├── .gitignore ├── server.py ├── README.md └── notebooks └── PMI-Custom_Model_Development.ipynb /.gitignore: -------------------------------------------------------------------------------- 1 | data 2 | .ipynb_checkpoints/ 3 | -------------------------------------------------------------------------------- /server.py: -------------------------------------------------------------------------------- 1 | from flask import Flask 2 | import requests 3 | import flask 4 | import pickle 5 | import pandas as pd 6 | import sys 7 | import os 8 | from dotenv import load_dotenv 9 | from joblib import load 10 | load_dotenv() 11 | 12 | # app = Flask(__name__) 13 | # @app.route('/') 14 | # cookies only needed if behind realm auth 15 | # get meter data 16 | # maximo_domain = sys.argv[1] # IP Address + PORT 17 | #max_token = sys.argv[2] 18 | asset_id = sys.argv[1] # "2112" 19 | realm_token = sys.argv[2] 20 | 21 | cookies = dict(LtpaToken2=realm_token) 22 | headers = {'Content-Type': 'application/json', 23 | "maxauth": os.getenv("max_token")} 24 | 25 | # Prereq 26 | # Create meter for each sensor value / feature associated with model 27 | q_endpoint = "/maximo/oslc/os/mxasset?oslc.select=assetmeter,expectedlife&oslc.where=assetnum=" + asset_id 28 | res = requests.get(os.getenv("maximo_domain") + q_endpoint, 29 | cookies=cookies, headers=headers) 30 | 31 | meter_data = {} 32 | print(res) 33 | meters = res.json()["rdfs:member"][0]["spi:assetmeter"] 34 | for m in meters: 35 | meter_data[m['spi:metername']] = m['spi:lastreading'] 36 | # convert meter values to pandas dataframe 37 | meter_df = pd.DataFrame.from_dict(meter_data, orient='index') 38 | 39 | ''' 40 | clf = load("lin_reg.joblib") 41 | clf.predict(meter_df) 42 | predictions = clf.predict(meter_df) 43 | print("RUL:" + predictions.join()) 44 | ''' 45 | 46 | threshold = 50 47 | predictions = [62] 48 | 49 | # update RUL in maximo DB 50 | requests.post(maximo_domain + "/maximo/oslc/os/mxasset/" + restid, 51 | data={"expectedlife": predictions[0]}, cookies=cookies, headers=headers) 52 | 53 | # create work order if RUL below threshold 54 | if predictions[0] < threshold: 55 | asset_num = str(2112) 56 | site_id = "DENVER" 57 | wo_endpoint = "/maxrest/rest/mbo/workorder?_lid=" + username + "&_lpwd=" + password + \ 58 | "&_compact=true&_format=json&description=FixTurboFan&siteid=" + \ 59 | site_id + "&assetnum=" + asset_num 60 | requests.put(maximo_domain + wo_endpoint, 61 | data={"expectedlife": predictions[0]}, cookies=cookies, headers=headers) 62 | 63 | # NOTES, remove below 64 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # IoT - Create Predictive Maintenance Models To Detect Equipment Breakdown Risks in Maximo 2 | 3 | ## Description 4 | Instrumented, connected assets generate volumes of operational data - structured and unstructured - that can be used to identify risks if the organizations have analytic tools to convey this insight to personnel responsible for asset operations. 5 | 6 | In this journey, we'll show how to build and apply custom machine learning models to identify risks and suggest proactive maintenance to avoid service disruption. These models will also estimate how long a mechanical asset can be used before needing maintenance/replacement. 7 | 8 | We'll also show how to import the custom models into a Maximo instance. Maximo is a system used for managing assets and workflow processes. This can be used to increase efficiency by automating processes such as work orders, notifications, anomaly detection, etc. 9 | 10 | 11 | ## Included Components 12 | * [Maximo](https://www.ibm.com/products/maximo) 13 | * [Maximo APM - Predictive Maintenance Insights](https://www.ibm.com/us-en/marketplace/predictive-maintenance-insights) 14 | * [Watson Studio](https://cloud.ibm.com/catalog/services/watson-studio) 15 | 16 | ## Application Workflow Diagram 17 | 18 | 19 | ## Flow 20 | 1. Build custom machine learning model in Watson Studio, and export custom model as python package 21 | 2. Publish sensor data from field assets to Maximo 22 | 3. Periodically pull asset data from Maximo to Python server, and generate predictions based on packaged custom model 23 | 4. Publish predicted "Remaining Useful Life" value to Maximo 24 | 5. (Within Maximo) Create "automation scripts" to execute when "Remaining Useful Life" is updated. If RUL below X number of days, create a work order to have asset replaced / updated. 