy7WGKzwd3SpUXv3oAvMiMvbNMa36Fag+bqOtPV2UZzQBKYynJFRsPbFSxXO44fmnlkkbHQ0wK0bhMgnrTJGODipJbdgS1Vix6kc0AVbhiVz0NVN/8OPxqxcg88VSLAD0oAnDjGSOlMfnDDpSBuAM8d6FIwc8igCNj8xqNsN90U5uM4qLfjpTAjYfNzQpKsO4pGJNIrYoAsxkiStaxm2sM5rGD4YEVftX+YFjTEP1m3xKJlHBqin3etdBNGt5YMMnI6VznzRuQx5BoAmWQ9DTwfaoFkBGTT1f3pgSscrxUXQ4PpT92SMVE5yxoAhl71VkBKGrUnSo0heeVYxyW6UAcDej/AImFxn/no386KuavbeRrN3GRysrD9aK+fqfG/U2Wx2Gkt+6ugf8An7n/APRjV0+ir+5Zu9cpphxBde95P/6Nauw06Py9NDZwa9al8C9DN7kVySGY9qqMcqT0qzO2eSeKoyuenatBAzKOtNWQbeemaZJgHjpiljXcmDVR3Al3E9DxTlUFxSICDjFWYYsuOM11wRI/acALUbgJ9849annlW1U45bHFZTu80mXbqa7YRJJmmXd8vNT2u5mwOlVE4OR1HFX7UYQkcE9a2loiW7FyJexq26FYwM81Bbr69qnBaRsnpXHJ6mEpDY17VpWsfSq8EWW5Fa1nEGbFctadkcNWRqabDtXd2qprX/IS0/8A66f1Fatunlwj1rK1nnUdP/66f1FfGZ874ST84/8ApSJwcr116P8AJm0oqQCmgVIq11YirZWMaNO4oFPC0AVIq183iqx7lCiKgxTiaOlRu+DivBqTc2evCFlYfThUYb1p4YEVEYspjs0xnqKW4CnAqs1ztyc05NvQuNNsp3s2ZTms0XYjk55FNu7rM5GeaoPLwfWsORmnsjXlkSTBHSkSUIpz1NZcM2CGc8elaClWTPFHIw9kToc/eOAauW6RbcMc5rMMoG3nipIJt0m0Hijla1JlS0NF7dCw5qHUItmnT55/dn+VSJJk4PNJfqTplwf+mbfyrT2z9nJPszjlR95FXQR/xKYfx/nWuFAWs7w9HnRoT9f51psMcV6mXVrYeC8l+R52JpXm2V3gViMcGnhMU1wxbinr2zX0VCueTUpCMnFIq4NT9qaVr2YzUkckqVnchkXdxVeaDPI6VbIwaQjIrWMrEK6ZkyxAbuKxrsIproblCM46Vg3sfzkHvXo4eV2dUJmDeAHoOKwbtRzxXRXC87aw75dpOORXsU5HbTkYjgCX5uOKhmK+X60t05LEniqpkAUg10SV0dsXdFWWVA3IqNrsKCVUdOpqF3LO314qGQHyzn6V5lVGqHTXzuwUgYqFJ8E5XHNMIAkXPpSyDCg1y2GWYrkLJgHtVpZhKdo5NZMaM8vA68VqwKtmgMnLUAX7e0AXc3AqxLcpAoCc1kyajuICnA70NMH/AIqALpuzKcMcCom/MVGmNvNCnDHByPSgC1BKy4xWzZ3rJjJz6isOM96sIx6igDsLaRJY8Y4pdhST1HasjT7shdpNbkEySoMjkUAQuuQSKrup7d60fLVlOKqTRMpGOlAFbkcGpEXPekPHvSjjvigBwG04zU8bDHBzUAyT60+MDr3oAt56U1jSbulKTnrQA3HIIJpdx3cE8UDocfhQgO3mgCx9p+UK1NfynXg4NQygKuaozT/N8pxTAmubcOpKtWW0WDhjVg3T/wAR4ppeOXIyM0AQlQBhSKcEYLyKgmgdPmBpsV8UbbN0oAlfGOOlV2H92rexZ1zEfwqs0ZDEdDTAgPFJ0pzZ6Gkx+NACg84qzbvyATVLPzZqe3bnJoA6PT5VKgE8Vm6taeRP5gHytTrWRggwO9aVzH9usGU/eA4piOcABHFKp29ahw0blG6qakyMcigCbfnBX8aRjk+5qLeduB3oDEHANMBsmc4HJrZsbVLOH7RP9/GfpUFhYZJmn6DnmotUv/Oysf3cYoA891mcy65eOTy0zH9aKg1AbtSuSeplb+dFfP1PjfqbLY7HShvadcZ/02f/ANGNXbACOzA6cdK5Dw1EZbq5PZbyf/0Y1dXM+cknjGBXrUvgXoZvcz5XzkGomI/SlkOZDUQz3NaCHBQVpwGCABUQbnip4juOTVw3Aq3N+9tMVRQePSmLrMyHKhR+FQ6kP9L/ACrfNroNnZ2/26GQyyJuJU//AF67IvlJMVtVlkbLorfhVqB0udNklZVV1lA4HtSanpEcNst9p7tJaSHv/AfSksUzo8uP+e6/yNdkGmk0SyWEbuPTrxV2MdADzVaGM7QT1PJq9bIC+SOB1q5yOeciwjmKIbvvE81dh2leKoMN8ue3ar1vwwz0xXHNnJKRft0zWvYwguKz7dMkEVuWSY615ledkcVSRbxgAVj6yP8AiZad/wBdP6itrq1Y+tD/AImem/8AXT+or5PPHbAv1j/6UjbL43xF/J/kzdA4pyjmjFPArmxdY7aFIcBUnQU1R3pSa+Zr1XJ2PcpQshGJAqrLIA3Jp0z+tUJJcyYzU0qdzupwuWlmLHANLNdCNcA896omUR555qnLck9TW/s7HRGjzMuyXIPJNUri6A6GqUt4F4qpJPv71HszpVGwtwd8ysePemSTRJCdv3u9MkbfHg9e1Z8isMgnij2Y/ZloS5HzNgGrFpdllCg9ayWk3YGelSR3Kwj5RnHej2Y/ZG/KzbAF7daktXUdHG6sqK5Myb92T3FSQuRJxSlT0JlS0OhjlCsATk+tWrqYNpFyP+mTfyrDSYjGTirkjltMn7Dyz/KuCvTtB+hxzpl/w4M6FB9T/M1oy+3Wsrw5KF0eIH3/AJ1pSP3FGEquNOK8kefWp3Grk8mnAVFG5ZiDVhRjGa97D19Ty6lLUULxTioIp3U0u3nmvo8PWucdSkV3SojxVphVdl5r0oyuedVhZla4XMZxXP6mjIuTXSOuRzWLq0ZMee1dtCVmYxlZnK3jbULZ5rn7qfOQe9b19H1Fc9ex/PXtU5HdTkZN3yDgVlzbiDWpcg4Pes5nVVO6u+Luj0abKnl7TvHIqpdMpbCnB61YuJd0eI6gSLzCA3p1rhrI6EVip6mp4rczgVOsQk+RBwOtWHdLSHp82OK47FEQRLOPLYLdqzLi6kmk56VO0huGLOaqSnY3FKwEkRyD61bgB289qoQthwc8mrQcgmkBZ88jiprchxkHms4bmepYy0b9eaQGsARip1k6YrOhuG3fOeKuxDMm5fmWkBoxOVwVrcsZ9+AODiuejf0rRtZPLcOOvpQBuh2Sp96yD5uKgQ70DL1PagfLnPJoAWW1BBKGq5iYDJFWCxABzinLKp+U0AVVJDcU9Mg5NTNAGPy0ht2CigBUz3oJ3HFOCEYJ6CkbCnJoAaMinjnjNNPJFMPBOelMBlzIV4BzVCTLcip3cljkVEW65oArO2VO6qrkqQVNWpcHOKpyjnAoAspeDhJOfemz2yuNyHJNUZH7ClhuHjbk5FMB8UstnJxk+1aStFdxZQ4k71VEkVwMNgNUTJJZv50XTvQBLJH5bFWqLjHFWty3kXmJ97uKqsMHjjHUUwGHjvUsRzjFQP8ASpI+2KANKGYAY6Vq2UwVgp6GsIPgDirttcDd06UwHazYhZPOjXg9aysdga6tQL20KN1xXPSWskd0YlXnPBoEQbGKhUGTWjZ6cBiS44UVZt7WK2j8yUjf6VVvL4yAovAoAZqeo7l8mD5UHGRWOeMAnOTUsikZzzUcQ3XKKw74oA47U1xqlyP+mrfzoqXVT/xN7rcOfNb+dFeBU+N+pstju/B8UkWn3EtxtaU3cwdlGAT5jZxWjd3AzgcCquiNs0mYA9by4/8ARrVJKcjjk55r1qX8Nehm9yPbnk0zjrnipJMlvbFNP3QAa0EJszjFWIlHY1Go9ecVYhAZ+FwO9aQ3Ax9TUregHrxV/wAQD/R7DvmEcD6mqerEfbicgdOtbN1Yx6nDZGO+t4gkQViXGQcntXbGyabJK2k4l8Paklw3yKVZR780aYudIlJ/57D+VSanPY2WkjTNPcSln3TS7cZIpNNH/Eom/wCuy/8AoNdEb2b7szk9B8QY/QVoLiOIAdWFVYYwWXPAByamVg0pf3wBUzZxTkWYU3Y9KuwJuOKrwnsBV+BDwQa45yOOcjQsxnit62AEY45rJsl6HPPpW1EPlry68rs45O8h61ka3/yE9N/66f1FbK1j65/yFNM/66/1FfI59U/2Rrzj/wClI9bLofvU/X8jfApwpVXNOAwK8fF1j16FIKhkfBxT5H2Cqkr9ya8unFydz1IRIp5e1Zs0u05NTXEmMmsW/u+oU16tKnoenQpXdiSW7BJyarvdD159Ko/aM59ahefa1ach6So2LM046mozKCoIqhJMWzg1D9r2jBo9mV7I03mQRg96gmZZFPlnJxWTc3TbvlPFPtJm8wFjxT9mP2RJGHMh3Z4p7Hah5qaZWkT9zjJqpJbSAb3BFHsw9kXrO4UJgcHvV6OXByDWEsmMDr/StCCVlXhqUqegpUtDXjmwoLc1be4L2MoB42GsT7QeABVtZi1u6/7Jrz8TT/dy9GcNWnob+hORpceO2f51qpJuODWHojEWKfj/ADrTG8tkGvNpQ/cwfkjzZQujRjVQeKm25+lV7eTd161cXFdFGo4uzOGpSQirzTv8acFGaP4uK+iw1Y4qlIiIqF1qy3UVC3U179CqmeTXplVxWXfQFk4rXcHvVaVNwr0qcrO55E1ZnE31qzE7RXP3tlJnniu7vLQliQKyLm0wp3CvXozTOmlI8/ubScZwMisO6tZ9xVlIrvbuzCsWU1hXyMoJI/SvVpM9OkzmV0+RUywpHtJBgoM/StKeVzFgGn2ke073+Xjqazro7EVB/olpvK8nrxWNcztJJk9K2L+bdKQeV7Vn4BzlOK4LFFJ2+Q444qkzE/erVaNPmOMZGKomBuSGqbANiT+JuMVMrjGd2aidWji3SfpUEcmW46Uhl/zdw4605CS2c81ViJyasKxHIFSBZTO7Jq7DK8fTj2rOjdmb0qyrsTmpA2baQSYXGGq8gYcZ5BrChkfcMHBFbNtL5i/MeaAOgspN6BDwcdasH5eCck1lws0Wwg9q00cSxhj1FAB94cnpUbDDcVKcY4phGetAArsp9KsRT/3uarrkdKeCxIoAuB0fg8U2WHeMioQWBJp/nHGKAI2RgOBUM5KqP1q4swCcionMb9aAM4nmopRWlJBGx+U1RmtZAPl5FAFKT7p5qm45NXJIXU5YGoXVQOlMClt5pjE5wOtWHAA461XJI5NACZI56GrEN5n93KflqmxJzULZB4pgaqObeTcnKE1PcwiRBLGfvdayre72N5cvIPetO1YMCgbIPSgCu2R940qZ6g1LNBjnGBUaRkn5fmpgWYzkU9MrJxToLSVh/dFaENrDFy5yf7tAFrTfML4q7fQLFbmcKC4rPF8kTARL+HpV63n+0ZWUZUimI5y4uGmfJNV2GOS2au6pZNbXBKr8jVRJxwe1ADHP4060QPfJx701uhx+FXNJVWvenAGPxoA4PxEAniO/VPuidsfnRUevHPiC+J6+c386K8Cp8b9TZbHf6K/+gzA9ruf/ANGtU7tlmHaqekNixuAf+fuf/wBGtVwbQvTNevS+Behm9xF5Byc0gAz0pcfKSOKM/KK0EOVPmz1FTD93GW6CmKcLgd6r3UjCPbmtKe4GbdMZLhieeaiHfnrVlAmfmXJqUBD/AMswK9OkiGVUGe9dBpoP9izlRnEy/jxVa3gi/iRc1oW/7qMgfKmckVrU2OebHN8kAA6tz9KkgT5gW9KbH85yematLHluK45s4akixbY3c9K0oR8wxWfEnPHWtW1UHFcNSRw1JF+0XD81twjK1m20YxmtOAfJxXl1pmVJc0x4HPFY+uj/AImml/8AXX+orbUVi68P+JtpX/XX+or4nPKl8PJea/NH0mAp2mmdGFxSMacTUUrYWvHrS55HsQjZWK87Z78Cs+5lypwcDtU1xLjOKzbibJ56AV00aZ6NGBXu5yinnrXO3UpZjV6+uCc5PWseRtmcmvUhA9yhBQjdiCQpnPWmMX+8Khll5zUL3GeFNa8pv7RC3DnbkDB+tUhKxkwRT3k9TmoZHXuMU+UPaIkOG470pYonHWoA2CDmrBQyoNvOetHKL2iJLe8ZSPmrWjm8yI8bsdaw1h8qQZrUtX+QheAetPlF7RCGwE0pdGx7VYjh8khSv45p6MqN8p/GlMgLZLZNJx0GqibsWol39BirmxVt5OOdprOSUA8NVxZQ1u4zn5TXBiqf7qXozCvH3WzX0aQLYID15/nWmD7/AIVi6Xj7GnPr/OtBJAD1ya82hTvhoei/I8nlvFGnCxVhnj3rRjOVGDmsmCXJGea1IWBAxxWUqdnc4qqsTinAUwU4GuyjOxytXFZaryD5qs9RUbLg17uGq3PPxFO6KrrmoGHNW2FV5RX0FOV0fP14W1KVzFu7ViX0GSRjiuhcZWsi8TLGvQoSaZhTepzF3bjmsTUIAYyCB+VdLfIAcdKyLpFII617tGR6dJnIPZb5iox61DORGRGoGAOea3LrbDG7bQDjFcxeyOjZUA1vWV0ejEgu2LjIAGKpI7FiKlkvWJwyqDVb7VGJMOu3Jrz2ixCzFipFQTBEIJfBNTzzRB/3ZySKqMnmScjOKlgPlOYcdqrRxL5o7E1NIwwB0AqFmCyArzUMCx5Pl52jNOSTHysKIpD371LIgZMDrUMYgOehxU0eAOtQBSoAqWMAdahgXocFcHrVy3b5woNZ6PkZFaVqvAYDJoA3bSbcArc46VowPjg8iuejfa2S2DWvZXIZQG60AabD5Nwph6U8MNtIVxyKAGZOakT3pu0ZpQdrdM0APBzmg54waB83UYpQpJAoAjZvl5qs78HFTz4Raq9s0AAmZevWlW5dTyaiwCfmNI/OaYEz3UcnDKCajkjt25PFVHBNQs+BjOKAJpbe3b7rioJNPUjKMxNRPK4GNxqBpnHQLTAWSxcNwGNRGxk7IRSi5cD5sCmSXrH+EUANbTpOc5qa0jeBxucgCqT3L5yCKgkuGPXFMDsreOG8ixvGaGhgtV5GSOlcpY6hLbzBsnb3FdZFJFqNorr96gCM32QAhwPpURmctycCmzRbGA24phGcZPNMCwjEtxzV61n2cHsazlPHTmrEJ+bO760AbVzAt9ZHaPmxxXJzRNE7I3VTXT2c/ln5jx2qlrtkNn2mHk9wKYGH1UccVoaOpAZ1H3Tk1mBj1zxitjSmC28hJwCOtAHmuutu1+9IHWZv50U3VudYuyDn98386K8Cp8b9TVbHc6Wc2dwP+nuf/wBGtV1d2Mdqz9KI+z3AP/P3P/6NatAMeB2r16X8Nehm9yQH5cGkWPzG9hTcndjH41Oi/LkGtBCH5ASegrLnn8yU+lW7+48uPaO9Zq8j61pT3EyZcYyanUcA1WQBmxjNW1IGMcYr1qJDLcJJYYq3JJwqr+NVIn8tfMJ4PSnRybjnuaKhyVDRt14rRgj6Vn27DjNacJ6VwVGzgqFqKJd2RV6CMqflqrA2WrRiOa86o2cFRsv24baK04Pu1mwHtWlAcDmvGxMnc2wiXMTgVieIBjVtK/66/wBRW2DzWJ4g51bSv+uv9RXw+cyfsX6r80fV4NK50BNU5296uEZGKqTxEKTXLTV5ano07XMy6k5rKuZdqnnrWncoTWRexfLXr0Uj16CWhkXL5frWfKxJJPIq9NGSTWddHapVetehFI9Go2o6FC6uQgKgVSWfa2SalmgJySapTRYHGTWlkc3NIllcy/MoyKYdxGAKhilkTC44q4VVFDMeTRZC5pDkjzirgbYnC5NVovvZHAqySgTJOT6U7InmkNZwzAZwTVmFiowKp/KW3txjpTmmbrEMLRZC5pGj90Ak9aFdC3BqgkxK/MacpKpkDJpNIqMpXNSExnjPNWU+RW9xWVaZMnzc1qAkqQfSuPEpeyn6P8jtu3TdzTsJgtuq7sHJ4q8khHOaybYfuwauxHHSvOwyX1Wn/hX5HFGC9mn5GtBISRmtW2lwRzWHb89a1bVSMYrOaRw14o11ORTqamdozTqxWh5Y4GkdepoFDHivSw8mY1EmtSJ+lV5V4qdzxUMp4619PhpNo+fxKRWNZd+BzzWm561lXhyTXrUfiPJXxGDd7dxOcmsySIOrNWpdkE4HWs29kENuQOCa9yk3Y9OizldVcsSiHgGuZvmcN1rob91DkjnNc3qDAt8o7121PhPThsUmK4O9uagnRWVTmkc/vDmoJZcuBngVws0GykALg81E1yY269ajkuCXx2pvEmahjJftWVOaRXB61XkAC+9N+YLx0rNga0eNowc1OgJ71mW8hx1rRikUR8nmoYywnuM0/wAvPQVCj54BqQM2cVmwJ4o5OFxmtSEoiAHqKq2XHOeas7Rvznr1oAsIQ54q5BlcEHmqEeFk4qxG+1uaAOgtLgSYVqvKo3e1YdpIA24HNbkDCeMkHGKAGNxIfSk5JqQjI5pjcCgAGTTgSGxTA42+9LuyeKAIpeVPeqvJOankODxUTDBoAYTzUcg7gVI6jtUTA9jk0AQtktjPFV5OuKsOcKexqu59aYFd+KrtzViUc8VXYUAREcc1E/SpH6VC3SmBA7YqHqeakc5NQbvnwelAE5OenHFaOj6g1rNsZvlNZg+7Sr8pDDrTA72RFuIA6cnFZzxtG/AqDRdS+UIzVtXEInTfHzxTAzBz938akRtrYNIRtyO9HQ5oAtxy4bk8VpQP9pjaJuQRWKjAjmrcEzQuGU44oAytTsWsbraR+7PQ1eszs01jjqcVqXMK6pYlf4wODWcsZg02aOQ/Mp4NAHmF+d2pXJ/6at/Oil1IbdUuQOglb+dFeBU+N+pqtjtNK/1Nx/19z/8Ao1q0euMVm6Uf3FyP+nuf/wBGtWgoPavYpfw16Gb3JQC3Aq5gQwZbg4qK3TLgnoOtVtXu9sZReprQRl3M32iZj2BpFOelQFst8v41KgPAHWtKe4FyBcc96m2l24/GoozjANPkk8pTtPLCvUpMhjmlB+QHgVPEx4xWYjEfKTnvmrcDnNa1Ec1RG1A/yjNaUD7cZrIt2JWtKFs4GK4KiOCojWgbnIrQhfkVlQvitC3JJrgqROCpE2YDwCetX4W5rMtueavxHnNeRiYXIoy5Zl1axdfP/E20r/rr/UVsDpmsXXTnVtL/AOuv9RXxWeU7YdvzX5o+pwM7ysdIaZKu6M1J1FJXmyi4yPUi+plTw4HIrJvofl4roriIFaybyLbGeK7aMz0aFTVHJ3o8sH1rFmBOSa6HUEGTmsC7O3O2vRhO59BBKcCiyEZJqpIhLEjkVaJaQYNMYbRxziteYPZFKWPbyBg00AnGeasN853U0MBxxT5heyIjOUXrTGuznIpZYPN+YHn0qo8bRtzRzC9kXIpHeQMzZHpWhu2xgBeDWZBnzB6VfSYY2E9KfML2Q/Z8vA5q5EMqBUEakj27VZt0Jbik5aBGlqW7eMdhV4Rfu2J9KbZwk4yK0ZIsWrnHRTXHiZfuZej/ACCtNRi0hlkmYAT71dijy1Glwb7BD9f51ow2vPSvMw8v9mp+i/I81VUoISCLkVsWsXAJqCK3wwrSiTatZyldnn1qlx46UooFOArWnHmZxMUCmPTzUT16+Gp6nLXnZDHNVpDU7HiqzmvpKEbI+exE7kEn3TWTeP8AKeK1pThayLttxIPFelR3OGC1MWdfn3VhapNnIrpLsqkfHWuYv8Mxr28PqenSRzF8cAtXP3Zz9c10uoKrLgViXEC7844rsq/CelHYwpTtznrVBnDMc1uz26EMTWJNCok+Rq4Gyymz7ZCG6VJGM5PapWtXkI24pZojbx4I5qGMqscSetOWUlsYqAud3pT0BVs5rNgWlkRD83FKXy2FNU5CGapI2bHy9ahjLkcrq3y1rWbhly/Wsi3XHzSGrKzEP8nSoA245Ng+UcVbibd9ax4Lnt1rVtSH57igC8gXPPWlBw3tUSli1Sod2MjrQBegcKOK1bSfbjB4PWsZRsxjrWhbtsxQBs9ce9IwBHSmRSb4cDrUg+5jNAEX0HFHQ9KQ8Hk8UjMM8UARPy2ajbkGlyARmkPUkHigCNsVG4z0/GpyBvxUTdeKAK0vDYqu49atyLzk9arSc0wKrkd6gJGTxU8ij1qu/HQ0AQyMBVWVjip5T81VZGoAgZsnioz97mnsRnios4fmmBMvSpEWoRU0ZINAE1tK0U2VrrtKvRMApPGK4sNhjzWlp935MgXtTA667tlK7o/0rNYENzWhZ3Syptam3lrtO9OlAFQHAHrUkZbdzUR+7zTVYg5zTA07S48t+vHeruqWyyaXI8Y54zWIGIbHaukgcS6f5X94UAeK6suzWLtT1Erfzoq14hhC+I78DtO386K8Cp8b9TVbHS6TzHc+n2yf/wBGNWpGuTjP0rK0okRXXp9rn/8ARjVtWqZUOenavYpfAvQze5YciGHJ4rnbqbzpya09VuyqbB3rG7e9aCBeGAqcMFHvUUfL5PFPI3MauG4FuE71yeAKrzTbnJHbikefy0CD8ars+c4r0qTIJlerlu2CKzAx9a0bcjaoPWuqWqMZo2LdiSK07diAKybd+AO9acD889K5Jo4po1Im9avQvjpWXC43Cr8T7mGOlcM4nFOJt2bg45rSQ1h2zgMMVrwtlRXm1onFLSVzQU/KKxdcP/E10z/rr/UVrxHisfXP+Qppn/XX+or47Pqf+yt+cf8A0pH0GXzvNfP8jpkORSkUxDUuK8zEUrM9mjK8RjKCpzWVdqCGz+FarVRuI93GK4Yz5XY7qTszkr6DcxzWBd2xcnC1291agsSRWLcQoGIArvpzPeoVtDlzCQhAXmoGgIHzcV0pgjXJbiqVxGjttxW/tDs9qc3NtGVFVXhb7w5rbuLFC/HFV/sTqDg5FP2g/aGTEuW5OKkMHmNzVh7Uh84qzDb8g4p+0H7Qoi32OKbBA73G3kVvfZkcBsYp8dqvmEhcU/aD9oNhtNkQB5qxbwhatQwnYMipUgAOCOtJ1NBOorE9vD3FXXT/AEGXP9w/yqKGPCgAVakX/QZv+uZ/lXBian7uXozzq0rpljQ4w2lxfj/OteOM56VR8OoDpEXHr/OtxIwB0rgoT/2eC8l+R48p2VhiR9zU4oxTlFVGLlI527igU/oKQClPFevRpGMpWQ0nIqJ6fUTV7mGpHl4ipoROcVXbk1NIagY4r2Iqx4VZ3ZDP92sm8YFuK0Lh+DWTO3ykmu6jEKSMjUJeCo61ztw+M55NbF843Ek1g3s6BSp7171CNkenSRk3QLtwO9UbpPkGBV6UrtJ3VlTTqQwMuMcirrvQ7kZd4pGV6ZrIKbJDu5rWkmViS7gis64CM+Q4xXA2WS2xQAkdagvZVb71EcZz8pB+hqrdlg20ipYEaKjyYYdama0Rozg4xUUUTFlq1LkIAv41DGZklvIpz2qS3UYx3q0i5/8Ar1KY4x0wDUMCEkKu0VYt1xH83BqMQbfn60v2kFtvSpAuQEA81djlZGBTp3rMWQr0qzDMT1oA3IroMMHrV1BlQR09awrdzuJPStK2uiGCnlaANaHjBbmrSk5yBVVeUBB4qzE3y4NAGnZvhhVvPzVnwnYRWgwyoIoAjlGCcVDLkAYqw/KA9xUM/KjFAEB5xScDOaUH1pO4zQA1iByOtRk5GRxUjccjpUbgnpQBHIcmq0q1ZxhTnrUEhycYoAoucE5qvLxzVuVRu5FUrg88UwKsvrVRzlqssc5qs45zQBERhqikIEmDUhb5s1Vnb9+KALIb0qVTVeM4qwhAOTTAjdiHqaN2BzUU42yA9qVDQBv6XqBXarHmuot5xPHtbnNefxuUYMtdNpV7vUZPNMDQvLfy2yPu1U6Y+tbCkTxYfBrMuIjHJ7dqAFB/eCteCQ/ZAw6q1Yozn0atm35szjqSKAPNfEoDeJ9QI6GdqKj19s+Ib4/9Nm/nRXg1PjfqarY6HSVJW497yf8A9GtW7hY4cHjAzWNom4LckDI+1z/+jWqe+usJtzya9el8C9DN7lS5l82cnPAqHG48U1Oc5qWNPl5rQQoU4xQTsy/YUd9oqvPNvbaOgq4bgNaRpGz3pocqcUBgAcVGW4x3r0KbIZOvJFXrcjH0rNjNXoMda7lqiJI2bdulaUDY61kW7cCtOB8D3rCaOWaNKJs+lXoG96y4mP8AFV2A+lck4nHOJs2zY6mte3bIFYFs+GrXt5ORXnVonDVia8R5rM13/kKaX/10/qKvxOG6Vna2c6lpn/XT+or5PPo/7E/WP/pSO7LJfvbev5HSLxUobtUCkDmnqcnmuTFUj18PVHsKikHHSpc5HFRk5zXzNePJK56sJXMm8BJ46ViXKAsT0rorhOGyKwbvDyYHQU6dQ9WhOyMO8ky+FqtJyoJOK0Lm23txVeS3wB3rf2h1e0KJjJbjmpIo92Qak+zOX3c4qdY8LnvT9oP2hVntUVeF5pLe33Nhhjjir5iDY5zTCNsowuMU/aD9oMFnlQAM80kcLCQL1x1q/bqzAkdKaMLc/IM1XtCvaEqRZB4xTvJTzBubFSBSSakjg3NyM0nU0E6mhJCig+tTTgCwuOP+WZ/lT4odjdM1Ldxk6fccYHlt/KuHEVPcl6HFUncl8MrnSIfx/nW2OKx/DQ/4kcJ9z/Otj61y4Zt0oryR503qOAp4pgPpShs17NCnc55TSH5pC1ITxTCeK92hSOOpUFJqFzjrT2NQua9inCyPKrTInPNRt90041FM21K6oo8x6sz7uUcgDHFZV0zCA81euTuesrUHIj2mvTox1SOmnE56/YnPNYd393JGa2LrknPSsi6A2YFe7SR6NJGLczEBhisO4DZL9q6CVPm+as3UotsWQOM1jXZ1owSTkgjg9Kp3EWAdx6CrjjdIFHFMv4dlvyeTXDcorWYLRMyFhiq8146SYxu+tXAwSz2LwTWcY/3w3VLAtQ3qOoJXBqwjJIpKtk1RkaKNsCqzysW/dnFSM1iVHAPzU6OLPLGsyGdw3zHNaC3AOAvUipAmDZytIIU3EkVLFGEUNnO7rSEbztFSAwkoMCpoTtwaRI/m2mrSwDGfSgCZHGBgdatx/KRgVUi+WrMD55NAGpaXWGCv92tRQpAI6dqwI2w+a1bS424DdDQBsW5Jx6VpjBQYrJiG5gVPy1rQndGM0AG3KnNV5fQVZAwxPaq8+S3PFAFfBIzSZHennrUbL81ACYGzmmFcd6lC8U0qMc9aAIjj0qtJw3pVlgQOtVpATQBTnas6Unec1o3HyKccms6Q5HPWgCsx61WlzjirTrleelVnoArgHPNVLggTLV4jg1QuRmUUwJo2x1qzGeapDtVmJqAJZyHXPpUUcnFTEDafeqKkpMVNAGjHgjNWrWdopQQeM1QifOCeKsqR60wOw0+43RghuK0HhWeI45b0rktMu/JlCOflNdRbTgrlenSgCjIGjYqRyP4q17PBtQG6lh0qC5hjlhyo5qWy+WBVPXOaAPNNfGPEN8P+mzfzopNfOPEF9/12b+dFeFU+Nmq2Om0uTy7O6x/z9z/+jWqnJKZpie1RR3XlWtzGvVryf/0Y1SwJiLplj1r16XwL0M3uORc9OMU5nIpzKEjAJwTUYAPzN/DWghs8nlRE5+Y1TBJGfWid/NkJP3e1AGMYqo7gSADHSkx81KPc0mB2Nd1Nksco+bpVqA/MBVcYx1qaPK4Nd0GSzXg7Y4q/E3cc4rMt3zitCI4+7TkjCcTSibK5FW4Xz7VnQsynJHFXUbcPSuaUTllE1LVsDk1q2r46Vh27g4Ga1rUjjBrgrROKrE3bbleaoawc6lpv/XT+oq5avkfSqOsH/iZacf8App/UV8fny/2OXrH/ANKRWX+7iF6P8mdITyPSnhsE5qFTmn54rXEUrq500ag7d2FOUZqMc0/OOBXy+KoHrUapXvUZlwtYFxalHOeprpiARzWfdW/mN8teG705Hr0qhzptz3pUtA6k9xWs1sR2oS0YqSBgU/aHR7Qxhbtk/LxUM0O37oxW48B/hqtJaHPzU/aD9oY8aM0nAqx9nCg7lOfWrkcIU5xSTxO3I6U/aD9oZ0bYLKMirMEO+TgY+tP+yEYKrkmr9nZsZDkVXtB+0GRQ8+pq1Fb4arMdltkyetW1h2nheaHU0FKqQRW/bvRqEOzS7nP/ADyb+VXkQL1qtqh/4lV1/wBcm/lXLUvKD9DllU1KvhxsaDB/wL+ZrUL1k+HR/wASOD/gX8zWkTg114Gk3Rg/JHDWqWbRIHINPVs1Dk5p6DBr6PD0WcMqrJTTWPFGaYxr3KVNJHJUqCM3FQscmnOcVGTXZFHn1JXENVpWzVg1UnOOa2huYxV2Z8p3Oe2Kyb4hiQe1asoHJzWLe87gK9OgtTupxMS9AwQKw7gGtm7HUZ5rKkOcjGa9ynpE76aMyQbjzVK7TeCo5rTkjGev4VXkjVckjtXHWZ0I5K4hK3g+tQ3sTFwC2Rmtq9txvB71mXFrJu3MeM1xXGUZoyGXnAqhO+2Y4I4q/cxSF8jJAFZskDFiWBpDIGbfnI5p8QHQjmlCYG3+KpUt3UAnrUgRHC59akWXIXaPm9ac8alcn71JCuJcgcCkBa85o1UE/WrkUsb42nDVlytu+7zU0UZXawpAbCrz71YA4GaqQz7FBer0bLMMrQA4oNgx3o2FenSpI0IOT0px+b73AoAkhBAGelX4iCPWqQUhRtORViDJxnigDbsLjjYRW5bcrzXM2xIYFa6Gwl3R4YfNQBbP3eBVafJ61aUEHpxUEwwSWoAqY7mkK55qQjPApMYGKAI1AOaQp78U8cUhGRQBBIAF9aqy+1W2XPWq0xx05oAzpzg1SmUNyOKvTDPWqco64oAqOCOCc5qq4+bHWrUg4BqDqeKAIZBgYqjcjGPWr8x/Oql0PlFAEaj5c05TjNEX3eaUjFAEqOeM1VumxMGHFTA4qC5G5QRTAmgfcelaEWDWPBJg4zWlC2e9AF1ODkHpW9pd9nEbVz6c1NHI0MgKmgDt4nwMcEVbt7fO1lHHIrG067WeEBjzW5prYZ42bjPFAHk3iL5fEl+PSdv50VJ4pj2eK9SUdrhv50V4dT42arYmtQZr65QdFu5//RjVtxqOoHIrN0yPE143c3k/P/bRq1CdqgfnXr0vgXoZvcjlIJyap3j+XGUX7x61eJUBnbt+tZU7mSUse9aCIx2qRelMIAxgYp4PFVHcBwGMmlUd6TqKeD8tdcGSIn3/AMatqNxHFVYvv1bTC12QYi5BwRxV+Jx2rOhJHTmrcb45rp3RDVzSiORyeatxNis6NsrmrUT4HzVlKJzyiakD81qWrdKxIGy3FalsTxXDWiclSJ0Vs4GKr6uc6hp//XT+opIH2gZ61HqT7r7T/wDf/qK+M4gjbBt+cf8A0pE4SNq69H+TOkRsCnZ9KgD0/djmu6cLo4o1LExcKOaaZAOc1TlnPUGoDK8nQ14mJoNnfSxBoNcjtTfNCAlvvVSEu1cEc+tLvLYLc187iMK2z1qOIJGuV+YmiO5SQbc1UnxgkcVVaTYBsPNcf1RnV9ZRqySrjYB+NQlVGAOc1UilMnB61aBC4PU1P1Vj+sIesMZfmnPbo3CUwyZfcOKWKYq4GM571P1WRX1hEgtD5fHFW7dBGmMfjTEIYZzU69MfpS+rSK+sIevPQVIelMQYpzMBVxw03uN1la4zJ6Y4qrqZxpV1/wBcm/lVncPWqmqEHS7n/rk38q6XhGqUn5M5XXvJFbw62NDg/wCBfzNaW7JrJ0BsaJCPr/OtLNehluHbw1N+S/I4cTWtUkvMlBqRTUAanbsV9JRocupwusTFqiduaTfTWya7ErGU6l0NbJpuOKdTWOK0Oa12RtIUHTNU5pc54p8sgDHLd6pzShvu9a6IQN6cSCZ8KfWsa55JLcVrS/d9TWRdEkkEV6VFanfCJkXShs7DzWVLgHHcd61LlDyRWVMzBmXGa9aLtE64qyK0pVsjvVZ4mPQ55p0gJ6dfSnIm2Pc52n61w1WaIr3FvGIssMmsyYKY8HBX6VpXU8e3AP61nTELHlSMZ5rkKMK/Ty3ynC1mBmklPpWvqa/uyU/nWGJTHyaQFhwgI3Lg+tMmdAnDVAzCT5ix601wrEDNIBjCSRxtqdzsTZJ8pqSGPkY4p23Mh3DI96QFULtwatwZfFRm2LtlcirMMZjX5qAJnOV20+CVkIx0FNyuOvNCISeKANeC6ScBDwasmJgMN0rHj3L93g+taVpcBvkmP40AW4sp1qzsBXNRbdoyDuHbFSRk7uaALloO5bpWzZuFkBzWXAOBhau25wQD1oA3sgc9c1DKoK5p8P7yHjqKQrkYagCo3BzTCB1NTS7EOKi6MCKAGAdwfwpM4HWnBQ2SeDTScDmgCNxgdaquM5xVthn3qCQHbwMUAZsw5qlKOtaMq+tU51B6UAZ7gmosEKatOuAe9QMCFNAFZhkHNVbhMx1dI+U1Xl5jPFAFVOBShs0wHa2KU460AOHFMYZU0o56UD72KAKa/LJir0EhqldIY5g2eDUkcnpTA1onx3qyjbqzLeTJwa0YhyCOlAGpYXZgmA7V2GlSefIpFcKpG7Oa6Tw1eYvFjz+dAHEeJz/xVOo/9d2/nRR4n58U6j/13b+dFeHU+Nmq2NfTQBHdluv2yfH/AH8arhPmYA61TseY7nH/AD+T/wDoxqtuPIj3E/er16XwL0M3uQahKFIjHpzWfjJp8rbpCSc00D5a0EJjmnheKaBUoHHApxAI0LHGKJBt49KmiXC56UkgB4BrpixEcYyKtpwvzCqqenpVtMEDPNdUWSWIj6VaVQcVVTHGOKspmumMhFlGxgVbRugqmnNWIiQD3rR2aJcbmhC4DVp2zdOaxYDuI5rUgYDFclWJzTibcLk4pL1j9sss9n/qKrwPnFLeyf6RbH0b/CvjuIo/7C/WH/pSFh4WrfJ/kzofP+bHSnJdBjtzWM95sXOeTUaXGCOa9V0dDxpwNi54YAdGpqDy169aSOdZIx3OKrvKVYg8ZrjqUOYwTlBlhpQRz1FKkykYzVFpFVevNQeed/FeVWwh108RY02kQtg80GKMj7tUBLk5zVqKTKcda5vqZv8AWQCBZMjiplfLVXlbcPQ0yOUhwvWp+plfWS6WB+tQSNKOV6VZjUbQTUwAY4I4NT9TH9ZIrdnwN3StFHG33qm2FG1amiJ2/NU/U/IpYotrJSO9RK1BOauODKeKdhfMy1V9TbOmXH/XNv5VLkCqeoyD+z7gf9Mz/KtauFtQn6P8jCOI/eRXmiLQT/xJ4fx/nWju5rK0R9ukQ/j/ADq8Jhu5PFb5VRX1Ki/7sfyMcXVf1ia83+ZZBpc81ErDGaeGFenaxjFtkoFL9aaDgU2SQLxU7m6QrDAqvIemTSSz5HBqpJIcda2hBmkYDJpArnvVSWUk8CieXiqpl5Nd0IaHXCATS4HFZlxJnJzVi5kx0rIuZTzzXfRpnZGNhsrgZyeKyZ33ykAY96tvh1wDk1VeEhiScCux6I0KzoFUsTzWRdS5zyatXM7KWGeM8VmzOCh3da4ajLRl3cjBhgmq4nd/kbOKLpyzVDFKd2D1zxXMMZdRyH5dxxWRKHjyCuVrpHUPjPWkls4zFlgKAOV3lhgDFTQx+YwzVq4hjViEHeoElCHaOopAXo0Crk0ghJ+61VRcFQd3epYbrHApAXVXaQD2pxKlfmFReZleTUkZBGM0AOa3DbStJ9xiKtRxsU4oMKsfm60ARQn5jk1Iq/N1qNh5bY7U5W+bIoA1LO98siOTkVrRrvO4Ywa5xASQSK2bCcxkI/3fWgDVibaQKvRLuYGqka87uoPercDBG55zQBqWrHdtqcjDGqkOQwKnirUgJYMKAIZlB69arn36VYkHy89agYUANxjrSEZ4NKVz3o280AREdeagkBweatMvoKgmHFAGdKOtU5FwCK0JABVKYA8igClsqN1ypzVrHFQzJigDPZecConGBirciiqsgxQBQfh+lIW4xjrTp+HzTBjHNACjstICfNpFPz0qDMxoAW5h86IjuKoRuV+XuK1scVmXsXlSCQd6ALELHgmte3fKjmsOF8gYrStZMYzQBrpyBtrW0dtmpRsD061k275wa0dNI+3Ak96AOa8SHd4m1A+s7fzopniJs+JL/wD67t/OivEqfGzVbG5orLc280yE+W11MwyMcGRu1Pu38yQ4+6OlNs5NkFyo6m8n/wDRjUyU4OK9el8C9DN7lYLljTgDilUdacBxWghE4PIp/OeKTHy8VIg496aAkGVTJquW64qduE5quTzW0WIkiBPJq3EBiq0VW4gK3jIRKvHvViM1Aow1WkCke9bxkImQmpUbANRJxUykY5raMhE0Bx2rQhfkVQjPFWIm+YUT1RLVzZhkwMUty/zxHPQ1USTGKfO+Qua+T4ig/wCz5NK+sXpr9pDowtU+/wDIWW7+cjtT4JiRuY1nkLuOZBTt4xhZRTln2XfzS/8AAJ//ACJxywVR9PxRsW94UlGfumrk0oIDg8VgLLtUZIIHWphqasNoZSPTNc0s8y99Zf8AgE//AJE55ZfVfRfev8ye4usSHnrUK3XPWqc5y+95Av1pgdM5Ey1zyznAPrL/AMAn/wDImDy2t2X3r/M1obrjmrUV3hxzxWB5qZ/1y1LFPsbPmKaz/tjAd3/4BP8A+RJ/s6v5fev8zflmz3oglGetZBvdw6j86QXu30H40v7XwHd/+AT/APkRf2diPL71/mdlGymAEmgShSec4rnYNXl8oKsJf3Bp41WcMT9maj+18B3f/gEv8g/s/EeX3r/M6ESox96USqO9c4dVn3hhbsPxp/8Aa0//AD6t+tL+18B3f/gEv8g/s/E+X3r/ADOkWYdqQzDHJrnBq9wP+XVv1pTq05HNq35mj+18B3f/AIBL/If1DE+X3r/M3zOoHNZuoTZtZRnqprOfVpmGPs5FV5b6SWNlMR5GKmvm2ClRnGLldp/Yl29DSll9ZTTaW/df5mjptxt0+NfTP86vLJuxWLauUtlU8EVYjuCrA5r1srotZfRTVnyx/JCxFO9eb83+ZtebhcZ5pi3DAjNZhuSTnNSR3BPFd3srII0jZSfHU1FJNyTnNUhcHy+aa03ynms1S1OmNMlebgmq7zkiomlwBkmomcBSa6IwNo0wmkz3qoXw2ac8nFVySW54rqjGx0xgRXDk5rKugTxW4LbzlODWdcwhXKnqDXVSkk7Glijb2jy529RVPUBLa5Vz1FbtiuA2PWsjxAC0vA6Cic9QOWlm8xznqKzZ5WJYVauiY5PrVCZsg54rjmyinIM8HrS28AZyccimOrltyc1bjV4sO3dc1iMim/dKS3BrJvNQmhAxkqa1bt1mwM4rGu1KnyzyKAITdLNGT0b0qq7kEHHNIY/Lm3A8elTmEPGCKQEJZnkGeBUu4IPelERbHtQ0YPBJzQBMk5wM1oWq72GKz0CovPNWFuCCBHxQBtxsYuH6U+Uq2CvHFZf2t0jwfmqzDceYooAm2o3U804weWu8DIqNioYdjU8Urs21uVoAdCm75u1XYhnAaodhYfJwKkhBJwe1AG1Y3OP3bfdrTVMcjkVhQ/IQc1r2VwD8rmgC/CxUrmtE/wCrBrNU/MMn6VpRfNFQBG+T2qFx0zVo8HBFQOM5oAiIA7U0+op5J6YpCpPSgBjDK81BIuBkVZIAHvULg4xgUAUHJOeKpzDrV+QYOBVSRMHFAFFjtqJvm561ZkTg1VPy0AVZB8xqtJx2q7KMjIqo6ZoAz7lKq5y1aMqZB9azHJVzQA5cgmnQE+YTUSN8jc1JbNlc+9AFwDjmq9xB5qEGrkallzQwzQBiRN5chQ9q0YHqreweVLvHSi3l7UAdDasu3k1q6cQt6hbpnrWDbHgZNa9o2LmPnigDndfO7xBfH1mb+dFM1lidavD/ANNm/nRXiVPjZqtjdtM/6Uf+nyf/ANGNUkgyAe9R2S5N0c/8vk//AKMap3XgEGvXpfAvQze5Eoz0pduGNOwfTFBU1oIavFSKPmzSBODT1X5eaaAZK3aoeCancCocYarTAmj4HFWkHFQRCrUa1omIlTtVhF71Eijp1q1EMritVIRIgzgntUm3I49aYvpmpEGOBk1opgTL1qxCuOarIpzk1aUcCtVO6EThsUyebinKByxqvPzk04pNiauVZpQFzmo0myQAeTTJOW46VTuJ/shK/wATdPaujlTMpU7l+5v9n7sHjvVE3TLIPLNZkl18+XP1pgucMTms5UjGVI6mGdbmHEhOQKoyzFJSo6Vlw6g6uFVq0WZLiAsuAw6msJUTB0R8Un7zcx4qaS8A6Vi/athIJ6VVn1AlgFPFT7En2J0iXuBy1KdQC9ea52O53fxdKZPekfL+WKXsRexO40nVF3gZ4NbJuQeh7153pc77hk9K6aC+G1cml7EXsTfMue9NS4IbBPArPS6GzPamvc7QSpzS9iHsTWN2Duwai+1Z6tWUbsbRjjPWmPcAEYpqgHsTXe554NQ+b83Ws1rrjJqNLrLcHNWqJoqJsrcAZyack241mecMAjr3qUTgLweaTpGsaJpeaS2M1LFIQ2M1lpMW5zU0UjFsZrKVM2jSNhZexpryc4qtGxA5obJrn5Fc2UEPeXse1QmQngUpG480YAf0FaJJFKKRGcluaQD5qk2jnmlRQQcmquUTwf6rI61n3agztmtCPG3iqFyD5pzSp6SbBjbZRl8Vk6tGTM+RkVtW6gKx9aoXqjzGLYOacpagcbeWe/J24xWPcWyZ56niuzuUVuNo5rA1C0CsSorCTGZdvYqnDd6ZegL8idQMVOZTCMHk1Ru2MbGQt97moAztrOxVhgg5zVe6VBGd33qtBtxyrZ5q0LdJl+ZQaQHKSru+5UquyQcDmtS7tFgzsXNV7eHeDuH4UAUg7bR79anjAYgY5PWq8+6OVgAetLFK0akg5NAFmSD5sA0eUY8VDHI7HLNzU0dxu/1g6elAE20mPNIk+GG3jFPjnVzsPSneSobIoAvW5M0gZhWlFGMEj8KzonWPaqnrWhBIGwDxigC1CMH19qsKA+dqgGqiowYtnPpircCt160ASoSq4I5qaJiOe9NCqRyTmnxIAcE80AbdnIJIgG61q22PLI71gWsgjxnJ5rdgbePl44oAexJzx+NRbd2anUEHaTxTGTy29c0AVSuT1owTwO1SbQCTSDjpQBGRmoXG5ueMVZYZ5HBqJueooApSJ83Wq0y4q7Io68VXkA74NAGe4yKpOpGRitORBn5RxVKRSDzxQBSPHBFRSIApIqzKB1quzDBBoApSLhAT1rKvE2Sbh0NbTjbwazL5C0TY7dKAKK8hsdxT7NjtI96hgfD4PerWmx5kdD2OaANW3GIjxmkZMnpipIs7cZxUm098UAULq3DwEDk1jqSkmOmK6UqD2rD1K1ME2/nBoAu2km4Cti0f9+ntXN2c21tueRWrDdqjKe9AGfqfOqXP/XVv50VBeybr+cg9ZG/nRXiVPjZqtjqNPA3XYJ63k/8A6Map3AXiobDGLrj/AJfJ/wD0Y1WHIfpXr0vgXoZvcaKCKeuAMUbT6VoIYR6044C04IT2pHB6YoAhbIqPG5smpn5pgTLU7gTxDGKtIBVeIYY59OKtRHjmquBMgOeCMVZjHGTVZevFW4+BT5gHLj0NTR9cg8UxQD0qRQRlQKrnAmTae9S5APWoYuAARzU+3LdKtVBWJxwoJ6Gqdw2WOOlWpM7APSqDkmQ56VoqorFSQFEaTOB2FYd0zSk+Y3PY1p6lcZfYn3RWPO25fpWqrBYruMp1BINJuYrgAVEcs2M8VIhCjJNaKuKwqybFK4+b1qe1vWU7SeO+arSurLlar8od2c1ft0LlLuoAj548YrNV/MbrzV+O6SePyz1qpIixuexp+2iLlQ8ybEx3qqs0k9wAuSFqneXLbtobmpdMnKZzyaPbRDlR0MVx5EfHUirMd+xiDbu/rTtP01ZLXzJDmRx8oqlPC1uxG0jnkU/bRDlRu2uq5QKTx61cW782P5a5WKRgQOgNX4bllyBS9rEOVGt57Hr2oNwd4qvAwlT3qeSMbc9xT9tEOVEol3J81Ro43ZFVt5BwKegO7NT7ZD5TRhbcas4C8GqcTdMVYZ92COoqHWQ7FuIircIG7NUIWY4wQBV2AfPzmsZTTGXlpSxHTpSJyM07APXpXPdDI8seaX7688GnjbnC8UYHPrRzIBhXC8ClWMjmjODg1InLClzAPjj2qeeapT4EhzzV538oHPeqEhy1SpDFgweMfhVe8gbzN2OKlEuw09wXhYk5pOQHP3cfOQDms2a2Z+oI+tad9uzxmsty4b5ixFQ2Bk3emurF8ZArnr7duKOpPPFdlJI5Tbn86y7tCmXaNTxSA5SP92xHSrkLlEznPtUV26s5LJgZ7VGLqEKBnafegAkuQZCG4+tRLPGG+UZ9cUjpvkLp8wqu0Trll4HpQBdnghmUEAAmmLY24hbjmmJJtQZpj3GcgcA0AUZoSjZXpUatIwwBVuNGdj3ApZLYjlODQBHH+7OT1q0khOQRVZ9yL83J9qkjYFRjigC2A2FIHIqzDcZwD1qksrA8elORzwRQB0FrOhwj5HvWnGvGRwvr61zkU3QgdK07O/ZjtYcUAaYXDZqwkYfnoahT5gGHSp0JHNAFmEFeMitmzkbOD0rIiXfgitG3yCDmgDWZcgYpjjd+FSRndEDTdvBoArslMx2qdl5qJhtagBm0Ac1ER+FTHmozycGgCrMvPSqzKMnIxV2VOeKrOvNAFNxg4HSqs65q+4xVWVQckUAZroD1qpIuDxWg6EZzVWVeDQBTfBPNVp4A6Ng1bcAVUkbBOKAOdnzb3OG4571pafIBI0mRyKZqUCzpvHUU3TI/MQjuKANhZBngZFTCYHqtVoxtGMVYQA4zQA4lTyM1Vu44bmMxzlgMEgj1q2wK9Kq3EbYzQBzCyyRSGLpg8Hua0LRiw5PINQ6pbFWEydR1qKylyRzyaAJJ2DXMpH99v50VEfvvz/G38zRXiVPjZqtjudOUst1gf8vc/wD6Map9u1gKj09T5VyQf+Xuf/0Y1T7BnJzmvXpfAvQze4mz5qXaak2jFKK0ERBSKYyk5OanYH0qNsj0oAgKcc0xVJORUxyVpiA0ATxnpkc1ZUZIwKhiXirEYIWgCZQN3fNWEIxxmoQp49asIp29KAHJ16GpUJPWmrjFTIBt6UAOjTJ6nipl61HGCG4qVeOvFMAl3cgVn3EgiXB6mr83QNmse7YvIT6VQFC4yTkVnXDfMVx161oSuTx1qrLHlckUwM549qlgec01eQQammQVC5UKMZzVCITG4JOCRVa7uAibIjl+9WLqYwIFjc/MOfasiU8nb+dUBPZzss/Wr0+ZlDKee9ZMPEwOe1aMUpQfMODQBWu7QriQkH2qnFL5cu4HFXLmVxJ/sn1qlcIG+ZR+VAHa+HNQ8xRE7HJ+6T2rX1K3E0RYKcjrjvXB6VOyXCBW2nPFeh2VwJ0UN98DBB707gc00Mm4EA7R39Ks22A208mta9sWTLQDKH7wqkLZo+qFT6kUXAtwJs+b0qzhXGTVaGTaArdD3pyS7QwbgUXASaMIVIyQaliTd0BA96ckgKhcZ9Kk52+lICWEKPvdfapUdVc5HWqqAr1POasZxj1qQLKFQ2TxV62YM3U1nL0zxVy0Z+eOKkZpxH5SKcAc9eKjjHIxmpePWpAG6UikDtS4wRSlcDNIBjrlQRUsAAYDNN2ZBzSJ1AHXNAD5o2Jx1FVjAx/hNW5XdAMVAZZueaQFd7c4JCkmnxRP5LAg/jSmWZVJBFMimkdsOeKAMe/QhiAeQazGyD81a18uLhhnPNUJY8j3oAzJUAk5ORUF1CJOO2KvyRk9qikQllwO2KAOYuNOG5tvFY1/p+0ZxXYzxFZGqpNbLcxEEAGgDiUEkTcN8vpVveHj+Y4qxd6UYpie2azbhSjAJ+VAEpYbdqYJ96jRdz/ve3pUas0fJGTTt/8AFjFAF1CmDxtx6d6cWVlwOappL5nHpSxsySGgCUgd1qP7McllPFSCTfuyKswpiPI59qAKqIQeakWNV5LVadAydMGoii9AM0APiLE4HStCEBBxVGOTAwoxV2AlhzQBrWM/8L9K1UTJz2rBib5gORWrZTkMFY5B9aANO3HzZ7VehTLZyaqxL09Kv2/pQBdtmydpPFPYFXyDUcfyucYqZk5BPQ0ANkG0Aryahc5HSrJHy8VAx5OaAITnPFRt16VN1HSmcg0AV3FQMM9KtSc5zVdhnpQBUlFQSDjAq2ynPNV5lPpQBnyocnmqcinOK0JI+fm6VWkXHSgDPdBVOVetaUq4GRVORc5zxQBmyruUr2qtp58m6Iq7IME1SxtuFdfWgDZUZ7VKI/SmREsik1YVcDigBm0sMCleP5OeanVdvPWh8EYFAGLd2vmIVxwawI08m4ZW4Kn5a62eI44zXP6rbbGEy/jQBnZ+Z/8Afb+ZoqNTu3H1Zv5mivEqfGzVbHo2nZ8m5/6+5/8A0a1XEGetVNPBNvc4OD9sn/8ARrVeRSMCvYpfAvQze4hTd04pCmKk2nJ4oOO9aCISMimkfLU3rUbLkUAVzwppIhxUkiArxwaI0oAljU1ZTtUKrirCLgDNAEwBOMVYTNQKQMVOgPWgCQDB4FSqDgZqMDjjrUv8IoAljX5/apMAtxSRDC8dalQY96AK164WH3rEmJatbUXzwelZIXzGOeB60wKpUq2SKZLhlq1KgCgDk1VlXrtHTrTAoTIKpXJWIY7mtCZ1jBJ69qy5mD5ZqoClMGciqckbr09a0DIoJOOlQSSqDnFMRXjRg2eKteaEj55B6+1VfP64HFRNIXOB0oA3dGgt9Qn8mXkLyPer+q+GIxGbixz8g5jPeubs5nsplmgJDLXd6PrEWoQLvwsnRge9AHExRyR3ClV4B79q6yCeSOCOVT8w/Wo9U0tTIbi1GBnlaIlxaD07UXA6O0uxdxq4IBH3l9asERzho3249a5rz1tYjsJ3Y7VDZS3dxeLtZsk9KdwNO9VrNThcqPumswyvOfvY9q6W6WO5tvs+QXxyfSuZktJLa4O/OAePei4GtZfPGoPUCrqDjnmsqCUpIOa04/mUEGpAewzQpHRjzTsYNM/i5H0pDLMXpWhbMO1ZsRycGr1sMHikBpxk7hU33RUcZG0U8/NyKQCgt370qEspB7UoAHTrSgYFIBOSKFwDkinduKTbxzSAnJUqMio5ApHApV5XDc0sgAXigCEKoHNRsFAJXGKmVQeKguGWNCg5JoAx7kZuGc88cVQkBLCtV4+pNVJIfm3CgDPmUDqDTdmYxtHerZjyxzzTWUFQOlAGRdxsCeOtUjEVGDwa1rhfmNVZox5ee9AGPc2+7IPNZE+kESbwOK6Rl3Y4oeAMhx6UAcNd20iTcDio1ODtcVu38LK5BFY9zC68nigCLyhHICvQ1bgMZb5gKzGlkXhulSKzZGw0AaQRDIw4AoDhTgcAVmCVkm5bNW4pxI2GoAtiYGnxoHbikESr7inwnB4FAD1szksDU8MRDc5oRju5NTpkjOaAJ4kycntVuEMv49KpJkECr0PLAUAblhNvQI/Ve9aMQ4461i2x2sCK3LSTeucc0AWowpzVrZmMHsKrJw2KtpjyiKAEUAqearuvJqdTg8Co5QdxOKAKpyoNMPPWpmHHPSoW/SgCNiKgK5PtU7YxUTP2oAgcD1qBxxzVlwD1qB/m4oApS+1VnQEc8VblXBNVnHFAFGXCr61TnB61oSIAcGqk6grzQBmupNUZwBWk6jtVKZN2eKAL9i+6BavBM9KzNNk3Ls6YrYjBA5oAMELil2YFPIyfagY6UAVpErL1C28y3dSK2ipzz0qtcx7lJoA4EZQsvo7fzNFWJ49t1KP+mjfzorxKnxs1Wx6LpQHk3GRn/TJ//RrVobcvkdKpaSP9Huf+vyf/ANGtWigFexS/hr0M3uNHGaQrnrUpUCkIOK0EVmFMPAqcjPWomWgCvKPlznBoReBg0r46U8LjFAEsfHWrCjIqFF4qeMYoAkTAbpmp48jn9KhQfNVhBwaAJIxhqmC56UxBkipUBDEUASJxUwHyk1CuQalYELmgDMvGLEjvVPG1Oa0ZRnJxWey4Y56GmBCcMSMVSnbyVLE/hV+TCqR+tY91J5jH0oAzpnLOxPeqczfLirc2S3FVJ43JBHSmBWlwFqCTAUZqaTJzz3pqwmRxnpTAqkb1wooEXyehFXZI0UYWq0x2gBaAIwxUYPWtPTJymGThlNZQUyt6etWGl8h0CHjvRcDvrC6S7h4HPRqhv4kVXWLgcED0qh4auk8lo3OGLZrUu4ZC29RkNwRRcDKt7eWWbaBuY9BW8wi0m2BUB7hhgY7U1TDpdsJJeZ2+6PSq6M8spaT5pSc8+lFwF0u3unu3nkJ2H73uav6lCDAj4z/Srybfsqxx4z7etRXiGGxdJ+D2ouBzgJEv0rWgmDKorJY5mPHWr1uMACgDRcEMMU7aW59KYhwwJ5qxEAW570gFiTJyavwdarKvzccCrcAxSA0I8bRmpgMfSoEGQDU4yfpSAAuetSAKKaVI5pdhNIBWVfpTcLjqacY+9IqZBxQAuRTWbcMGnbABzTHUdqAGs5VcCqswPU81YZahYE5zQBVJ3Dmq0gPIq264zzVZ+aAIAuByKYyqTwKm5Aphww64NAFG4iA521SmA29OK2ioMeDyaoXMG5DigDIDDnNRyTBV4HNTmFlYg1DJGGyB1FAFOWBZ1JYVlT2YKkNz6VtSExR7T3qk4D8UAc1NaEMVI+lV5F8g/KM10VxakZNYk8BQkt0zQBXWJJBk8GmYZWwnrU0ZBap/JG0stAC29wRwefWr0QDjMZ/Cs1Y8Nx1q1EHT7tAF4HJHGCKsISOlRwvG6YI+ap064FAFiNMqCetXIEycGqsZwQM1pRL8ooAswR4xWnatscDtWfEema0IRgA0AaaDdg1dhAxg9MVQt3ycVfiFACNgLheuajkBPOaldfmzTHHyGgCo4+XrUB9DVl+lQN1oAjdeBioWX1qc1C9AETqKgc+gqdqhbrQBWkB71XYAirUneq7jAoAozKC1U5E+U1oyKCKqSjAOKAM2RcVUlGelaMgz1qrKg6UAVrI+XdDPQ1vpljXPfcmBXsa3rVyyA+1AFgcj3pwUZxSqvtUqgDtQBCYiM+lV5Y8qRV8rmo2jJzxQB55ertv5x/00b+dFTakuNUuR/wBNW/nRXiVPjZqtjv8ARxm1uOf+Xuf/ANGtV4AiqOkLm1uO3+lz/wDo1q0QoIAPNezS/hr0M3uKVyM9qiOalIxwOlNbAFaCImAqN+nFSHjOajIOc0AQSYzinKMYpjDLZ9KkUHaDQBMvSrCgHFQR89KmAzg0ASjhuBU8fNQo2etTKvpQBOnLelWYwCxJ9Krx8jBqyi7l9KABSA3rTpDg9c01VxJgc0sjBfegCrKSMheapSx+vFXXG7pwarXHyoc80AZl04+5nisyTG4jtVycEvz1qlcAiTFAFKRPnPNQSI+OelTvhZCSajeVm4xxQBRkgzwOBTcCHrV9VDjBqtdQ+nPNAFThs5NMcAZqVkUHHSq1wy5wjfWgCBpNilulQCYO1I6uxOORU1ra7kLY6daLgalhelXQL8pUZz613ek38dxZhzgyAYwa85R1XG08gVt6VeNblJB9NuaLgbU1rc3l8wkJKg5Fbi2qJCMDLKOD61DY3MV0gMeCw5I71PEskU7uWLK3QHtRcCjbO8EjPcEgZ5WrWq3KvCoD7i2Me3FV9XaKNQxcBx+tYrXnnOoGQuOaLgX1jEnzCrkCelZ9tNtGK1IGyOO9IC3GnHrVhEwfWmQgVZUenSgByg5xjmp4VO7PpUY+X3qxEO/rSAsqOlWVfC9M1Cg4qZR+dAD8kr6CgNkUFe5P4UItACrnbyaaNwz6U8jH0pCePagBDmkYZAqTtkUx+RkUAMfC1E5XFP4IwetMkT5aAK7gEVUeMZ4q4cbaqyHBoAgYkcVGRz0qVgTSb+Oe3pQBG8eV5qpISoIq+WDLkHmqk65U80AZcuQ2ar8N1HNXHjLDJ6VC0QA460AZ11DnvzWeY2jYcfjWxMm71zUOwhSGGaAMqR9/BqlPbB1I6itSeP5Dhcc1W8tg4IBIoAwZrNowSgxRDlVCtzmt3yUk3Eis24gEcg2A470ARbVTnFPiem5DUiblbgcUASqWRsrV21mLthuD61WALKMetS+UxYFeKANaFdzc8e9aNt1w3NZtpmRQj4GO+a1YsKAO9AF1IwauW44wTVSLqKuqMAELQBYh+STrxWtCQQMVkxEZ5HJrSsjkEUATkDcc1C4681O+3cDUUooAqP156VXYgNVllyDz0qrLQAxiM8VE5xT8nOKY/Dc0AR5G3pzUDg1K3WonJoAhcGqzZyQeatt905NQSDpigCs68VTlTNXmGaqyLkGgCk6Yqq6gtV91Gyqkic5FAGfPGAc1padJujAqpIgYHPWpdOfEhXOKANpBkVOsZxVaE5YAGrynjFADAuKbtzU+3NCxgt70Aec6quNXuv8Arq386Kk1jjWrz/rs386K8Op8bNVsdxo4P2Wf/r7n/wDRrVpDPaqGin/RJ8/8/c//AKNatAcnrXt0v4cfQze4mM9aY3HHWnvgHGM1GT8xrQRGR1NRPjGc1KTwaglxigCEcsfSpUHIzTFFTouaAHqPmGOKnVqhX73NSr1pASAVPHUQOeKnT5frQBZiQ96mORwKhjY7eTzVgcqMc0ACrlht/GklGFNPTI6cUyQZoArMnyljVOds8Crk3C9etUZBSAzLoEEnFUJXBIrWnX5TWVLFljxQBSkTexqFl5xWhJEoTjrVOTHTGPekBCcIevSoJZgzc8YqchQrM56dKy7qXe2AfpQBHcuXbg4qsItrEsc09VMjhTTrqMquI+T3oAhEqDIHWpXdhb/uu/WoFgVU3EZapEyq89KAFgt9kfmSNz6Vpae5DfO230zVIAyAAjirTzbI1SIc+tAG4h8lfMEjKf7ympH1W5kxHHKdvrWG0pKKrOc1astx4Y0AXrlHfaWkLk9ajVSjjJyKnmJSDAbp1qurZYHFAF22+961u2q5iBxzWDFMEA29a2bSYuq5NAGqDiMcYNSxsWFVUIP3qsxnnA6UAWMZxirUeRgVAoxjHSrCDkGgC5GegNT7QOhqsjDuCanB4B/SgB5Ge9Kp+brR1AOMClHXigBx6+1NPJp55FNA9KAHA/LTSM5p2OKTFAEO35jTHHBqU9ajkxigCrIMrxVZhng1ak46GoGBzntQBAwFQSHBwlWXXFV3UA7hQA19oj4+93qpO3y1aLZXP6VWceZnsKAKbk7QBUZGDzT5MrVZ3bd1oAWRB2NVpA1W9wYVG6ZIOOKAKax7shqhlQKh2c1cZipIVahbIU4GM0AZcibX4OKgmiZlrQubfcu4DmoljBXB4oAwZIGgkzjINKrlXw/StqaGN1K9TWRcWjBh7daAJRk/dq1btv4PaqlqrqeelXVTy/mTvQBajGGx0PatO1m3sFPUVlwkvgtjPrVuLaHBPUd6AN2JW43celX4RhetZttMsqj1WtGA0AWEA6mr1mdsg9DVRR8hHWrVpnzcHoOlAF6Rc9KjkXEfPU9KnIJTPeonPHzc0AUnGP61UlxmrsmACcVSmHcCgCEt7VE+Sc08k0xulAERxn5etRuBUjVGQKAInGRxVeT27VZbg8CoHHWgCu3FQspIOKncc1CxxkUAVHQ1WZKvP6VVdDQBRkXHSo4ABPk8GrMseBVfDBgaANuxbPbkVpKOKxtPkxOAf4hW3ChI5GcUALtOMijYSARUgHGDSp6A0AebayP+J3ef9dm/nRTta/5Dl5/12b+dFeJU+Nmq2O30XH2WfP8Az9z/APo1q0MelZmjHFvcZ/5+5/8A0a1aIb0Ne1S/hx9DN7jj931NRN3NO3D+Hg1Ez8kYrQQwjgmqsrc4qWSQKpI6iqwbe249aAJVOKmB54qFOT0qccdqQD14PIqdeKhU5IqcCgCRfap096iTip0IoAmT7w7irXXpxVYdsVZQ5UDvQA9OKibPmVMMcAUxyCD7UgKc3JIbtVRgMknpVlvvEZqs6nmgClNgriq8igKCBxV5o+DmqUq4J9KAMyZvnPYVDNtcZ9KtzwlwcCsq5mMZKgUgIJ5A7lf4VrPcqZDx06VYYmR+RgGh7dVORQBXDqM4HzdqZGh3EyHGacVzJQxyT7UAK+3H3cirf2eFrPcQM/Ws1S0km3tWhbWrPGzFuF7UAQsdq7VHHaqyyFQSOcVJfzBIgi/erPiZ8AHuaAL1vKZpdzDgV0NpEoVeea520OWKjrXQWw22oJPNAFq+KgKBxnrWaZSrbVNT3k5aJV71WjiOQ3U0AXLclmFb9i+MFqx7ZQqjdWtAUVRjqaANbfkDI/KrEbdMVQjyzDBq6mRjmgC5ESRjvVpCRj1qrED1q0gIwaALkOSOSKnAJ5zUcC5XOKsBSRQA0jNSDHA70BOKcuMZoACRjFA9qdgY5pQRSAZlqU9KecbeKTGOtAFdgc8Uxh6ip2HzcVE4pgVZF3NioGXk8mrjgDmoJF44oAqsvr0qvLx90cVa+tQyg4oAqnI7VA3GcdKtkfKQ1VCP3mKAKdzyMCqbAitSWMt2FUjFhzmgCuOtS7+AvemsmGyKTfx70AK0ec4HNVGQ5INXVcn2qOVBu570AUm5+WoHiK844q+0QRc1DnPBHegCk0OOVHNDWyuuSPrV8xhm4o2ZVgKAMh7UqDsHFLEGAxIOKvFCBg1Xmi9DQAgCowxzVqMAjJGKrRxlmGeKsJkHHUUAXrWTYwxiuhgZWjBxzXPQJuxxWxZPtIVjQBpKw24HWrsI/iFVFQZzVyADFAFtASu3vUcgweKkjPzdaR1AyTQBSmXNU5Mk8CtCYA8iqMpODxQBUcgHFRsacw+Y1G1AET8nimHg1K5GeKjOCKAIifmqGQc8mpmGORUTc80AQMOaicVO45qCTigCFl796rSAjrVonioJckc0AU5EOMg5qBuRxVpxnoKgePjigCa2PyK69VNdDBIWQMp6iuateHKnoa3LNwqhTQBfUMR0pyKNw/WkRsmnrwTigDzbXQP7evcdPOb+dFGuD/ifXv8A12b+dFeJU+Nmq2Ou0k/6Ncf9fc//AKNar4YjpzWVoqiCxliDMwS6mUFzknEjdTWiGB4zivZpfw4+hm9yVpNw6YqCR8c9aV22ng1Vmm2Ar3rQRFLJufjinIMdahUbjnvU6DsaAJUxjgVOtRKBipB2oAmXnpUq+9Rr2xUooAkU84qyoxjNVlBBFWVPHIoAsADAINWI+xA4qqgIjwRVuMDYO1IB+RngYNNmAWP0zTl+9ReKNgoAziqg5zTdueccVMV46UZAXB4oApzqCMrwKoTIFJBya0JByfSqN0/lxknn0pAZV3OYwVUVjSL5khJXArTkO5suetUp1P3RQBRZccdlqFn+brxU8oI4H41XlTK8daAGMRv+Wo2Pz0qId+DUwiBPzde1ADUj8v5sc1JHdGGF8nGe1MuZNmB1NVJcsnPWgBkoErK4+bJ/Kn+SImIYZOMimWxCHax+lWPNC5aQZOMCgCOzkQTnjkniuiCB4RsPSuNhlK3gJ/vdq6vS5N+7JyKAFulO4D0qe3TgZqeeLPzAZGKVNmwELigAxtIX1rQtzjaD1FUokLvk/hVyMFDnvQBqDCoCPvGrMDE4Dc1Vi+ZA2KuJxgigC/GQVGOKtoelUozwKvRrkCgC9CRtHGKsDPRScVDCvyjAqyPlHSgA+6OeaAOOBTs54IpAxDYNIBVU96XZSnIpvzH6UDHhRtpGGeppOwp3WgCPge9MdQeRUpWmMtAitIvFQMMGrTjPSoX64pgVX5OQMCoJcN0q2wPIAqs69aAKrjjmqjJ82c1e8vLkNUEkYVjnpQBTcnmqx+/zVi5OzoOtV1+c88UARy8dKrtVlwc1A4PagBgPNPZhs9TUasQcnpRu64oAYH3Ehvwo2gdfwxQ8QOH5pm/nmgByIeST9KVvkPTqKYsu6TnpUrnewZecDmgCMpvjyKrHhxkA1e2kDPY0wwA5JHagCsw3tlRinRRqv3xzT1iYZxU0ce/hhQBPAuxeO9W4gd4NVkQg8dKtxZK4zzQBrWkgZQrcmtGFNrDFY1qdsgOa3osOAxOMUATqBu4plwjcVMBlML1pr/NHnNAFJ8bDiqkw+Wr7L8h4qjKOaAM+VcVD1Bqef71QdjQBHimYFPPFMFADGxUTD0qRjmmGgCErmonX1qZj6VGwyKAIHA28Cq8nsKslSagcUAVXU1Ac9DxVp+aruvPNAESDbKMGteA7QMmsnHp1q/bkbVLE+lAGvG2cVbjHzVmwnDcGr9uQTyaAPONd/wCQ/e/9dm/nRRro/wCJ/e4/57N/OivEqfGzVbHT6Y2IbnA/5e5//RrVbBHWs6wc+XcgH/l8n/8ARrVYefapOa9ml/DXoZvcmmmAXNZ4cvISetNaUzN14pYxzWgieMfnU6Dt3qEdeKmWgCVQQ3PIqbb3qIHNTKeeaAJFOBUq8ioxipFPpSAlU5xU6npk1AvvU8YHegC1Gd3Q1ZiyRVVBg8cVch+7+NAEkf3qLwFlGKmjj+fIpJ+RikBQBB6imOgftxVjYEzkVE4ODg/hQBRmTqc8Csa8YtJkfcrUvpSilR3rHkkO0g0AUpVO7j8KqOu6Q7jVwvukGewxVaeMjkUAU5UA5HNNWESLyMVL5RPWnOAIwc9KAKgtSs3FFxiJDwM1diAYFm7Vm3waRjsPFAGezM8vzGnPGSvFKsBDe9WCnkxAv1NAFf7MHh3NwRUFxKqWu3qadNcsqHaePSqcq74l3dWoArxAtJmuw8OJvifJGa5cRFIuPzrd8PXB2uo4z3oA6K6cIu3viqwYyRKo65qN23zcHOKsxxEqGFAE0CsvatCKJSQTUNvDkDccCryLlsDgUATxsFwAKnUndUIQKBmnrIsf1oA0IxgAscVfiOAPesRGZ3BJrXtfujd0oA1YD8oqwehxVSE8cdKtLgrxQAH734U7FJxS96AHAndSg54pRzRjHSpGJ2NIp44pwFGMHigBpHNIwpS3rSP0pgQSDacioT8z+9TsM1A42tmmIiYYY1XZe9Tk+lRscjigCFkBQmqzxHb7mrTjK4FQnjrQBl3MZHWqbAgVsSoHPJqlPbkHigClnb1FNBDMcVJJ8v3hUaLjp0oAYY1PegqNvAp3lbjx3pyQsfloAg5PGKhdCG6VfFtI/QYpw0xn+8aAMmRCBkCnQN/DjitU6Vk4JqeLS1wcigDKGc7amVXLYArT/s9F5A5qWO12ryKAMxbcnrSGDb0rbFqpHShrAFelAGZFFwN1TeV5fNXBaFcfLkCkaAkYKkUAR2w3SCtu0PbrWfBAFPSr0B2MKANJFIb0ocdRilXG0E0pHHFAFMjqKpSrya0GTBJJqpKo3UAZN0pB4qr9a0LtMqaz2agBnGaY4+bApT1pDw1ADGGKjbrUh5phoAhII6UxlqXNRNkGgCNlxVdvlbkVbcZFQNz1FAFV1zVd0q46nFQOKAKu3B4qxCSUKk89qiK4NEL7ZefpQBrWzBkB9ODV2JttZMD+VMVzwa04nzgCgDgNb5129/67N/Oik1vP9uXn/XZv50V4lT42arY2LWdYvteT/wAvc/8A6MammQyN14qnGC015np9sn/9GNVuFRjB/CvZpfw16Gb3Jo09OKsIMVFHmpkGK0ESpz1qZRjpUa4qVcigB6DFSr1zTFqRcd6QEi+tSrk1GmMVKvAoAlVcVYTpUCepqdeRxQBYRhtq5ACQBVKMDbWlbr8gpAWIuvXFI6596fAASaeMY545oApTIVqvLhF3EcVflUNWZe5+6TxQBj3bF5Ce1Z8x45WtOdFU9azrlsNgc0AZm1gzE012Lrk8VM8JOXZuM9KjZVLbQKAKzLuOVPTrSJDvUnsKsMFI2gfXFQtlFKocD0oAgY4jYAYqkSyPgjirMnynBNV5XCnHWgAUKzFsYxVK7kkPUZA6Vb2hYdxPXtVSQlwQOlAGa6s74NTBN2NwwBVuOFclmpbkp9m/djmgDPnIYiNela2iL5Mpz90Cs23tyX3yVqWETBCx70Aa1golumwMjNbkUSqpU1naSqRkkYyavh8yH1oAl+6wHarEZywqrGM/eNSGcRD5T0oAvvOsa8jmmDExHaqZkMg3MasWxJNAGhbIRyecVqwfMowMVRg2+Vz1q7b52jFAGhAxHFXIxjk1TTKgGraZIFAEgHNLnnGKAOKUCkMdtz0pexFIOOTSnsR3pAABpc0gNKRSAY1NJ7VJiojwcGqQEbAg1C4z1qyVIqFl7npTEVmGBxVduOnFWnHHFVpBigCIv271FIM81IeDUDy4bGKAGMBmoJVParBO4ZximhN1AFB4w336T7KXA8ocVqR2O45erUdsiqccEUAZcVhjG7rUws13ccVorExxlacIsnpQBRW1p3kAVcMZWl8vJ5oApmEdutOWLIOaueSMUnlnPFAFeOBamEAPWn7cU8CgCA2/pThHgc9Km25704LQBWCHsKd5G9ssKtKvzYNOEQJ4NAFPyccClVCOtXDEBTWi+U4oAlg5j5qTHHFRW/pUwGCQaQyrIOcVVlFXpFw9VZwBTEZ0qllNZMibHINbjjGayryMht1AFI9aTnOTzTmOaQGgBhpjDNSZ55ppHpQBEVNR4I61OR7VGwoAhbmoWGRVgioWFAEDComXaeeanaoW5NAFaRevFQDhqtucKaqsD1oAnMnCt6HmtO2mVtvfNZEYJhIP0qaxnKy7fSgDl9aP/E8vP+uzfzopmrtu1m7PrK386K8Sp8bNVsaMBHn3mR/y+T/+jGq5Ect06VUg/wCPi8BGP9Mn5/7aNV2MA8DtXsUv4a9DN7k6ctU2OKiQHbxU6jPFaCHAdKlXmmAVKB6UAPUZqReetRhT2qUCgCRcVIOTUagVKvXigCZOOKmj4qBBg5NToc0gLMPJ61qW4ITHas62HIrWh4XFAE0KZ+tSFMA0QnDfdqU/Nx0oAqYwc81jXanz2JJAzW1c/u0OCayZyHXnOaAMmZQ0nynGOtVZoNyg9q0sDeTIv0NUp3y7Ac46UAZ00WEx2quEAQnvVoh3kOeB6VBISDt28GgCmf4m6ZqoweTPzYFaclnJIwEfINSxaBNKfmXAoAwnReQOvrVWSIAep+ldvb+HIQR5oyKt/wBjWqHCwqfqKAPPPJuJvlWMkfSmLY3QbhDj6V6ULCIfcjVcei0w6enPyDn2oA89XT5iSGjIUd6g+yPuOVYj6V6QLCMqVCAHvTTpMarhUVh9KAPPI4V3gMvArQjRVt+BXTyaFA6NujwfUVk3eizxvmBty+mKAILBtjZkHy9qsLODOWB+WqbmWH5ZlKjtxSptYYoA1d/GVOKgwZZPvZqu8g6JnPpmrdqh2Ak4oAuwKDHsPardqD5gXqKq24Ak5rRtlCPuFAF1QCeK0rXaVrMhbLEHvWnZKBnNAF1QePSraY24qunPQ1ZTg9KAJQAYxSheKQDApTUDHL0xikBx2oA4zSbTnigB3B5xS9aRQR1pwXNIYwgGmNg8VKVyOuKYVH1ppiIXGFqE9KstwKgZDyRVCIWByagcY4Jq0wOBUTKCTnimBQl4yEqlKAGAHWtGUAg9+fSoBbB2Bb+VAFaGNmfBHFaUUK44FPjjA6DipVHoKAGrCKkKcDFPVflp+04GKAIzEBjFIUyKlINAXHegCLZlQPSl21KBycDNOA4xQBARjvTMEDip2H41E4PHagCE9cigEnrT8AHFNPX0oAFODipQB171EFOakAoAkBz1qSIY60xRx6mplHAoAUgnpRjP1pyn0pyr60hjFUhs1Iw7mlPSk6qAaQELgk8VWmTPWrknA4qtKDjmmIoOMVRuot6nNaMqntVWVc8EUwMEjDHNJVi7Ty5cgcZqvn0oAD7004pSce9Gc9qAI2GetNIJGMVJ35pp9qAI8FV6VXcE1ZYjGCajce9AFRgRULCrbrjrzVduvtQBCV4qu4HSrLZJ46VXkx6UAV1+V8djSeZtnDLxg0kmRkg0oAaPJ4IoA5zUX36nctnrK386Kr3n/H9Pz/y0b+dFeHU+Nmq2N+35mvMj/l8n/wDRjVciXniqtqD515/1+T/+jGq4hYHgV7NL+GvQze5YQ44qVBz1qJBnqKmFaCHqOetSqKYExzmnrx3oAlGakUc1GPpUi0APBwakU56UwdKeAQOtICVDjrU0TCoc4AzUqY3cUAaVomDmtSJcLVG2TbGoB+9WjAmBgmgCaEMTU7J0xUcatnGeKsdKAKN78q4xWQ/3jkcVsXWDnisqXA7UAU7tN0RI4rGKgsSxxWxO5ZsdR6UyHTzcHceB6UAZiwyS5CD5fWrFto7PhpOlbkNgkajjkVa8vGAooAzodNijXdt6VOIAORxVwx+1GzIxjFAFTyeMryKZ5Z3cVe8v5ajCDPy9aAKvlg9OtIYDuq1s5pNmO9AFeO2Z5AsS5Yjk00w/K2OCDzVlUyfnOKfsXHBpAUDDng1HJYlz6VolcDJGaDHletMDDn05HUq6A++Kwr/QHD7rVvwrsjDnOagktqAOGS1ZGw/3qvxIwXArdudOjlydvNZrWjwMcjAoAIDzgjmtFfkC+9U7VN0mewq7nc3y9qALUB+bOK14OV6Vl267ACRkmtGA9KALkQx0q3H05quoJWrEQ+XmkBOrKCKcTnnFMwBTwRioYwzlaOlA5JoI4ApDFzmjcQaQcUD3oADnPNNIOeKkJz06Uw5xxTQETjHWmHmpmXI5qJvl6VaEV2yDUbAnkVYcBhxwaUR9MimIrGIkcrTRGR2q2wycYpPLoAhWOpVjAqVUxQRg0gGY5xSqM9akxSYoAbj2zQFA6rS4xSgUDG7SG46UhGTUmMUjDNAEDCmsDxmp2XIqLBJx2piIWXOcdab97g9anK4pAuaAIPu96kQ54qXYOmOaPLw3NAAFwamXpzUYX0qVVwPWkACplGRUafMakXikxilaTpxS0fhUgREVBKCassMNwKgcdatAUpQwqpLkdauyqTVOUHNMRn3EQkhPrmssx7Sa22XB5GazLuPDEigCtmjOKTNOGDQA3Pzc0nBXin8/w80+O0llzgEfSgCq+B1prJx0rQGmsfvMv4006Y+Ov65oAy2XORVeQdq0Li1lToufoKpyKQPmyKAKrZVeKqynjirbnB67qrv3waAKMnzCi2deVPJp02c8GoYnKzgr170Ac7fDGoXHH/LRv50U7UOdSuD/ANNW/nRXiVPjZqtjdtv9def9fk//AKMaryjIHNUrb/WXn/X5P/6MarqcgV7FL+GvQze5Oi1MpqJMA1Mik1oIfjNSIoNRqDmp1GKAFXg1IDmmgc04CkBKOad7GmKPSnrx97rQBIB61Paxl5h3weRUI5H0rQ02I/6z14xQBqQqAuKtwrk81ViHOauxjgYoAswjmpyuahjG2rC8qaTGUZcYNZdx3A4Nasg681X+yhm3NyKYjPtrLMm5161eWBU6LirKR7V6U8R5GT1oArhFJ54pfLx1GKnEeTzSlN30oAr7M9DQUOOKnCUwjLYoAh2kDmmbQG4FW9nHFNK89KAK+zLdKYyDdjFWipz0pNgJ5oAr+UD2pGiAIwKs7cUgX1oAgCH0yKDFxU+wg8U9kGcUAVGjGBgVG0Oe1XCADikK5oApG3FVrqzEy4xWmyHNL5XHNAHOS2Rt14H402JdrCt+a088YHas+S2MTYxQAsZwMVbtcD7xqntKtk1PC3zZpgaiuABg5q1DlulVIxuUYq1CGBpMCUjnmnqKPr1pRWYxwHy+9JmlHJoPWkMTGPeg9cUlOHUZoAQDANITingdaY+fShbgMY5FMxTzmmEnNWhDVUE9KkKnjFCnB6U/JJwKGwIsZ7U8IKfj1FBHtSuBGRzQcDrT8c0jAZouA0e1Iadg9qCaYDMZoNO96THegCMtg07eDikbBoAApiHe56UnUYA/Gm5PfpTloGIYwPekCin/AEpBQAmAB05pMEjkU4dDzTutADQlPC4p23jmlA9Km4CDB6U4DFKoA5IpG55FK4xM89KcvvTMU9V7k0MBr8e9QScVO3XjpUMoqkIpymqsq5q5IOetVpBVCKjg9qqXEYZSCKvOKgdS3OaAMKRNkhB6U+GFpmwnT1q/LZGdhj15pksiWUflp96gByxRWyZfk1E9+QxEY49jVSSQyn5jUYOKAJmupmb/ABpovJ4+QfyqPO49aYeO+aALKaq54kXI9zUjG2u0wwAJrNYbuvFRM237hxQAt9pzQqWi5FZTKVGCOa2YtR2DbKNynuabd2UNzF5lt97uKAOdlXDcVXU7WOPzq3cxlXKvwRVVRjNAHO3jZvpz/wBNG/nRTbvi8n/66N/OivDqfGzVbHRW/wDrbv8A6/J//RjVoRLlRms+25ku/wDr8n/9GNWhFxgV7FL4F6Gb3LCrzUoGKjj9alX3rQQ9eCKl70xcU/PpQA9RinLTFGafj0oAlWpFwKiUHFSdhigB2dwAA6mtmzj8qIetZdnH5twFPQc1tKpDfSgCxGDnNXYBxyKqxgnFXY+CKALEeB1FWUA2nFQxrkVOABUyGUZFw2KeoxTnXMhoA55piAL7UpUmpFXil2cUrjIcYPNLj0pxHzUbaYCbeKYVNS4oxQBDtNG3PepccUmBQIhIxTGVh0qYjnpSbc0wItgIGaftANPKdKUgUAMA9aTAOc08D1pduTQBAIwelKUU9alIwvAo28ZpgRGLmlKZ4qUj0pQvrSAjEWF7VXmtRKpq4FpGXHSgDn7hGVtvpTU4xj8a1rm1Eilu9UfK25BFMC5bONoArQT0PSs63U8DFaMQzik9gJgTt5/ClPzZpvJOMU/GPxrMoFGFxSkfKKeq5pSvGKm4EQpSaf5eKNlFwGq2KQ8nnrTmT0poAHWgBu3DVGAcmpTy1IRiqTENx2HUUDn604jHSm4wRTAeOfrSlaBwaU5JqBjStJUgFNIp3AbSEU4ikpgNPFNNSEUnencRHtpMYNPxzRjnmncCIDk5pRgVKRmozRcApaaDmlpgLt54NOAx1pozTx70mA7rSqKBTgRUMYuSBimdOtPxmkfk0kAxuaBweelBPtR1qgFcAc9qgkAKjFTdRUbjjFNCKsvTFVZFzxVuQHNV3B3VYisyfpULL6VYkB5qMJyBQAKgiiLHr2rBv0Pn+YO/Wt+7baFVe1ZtwgZfmHWgDGLEDihjRMvlSEH7vY1EW5oAfkd6RgCMjrSBgVJNNJHagAP3TmoW96nOTjNI6gjApgUJcevFRQXslrJ8pyh7VLKNrncKpSEDNAGjcwR39rvj4cc1gvG0Umx+tWra9+zT8nKnrWje2iXluLiAgn2oA8+vOL6cf9NG/nRRegi/nBGD5jfzorwqnxs1Wx0Vr/rbv/r8n/8ARjVoRY3AGqFrnzrw/wDT5P8A+jGrQi55r2KXwL0M3uTrgNgDipl5qNc4qRRWgiQEDgVIqmo1HNSAkUAPXApwI70wHPXincDoaAHh8UoY7gaZ/GPSnoCzhV6saANTTUwDIR1PFaaNuFQQKIownpVhcdqALUXarsdU4uQM8VcjHTHSgC3GCB6VOpHeoozkc1MMbazkUQSAA9KRVyalI3Um3FO+ggA7UuKUijFSMZgZpcUuzjNL0p3AaFxTWGakFJjmi4EfbFJtwKk20EU7iIqTb6U8rS7MU7gMx60HFPIzSBaLgMHPalxTiM0baAGkZNKFGadilI4z3ouAxVp23NO28UdBSuBGRikpxFJimBGRniq8ttk5FW8CjaCOKq4inGjJKM9KvJ09KaUHfrUirSbGOUfLk1IpBHSmhTSgelZMY/dxS5yKawwBQg65qQAn3oyccUY+anYHagCPcwpnrmpitNYJj3qkwIs80tMJO72pwOe9UIfjJpp+9Sg4oIyN1IYd+tOHIpnQ08EUMB4ptKPejIJ4qQExik204D5qXii4ERGBzSVNtBFMZe1NMBmOaQinYpQKdwI8ntTXFS4xmozzVJiI160ppSMUDpzVAA5WnCmilzzSAerZ4pwPOMVGAM0/OCDUtDHg4pSO9JnsKUc1ADCKQ1IajNUgCmMcdacelNPKmqQiCXGM9qrSYPQVZK5Qg1XcY+lUIgcUkQ+Y+1OfHY0+IYjY0wKM3zMaqSjdgGrbdScd6hk57UAZd5CJIeByKymbGc8EV0LgLwRwax72zIJdelAFfd8oxTQQW64pmQkmO+KQsPSmBNu7Zp3QVBv9KUSAKaAK91zz3rOkB5yeKuTtuqnLyuKAKb4HJq5pV8LebbLyjVRl44pkZzx70AZerxodZuyh+XzWx+dFVbyRjfT5/wCejfzorw6nxs1Wx0Fof3t4P+nyf/0Y1X4s9Kz7bia8/wCvyf8A9GNV+MHCkcCvXpfAvQze5cjz3NSIahUj61IARWgiWpAPl61GuSKkGCnrQAU8fSmj3pN+PzoAlJ+b2q9p0W6bzD26VnBt8gUdTW7aQ+XEuOvegC2uNvPWrEajAxVdVyPxqzFxigCzDyOavRDA4qjHV+EfKOKALMRqUc1Egx1qUcjis5FD1GRzQR0oUHmn46ZqAI8ZpR+tPBx2pCR3FFwGkUmMjpTvoaADQBHRz0p5GKTj8aYDCKKfjNIVxTuA3FBpQKUigBuMUEU7jvRii4DQMUAHNO204KM0XAjAwaWnFcGgii4DRSMOaXFHbmgCMj0oIp9JVXER45pwAx7UbfmpxFADCQRUi4AGKTaNuKUDHakMl34xxzSEnOcUmc4NBHpmpAcH9RSlueKZyBgUEmlYBc805evNRjrUoHHNDAUjP0pVt8ctyKfH71IfQms22tBjBGmM7RzUT2oL5X0qS4bbF8tNjZhGGbmmr2uBXaNlPNN4C4q4CJelV5UCk5q1K+jERU/gdKb9aUCqAcGx1FPHT0pgyOetPxnk1LAUUUh9jThSAaeKQjPI608DHWmsMnjgUANIxSe9ONMNUAhPNMIPanUA5+tUIjOe9J2p7D1FIABTAQdKTNOxigD1pgIvHNP4K03HPFKcfjSAdn0pyHg1Cc5GafF1NJrQZN1FMPPanjoabg461IDDwOajA25z3p7nimEE81aEMxzjtVaQEEirDZ61DMvzA1QFRlINSRn9w2etIy+hpVHUH0piKTgjOOlR4DDmrD45qE4I6YoArSqTkEfSoY4RO4iYdTVlxkk88UWKf6YuDyc0wM7W9JiigMtuDletc7vx1r0YZffE6KV2kHIrz7UYlg1GZE5UHIxQBH5mOKaXqLeDyaYZOaAFZxznpVOZ+pqSQkHPX2qrKOM/pTAryHLZFMiOWJ9DRISRxxTYTh+aAMS75vZ/+ujfzoouz/ps+P8Ano386K8Kp8bNVsdBbf668J/5/J//AEY1Xo5OnpWZYs7G5aZNkhupiyg5APmNkZ71oRcjNevS+Behm9y8hqZRlarxyblPrUqsfWtBEqnDZJ4p4IHANRrytOXFAEp4X3pNwA6U0EZPeiJTI/AoAv6dD5sm9lwBW0nyrxVSzQLCF6VcTOMYzQBJGCeRVmOoEXnJ6VYhXnrxQBZiB4q7GDkVTh+ZsdMVcjGW4PSgC0owaljODUAJz1qdKzkUSZ9KUZxzQKfgHrWQDcUdetOJ7U09OlACYFJ9KXtRTAMZxmmkAGnYoNADcZ6UYpT1opgJijFLS4pAMxx70AZFPA5op3AaBRTsetJgDpSAbQeaU+9IeOgpgM5oIOKDntRnI5qgG4pSM0pHpSUxBigDDDJpc5HNKoGaQxD96nYwCaQjqaUEZ4pAIBhace2aVuoDfhilx+NIBpHpRj1p4xnpQycilcBhwOlPXleaaVAbg04ChgTImRTyPmpIm+Xmnggg1k73GV5SXOMd6WX5YwtKo3S+1Nm+ZufpVdbANhO1zk9akmj3rVU8NUiXGeHOatp3uhDHi2tk5xTfpV5WVx0zUcsA6rxSU+jCxAKWjG3qabmqAcBinbqTdxRzmkA6g9Kbn3pSecUgEaozxzT3PtTOD3q0A0/NSHA5pxHoaQj1qgGn2pM4NOPSm4zTEBpp560p4pu7BximA8Yo60DBHNJ34oAXZk+1PHBpATSjipAkB4ppJ7UZpuTnjpSGKcGozwad0pGJ29qaERyOShqu4yuSc1Z4KHNQFQUq0BVPfFOizz64ppGCcU2NmEgyaYiKT7xBHNRFcmrM6/vs+tQc5OKAIHG3kVFG7JKCuBg81YK4HzCq8gGc4pgak0nl2UrxOVBXOa82uWZbmTLFmJOSa7CWSXytnmNt9M1y2tRCOTzBxu60AZ7MN386az5PFRM+BjPFM3kfSmA525qtIalkIxkE1WkJ280AQyH0NJESTTHyFOKWIZbg/WgDGuv+Pyb/AK6N/Oii5/4+5v8Aro386K8Kp8bNVsb8OfOvMf8AP5P/AOjGq7G3pVCEn7Ref9fk/wD6MariMBivXpfAvQze5bVuRgVOCO9VULE8VMr+taCJt2MYp6lcZDVAGp4cA+1AEoIxxWpYQHaHIrMtYjcS+wroLcKqhRxigCeMBetToT24qNQT34p64BxigCwgJXHSrMWMAA9O9Vo+AfWp4uV9DQBchPPTj1q1H14FVYThKuRfdzQBIDzU6VCvJqdBiokNEy8in4NRq3pT81ixigZ60MMUUe9IBneloopgFIaWjFADaWlxRimA0UoPrRRx3oAMij6UuAaKQCEntSHinU00wEOTSDPejNGaYCY3fWkwc4xR7ilUE80wD7tMbrwacQTSCmgDpTlz6Ug60/NJgN5INKmM/wBaYzcHFKgOKOgEg5zkfSj9KB94U7FSAi9aMHNOAIpMZpANIw3IpwFAXNBBFAD0bY3PSpipeMkVWIJp6SsnB6VLXVDHxqUhO7g1Fgs1WN6laEA60rgVpIjngcVXKbWNaO4U2SJHIIFVGdtwsZyyuj5WpmmaQjNJNbYbKiq53KeRWtoy1QiyTu4xj3pCuDgnj1piSEipVPrS2AAOcDpS5GeaXGelIV4qQEJB6EUueaYOD0pSM0wFJz1pmRnpQfpSZx2pgDUnX60Hmg0wEOaTHBpaKYhvb1puTTyuelJgjjrTAbS/SnHtmm8dqYAGpwPrUamn7qGA/jFJmkQ4BzSHJFTYB2T+FNA55NAOKMZ60wGkdajxjipeKiIw1UgKzrhjVckg596tyr8xNQPgD1piHyoHhDLzVPow5q3DINpVvTiq8i7W56UAQPliagkBxU8jHPBqDPy8mmBVlBYYzWNq9t5lqRj7tbT9cms+5UsrA9DQBxW75ivpSFs8Ut6hgu3AGOahL/LkdaYDnb5eKryyFV6UOzHkGq8svy89aABpSRSQuzSYXkniqzvlwOpPpW9olgVzc3AAVex70Ac1dQFLyZT1Ejfzop2pXAk1S5ZehlbH50V4VT42arY1oj++vfX7ZP8A+jGqxGflB71UjP8ApN6P+nyf/wBGNVlD0r16XwL0M3uW4zk81KGBqurYpVcdq0EWQwHJNGc9DUG4/hUtoPOulU+tAG9pcJWHcRya1YkIHSorePYgHbFWkPNADwCR0xUkYGeTQDmnKBmgCRRtbINTr0qBVLVOq5XFAFyFcx9auQ/cqtEoCD0qwq7VGOhNAEy9qmU1ChAbmplqJDJUH5U+o1anA1kxj80maKKQB16UUdBRQAUUEYoxQAUUuKSgApNtOFBoASg0UUANJpP4aUmkJFUAg560wn5sCjvxSAc1QBTx0pAtOXvSYCAc80hA3cU4d6UANRcAWlPGTSgYpjHmpAaoznHWpguFxUYA3ZxxTuvIpsB2MU4AUioT0pSpHUVAD8DFJtxyKcqEJk9+lNZTtFSMB0pD6UzkGnck1QCUjNnrTiopGQCjQQ0NgcUbz60jZpgXmqsBKGyOTR5xXjqKibg0AZU9qLIC0riTABqvdKAdoqPJB4ODSnL9etNRs7gQYMbcc1Ov3cio2XBzSqxHtVvUCUNTs4pi0pORUAOznrRtzSbgBgUAmgBrKRTQfXpUm7PWmkA9OlNARk85FB54pxTH0pCKoQDikFHU0uaAA0Y7mnAZFIw9KVxjM9c0mOOBUhHFNJxTER/SjPrS4+bNIaoB+4EUh5PSm96cKQCHgZFIWyvNO4qM0wHKR0pHX5higNz0pz4NHUCvKOeKruMHirLjnrVdwPWqEVmJRs9anUrcx7f4qhcZzgVEjmJsqeaYDHjZGIIx71A546VpBluUw5ANUrm2deQMigClKMrzWZcyYAxWjICQQeKoXGn3U0DSwxMyDvTA5PXFxMHUdetZRbCmtfVyTCVIO8HGMc1iyWlyi5kjZcjIyOtAETyAHk1XYtK2EBJ6DAq3Bpk9ywypC561o7bTSoyTh5RQBFY6ZHbRCe76nnBqO+1YzYgh+VBxxWdqGqS3MnykhTxVaEgN75oArTf8fEuefnb+dFNlbM8n++386K8Kp8bNVsbaH/TL3/r8n/8ARjVbj68VVQf6Ve+v2yf/ANGNU65GMV69L4F6Gb3Jy2DzS7vSo9xpucVoInMg4rV0OPfMZCOKwi3P6V1ui24is0z1bmgDXQEgYqyi1EowowasIcDJoAeBxSjGaTJK04LkZoAepwcjpVpDkZFQpgrgirCLheKALi5EYNTj7ozUCklFAqwp+TmgCRcdakU1Gp6U8YzUsZKOlP8A4ajFSDpzWbGOzS0w9ad9KkBetKBSCnAYpAA96KMUopANxRilbNJyaADFBopDQAUhpTxSE0wENR9aVjScVaAMc0MPSnDmii4AoxQRQuQelOPJpAIoApeBRilxSAQ0wrzTySKM7hQAKuanhi3n2qJRzwatJhV4qJMY5VGfpSllztNHYEVGB+8LGstxkm5S2KJEBGRVdm+bI9anDblH607WEQMozShQBUpiDdKa0JA4quZAR4GaUgYpQuOtGQWxQBFn1FRscHpUrioW9q0QhjHNJnA4p1IVqwAcrxS/Smg4pcelACbTUTbhU3PakYZFNMBqP69amGM1XCkNUqknrRJASFetIvHWnUhFQA0/MDUeDipSvFNK4WqTAci5XmlaE7eKFGFGOakRjnA6VLbArFSOtJjb71JLkPx0pFwR1qr6ANFO2gnNOC4FKAO/FK4DD6UwrUpximn2ppgREccUmPl5pzChQTx3q7iGqKO5p23saNuM0XAjzTHwBxUhXFREE5qkAiN0qU81EoYN0qUgg0ARzDDfhVbFWpRuHPFVjwtNbCIWJ54qu67hVo5IqFsCmBTdipyvFSx35C4cZxTJlIPXiqknGaANILa3Mq7sAntW1BEkcCxxY2L1FcY0hjcMp5FdRaK0lvHcW7ktxuX1qZbDRVn0fR59TLNEhuO61UuI9L1G4l0x41WaMfISKde3MMPiqIfddl+b9aw/E0p0rxDa3yjCkjcRQloByOvXkllJLbRIFMbbeBXJXMsksh8xs16d410UXduur2C74pFBkA7e9eY3cbK2VBIq07iKuOamgHzc1VLkNgVPC+Oc0wIJf9fJ/vt/OikkOZpD/tt/OivBqfG/U1WxuKQLy9/6/J//AEY1WBg5waq8fbL3/r8n/wDRjVIhxXr0vgXoZvcn/hprNim789aYx960ESwnzLhE9Wru7NVWJVHYYri9GjEuoLkZxzXcW+A444oAuIvy8VYU5wMYqKP3GKl2DruoAkB+bAqVQCvPWoQMYqaPvmgCWMDvVgcDiokAwDU6Zx04oAswj5QcVOCDH071FCMD2qdR8ppAC8CnA03mnocdRmgZKh4qTqMCowc9sCpUFZMYtOFJilFSAuKcB60gpwFSxhj0oxSgUhHNIAIpCMdKCOaRjQIDTRSnlqCKYATjpUT+1ObNMFWgEFLigDLU4CmAg4p3FKFo4FSAg4Jo255NLil6UAJjFOJ44pM0UgG5z1o4BpccUuMUwHRgBuasbfl4quo5BqyOgIrKQxdpVRTU5VqezYXimj/VmoGQN8ppVLDp0qRYw3bNKYm6iruhCLOQcHpU6yIwqsYccmjZjBzUtJgSSY3cUwLg5pcd6bzupoYrdOlQsvXipXJppfHaqQivtwDmkIwKmypHIphGRWlxENOAPal8s5p3Q1VwG4x1pCSOlPY0080gGHmjoPen4oNO4ArZp2aj5zTgQaTQEmKQqCOaAaWkA1VKnrTj7cUUhpAI5yvNRYwakbGOaZ0birQEitS5BNR5wKTPelYCQrTGpyvnrTiARRsBCabz24qUpTNvzVVwAD1NBx2pwSjbSuBEw71Hj0qZuKjq0IWJTtJNWPKV8etQxnHBqdGw9RK4yrMMEj0qkBnOa0Jj+8PvVF/lY1rHYRC3y/SoyA1SydKi4qxEUiFh2rOlQhjnpWmTVaaEsuRQBkyfK2cHFaugasLec28pwr/dNZ0ytzmqMpKEEcMOQfShq6A3fFMKR67p9wBgu3zGsj4iOq2ttIOfm4rbg8jxJpiQXEmy6gIIbPWn6vHpC20UeumNljwB5hzn8BUR00Gcj4O8UCKT7BqAV7SY7VZuQp9Kh8d+B47KBtS0osYWbLIOig11upeEdJ1XRVbRkSJ1G6Mx9Ca46/8AFWoQ6WukXq/voXIc7SNy+lC1d0B51c2pVCMYOeKqJmKQDPOeldfcabHPA1zbHdH1weqE9q5e6tmjkLEZwetaCKxbLvx/G386KYr7tx/2m/maK8Gp8b9TVbG/gfbL7/r8n/8ARjVJngdqiPN5fA/8/k//AKManA54GeK9el8C9DN7jz9aRvuimk012xgVoI2vDaBrh2PbpXZ26luelcn4ZUeWz8cmutgGV5yKALi+mc1IF7EVGnHSpUbn5qAHrkcVMme9RoPmOanRPTmgCRFGeKnQnOAOKjjzuORwKmXJYbeB70AXIx8g5FPHFMhBA7U/PPNAAp3E9acvBqNQQ1TAZHAOaQyVAfwqVKjRcdalB9qykMcKdt4zSA08fdrNjEFPFG3AoxUgB4pp9qdt9TSEY+lACDpTSR3p2cDpTTTEISO1NY0pOM8U3k1QCE0mM9Kdil+lUAwe9OUUoHrS0gExmloFFIAoHWlPFFIBCMUUUUwFFKRmgDilApAC+9SRthuelRdelLk7qTVxlvaH5FKVUDpUCMdtP3Z6k1lZjJRx06UnmKDUbHB65pPc0WAmyrNzTSARjFRZ54pyvg80WsAbDimnipTICtRNTVwEbkVAW5xU5+7URAzVoRGfajJpzD0ppPNWIQMR1pwINMI45pwpgOwDTWTvSg0ZpAMOT2xQSfSn7waTincCE5PHSkB5xUzKMe9M2d6pMByNmn0zlaNxzg0gHdelG7Io6UH2pAIy7hTMFaec4pN3Y800BGTSfjTitN281QDl4HNOVhim4OMU5SKTAeGBpWUEjFMWnKNpz1qQFC47Uxs7unFT7sjpTGj3DrST7jIG5FRleamK44pmc1omIYcg8VKjAdaZzmgdaHqA6Ubm4H41RuExJV1iSMGoZkBINVHQTKL/AHeKgJ+X3qzJEy/SqjE7/atRCD86dkAHvSZ20ZB60wKN3b4+bs1ZFyvJFdFKEZMP17VjXUYOc9aAObu2mt5N9tK8bjoynBrkNR1O5uNQb7ZPJKw6F2zXb3iKYzt69815zr6tHqLFaAOk0jxRqWlGM210/lI2THu4x9K7LXYLTxjoH9s6WM3kY/fIv8WOa8WTUJR94bRjFdD4P8XyaBrCSbybSQ7ZouxBpOPVDJrTUHtZRvY4JG4Z4x6VcvtKjvrZrixzIh5ZV6p+HpVfxJLp8viS6fSZkmt5v3mV6KTzgVQXVZtNZfJZvqP8O9UIxXgaKWRPlO12GQfc0VJPqMl1cyzufmd2J+QDv6CivAqfG/U1Wxqn/j9vv+vyf/0Y1SjioST/AGhfY/5/J/8A0Y1SjpjvXr0vgXoZvcUmoZGqRgRVd85OelaCOp8MkfZOnU9a6yDKpyc1y3hpcWYKnjPeuqh+YDGKALUeOpqfPTAzUSD2qwgoAkT7wIHNTJhW9c9cVEmN5+lSoSPu96AJlYMCI8ge9TRjLLz0FRJ61NAf3h+lAFlDzxxTmP70Co1PBp6rlhxQA8DDVMn400KM1KuRUtjHqKlXFRqCeTUiisWMcAD0qQjC00LmnYrNjA9BQDRRnk8UAIab170pPFNzTQCnI6Gmk4oLU3HrVJCDNDL0OeaMc0vWgBKBS0hoAWlA9aTtS896ADA7UAZp2B2pMUgDGKSjnvS4HegBKXFGMUp60AA4HFKM0Z5oFIYKOtKBzSU4D5qQAMjOKcSTS7KXBpXATPag5PFKFy1OIOakZGBilzTtvvSYz0p3ASkanY5oK0AMyT3ppHtTwooIxTEREVE33qmbFRkZq0IQ0gA/KlIzRiqACcUCgdKWkAm0CmGpG5WmAZpoAXnrTs44po4ozTAdwe9GBTRS9eKQAM03mlJ5wKQnacGmAuTjmkAx0pcqeKcFA6UXAYeaTaacRjpTRnNACsORzQoOaPpThmgAC4pwOaQ5oGfpSACT2NKCR1o70p5NIBDgioioGeKkJ54pu7HUU0BHS4GeKkUhuMU0ptOaq4DQMk0x1+XNSDG45of7lPqBWdQwwazrqIxHjpWnjAqOSMOuCM1qmIyC1IW6VLd2xiG9ORVQyggHpViHv15qldrvQ7eoqw8nGc1Vll5xTA5u9cLIwavOvEzGPUwQeDXo+uRYQyL29K8x8VzHz43/AAoAyp8lSRVJbkxMB0561ciuvOwGCqMY6VSmXynbv70AadveKkg2sD/KmX2r7UPO6TOFArCmutowh/CrejWD3d2s05xHH8zE9hQBo2gd7cM4wxLEj8TRU5uA7u0HEZdtoPpk0V4NT436mq2Nl3Ivr7H/AD+T/wDoxqlEmRmq0j41C/H/AE+T/wDoxqli6fWvXpfw16Gb3JWfCjNVHfr35qw5+U5qm74zj1rQR23hoq2nrn1rpYW2DiuV8MSA6cPrXTwnjFAF6NyCKuxZPNVIF4GatRkZIzikA/d8/FTRmogdpxjrUiH5uaALC+1WLZcs30qEHaBirEORu+lAEmPkqwinApqp8oqXGAMUhj9lOAPakXkc05RzWbYxygjrUydeaaop4FZtgSACjFA6UCsxgaYc59qeelRsaaACaYaM0hqxAM96UUH3o6UAJnFKDRRigBTSAc05RmlIxSATHNKRQM0tIBAKWjFAOKBiEZox60+kJoASnYGKaDTqQDf4qUCjqcUoO080AApw9aCM0oFIYq5NLz2owe1OP3akBgJzStnGaNpFIRxQAufl6UlJS4NABg5o6ingUbcUXAg5zR9akZcU0jiquIjYUzvUhqM/eq0AYoxS0UCGiilxnpQBjrTAMYpuPSn9aQUAMxRinkU007gNpSflpcUYx1oAaRgZHWmsvGc08gdaYynGaaAZzT1kPekAzzSHrVAShs0vBqHnt0o381PKBPt4zQVIGaiDH1qRZMDmlZgKOnNIQeopwYDrTsBhSuBHSjvTyoIwKQptpXGR9O9I3PWnke1J0qhDOg460oYr1oOCcim/eHNMCQYJpHX5eKYo+apN+AcilswKxGMd6jc4Oc1cG1/aopYT25rRS11AqHawKsPlNZl7abAWiGV/lWlIpT71REgghuhrZEnPPN8pAxmqc8oC5q/qtmYGMsXK1gz3B6t0qgI7xxLEVPQ15Z43j+zTKPfPNeg3l0qsTkY964Px1NFcWiOpBdTTA5CO4CrknvUN1dtI2F6dzVUszcVrado5kAmuPljxnJ4oApWWnTX06pGOp5J7D1ravb2HSrYWtud7gfMfU0k+sQ2ML2+nKORhpe5rnpJGmkLuckmgDesCXs0Y9SW/maKXT/8AjxTjuf5mivAqfG/U1WxvTN/xM9Q/6/Jv/RjVKrnAxVa5cf2pqA7/AGyb/wBGNQjMBXr0v4a9DN7lmV/l5qpI2BUsgO0HNU5mxzmtBHbeFZN1hj3rrLYgVxHhOUGzb612Nmc96ANeA/LVtDVCJsLwKuoehNAFhdo61KmCeKhUhhU0fy470gLCnBHfmrdoM789DUEMRY5xWjbxBUIpMBxQKopwGcUrDihOelT0GP6Cnj7vFJ1OMU4Dnis2McoNSCmgY604nArNjHgigkZqPfSZ+aiwCluaTGaQnNLTEJjGaTtThSYoATFFOxQBmgBAKKdijn0oAMUtKBS4xUjG0oxuo70DrQAGlAG2lI9aXtigBMZ6UmPWnYxQTSGMIoxg0480mMUxCDrmnYzQKcOKQAKcBmkFPBqWMFXmnFaM8UbqkBp9KNvy9aTq1J360wFC0Y/CkBIozzzQA4D3p2KbgjkUvNIBCAaYQBTgOTmkYU0BGRxxUZWpu1MK81aYhu0YpCKfjimY5qgExQc048UYzQIbijinYpCKADimkU76UnNMBg460dacRSU7gJ2prcrTyM00jimAirxSFeafjpSEU7gRYyaMYp/H0prD0NO4EbEq2RTlkz1oYZX3pqiq0aAnDAmnA8cVEopc44qLASF8UebUJY0gcdKfKBZD8c0ZVqr+YOhNKGz0NLlAsmMbeKZ5JOcUxWbHPFJ5xANKzAcEIIpSpJ4pscp3DPNSiZQxpO4FZlcdBSLIydala5jGckVFJfQKPmxWiu+gBlJPvCq1xZkqSnP0pG1S3XtTf7bgUVaUlsLQzbhJQhSRCR9K4fXI5baVjBGzKfavSBqdlcnBK++ahnis5lZVWNs/nWqk+qFY8SvLPVb/ADtQqntVWbwWZbNmupcHH8RxivQNfF1ZRuIIMA/dIFeeardap5gMkb4/SqEctJY6bpTs0jCV16DsazNR1eS5BjQ7Y+yjimaqsq3j+aD8xyM9qz2y2DjpQMRmJbNIvUH3pyQu5+VCa09P8P3l9MsaxkFugxzQBoWH/Hkn1P8AM0VfXTGss28nDRswOfqaK8Cp8b9TVbDb2QDWdR/6/Jv/AEY1AnyorI1LVAuvalsw8bXczKwPBBc80sOowOvLbTXr0v4a9DN7mw8u4Cq1w3HBquLkN9xgfqagnmYj5a0Edl4Rl/0dhnvXcWkoGOa858H3GUkB4967e0lGQxz9aAOnhlBwBVxGrGt5htBBxWnFJlOuKALy4C8VcgQuy4rOgO5gAc81vWMGxeaQFmCEqvNW4wBUQboKlUHceKiQxWHrTV4binO2EySOKhEy7gMc0lewy0n3qfUAc54FPUH1/CoaAkLjHFAJIpm008cCloAAZp+CD0pBz2p2OKkBMc0Ac0d6cD8xpDExzRjnmj0zS4GaAGgDNKBQBSgc0AKBQOKXFGKkBB1pcUuKQigYY+alVcUvUUUXEJ1paKKQwpCKXFGKAExRS4paLgIBS8UY70oGaQABTh1oxSr3pAKaMClA4FKVqQI2HpSbOKlAwKVVz1ouBDt7UoUDrUjAdutNx0zTuADP4UmCKdxQw5oAZ1pG9KfgA9Kay5NNARkUnSntxTaoBhHekpxGKTiqEIeO9Nx70/HFJigBMUlLjJzS4pgIBSd6dim45oAG5pu31p7D0ppB700Awj0pccU7FKF4ouIYaTFPxikp3AbsB5NNIHWpcZ5pjUJgREZoC5qRRmgr8tVcAVcLTXHzVKF/dmoz0pJ6gRkUwr3qbHFMIP4VaYEW3NKmehp+zNKEP0p3EMO7oTUUgYd6tYHemSJnpQmMgj3hhlql6saai/MKnCdOKcmIzZUO881Wkiyp3citF0OTx3qFkB6itkxGPLAQelVnhJJz2rXmXmqrJ1zVCMZrY5OMjPcVA1zcQcBiR61svFtjOO9Z00PXigCCXWmMIWZTj1Fc9qWtaeZtktsjH1YVrXEWw56juKxNZ0iK7gM0Iww7UAch4t/siZEl+4cdErn7CbRIpBu3nPqRVjXdNzbSHBDR8iuNbhqBnf3F9oNuge3himbuHFQnxglsdtom1COVQYxXDZNWrUMTkjIPFAHTJqMl2DO3WRmPP1NFRWqBbdRjoT/M0V4FT436mq2MHXgE8S6mqgKovJgABwPnNUK09fikbxNqhVGI+2zcgf7ZrP8AJl/55v8A98mvWpNci9DN7jQ7DoxH41ILiTGN5pvky/8APN/++TR5Mp/5Zv8A98mtLoRv+HdVnsyxWMSL3ya6u18WyIwD267fTNcpo8Jjt9zDGexFaSr0OBii6A7G38VwtzJEwPtWva+J4XwDHIc9OK4WBAWX5gPxrqtDs4nmDSzxhV9ZAKLoDuNHv4X+Z4ZAeo4rpYdQU42xtXPWMlhGu43UIwMAeata0N5Y8ZvLcf8AbZf8aV0BoreSOTiPBHrU8U0jsQeOOlZw1CxBOL62H/bZf8atR6np2R/p1tnH/PZf8aV0MsOp8k+9JDHnGRiozqmnlSPt1r/3+X/Glj1TTguDfWv/AH+X/GlzIC+o6HNSAZ7fjVNNU07bzfWv/f5f8aeNV04H/kIWuP8Arsv+NZOSGW+aeOlVRqumnn+0LX/v+v8AjS/2rpv/AEELX/v+v+NRzIZbXIpSTnpVT+1tMH/MRtf+/wCv+NL/AGvppA/4mFr/AN/1/wAam6AtU5euapnVdN/6CNr/AN/1/wAaUavpn/QQtP8Av+v+NF0Bcxmm7fmqt/a+m/8AQQtP+/6/40f2vpv/AEELX/v+v+NK6GWsUoFVP7X03/oIWv8A3/X/ABo/tfTf+gha/wDf9f8AGi6AuYoqp/a+mf8AQQtf+/6/40f2vpn/AEEbT/v+v+NK6AuECjFVDqum/wDQRtP+/wCv+NH9r6Z/0EbT/v8Ar/jSugLQ4pc+1VBq2mf9BG0/7/r/AI0DV9Nz/wAhG0/7/r/jRdAXMUY4qodX0zHOo2v/AH/X/Gj+19Lx/wAhG1/7/r/jRdAW6Qg9+Kq/2vpp6ajaf9/1/wAaP7X03POo2n/f9f8AGi6AtEYox7VVOsaYOf7RtP8Av+v+NKdZ0wrj+0LT/v8Ar/jRcC2ACKUDFUxq2mY/5CNp/wB/1/xp39saZ/0EbT/v+v8AjSugLlAFVP7X0z/oI2n/AH/X/Gj+2NM/6CNp/wB/1/xpXQF4Dil6VTXWNM/6CNp/3/X/ABpTrGmf9BG0/wC/6/41NxlofezTu1UhrGmZ/wCQjaf9/wBf8aX+19Mwf+Jlaf8Af9f8aV0Bao+tVP7Y0z/oI2n/AH/X/GkOsaX/ANBG0/7/AK/407oRc+lLtzVNdX0s9dRtP+/6/wCNK2saWOmpWn/f9f8AGi4FljSEZOKqHWNM/wCgjaf9/wBf8aRtY03PGo2n/f8AX/GndAWCMHmmnFV/7Y0w9dRtP+/6/wCNN/tfTf8AoIWn/f8AX/GqugLJFJ+FVzrGm/8AQQtP+/6/40z+2NM7ajaf9/1/xqk0BbHNJjjFVhq2m/8AQQtP+/6/40v9raZ/0EbT/v8Ar/jRzICxjjFLt45qqdX00f8AMQtf+/6/4006vph/5iNr/wB/1/xp3AtUhqt/a+mf9BC1/wC/6/40n9rab/0ELX/v+v8AjTuhFon0pCePWqjatpoORqFr/wB/l/xpBq2mk/8AIQtf+/y/407oC0eelOGcc1SXVtNDH/iYWv8A3/X/ABpw1bTh11C1/wC/6/40XQFoc96CAOtVv7X0wLxf2uf+uy/40h1bTSOb+1/7/r/jRzIC0fSk2etVv7W03I/4mFr/AN/1/wAaQ6tpx4GoWv8A3+X/ABougLBXHSkC9qr/ANqaaOTqFqf+26/40f2vpva/tf8Av8v+NVzIC3yBimEdqrnV9O/6CFr/AN/l/wAab/a2m/8AQQtR/wBt1/xoTQFrbSEdqrf2tpw/5iFr/wB/l/xpf7V04/8AMQtf+/y/40cyAmxilI96r/2rp2f+P+1/7/L/AI0v9p6d/wBBC0/7/r/jT5kBOq5pCvzVCNV07/oIWv8A3+X/ABpv9q6dn/kIWv8A3+X/ABo5kBMo+c1Koy2KqDU9Nyf+Jha/9/l/xoTVtO3f8f8Aa/8Af5f8aHJCHMCGP1qN0psmqadu4v7X/v8AL/jUbanp5H/H/a/9/l/xrVTQWI5lzUBXIwBUsmoacf8Al+tf+/y/41XbULAdL62/7/L/AI1tGce5NiKZKpSpjnFW21CwPW9tv+/q/wCNV5bzT2Bxe2//AH9X/Gq5kIyr1AenFZkuY0we9as1zZMcC7gP/bUVQvZbMYH2qA8dpBRzIDiNf0/bI0q/NG4wRXlN9C0F5JGwwVNe33kltOjRmWIq3T5xwa8p8V2Bh1RjH8+4ZJXmjmQznx1Ga2rK1VtPMgkjBVwNpPNZCQyFh8j9fStvSUKeakqDGMgMP0o5kBfh/wBX0/iP8zRT0U4OORub+Zorwanxv1NVsfU3gDwpoF/4VFxe6NZTzPd3G6SSBWY/vm74rpf+EI8L/wDQv6d/4DL/AIUUVyFB/wAIR4X/AOhf07/wGX/Cj/hCPC//AEL+nf8AgMv+FFFAB/whHhf/AKF/Tv8AwGX/AAo/4Qjwv/0L+nf+Ay/4UUUAH/CEeF/+hf07/wABl/wo/wCEI8L/APQv6d/4DL/hRRQAf8IR4X/6F/Tv/AZf8KP+EI8L/wDQv6d/4DL/AIUUUAH/AAhHhf8A6F/Tv/AZf8KP+EI8L/8AQv6d/wCAy/4UUUAH/CEeF/8AoX9O/wDAZf8ACj/hCPC//Qv6d/4DL/hRRQAf8IR4X/6F/Tv/AAGX/Cj/AIQjwv8A9C/p3/gMv+FFFAB/whHhf/oX9O/8Bl/wo/4Qjwv/ANC/p3/gMv8AhRRQBzFx4R8PD4p2VsNFsRA2lSuY/IXaWEijOMda6f8A4Qjwv/0L+nf+Ay/4UUUAH/CEeF/+hf07/wABl/wo/wCEI8L/APQv6d/4DL/hRRQAf8IR4X/6F/Tv/AZf8KP+EI8L/wDQv6d/4DL/AIUUUAH/AAhHhf8A6F/Tv/AZf8KP+EI8L/8AQv6d/wCAy/4UUUAH/CEeF/8AoX9O/wDAZf8ACj/hCPC//Qv6d/4DL/hRRQAf8IR4X/6F/Tv/AAGX/Cj/AIQjwv8A9C/p3/gMv+FFFAB/whHhf/oX9O/8Bl/wo/4Qjwv/ANC/p3/gMv8AhRRQAf8ACEeF/wDoX9O/8Bl/wo/4Qjwv/wBC/p3/AIDL/hRRQAf8IR4X/wChf07/AMBl/wAKP+EI8L/9C/p3/gMv+FFFAHO+P/CHh208A6tPbaJYxSpDlXSBQVO4dDit638E+GDaxE6BpxJQZP2ZfT6UUUASf8IR4X/6F/Tv/AZf8KP+EI8L/wDQv6d/4DL/AIUUUAH/AAhHhf8A6F/Tv/AZf8KP+EI8L/8AQv6d/wCAy/4UUUAH/CEeF/8AoX9O/wDAZf8ACj/hCPC//Qv6d/4DL/hRRQAf8IR4X/6F/Tv/AAGX/Cj/AIQjwv8A9C/p3/gMv+FFFAB/whHhf/oX9O/8Bl/wo/4Qjwv/ANC/p3/gMv8AhRRQAf8ACEeF/wDoX9O/8Bl/wo/4Qjwv/wBC/p3/AIDL/hRRQAf8IR4X/wChf07/AMBl/wAKP+EI8L/9C/p3/gMv+FFFAB/whHhf/oX9O/8AAZf8KP8AhCPC/wD0L+nf+Ay/4UUUAH/CEeF/+hf07/wGX/CuY8IeEfD1xqXiVZ9FsZFh1V0jDQKdi7F4HHAoooA6f/hCPC//AEL+nf8AgMv+FH/CEeF/+hf07/wGX/CiigA/4Qjwv/0L+nf+Ay/4Uf8ACEeF/wDoX9O/8Bl/woooAP8AhCPC/wD0L+nf+Ay/4Uf8IR4X/wChf07/AMBl/wAKKKAD/hCPC/8A0L+nf+Ay/wCFH/CEeF/+hf07/wABl/woooAP+EI8L/8AQv6d/wCAy/4Uf8IR4X/6F/Tv/AZf8KKKAD/hCPC//Qv6d/4DL/hR/wAIR4X/AOhf07/wGX/CiigA/wCEI8L/APQv6d/4DL/hR/whHhf/AKF/Tv8AwGX/AAoooAP+EI8L/wDQv6d/4DL/AIUf8IR4X/6F/Tv/AAGX/CiigA/4Qjwv/wBC/p3/AIDL/hXMfEHwj4etPCLy2ui2MMn2m3G5IFBwZVBHT0oooA6f/hCPC/8A0L+nf+Ay/wCFH/CEeF/+hf07/wABl/woooAP+EI8L/8AQv6d/wCAy/4Uf8IR4X/6F/Tv/AZf8KKKAD/hCPC//Qv6d/4DL/hR/wAIR4X/AOhf07/wGX/CiigA/wCEI8L/APQv6d/4DL/hR/whHhf/AKF/Tv8AwGX/AAoooAP+EI8L/wDQv6d/4DL/AIUf8IR4X/6F/Tv/AAGX/CiigA/4Qjwv/wBC/p3/AIDL/hR/whHhf/oX9O/8Bl/woooAP+EI8L/9C/p3/gMv+FH/AAhHhf8A6F/Tv/AZf8KKKAD/AIQjwv8A9C/p3/gMv+FH/CEeF/8AoX9O/wDAZf8ACiigA/4Qjwv/ANC/p3/gMv8AhR/whHhf/oX9O/8AAZf8KKKAOY8C+EfD11puptc6LYysmq3KKXgU4UPwOnQV0/8AwhHhf/oX9O/8Bl/woooAUeCfDAGBoGnAf9e6/wCFFFFAH//Z"}}},{"cell_type":"markdown","source":"# Imports","metadata":{}},{"cell_type":"code","source":"import numpy as np\nfrom tensorflow.keras.models import load_model\nfrom tensorflow.keras.preprocessing.image import load_img, img_to_array\nfrom skimage import exposure\nimport matplotlib.pyplot as plt\nimport matplotlib\nmatplotlib.rcParams.update({'font.size': 16})\nimport pandas as pd \nfrom matplotlib.patches import Rectangle\nfrom ast import literal_eval\nimport tensorflow as tf\nimport cv2\nfrom tensorflow.keras.models import Model\nfrom tensorflow.keras.layers import Input, GlobalAveragePooling2D, Dense\nfrom keras.preprocessing.image import ImageDataGenerator\nimport warnings\nwarnings.filterwarnings(\"ignore\")\nimport os","metadata":{"execution":{"iopub.status.busy":"2021-06-01T20:13:06.065735Z","iopub.execute_input":"2021-06-01T20:13:06.066166Z","iopub.status.idle":"2021-06-01T20:13:06.075823Z","shell.execute_reply.started":"2021-06-01T20:13:06.066128Z","shell.execute_reply":"2021-06-01T20:13:06.074719Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Same processing as original notebooks","metadata":{}},{"cell_type":"code","source":"df_image = pd.read_csv('../input/siim-covid19-detection/train_image_level.csv')\ndf_study = pd.read_csv('../input/siim-covid19-detection/train_study_level.csv')\ndf_study['id'] = df_study['id'].str.replace('_study',\"\")\ndf_study.rename({'id': 'StudyInstanceUID'},axis=1, inplace=True)\ndf_train = df_image.merge(df_study, on='StudyInstanceUID')\ndf_train.loc[df_train['Negative for Pneumonia']==1, 'study_label'] = 'negative'\ndf_train.loc[df_train['Typical Appearance']==1, 'study_label'] = 'typical'\ndf_train.loc[df_train['Indeterminate Appearance']==1, 'study_label'] = 'indeterminate'\ndf_train.loc[df_train['Atypical Appearance']==1, 'study_label'] = 'atypical'\ndf_train.drop(['Negative for Pneumonia','Typical Appearance', 'Indeterminate Appearance', 'Atypical Appearance'], axis=1, inplace=True)\ndf_train['id'] = df_train['id'].str.replace('_image', '.jpg')\ndf_train['image_label'] = df_train['label'].str.split().apply(lambda x : x[0])\ndf_size = pd.read_csv('../input/covid-jpg-512/size.csv')\ndf_train = df_train.merge(df_size, on='id')\ndf_train.head(3)","metadata":{"execution":{"iopub.status.busy":"2021-06-01T20:13:08.380600Z","iopub.execute_input":"2021-06-01T20:13:08.381474Z","iopub.status.idle":"2021-06-01T20:13:08.613109Z","shell.execute_reply.started":"2021-06-01T20:13:08.381418Z","shell.execute_reply":"2021-06-01T20:13:08.611782Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def preprocess_image(img):\n    equ_img = exposure.equalize_hist(img)\n    return equ_img\n\nimg_size = 299\nbatch_size = 16\n\nimage_generator = ImageDataGenerator(\n        validation_split=0.2,\n        #rotation_range=20,\n        horizontal_flip = True,\n        zoom_range = 0.1,\n        #shear_range = 0.1,\n        brightness_range = [0.8, 1.1],\n        fill_mode='nearest',\n        preprocessing_function=preprocess_image\n)\n\nimage_generator_valid = ImageDataGenerator(validation_split=0.2,preprocessing_function=preprocess_image)\n\ntrain_generator1 = image_generator.flow_from_dataframe(\n        dataframe = df_train,\n        