25 | 26 | # Steps 27 | 1. [Deploy Cloud Services](#1-deploy-cloud-services) 28 | 2. [Create Custom Model](#2-create-custom-model) 29 | 3. [Register Assets in Maximo](#3-register-assets-in-maximo) 30 | 4. [Create Automation Scripts](#4-create-automation-scripts) 31 | 32 | ## Install Prerequisites: 33 | ### Python 34 | ``` 35 | # OS X 36 | brew install python python-pip 37 | 38 | # Linux 39 | apt-get install python python-pip 40 | 41 | # Install python packages 42 | pip install pickle scikit-learn flask 43 | ``` 44 | 45 | 46 | ### 1. Provision services via IBM Cloud and IBM Marketplace dashboards 47 | Create the following services 48 | * [Maximo](https://www.ibm.com/products/maximo) 49 | * [Maximo APM - Predictive Maintenance Insights](https://www.ibm.com/us-en/marketplace/predictive-maintenance-insights) 50 | * [Watson Studio](https://cloud.ibm.com/catalog/services/watson-studio) 51 | 52 | ### 2. Create Custom Model 53 | Navigate to your provisioned Watson Studio instance 54 | 55 | 56 | Click "Get Started" 57 | 58 | Create a project. Select "Standard" project, and give it a name. 59 | 60 | 61 | 62 | Select your created project and click "New Notebook" 63 | 64 | Now, you can create your own notebook for a custom use case, or use one of our predefined notebooks. Here, we'll upload our preconfigured "Advanced Regression" notebook, which can be seen in further detail in the [notebooks](notebooks/Advanced_Regression_Turbofan.ipynb) directory of this project. 65 | 66 | We'll do so by selecting "From File" in the menu, then entering a name, and the path to the local notebook file within this repo. 67 | 68 | 69 | In this notebook, we'll be exploring a public dataset supplied by NASA. This dataset was generated by collecting sensor data from Turbofan engines. Each Turbofan has 20 sensors, as well as a "Cycles" field that signals how long a Turbofan has been running. 70 | 71 | 74 | 75 | 76 | 77 | 78 | After going through each step in the notebook, we can then export our model for consumption by Maximo PMI. This can be done by following the steps in the notebook at [notebooks/PMI-Custom_Model_Development.ipynb](notebooks/PMI-Custom_Model_Development.ipynb). You'll need the credentials that were included in your Welcome letter for the Maximo PMI instance. 79 | 80 | Once this has been completed, we'll see the custom model available in the Maximo APM View. 81 | 82 | As an alternative, we can also export the model by leveraging the `pickle` python package. This can be done by adding the following command to the notebook 83 | 84 | ``` 85 | pickle.dump(model, open("model.joblib", 'wb')) 86 | ``` 87 | 88 | ### 3. Register Assets in Maximo 89 | Now that we have our custom model built, the next step is to register assets in Maximo. This can be done by navigating to your Maximo instance. 90 | 91 | 92 | In the "Find Navigation Item" field, enter "Asset Templates". 93 | 94 | 95 | 96 | Click "New Asset Template" in the left hand menu. 97 | 98 | 99 | Give the Asset Template an ID (We used "Turbofan" here). We can also provide asset type, description, and other optional fields. Be sure to also enter "Turbofan" or some identifier in the description so we can use that as a search query. 100 | 101 | 102 | 103 | Directly below, select "Add Row" in the "Meters" section. 104 | 105 | 106 | 107 | In this case each meter will correspond to a sensor value. For simplicity, we'll add a single meter first, which measures the engine pressure. 108 | 109 | 110 | 111 | Save the Asset Template. Next, click "Generate New Assets", and create one or more assets. 112 |

113 | 114 |

115 | Now, we'll update the sensor data in one of our assets. To do so, enter "Assets" in the "Find Navigation Item" field. 116 |

117 | 118 |

119 | 120 | Next, we can search for our Asset(s) by entering "Turbofan" in the description. 121 | 122 | 123 | Select an Asset, scroll down to the "More Actions" field and select "Enter Meter Readings", and then a value for each Meter. 124 | 125 | 126 | 127 | 128 | These meter readings can also be updated via an HTTP POST request, we'll demonstrate this functionality in the next step. 129 | 130 | ### 4. Create Automation Scripts 131 | Now, we'll leverage a set of python scripts to periodically query our registered assets, and collect the latest meter data. As each asset's associated data is collected, we'll load our custom model generated from Step 2 to estimate the remaining useful life of the asset. 132 | 133 | In this particular implementation, we'll collect the asset data via the Maximo API. 134 | 135 | To do so, we'll need to authenticate with a base64 encoded set of our Maximo credentials. These credentials are then provided in the following HTTP request as the "maxauth" header. We can query the API to get all meter readings associated with an Asset by sending a GET request to the following endpoint like so. 136 | 137 | ``` 138 | curl -H "maxauth: " -H "Content-Type: application/json" ${MAXIMO_URL}/oslc/os/mxasset?oslc.select=assetmeter,expectedlife&oslc.where=assetnum=${ASSET_ID} 139 | ``` 140 | 141 | The database query is embedded at the end of the Maximo endpoint: 142 | The `oslc.select` value determines which asset attributes will be returned (meter readings in this case) 143 | The `oslc.where` value allows for us to only select data that matches a specific condition (matching our asset id) 144 | 145 | After receiving the Asset data, we can estimate the Remaining Useful Life for our asset by loading our previous model using the `pickle` library 146 | ``` 147 | classifier = pickle.loads(load('model.joblib')) 148 | predictions = classifier.predict(meter_data) 149 | ``` 150 | 151 | Finally, we can update the value for remaining useful life (expectedlife) in the Maximo database with the following POST request 152 | ``` 153 | requests.post("/oslc/os/mxasset/" + restid, 154 | data={"expectedlife": predictions[0]}) 155 | ``` 156 | 157 | The full Maximo asset query/update workflow can be executed by running the included `server.py` script like so 158 | ``` 159 | export TOKEN=$(echo ${MAXIMO_USER}:${MAXIMO_PASS} | base64) 160 | python server.py ${MAXIMO_URL} ${TOKEN} ${ASSET_ID} 161 | ``` 162 | 163 | In conclusion, here we have demonstrated how to create assets in Maximo, query their associated meters, and pass the metered data through a custom machine learning model. This allows for organizations to predict of when their various assets may fail, and proactively take action beforehand. 164 | 165 | ## License 166 | This code pattern is licensed under the Apache Software License, Version 2. Separate third-party code objects invoked within this code pattern are licensed by their respective providers pursuant to their own separate licenses. Contributions are subject to the [Developer Certificate of Origin, Version 1.1 (DCO)](https://developercertificate.org/) and the [Apache Software License, Version 2](https://www.apache.org/licenses/LICENSE-2.0.txt). 167 | 168 | [Apache Software License (ASL) FAQ](https://www.apache.org/foundation/license-faq.html#WhatDoesItMEAN) 169 | 170 | 171 | 172 | 188 | -------------------------------------------------------------------------------- /notebooks/PMI-Custom_Model_Development.ipynb: -------------------------------------------------------------------------------- 1 | {"nbformat_minor": 1, "cells": [{"execution_count": null, "cell_type": "code", "metadata": {}, "outputs": [], "source": "'''\nLicensed Materials - Property of IBM\nIBM Maximo APM - Predictive Maintenance Insights On-Premises\nIBM Maximo APM - Predictive Maintenance Insights SaaS \n\u00a9 Copyright IBM Corp. 2019 All Rights Reserved.\nUS Government Users Restricted Rights - Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM Corp.\n'''"}, {"source": "# Maximo APM PMI - Custom Model Development", "cell_type": "markdown", "metadata": {"collapsed": true}}, {"source": "1. [Introduction](#introduction)\n1. [Install Maximo APM PMI SDK](#install-maximo-apm-pmi-sdk)\n1. [Customize Out-of-Box Anomaly Detection Model](#customize-anomaly-detection-model)\n 1. [Package Custom Code](#package-custom-code)\n1. [Setup the Model Training Pipeline](#setup-model-training-pipline)\n 1. [Test Custom Code](#test-custom-code)\n1. [Train the Model Instance](#train-model-instance)\n1. [Register the Trained Model Instance](#register-trained-model-instance)", "cell_type": "markdown", "metadata": {}}, {"source": "\n## Introduction\n\nMaximo APM PMI supports custom models in addtion to the 5 out-of-box models. There are 2 approaches supported by PMI for the custom model development. The first approach is to develop/train your custom models with tools of your choice then deploy to Watson Machine Learning (WML). PMI supports using WML for scoring. With this approach, you can use any [machine learning framework supported](https://dataplatform.cloud.ibm.com/docs/content/wsj/analyze-data/pm_service_supported_frameworks.html?audience=wdp) by WML with Maximo APM PMI. You can see details on using this approch in the notebook \"PMI - Using Models Deployed on WML\".\n\nThe other approach is more native to PMI which only supports pure Python machine learning frameworks. You develop, train, and deploy your custom models with PMI SDK. In fact, the 5 out-of-box models fall in this category. In this notebook, we'll talk about this approach.\n\nNote that Maximo APM PMI UI does not yet support custom models. Only the out-of-box models can be visualized on the UI. The generated prediction results are stored in data lake and can be added to dashboard cards, used to calculate asset health scores, or retrieved by API.\n\nBefore you can start, similar to using the out-of-box model pipelines, you need to install PMI SDK first.", "cell_type": "markdown", "metadata": {}}, {"source": "\n## Install Maximo APM PMI SDK", "cell_type": "markdown", "metadata": {}}, {"source": "To install the SDK, you need your Maximo APM PMI instance ID, API base URL, and your API key. The Maximo APM PMI instance ID and API base URL can be found in the user welcome letter. For API key, request to your Maximo admin to create an user account first to generate one for you. Create one environment variable for each here.", "cell_type": "markdown", "metadata": {}}, {"execution_count": null, "cell_type": "code", "metadata": {}, "outputs": [], "source": "# The code was removed by Watson Studio for sharing."