directory='../input/covid-jpg-512/train',\n        x_col = 'id',\n        y_col =  'image_label',  \n        target_size=(img_size, img_size),\n        batch_size=batch_size,\n        subset='training', seed = 23) \n\nvalid_generator1 = image_generator_valid.flow_from_dataframe(\n    dataframe = df_train,\n    directory='../input/covid-jpg-512/train',\n    x_col = 'id',\n    y_col = 'image_label',\n    target_size=(img_size, img_size),\n    batch_size=batch_size,\n    subset='validation', shuffle=False, seed=23) \n\ntrain_generator2 = image_generator.flow_from_dataframe(\n        dataframe = df_train,\n        directory='../input/covid-jpg-512/train',\n        x_col = 'id',\n        y_col =  'study_label',  \n        target_size=(img_size, img_size),\n        batch_size=batch_size,\n        subset='training', seed = 23) \n\nvalid_generator2 = image_generator_valid.flow_from_dataframe(\n    dataframe = df_train,\n    directory='../input/covid-jpg-512/train',\n    x_col = 'id',\n    y_col = 'study_label',\n    target_size=(img_size, img_size),\n    batch_size=batch_size,\n    subset='validation', shuffle=False, seed=23) \n","metadata":{"execution":{"iopub.status.busy":"2021-06-01T20:31:12.132112Z","iopub.execute_input":"2021-06-01T20:31:12.132463Z","iopub.status.idle":"2021-06-01T20:31:14.703156Z","shell.execute_reply.started":"2021-06-01T20:31:12.132432Z","shell.execute_reply":"2021-06-01T20:31:14.701975Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def load_process(img, img_size):\n    img = load_img(img, target_size = (img_size, img_size))\n    img = img_to_array(img)\n    img = img.reshape((1, img.shape[0], img.shape[1], img.shape[2]))\n    img = preprocess_image(img)\n    return img","metadata":{"execution":{"iopub.status.busy":"2021-06-01T20:13:19.630044Z","iopub.execute_input":"2021-06-01T20:13:19.630349Z","iopub.status.idle":"2021-06-01T20:13:19.636304Z","shell.execute_reply.started":"2021-06-01T20:13:19.630319Z","shell.execute_reply":"2021-06-01T20:13:19.635223Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Main Grad-CAM algorithm","metadata":{}},{"cell_type":"code","source":"def grad_cam(input_image, model, layer_name):\n\n    desired_layer = model.get_layer(layer_name)\n    grad_model = Model(model.inputs, [desired_layer.output, model.output])\n\n    with tf.GradientTape() as tape:\n        layer_output, preds = grad_model(input_image)\n        ix = (np.argsort(preds, axis=1)[:, -1]).item()\n        output_idx = preds[:, ix]\n\n    gradient = tape.gradient(output_idx, layer_output)\n    alpha_kc = np.mean(gradient, axis=(0,1,2))\n    L_gradCam = tf.nn.relu(np.dot(layer_output, alpha_kc)[0])\n    L_gradCam = (L_gradCam - np.min(L_gradCam)) / (np.max(L_gradCam) - np.min(L_gradCam)) \n    return L_gradCam.numpy()","metadata":{"execution":{"iopub.status.busy":"2021-06-01T20:13:26.889975Z","iopub.execute_input":"2021-06-01T20:13:26.890398Z","iopub.status.idle":"2021-06-01T20:13:26.898946Z","shell.execute_reply.started":"2021-06-01T20:13:26.890351Z","shell.execute_reply":"2021-06-01T20:13:26.897871Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Blend\n\nSuperimposing gradCAM heatmap on the original image","metadata":{}},{"cell_type":"code","source":"def blend(img_path, gradCam_img, alpha, colormap = cv2.COLORMAP_JET):\n    origin_img = img_to_array(load_img(img_path))\n    gradCam_resized = cv2.resize(gradCam_img, (origin_img.shape[1], origin_img.shape[0]), interpolation = cv2.INTER_LINEAR)\n    heatmap  = cv2.applyColorMap(np.uint8(gradCam_resized*255), colormap)\n    superimposed_image = cv2.cvtColor(origin_img.astype('uint8'), cv2.COLOR_RGB2BGR) + heatmap * alpha\n    return heatmap, superimposed_image","metadata":{"execution":{"iopub.status.busy":"2021-06-01T20:13:26.900640Z","iopub.execute_input":"2021-06-01T20:13:26.901077Z","iopub.status.idle":"2021-06-01T20:13:26.919517Z","shell.execute_reply.started":"2021-06-01T20:13:26.901032Z","shell.execute_reply":"2021-06-01T20:13:26.918572Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Loading model of notebook 1 (Opacity vs None)","metadata":{}},{"cell_type":"code","source":"trained_model = load_model('../input/image-label-classification-opacity-vs-none/model.h5')","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Results","metadata":{}},{"cell_type":"code","source":"def plot_results(model, gen):\n    n = 50\n    train_dir = '../input/covid-jpg-512/train'\n    fig, axs = plt.subplots(10, 5, figsize=(20,60))\n    fig.subplots_adjust(hspace=.5, wspace=.1)\n    axs = axs.ravel()\n    gen.next()\n    classes = list(gen.class_indices.keys()) \n    idx = gen.index_array\n    layer_name = 'block14_sepconv2_act'\n    for i in range(n):\n        sample_img_path = os.path.join(train_dir, df_train['id'][idx[i]])\n        img = load_process(sample_img_path, img_size)\n        pred = model.predict(img)\n        grad_cam_img = grad_cam(img, model, layer_name)\n        heatmap_img, result_img = blend(sample_img_path, grad_cam_img, 0.5)\n        axs[i].imshow(result_img[:,:,::-1]/255)\n        axs[i].set_xticklabels([])\n        axs[i].set_yticklabels([])\n        if type(df_train['boxes'][idx[i]])==str:\n            boxes = literal_eval(df_train['boxes'][idx[i]])\n            for box in boxes:\n                axs[i].add_patch(Rectangle((box['x']*(512/df_train['dim1'][idx[i]]), box['y']*(512/df_train['dim0'][idx[i]])), box['width']*(512/df_train['dim1'][idx[i]]), box['height']*(512/df_train['dim0'][idx[i]]), fill=0, color='y', linewidth=2))\n                axs[i].set_title(f\"{df_train['study_label'][idx[i]]}, {df_train['image_label'][idx[i]]}\")\n        else:\n            axs[i].set_title(df_train['study_label'][idx[i]])\n        \n        axs[i].set_xlabel(f\"{classes[np.argmax(pred)]}, {round(pred[0][np.argmax(pred)]*100, 2)}%\")","metadata":{"execution":{"iopub.status.busy":"2021-06-01T20:34:44.734868Z","iopub.execute_input":"2021-06-01T20:34:44.735242Z","iopub.status.idle":"2021-06-01T20:34:44.749608Z","shell.execute_reply.started":"2021-06-01T20:34:44.735210Z","shell.execute_reply":"2021-06-01T20:34:44.748385Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Training images","metadata":{}},{"cell_type":"code","source":"plot_results(trained_model, train_generator1)","metadata":{"execution":{"iopub.status.busy":"2021-06-01T20:34:46.970326Z","iopub.execute_input":"2021-06-01T20:34:46.970703Z","iopub.status.idle":"2021-06-01T20:36:03.230381Z","shell.execute_reply.started":"2021-06-01T20:34:46.970667Z","shell.execute_reply":"2021-06-01T20:36:03.229265Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Validation images","metadata":{}},{"cell_type":"code","source":"plot_results(trained_model, valid_generator1)","metadata":{"execution":{"iopub.status.busy":"2021-06-01T20:36:03.231917Z","iopub.execute_input":"2021-06-01T20:36:03.232236Z","iopub.status.idle":"2021-06-01T20:37:18.974720Z","shell.execute_reply.started":"2021-06-01T20:36:03.232205Z","shell.execute_reply":"2021-06-01T20:37:18.973586Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Loading model of notebook 2 ('Negative' 'Typical' 'Indeterminate' 'Atypical')","metadata":{}},{"cell_type":"code","source":"trained_model2 = load_model('../input/classification-model/model.h5')","metadata":{"execution":{"iopub.status.busy":"2021-06-01T20:37:55.813264Z","iopub.execute_input":"2021-06-01T20:37:55.813623Z","iopub.status.idle":"2021-06-01T20:38:01.498232Z","shell.execute_reply.started":"2021-06-01T20:37:55.813592Z","shell.execute_reply":"2021-06-01T20:38:01.497189Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Results","metadata":{}},{"cell_type":"markdown","source":"### Training images","metadata":{}},{"cell_type":"code","source":"plot_results(trained_model2, train_generator2)","metadata":{"execution":{"iopub.status.busy":"2021-06-01T20:38:01.500397Z","iopub.execute_input":"2021-06-01T20:38:01.500707Z","iopub.status.idle":"2021-06-01T20:39:19.367015Z","shell.execute_reply.started":"2021-06-01T20:38:01.500677Z","shell.execute_reply":"2021-06-01T20:39:19.365462Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Validation images","metadata":{}},{"cell_type":"code","source":"plot_results(trained_model2, valid_generator2)","metadata":{"execution":{"iopub.status.busy":"2021-06-01T20:41:20.192417Z","iopub.execute_input":"2021-06-01T20:41:20.192766Z","iopub.status.idle":"2021-06-01T20:42:36.977502Z","shell.execute_reply.started":"2021-06-01T20:41:20.192736Z","shell.execute_reply":"2021-06-01T20:42:36.975789Z"},"trusted":true},"execution_count":null,"outputs":[]}]}