}, {"source": "Then, install PMI SDK with `pip`. Note that we have to upgrade `pip` first.", "cell_type": "markdown", "metadata": {}}, {"execution_count": null, "cell_type": "code", "metadata": {}, "outputs": [], "source": "# The code was removed by Watson Studio for sharing."}, {"source": "\n## Customize Out-of-Box Anomaly Detection Model", "cell_type": "markdown", "metadata": {}}, {"source": "Let's use an example of customing the out-of-box anomaly detection model as an example of how to create custom models. Let's say we want to change the algorithms used by the anomaly detection model to be solely composed on covariance structure based ones. You can reference the original out-of-box anomaly detection model in notebook \"PMI - Anomaly Detection\".\n\nA model template is mainly composed of two Python classes: one asset group pipeline class and the other estimator class. The pipeline class is responsible for settin up the model processing pipeline. A pipeline always starts with a data loader, which can load asset data and/or sensor data. A pipeline also contains one or more processing stages, each stage can be feature extraction, data pre-processing, or estimator (which learns from data). There can also be post-processing summarization at the end of the pipeline. In most cases, a pipeline has one estimator and a few pre-processing or post-processing stages. However, it is also possible to put more than one estimator in a pipeline, in which case, they use the same inputs.\n\nEstimator class is for the machine learning itself. It takes the input, features, to train the models, and with the models trained, to generate prediction results.\n\n![asset_group_pipeline.jpg](attachment:asset_group_pipeline.jpg)\n\nTo create custom models, you implment two Python classes inheriting from PMI SDK base classes. In this example, since we are modifying the anomaly detection model, we will create a estimator class inheriting from **```pmlib.anomaly_detection.AnomalyDetectionEstimator```** and a pipeline class inheriting from **```AnomalyDetectionAssetGroupPipeline```**. Since we jsut want to change the algorithms used, all we need to do is to override estimator's method **```get_stages```**. However, we do need to also create a custom pipeline class as an entry point to the customized anomaly detection model.", "cell_type": "markdown", "attachments": {"asset_group_pipeline.jpg": {"image/jpeg": "/9j/4AAQSkZJRgABAQAASABIAAD/4QBMRXhpZgAATU0AKgAAAAgAAYdpAAQAAAABAAAAGgAAAAAAA6ABAAMAAAABAAEAAKACAAQAAAABAAAEeaADAAQAAAABAAABggAAAAD/7QA4UGhvdG9zaG9wIDMuMAA4QklNBAQAAAAAAAA4QklNBCUAAAAAABDUHYzZjwCyBOmACZjs+EJ+/8AAEQgBggR5AwEiAAIRAQMRAf/EAB8AAAEFAQEBAQEBAAAAAAAAAAABAgMEBQYHCAkKC//EALUQAAIBAwMCBAMFBQQEAAABfQECAwAEEQUSITFBBhNRYQcicRQygZGhCCNCscEVUtHwJDNicoIJChYXGBkaJSYnKCkqNDU2Nzg5OkNERUZHSElKU1RVVldYWVpjZGVmZ2hpanN0dXZ3eHl6g4SFhoeIiYqSk5SVlpeYmZqio6Slpqeoqaqys7S1tre4ubrCw8TFxsfIycrS09TV1tfY2drh4uPk5ebn6Onq8fLz9PX29/j5+v/EAB8BAAMBAQEBAQEBAQEAAAAAAAABAgMEBQYHCAkKC//EALURAAIBAgQEAwQHBQQEAAECdwABAgMRBAUhMQYSQVEHYXETIjKBCBRCkaGxwQkjM1LwFWJy0QoWJDThJfEXGBkaJicoKSo1Njc4OTpDREVGR0hJSlNUVVZXWFlaY2RlZmdoaWpzdHV2d3h5eoKDhIWGh4iJipKTlJWWl5iZmqKjpKWmp6ipqrKztLW2t7i5usLDxMXGx8jJytLT1NXW19jZ2uLj5OXm5+jp6vLz9PX29/j5+v/bAEMAAQEBAQEBAgEBAgMCAgIDBAMDAwMEBQQEBAQEBQYFBQUFBQUGBgYGBgYGBgcHBwcHBwgICAgICQkJCQkJCQkJCf/bAEMBAQEBAgICBAICBAkGBQYJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCf/dAAQASP/aAAwDAQACEQMRAD8A/v4ooooAKKKKACiiigAooooAKKKKACiiigAor8pP2tPGv7SPiH9tDwn+z18HviXe/DnRrjwPq/iS8l07S9K1Ce5urTU9Os4VY6nbXISNY7qQkIFJbGTxXO/8Ki/bW/6Og8Uf+E14T/8AlXQUotn6+0V+QX/Cov21v+joPFH/AITXhP8A+VdH/Cov21v+joPFH/hNeE//AJV0D5Gfr7RX5Bf8Ki/bW/6Og8Uf+E14T/8AlXR/wqL9tb/o6DxR/wCE14T/APlXQHIz9faK/IL/AIVF+2t/0dB4o/8ACa8J/wDyro/4VF+2t/0dB4o/8Jrwn/8AKugORn6+0V+QX/Cov21v+joPFH/hNeE//lXR/wAKi/bW/wCjoPFH/hNeE/8A5V0ByM/X2ivyC/4VF+2t/wBHQeKP/Ca8J/8Ayro/4VF+2t/0dB4o/wDCa8J//KugORn6+0V+QX/Cov21v+joPFH/AITXhP8A+VdH/Cov21v+joPFH/hNeE//AJV0B7Nn6+0V+QX/AAqL9tb/AKOg8Uf+E14T/wDlXR/wqL9tb/o6DxR/4TXhP/5V0B7Nn6+0V+QX/Cov21v+joPFH/hNeE//AJV0f8Ki/bW/6Og8Uf8AhNeE/wD5V0B7Nn6+0V+QX/Cov21v+joPFH/hNeE//lXR/wAKi/bW/wCjoPFH/hNeE/8A5V0B7Nn6+0V+QX/Cov21v+joPFH/AITXhP8A+VdH/Cov21v+joPFH/hNeE//AJV0B7Nn6+0V+QX/AAqL9tb/AKOg8Uf+E14T/wDlXR/wqL9tb/o6DxR/4TXhP/5V0B7Nn6+0V+QX/Cov21v+joPFH/hNeE//AJV0f8Ki/bW/6Og8Uf8AhNeE/wD5V0B7Nn6+0V+QX/Cov21v+joPFH/hNeE//lXR/wAKi/bW/wCjoPFH/hNeE/8A5V0B7Nn6+0V+QX/Cov21v+joPFH/AITXhP8A+VdH/Cov21v+joPFH/hNeE//AJV0B7Nn6+0V+QX/AAqL9tb/AKOg8Uf+E14T/wDlXR/wqL9tb/o6DxR/4TXhP/5V0B7Nn6+0V+QX/Cov21v+joPFH/hNeE//AJV0f8Ki/bW/6Og8Uf8AhNeE/wD5V0B7Nn6+0V+QX/Cov21v+joPFH/hNeE//lXR/wAKi/bW/wCjoPFH/hNeE/8A5V0B7Nn6+0V+QX/Cov21v+joPFH/AITXhP8A+VdH/Cov21v+joPFH/hNeE//AJV0B7Nn6+0V+QX/AAqL9tb/AKOg8Uf+E14T/wDlXR/wqL9tb/o6DxR/4TXhP/5V0B7Nn6+0V+QX/Cov21v+joPFH/hNeE//AJV0f8Ki/bW/6Og8Uf8AhNeE/wD5V0B7Nn6+0V+QX/Cov21v+joPFH/hNeE//lXR/wAKi/bW/wCjoPFH/hNeE/8A5V0B7Nn6+0V+QX/Cov21v+joPFH/AITXhP8A+VdH/Cov21v+joPFH/hNeE//AJV0B7Nn6+0V+QX/AAqL9tb/AKOg8Uf+E14T/wDlXR/wqL9tb/o6DxR/4TXhP/5V0B7Nn6+0V+QX/Cov21v+joPFH/hNeE//AJV0f8Ki/bW/6Og8Uf8AhNeE/wD5V0B7Nn6+0V+QX/Cov21v+joPFH/hNeE//lXR/wAKi/bW/wCjoPFH/hNeE/8A5V0B7Nn6+0V+QX/Cov21v+joPFH/AITXhP8A+VdH/Cov21v+joPFH/hNeE//AJV0B7Nn6+0V+QX/AAqL9tb/AKOg8Uf+E14T/wDlXR/wqL9tb/o6DxR/4TXhP/5V0B7Nn6+0V+QX/Cov21v+joPFH/hNeE//AJV0f8Ki/bW/6Og8Uf8AhNeE/wD5V0B7Nn6+0V+QX/Cov21v+joPFH/hNeE//lXR/wAKi/bW/wCjoPFH/hNeE/8A5V0B7Nn6+0V+QX/Cov21v+joPFH/AITXhP8A+VdH/Cov21v+joPFH/hNeE//AJV0B7Nn6+0V+QX/AAqL9tb/AKOg8Uf+E14T/wDlXR/wqL9tb/o6DxR/4TXhP/5V0B7Nn6+0V+QX/Cov21v+joPFH/hNeE//AJV0f8Ki/bW/6Og8Uf8AhNeE/wD5V0B7Nn6+0V+QX/Cov21v+joPFH/hNeE//lXR/wAKi/bW/wCjoPFH/hNeE/8A5V0ByM/X2ivyC/4VF+2t/wBHQeKP/Ca8J/8Ayro/4VF+2t/0dB4o/wDCa8J//KugORn6+0V+QX/Cov21v+joPFH/AITXhP8A+VdH/Cov21v+joPFH/hNeE//AJV0ByM/X2ivyC/4VF+2t/0dB4o/8Jrwn/8AKuj/AIVF+2t/0dB4o/8ACa8J/wDyroDkZ+vtFfkF/wAKi/bW/wCjoPFH/hNeE/8A5V0f8Ki/bW/6Og8Uf+E14T/+VdAcjP19or8gv+FRftrf9HQeKP8AwmvCf/yro/4VF+2t/wBHQeKP/Ca8J/8AyroDkZ+vtFfkF/wqL9tb/o6DxR/4TXhP/wCVdH/Cov21v+joPFH/AITXhP8A+VdAcjP19or8gv8AhUX7a3/R0Hij/wAJrwn/APKuj/hUX7a3/R0Hij/wmvCf/wAq6A5Gfr7RX5Bf8Ki/bW/6Og8Uf+E14T/+VdH/AAqL9tb/AKOg8Uf+E14T/wDlXQHIz9faK/IL/hUX7a3/AEdB4o/8Jrwn/wDKuj/hUX7a3/R0Hij/AMJrwn/8q6A5Gfr7RX5Bf8Ki/bW/6Og8Uf8AhNeE/wD5V0f8Ki/bW/6Og8Uf+E14T/8AlXQHIz9faK/IL/hUX7a3/R0Hij/wmvCf/wAq6P8AhUX7a3/R0Hij/wAJrwn/APKugORn6+0V+QX/AAqL9tb/AKOg8Uf+E14T/wDlXR/wqL9tb/o6DxR/4TXhP/5V0ByM/X2ivyC/4VF+2t/0dB4o/wDCa8J//Kuj/hUX7a3/AEdB4o/8Jrwn/wDKugORn6+0V+QX/Cov21v+joPFH/hNeE//AJV0f8Ki/bW/6Og8Uf8AhNeE/wD5V0ByM/X2ivyC/wCFRftrf9HQeKP/AAmvCf8A8q6P+FRftrf9HQeKP/Ca8J//ACroDkZ+vtFfkF/wqL9tb/o6DxR/4TXhP/5V0f8ACov21v8Ao6DxR/4TXhP/AOVdAcjP19or8gv+FRftrf8AR0Hij/wmvCf/AMq6P+FRftrf9HQeKP8AwmvCf/yroDkZ+vtFfkF/wqL9tb/o6DxR/wCE14T/APlXR/wqL9tb/o6DxR/4TXhP/wCVdAcjP19or8gv+FRftrf9HQeKP/Ca8J//ACro/wCFRftrf9HQeKP/AAmvCf8A8q6A5Gfr7RX5Bf8ACov21v8Ao6DxR/4TXhP/AOVdH/Cov21v+joPFH/hNeE//lXQHIz9faK/IL/hUX7a3/R0Hij/AMJrwn/8q6P+FRftrf8AR0Hij/wmvCf/AMq6A5Gfr7RX5Bf8Ki/bW/6Og8Uf+E14T/8AlXR/wqL9tb/o6DxR/wCE14T/APlXQHIz9faK/IL/AIVF+2t/0dB4o/8ACa8J/wDyrrxL9obUv24v2cfhmvxhs/2h9c8QnTdb8PW0um3/AIe8NR291BqOtWOnzxyPbadHMuYbl8NG6sGwQaBODP3qooooJCiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooA/9D+/iiiigAooooAKKKKACiiigAooooAKKKKAPyk+Nn/AClJ8Jf9kn1//wBP2i19J182fGz/AJSk+Ev+yT6//wCn7Ra2/wBqP4/eGP2Vv2cPHH7SHjEj+z/BOi3erOhOPNeCMmGEf7UsuyNR3LCg2jsfO3wr/wCCm37JXxl/bb8Y/wDBPvwLq91N8RfA9vLcahFJb7LOQ25iFxFb3G8+ZLAZVEi7RjD4J2mv0CAJ6V/nXfDj9pX9kj9lz4V/su/8FA/C/wATdB13452fj3Ute+J+mWk5/tG40vxlI32+OZNvzGygRF2EkK8jEdK/eL/grb8aPiX4H/bP8JXPx++IHxJ+Gn7LF94TWa28V/DIuiJ4jlnYh9YubeKaZLYQGNol2FTkEA/vCAFI/puAzwKK/mJ/4KKfFf41eC/gt+zZ4g8H/FX4g+KP2b7mzum8f/Ej4d+XL4jvEWFP7NvLia2jZo7ZzvNw8aDJVgw8wKp5j9rj9vK9/Zy/4Jj/AAntv2Mfjj4h+J8Hxc8cJ4XtPiJPbnWPEWnaZKzzXUcdtHCks+o20YEESPEs5JzgNtIB3P6ncY615V8Nfjj8IfjFqvifQ/hb4hs9dvPBeqyaHrsNq5ZrDUYQDJbTAgYkUHkDI96/mV/ZF+Nnxu0P9vfwZ8A/gN41+O3j/wCFHxI0XWdO8U3vxN0TULa48PanFaSS2ep6fqN3ZxLCZHUr5f3FcgEMSm23/wAERv2O/E+i/tc/tIeOZfjN8QL2H4dfF3VNLudLuNRiax8SOtuU+2a3H5Obi5O8MXUoNyqccUBc/qS8XeLvCvgDwrqPjnxzqNto+i6PbSXl9fXkiw29vbwqXkllkchVRVBJJr8VJv8Ag4V/YKmW58SeGdG+Imv+CrOWSKfxlpnhK+n0BPKbbI/2r5XKKQckRfhXnX/By74g1Oy/4J76D4Klu5bDw34w+IXhzRvEtxExXbpMjzTSh2BGEMkMZOTjgA9a+r/23v2mP2iP+CdnwsST9lX9n/T/ABv8IfA3haS9v7uDxDaaNFplpYq4a2isnhkedVtkWTcgO/cRgsCSA2fpb8NfiP4J+MPw80P4sfDTUI9W8O+JbGDU9MvYgwS4tblBJDKocKwDIwOGAI6EA12+DnHev5df24/2kv2m/wBovQv2D7f9jTxjefA9f2gPtn2uHTnEkFnZXml2s3lmJESOd7SKWU22UULLtYbSARt/FTwN+09pH7W3wl/4IvfDD4/+OLTRL3w9rHj/AMWeO7y5im8W3dgLl4bbTLa+MYESJIhy6rna/IITaQLn9NxGODRX84vwP8b/ALTX7En/AAUT8a/8E6/GvxS8QfFXwR4h+Fd/488Lap4nmS51vR7uyaS3kt3u0RWljYo7KWwAQm0A7i3F/Cz9p79ojVf+DYDUP2oNR8a6xcfEWPwbq14viR7uRtTFxFqk0SSi5z5m9IwEDZyABzQFz+m6lIxwa/l5/ab+IH7Z/wAXf2iP2HP2ePgt8Yta+HY+LPw3vbzxTqlmRcTXPlabZXVxciKTMbXzL5ghncZheQuORg+4fsXa1+0N+x1/wVr8R/8ABNvx58UPEfxa8BeJPAC+O9BvfFtwt7q2mXMN4LWe3a5VVLxSfOQMKowmFDbiwFz9aP2zv2zfgp+wb8EpPj98fH1BdCS/tdMVNLtWvLqS5vGKwokKlc5IPcfiSAfnD9mb/grx+xt+1D8a4/2btAn8QeEPH91bvdWWg+MdGutDvL2KMFna2FwNkpVVZigYMVViAQpx8bf8HLOoS6R/wTi07VoLaW8e1+IHheVbeAAyylLl2EcYOAXbGFHqa4zxB4c/az/4Kef8FI/gL+0R4j+B3iL4J/D/AOA97qOtXOp+NGt4NZ1S5u1RY7C3tIJJHWEPEpcsSuGckg4DAmz9m/2ef2qfCH7SPij4ieFPC2ga/osvw28QS+HL2bW7BrOG9niB3T2DMzefbnHD4XIKnGGFfT2DjNfyOfGP9sD9qbSf2IP+Ci/jjTfiDrsGsfDf4mtpfhW8S8lE2j2QvLOP7PZODmCLY7LtTAwT6knT/am8Pft4/wDBO34PfCf/AIKUap+0Z4r+IWu6t4i8P2vjTwtqTRJ4Zu9P10Dfb6dp8aAW5gyERzlyf3vysNpA5j+s8DPAr5H/AG7/ANrTR/2Fv2SvGf7WGv6LN4itPB1tBcSadbzLbyz+fdQ2oCyurquDMGJKngEYzX4Gft/ftqeJ/iP/AMFRPF/7G3i3x38UvAPww+GHh7TrqeL4S6Td3ms6rq+qRpcq95d2lvPJbWsMMgVEKhXdTySTj5n+LXx4/aX+LX/BCP8Aa68C/tDTeKddsPA2sWVj4S8T+MdIuNG1fWdCuNSs5bd7uG4jiaSeAqVkkwSdygngUBc/qG0f9snwXq/7RnhP9mVfDniEax4t8Ir4wh1RbEvo1vbNn/Rpr7cFFzkcIFI5XkFgK+vsZr8LPBPxo+LVj/wWI+Bf7Ptn4j1CPwPqXwG/ti60JZ2FhNfxXCxJdPBnY0yoAocjIUYHGa/Mn9mz4Wft7ftyfBH9ov46a5+1F478KRfDPxr4ysvBul6LcrEkc+lM9yv9ozsDLcWoXy4YbZWVYkDFTlsUDbP7CKXGOtfyIfHj/gqz+0f4q/4J3/scQp4m1rwr4n/aDknh8W+J/CmltqOuRWGhsIr19LsoI3P2u8JVt0aZj+YqAvT6B/4JrfH743aN/wAFEB8APh34h+L3xE+B/iTwrc38uofFfRL+2vdD12yfcEh1C7toPMguYuBGeN7AAZXJBcx/TbRRRQUFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFfEP/AAUW/wCTTNU/7GHwh/6lGlV9vV8Q/wDBRb/k03VP+xh8If8AqUaVQJ7H7TUUUUGAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAH/0f7+KKKKACiiigAooooAKKKKACiiigAooooA/KT42f8AKUnwl/2SfX//AE/aLXhP/BTH9iTxH/wUI/Z0tP2aLHxWnhTQr/xBpl/4iY2zXEl/pVjL58tjGVkj8ppZFjbzDuA2cqc17t8bP+UpPhL/ALJPr/8A6ftFr6ToNorQ+KP2oP2CP2dP2kv2ZvGX7NzeFtF0O38UaLPpVte2um2wmsZCmLaeMoiuTBIqOAGBbbjPNfAw/wCCf/8AwVO8LfA7QvgH8Of2l9AbQYvBtr4Q1a21vwkmoJi2ie1N5Zs1x5nmS2xjWRLhnQuhYcEAfuhRQOx+NvhX/gnp+2L+yr8B/hb8B/8Agn18cLHwto/gDRLjSNSs/FOgRata6tPc3D3T6hhJY5beVZJJAsSu0ewhTnBJ8j0v/ghV4RtP2DbT9l2b4i31t8QtP8bSfEuy8c2FlFbGy8VSMG8+209X8tLUKBH5Afp8wIIAH73UUByn5u/s4fBD/gp34d+K+neK/wBrT4+6H4w8LaXbXETaFoPhePSzqM0kZSK4vLp5pXQxMfMCQhVZhg8V5P8AAv8A4J5ftIfs0/t0+Pvjz8HfixY2/wAJ/in4lk8WeJfCF5oyzX738tu8bpbaiZP3UbysshYIG2oEweWP69UUDsfOv7WP7LPwf/bT/Z98Sfs0/Haxa+8N+JoBFN5TBJ4JY2EkFxA5B2TQyKroSCMjDAqSD+KGrf8ABJr/AIKnXHwI1H9jC2/a6tdQ+Eeo6e+hn+1/CsF3ryaO6eUbP7YZsv8Auv3YkMgKjAXaBiv6OaKBWPyX1T/glzbW2s/sjJ4Q8XmPSf2WfMiVLy133OsQtp8NinzxuqQOPK3t8rjnA6ZO3+3Z/wAE7/Hf7Rfxu8Aftf8A7MPxFb4WfGD4c29zp1lqstgmp2F9pd4SZrG+tHdN6ZZyrBuN7ZUnaV/U2igLH5N/sg/8E3/iB8N/j/4x/bK/bS+IafFj4reMNFHhcXNpp66VpelaEGDvZWNsrOw81gDJIzAnsMlmb859Q/4IN/tgR/steIf+CfXhP9p1tL+BE8l9caNo6aDH/aiC5la4hsb6/EwaWxjnbzJUj2mU8fIpwP6fKKA5Ufym/wDBQr9mf4tXn/BRD9hX9nL4LfEWbwL4x8LeBdfstM8UW1otwi3Oi6fbLuls5H2yW1z5JSaFmOY3Iye/6nfsUf8ABOr4r/Br9pzxj+3L+2D8Sovil8WvFekweHre6sdNXStM0rR4HWU2tpbB3OZJEVnYkdDwWZmP6rzaZplxfQ6pcW0Ml1bBlhmaNWkjD/eCORuUN3wRnvV2gLH58/8ABS79iDUv+CgP7O1h8CtK8RxeF5bPxLpGvm8mtmulZdMmMrQ+WskZBkBwG3YHcGv0MuJfPuHnAxvYtj6nNQ0UDP5cv+Cmn7FWo/skf8EuP25fHeoeIotdX4x+J4fGMMMVs0BsEuNSsUFs7M7+ayn+MBQf7teq/Dv/AIJJftWfHnw78G/DX7Yn7QknxA+DPw9m0rxLpXhsaLFaanfzW0CyWMGq6ish8+O2DGLOwtImc7WO4f0WX+nafq1nJp2rW8V3byjDxTIskbDrhkYFSM+oq2ipGojQBVUYAHAAHQCgVj8kP2lv+CePx81X9sSb9vX9hX4q2/wx8f65o0Og+J7PV9KGsaPrVpanNs8sIlikiniwqh1J+VFxt+bdB8Z/2A/2tf2pf+Ce3xO/ZG/ak+NFj4p8XfESeF7fXbfQUsNP0i2hntZltILOGbzJUzbsTJJLvLSHsAK/XeigLH5w6T+wZqmmf8FAvh3+20fE0L2vgX4bN4BfSPsrCS5kMwl+2LP5m1Fxx5ZQn/arO/Y0/wCCfmq/sp/AP4vfBW/8Uw63J8UPFPifxFHdxWrQLZJ4hjCLCyNI5kaDqzBlDdgK/S+igdj8J5v+CMGqRfsJ/Av9nbwp8TZfDfxS/Z6vH1Twp450+xDRx3ks8ssqy2Esp8y3lV0WRDJkmMHJUsh+1P2Xvg5/wUa8K/E2bxp+2L8b9F8b6FHp0lna+HvD/htNJtzcu8bC+nuXmlmaRFRlWJNsfzk9hX6CUUCSCiiigYUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAV8Q/8ABRb/AJNM1T/sYfCH/qUaVX29XxD/AMFFv+TTdU/7GHwh/wCpRpVAnsftNRRRQYBRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAf/S/v4ooooAKKKKACiiigAooooAKKKKACiiigD8pPjZ/wApSfCX/ZJ9f/8AT9otfSdfNnxs/wCUpPhL/sk+v/8Ap+0WvpOg2hsFFFFBQUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAV8Q/wDBRb/k0zVP+xh8If8AqUaVX29XxD/wUW/5NN1T/sYfCH/qUaVQJ7H7TUUUUGAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAH//0/7+KKKKACiiigAooooAKKKKACiiigAooooA/KT42f8AKUnwl/2SfX//AE/aLX0nXzZ8bP8AlKT4S/7JPr//AKftFr6ToNobBRRRQUFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFfEP8AwUW/5NM1T/sYfCH/AKlGlV9vV8Q/8FFv+TTdU/7GHwh/6lGlUCex+01FFFBgFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQB//9T+/iiiigAooooAKKKKACiiigAooooAKKKKAPyk+Nn/AClJ8Jf9kn1//wBP2i19J182fGz/AJSk+Ev+yT6//wCn7Ra+k6DaGwV+enx3/wCCi3gr4HfGzVPgNb/D/wAY+MdX0XT7DUb2bQINMa1hj1Pzvs6F77UbNy7C3kJCoQMDnmv0Lr8Dvjz/AMpG/iz/ANiz4L/9B1avqODckpZjmEMJWbUXfbfRN9U+x5HEGYzwmFlXppNq2/qfTH/D1fSP+iH/ABM/78eHv/l5R/w9Y0j/AKIf8TP+/Hh7/wCXlfNFFftn/EG8t/5+T++P/wAifnf/ABEDF/yx/H/M+l/+Hq+kf9EP+Jn/AH48Pf8Ay8o/4esaR/0Q/wCJn/fjw9/8vK+aKKP+IN5b/wA/J/fH/wCRD/iIGL/lj+P+Z9L/APD1fSP+iH/Ez/vx4e/+XlH/AA9Y0j/oh/xM/wC/Hh7/AOXlfNFFH/EG8t/5+T++P/yIf8RAxf8ALH8f8z6X/wCHq+kf9EP+Jn/fjw9/8vKP+HrGkf8ARD/iZ/348Pf/AC8r5ooo/wCIN5b/AM/J/fH/AORD/iIGL/lj+P8AmfS//D1fSP8Aoh/xM/78eHv/AJeUf8PWNI/6If8AEz/vx4e/+XlfNFJR/wAQby3/AJ+T++P/AMiH/EQMX/LH8f8AM+mP+Hq+kf8ARD/iZ/348Pf/AC8o/wCHrGkf9EP+Jn/fjw9/8vK+aKSj/iDeW/8APyf3x/8AkQ/1/wAX/LH8f8z6Y/4er6R/0Q/4mf8Afjw9/wDLyj/h6xpH/RD/AImf9+PD3/y8r5opKP8AiDeW/wDPyf3x/wDkQ/1/xf8ALH8f8z6Y/wCHq+kf9EP+Jn/fjw9/8vKP+HrGkf8ARD/iZ/348Pf/AC8r5opKP+IN5b/z8n98f/kQ/wBf8X/LH8f8z6Y/4er6R/0Q/wCJn/fjw9/8vKP+HrGkf9EP+Jn/AH48Pf8Ay8r5opKP+IN5b/z8n98f/kQ/1/xf8sfx/wAz37xF/wAFffAPg7R5fEvjP4P/ABH0rSrUx/abya20Jo4EkkWMO4i1l5CoLDO1GOOgNfrvNE0ErQv1QlTj2r+W/wDbO/5Nd8Y/9e0H/pVDX9TGof8AH/P/ANdG/ma/KOP+FaGU4iFHDybUo31t3a6JH23C+dVcdSlOqkrO2n9Mp0UUV8EfThRRRQAUUUUAFSwRedMkOcb2C5+pqKrVj/x+w/76/wA6APxu8E/8FFf2qPijoLeNPh38F9Bn0Sa9v7W0lvfGUttcSJYXk1kZJIU0KZYy7QMwUSPgEcmuu/4bU/bbH/NEvC3/AIXM/wD8z1fMf7Ef/Jtekf8AYU8Q/wDp91Cvq7tiv58zXxJzOjiqlKDjaMmlp2bPzHGcV4uFacI2sm1t5mYf21f22/8AoiXhb/wuZ/8A5nqX/htT9tsf80S8Lf8Ahcz/APzPVpH86O2K4P8AiKOa94/cc/8ArhjPL7jMP7av7bf/AERLwt/4XM//AMz1L/w2p+22P+aJeFv/AAuZ/wD5nq0j+dHbFH/EUc17x+4P9cMZ5fcZh/bV/bb/AOiJeFv/AAuZ/wD5nqX/AIbU/bbH/NEvC3/hcz//ADPVpH86O2KP+Io5r3j9wf64Yzy+4zD+2r+23/0RLwt/4XM//wAz1L/w2p+22P8AmiXhb/wuZ/8A5nq0j+dHbFH/ABFHNe8fuD/XDGeX3GYf21f22/8AoiXhb/wuZ/8A5nqX/htT9tsf80S8Lf8Ahcz/APzPVpH86O2KP+Io5r3j9wf64Yzy+4zD+2r+23/0RLwt/wCFzP8A/M9S/wDDan7bY/5ol4W/8Lmf/wCZ6tI/nR2xR/xFHNe8fuD/AFwxnl9xmH9tX9tv/oiXhb/wuZ//AJnqX/htT9tsf80S8Lf+FzP/APM9Wkfzo7Yo/wCIo5r3j9wf64Yzy+4zD+2r+23/ANES8Lf+FzP/APM9XUfCX9t/47eI/wBofwf8C/jL8MdK8MQeNItVa01HTPEr6uY5NKtftbpJbvpVlhXX5QwkOD2rKP515NY/8n7fAP8A7m//ANM1fQcLeIGYYvMKeGrNcsm72Xkz0sn4lxVfEwpTtZ+XkftBRRRX7kfogUUUUAFFFFABX5veJP8AgpHoOmfEHxV4A8I/Cjx54rHg/WLjQr3UdKg0YWb3lqqNKsRvNXtpmVRIvzNEoPav0hr8O/gz/wAlL+OY/wCqp6//AOirOvkuNM/rZbg/rNFJu6Wt7a37NdjxM/zKeEoe1pq7vbU+k/8Ah5Jf/wDRBPib/wB+fDn/AMvqP+Hkeof9EE+Jv/frw5/8vqyT+dFflH/EW8w/59w+5/5nxn+u2J/lj+P+Zrf8PJL/AP6IJ8Tf+/Phz/5fUf8ADyPUP+iCfE3/AL9eHP8A5fVkniij/iLeYf8APuH3P/MP9dsT/LH8f8zW/wCHkd//ANEE+Jv/AH58Of8Ay+o/4eR6h/0QT4m/9+vDn/y+rJPFGOKP+It5h/z7h9z/AMw/12xP8q/H/M1v+Hkd/wD9EE+Jv/fnw5/8vqP+Hkeof9EE+Jv/AH68Of8Ay+rJPFGOM0f8RbzD/n3D7n/mH+u2J/lX4/5mt/w8jv8A/ognxN/78+HP/l9R/wAPI9Q/6IJ8Tf8Av14c/wDl9WSeKMcZo/4i3mH/AD7h9z/zD/XbE/yr8f8AM1v+Hkd//wBEE+Jv/fnw5/8AL6j/AIeR6h/0QT4m/wDfrw5/8vqyTxRjjNH/ABFvMP8An3D7n/mH+u2J/lX4/wCZrf8ADyO//wCiCfE3/vz4c/8Al9R/w8j1D/ognxN/79eHP/l9WSeKMcZo/wCIt5h/z7h9z/zD/XbE/wAq/H/M1v8Ah5Hf/wDRBPib/wB+fDn/AMvqP+Hkeof9EE+Jv/frw5/8vqyTxRjjNH/EW8w/59w+5/5h/rtif5V+P+Z0/hn/AIKRaBq3xE8KfD3xb8KfHnhP/hMdYg0Kx1HVoNGNmt7cpJJEkps9WuplVhE/zCJgCOcV+kFfh58Yf+Sq/An/ALKlof8A6TX1fuHX6vwXn9bMsH9YrJJ3a08rd2z7TIcyniqHtaiSd7aBRRRX1p7QUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABXxD/wUW/5NM1T/sYfCH/qUaVX29XxD/wUW/5NN1T/ALGHwh/6lGlUCex+01FFFBgFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQB//1f7+KKKKACiiigAooooAKKKKACiiigAooooA/KT42f8AKUnwl/2SfX//AE/aLX0nXzZ8bP8AlKT4S/7JPr//AKftFr6ToNobBX4G/Hn/AJSN/Fn/ALFnwX/6Dq1fvlX4HfHn/lI38Wf+xZ8F/wDoOrV+g+F3/I7pekv/AElnzHGf/Iun8vzQtGO1FFf1efiB9G+GP2Ufjb4w8NWHivQ9Ptms9Tj8218y8t4nkTOOEdw34V4/418CeMPhzrr+GvHGnTabeoN3lzD7ynoysMqyn1UkV9HftD8fB34QuvBGi3GD3H7yOuq+JGq3Xjb9kbwH4t8aeZfajZazPp8UrEmee0G7KbzlifkCg88gV8VhM7xd6dWtyuE5yjZJpqzlbW7T+HXRdz6Kvl1D34QupRipau6d0r9NN+7Phqth/D2vR6CniiSzmGmyzG3S6KHyWlUbjGH6FgOcV+m/hX4K2Hj2z1LQfFHwqtvCGlvp089hffay2oRSxrujMqvJ5j7u+YwOxrz1PiL4NsP2MtE1C98GadeRf21JZm3klmCGdbb5rvIbd5j9xnaM8VP+uPtJKFCnd8yT1i902tVJq+j0vp9wf2BypyqTsrNrR9LdGr9T88qK++vgT4LOv+GtDNv8JrbV9OuJFi1DWNRu/LefdIQ72wd41UIvQKGyQec18v8Ax58FaT8OfjH4i8E6DuFlp92UgDncyxuquqknk7Q2Mnk4r2cDn9Kvip4RK0opveL0Tt0ba32djgxOWTp0Y176Pya6X6pX+RzN78P/ABRYeA7L4lXUCro+oXUlnBLvUs00QJYFM7gMA8kYri6+uvFn/Jk/hH/sY73/ANFyVT+EHhTwD4b+Duu/Hrx/pK+IDZ30Wl6dp8rskDTyKHaSbbyQqngdOD6gjGlnrjRnVqRu1NwSW796y3dvV3SNJ5bepGEHZcqk2+ml2fKNFfY3jvw78M/HvwQsfjn4Q8ProF3Z6ymlanp9nI7Qzq6+Yrxb8lGIIXjPXvgV9E+F/grYePLLVNC8UfCq28IaY2nTz2F8LstqEUsa7ozKryeY+7vmMDsa5cTxhRo0/aVYNWbTTcU01a+nN729/dvp9xtSyGpUlywknomtHrf5adtT8sK9I8a/DHV/A/hfw14q1G4hmh8UWkl3bpHu3RrGwUiTIAyc54zXvnhHSvhr8LP2fNJ+LXjDw5D4o1fxPfT29rDdyOkEFvbZV2wnJdiDz7jpjn1H4zeGvDPxWufgr4W8GwSaPpWuW0kUULt5r28ck0ZdQxPzbPm2k9RjNTiOJmsTCMYtU1KactLPkjJu2t9Gu2tmOllH7qTbTlaLS105mrdLbM/Oqiv1qf8AZ78Far4o1P4dal4I03RPDsUU0Nnry6nG1+s0S/JNKnnEkSMOUKfLnnvj518BWHwe8Dfs0Q/FXxp4Yh8Raz/blxp9ussjpE/yA/vdpOURQxUYPzEfhGH41o1YXpQbd4pJOL+K9teay2d03dFVeHqkJWnJJWbvr0tfpfrppqfD9Famt31nqetXepadapY29xM8kVtGSyQozErGpPJCg4BPpWVX2MG2k2rHgyVnY+ZP2zv+TXfGP/XtB/6VQ1/Uvf8A/H9P/wBdG/ma/lf/AG2pbiD9lDxvPZxefMllE0cW4J5ji5hKpuPC7jxk8DOTX7c33x0/bpnvZjD+z5bwku3E/jPTRjn/AKZ20lfzp4z/AO/Uf8P6s/V/D7/d5+v6I+3KK+Gf+Fwft/y/6j4F6BH/ANd/HEY/9F6TJS/8LO/4KLS/6n4NeCov+uvju6J/JPDrfzr8cPv7n3LRXw1/wnf/AAUjl+58Mfh7B/v+MNRk/wDQdAWg+KP+Cl0v+r8E/DOD/f8AEesSf+g6QtAH3LRXwz/av/BTib7mg/CiD/f1XXpP/QdPWjd/wU6m6x/Ce3+kmvzfzSOgD7mq3Yg/bIGHQuuD9DXwf/Zv/BTWbk6x8K7f6afrs387uOv5kf2R/gb/AMHCcH/BT34r+IPgl4itvDHw5m8c6lLqt14jhuW8JXrvcs0raXplzJJduGJ+VraROeGnFAmz9Xf2Ixj9mvSP+wp4h/8AT7qFfWKI0jiNerEAfU18i/sKLep+y7oKam8clyNQ18TNEpRGkGuX+8orMxVS2SoLMQOCT1r7As/+PqIf7a/zFfx9nqvj63+OX5s/DswX+01P8T/M9V8XfBfxL4Is7m412/0tZrQKZLWO8ja4G7GAIvvE8g/TmvIQR0r69+Olx8NL74wX2iX+mXMWoy3lrHNfG6HkhGEW8+T5fH7vI+/15rtfizqnhTS4vFHgfULW7lt7e3ZLC2j0qOKGzdMGOZLpW3FCPvMfvZ/CvoMdw3QdSt7GaUYO3fX3t78ttuif+XpYjKqfNPklZR0766+nbpc+DcitfUtB1bSLOyv9SgMUOoxGe2YkYkjDFSwwTxkY5xX3fb6DfQeF9a8F+Lp4rw2nh17oWtvpiRWts4iDxOt3wWl9SAdxzzxXHeJPFvijVtF+GkepBJNNvVt5Lt/s8YTfHdhQC4T5BjGVBAPcGoqcKRpwbqTd7aaLfmUXe0n3737rTWZZPGMW5N39F3trr/X5/FXbPrQCDX2DqfhDUdOuPi5qd9pzQ2sayLBI8W1AzXQZfLJGPuc/L2rrdR1a01P4l+DfhPqNtaRaJd2emz3SCCMPcTeVuXfJjdgsqrgEZGQc5rCPCzWk52baS06uUorrotLt/mZrJ/5pW6LTza/Q+EcjtXongr4Z6/47sL7VNMns7W204xrPNeTrAimXIQbm45Kmvd/i34j0HU/B+raNqcN7c6ja3sZtZJNLjsUswGKvD5kbHKMvCg9SM1g/B2TTovhD4zfVdLfWYPP07NrG7xsx3vg7owWG3rwKmjkVGONVCc+aPLJ9tlJ2dubqul9PPQVPLoKv7OTurN/cn2v26HgXibw7P4W1ZtHurm1u3VVbzLSZZ4ju7B14yO47VW03QtW1ezvdQ02AzQ6dEJrlgR+7jLBQxyemTjjNfU/wd0vw54k1PxFoHirRxpHhSWOCaUzuwktJt6rDsmkXeTKcgjuPpz2Wn+KvH2la18StKhs4rK40+0V7W0it45AixyJHHtBQ718rB5BBJzjNb4fhqlNRrSk1GXMkkrvRSfW2mmm19VZWZrTymErTb0d9l2v39Pnr2Pg/IoBBr6u1fxfrfh74KeHrfS7e1jvdfuNThu5pLeMyFfNAMYLL8gJfnGCMDGMV6l4p0W8j8DeMfDHiu4jvp9G0+OVY4NMW1tbSYFSvkXGAznBweORnmsqXDCmnyTd1FS2VtY8yXxX2XRP0sZxylSTtLZJ7d1e2/wCh+fxZV+8cVs65oGr+G71dO1yA287RJMEYjOyVQyNwTwQc19V/FLWPGvgy8PhLwLpMLeHrbS7e5if7EkqsPLWR7oylSd4kzyWxxyDmud/aa8Q+J9U1zT7PUVH2D7DZXETiBEDSyQDfiRVBIzn5c4HYDFZY3IadClVcpNyg0vh01vfW+2mj/AmvlsKcJtt3jbpp18/Lf8D5fxXktj/yft8A/wDub/8A0zV60fzr5h+Ith8WtS/bF+BFr8EdU0jR/EJk8VmK51yxuNRshENI/eq1vbXVlIWZeEYTAKeSrdK34C/5G9D1f5M04b/36n/XRn7y0V8Nf8In/wAFL1H/ACPvwwb/ALlTWl/92A0f8Ix/wUwH/M7fC9v+5a1pf/c4a/qk/ZD7lor4a/4R3/gpivTxd8Lm/wC5f1tf/cuaP7D/AOCmQ/5mb4Wt/wBwXXF/9yZoA+5aK+Gf7I/4KaL/AMx74WN/3C9cX/2/NH9nf8FNV6at8K2/7cNdX/26NAH3PtYDdjg96/Dj4M/8lM+Of/ZU9f8A/RVnX44fFnw9/wAHCkv/AAWV8cah+yJPbRWPlaKut3AS6i8BOo0+DAlj1JpC8wXiT7Lm4B+7gV+sH7K4+IK6j8XE+LDac/icfEjW/wC1G0cTLYG68iz802wuCZhFu+6JDux1r828Vf8AkVr/ABL8mfJ8Yv8A2P5r9T6uxXpfhH4QePvHGkf274ctEltvONuGeeKItKACUUSMpJwR0rzQ/nX1L4bg8ESfs+6fN46uru2t4fEE7p9iiWSR2ECZXLsoTjo3PPavwnJMDTr1JKrsk3uo9V1adt+x+dZfh41JNT2Svvb8T5r1jSNT8P6nNo2tQPbXdsxSWKQYZWHY1m5FfZVj8RNK8VD4gfFyfRbW6mt10/7FFeoswi+byFZs9WwAzAHBIxyK2dIufAfh/wAF+HvFuurplsfEUtxdX4l0xrpZgJipghK8QKi8BRznn6+suGaU5Xp1ly2bu9Pd5+RPVpav+r6Hb/ZMJO8J6av5Xsj4dpAQa9c1nw98L7i41bUdK15rMJcTmzsZLSVnaNWPlKZMgKWGB8w4719U6vZaNffFbUvhVc+GtPttFbS3umuUtgsyuLYSC4E3GAH+TA+X8c1y4LhqVW95papK2t27222WnX7jGhlTnf3lul33v2226n59cD60cdK+s7q70Wy8F+BPB0Wl2Ef/AAktsqX1+8Km42G5KZR/4GAJO7r09K6i9i0rXvFXjP4Z6h4dsbHSdCsbuW1nitxHPA1qP3UjTfefzep3H5s8e+seGb2Sqau3R7uPNb0t179LalrKb7S7dOrV/wCmfEhwOO9HFfU3j+9TwPp+m+A/Dvh+xurO90SG6luZbbzZ5ZJoy0k6zD5l8o8jHAxzxXqHim2+EXgq5PgTWDZjTzpaFYU02SS8eaSHetyt4Op3nPHygcY4pw4Yu5J1UuWyd9Fd3srvdabrXyCOUXbTmla1+mr/AK/4B8EnA471NbQPd3MdpDjfKyouTgZY4GSeAMnqa+p4YtJ8T/CBtL8DQ2EV5pmmGbVbS7tMXblWy13BdEZPBGFyMD1zWr4rt9D8VeE7fU/AcGnppGnXGnxXtlJZ+Tf2kjsEO6Yj96srZ3ZJ+nBNZrhz3eZTvonp13vbX7Nteq7Wuyf7K0updL/5/d16+R8qa/ol94b1i50LU9n2i1cxyeW4dcj0Zcgj3FZHvX2v4kk0HwLB8QNe0/RdPuJ7DXre3s1uIFeOAOjZKpgDA7L0zzjitmw8HaBqfjm68R6dpthbTz+FrbVIknizYwXlwVUuYgDx1wMEA8kV1S4TcqnJTnrd6a6K8lv1+Hy6G0slvLljL/hrtfofB/HY5or374+a1o15qek6NpcViLiysl+3y2FuIIpLtyd5U7VLKAFxxgHOK8Cxxmvm8xwioVpUoy5rdf6+48rFUVTqOCd7HgfxhH/F1PgR/wBlS0P/ANJr6v3Dr8EP2nLbx1eeJPgzafDK9sNO8QSfE3RBY3WqW0t5ZxS/Z77DT28M9tJKmM/Kk0Zzj5ux/Qr/AIQ7/gpb/wBFE+GP/hH6z/8ANJX9BeFf/Irf+J/kj9N4O/3P5v8AQ+5aK+Gv+EO/4KXf9FE+GP8A4R+s/wDzSVydn8RP20Phb+0X8M/hr8btf8FeI9B8f3Gr2b/2FoWo6Xd20mn6dJfI4ludXvo2Vimxl8oHByGr9JPqz9EaKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAr4h/4KLf8AJpmqf9jD4Q/9SjSq+3q+If8Agot/yabqn/Yw+EP/AFKNKoE9j9pqKKKDAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigD/9b+/iiiigAooooAKKKKACiiigAooooAKKKKAPyk+Nn/AClJ8Jf9kn1//wBP2i19J182fGz/AJSk+Ev+yT6//wCn7Ra+k6DaGwV+Bvx5/wCUjfxZ/wCxZ8F/+g6tX75V+B3x5/5SN/Fn/sWfBf8A6Dq1foPhd/yO6XpL/wBJZ8xxn/yLp/L80LRjtRRX9Xn4gfYLfGb4CeKfh94U8K/Ezw7rF7d+F7NrVGs7mKGGQOwY5zl8HA9COa5D4gftFXfivxF4Zk8N6PBo3h/wfNHLpmlKxkQFJFctK5ALs+0AnHHPUkk/NtFeDR4bwsJ89m97JttLmveyvZXu/PU9Opm1aUeXRbbJXdtrv5H6BW/7UHwL0z4jan8W7Dw3rM2ua3BNFcNcXcTxW3mx7CLdBgkEgD5zwudoHSvFfAXxZ+GUXwVk+DnxU0m/vIbfUG1Kyn0+ZI2Erx+WUk3/AMPfIyee2K+Z6KxpcK4SEeWPN9nXmd1y3tZ30sm0aTzqvJ3duvRdd/yPsy0+P/wg1rwx4RHxF8Pane6t4Mt0trWK1ukisZxEwZHlUjeG+UbtvX1I4Hgnxr8e2HxR+Kmt/EDTIJLW31SZZUilILqBGqYbbkdVzwa8uorrwORYfD1fbUr31W7aSbu7LZamGJzKrVhyTtbTp2Vlf5H1b4R+MHwbl+C2nfCT4qaJqt8NMv576OXT54ogWmBAB388BiMYqz4b+Nfwd0TTdf8AhfdeH9RufA2tyQXMUL3Ef2+1uolUNKkmNjBiBwegHfJFfJNJWM+G8PLmu5Wk725nZO97rXR37Gkc3rK22itstrWs+6sfVvib4/8AhPSNF8P+Bfg7os1joOhapHrMv9oyLLcXt3GwKmUoNqqAMYHtwMV6rb/tPfAvTPiJqnxY07w3rMut67bzxXDXF3G8Vt50e0i3QY3AkAfMflXO0DpX5/UlYVeEsHOPK79b+87u7u7u+uv+W2hpDO8RF3VunRaW2t2PqLwD8YfhnN8I4fg78adHvtQstMu3vdOudNlSKaNpAfMjbzONrEk59/YGrfxL/aM0nxBq3gXWvhvox0JvBIcW9vI4ljwJVeMBvvNlUw5bBJJxXynSV0/6uYX23tmne7dru15JqWm2t3cy/tatyezTXTWyvpqtd9D618YfET9lrxrrF9481TwprUetaiXmmtIruJbI3DjLOG2+aFLfMQMfSvM7r4n6bc/s+2XwdFrKLy11mTU2uMr5RR4vL2AZ3bs89MV4vSVpQyKjTjGN5NRaavJu1rpWv6kVcyqSbdkrqzsktxaKKSvZPPPmT9s7/k13xj/17Qf+lUNf1L3/APx/T/8AXRv5mv5aP2zv+TXfGP8A17Qf+lUNf1Mah/x/z/8AXRv5mv5y8Z/9+o/4f1Z+s+H3+7z9f0KdFFFfjh9+FFFFABRRRQAVbsmY3sJJzh1H61Uq1Y/8fsP++v8AOgD+f/8AYjGP2a9I/wCwp4h/9PuoV9YAlTuHBHIr5Q/Yj/5Nr0j/ALCniH/0+6hX1d2xX8e5/wD7/X/xS/Nn4ZmX+81PV/mX9V1XU9cv5NV1q4ku7mYgvLKxZ2wMDJPJ4GK6S8+Ivj+/0P8A4Rq+1q9m0/aE+zvM5jKjopUnkD0PFcYfzo7YrgjiqsbtSeu+u/r3OdVppuz3O1b4k/EJtKTQ21u+NmkbQiDz32CNhtKbc42leMHjFZf/AAlvij/hHx4TOoXH9mB94tfMbygwO4EJnAOefrz1rnj+dHbFVLG1nvN7W3e3b0G68+smdpqXxH+IGsWTabq2tXtzbvH5LRSzuyGPIO0qTgjIHX0Fc9f61rGqXMV7qV1LPNAiRxu7ksiRjCKpPICjp6VmH86O2KVXF1Z/HJv1Yp1py+JnYa98QfHfiewTS/EesXl9bRkFYp5ndAR0OCTkjsTVPw94y8XeEvN/4RfU7rTvPx5n2aVo9+3ON20jOMnFc2fzo7YpvG1nP2rm+bvd3+8brz5ua7udTrHjjxn4ht5bXXtVu7yOdleRZpWcO0YwpYEnJUdM9KltvH/jmz1hfENtq93HfLEIBcCZvM8peAm7Odo9DxXIn86O2KPrtbm5ud39WDrzvfmZrahrut6rbRWep3ctxFC8kkaSOWVXlO6RgCeC55Y9zW/ffEn4g6npv9j6lrd9PaeV5JhkndkMfHylScEcDr6VxR/OjtilHF1Ve0nrvruCrzV7M6tvHfjZvD//AAijatdnTMbfsvnP5W3025xj26VU1LxZ4n1jSbbQtW1C4ubKy/1EEsjMkfGPlUnA449hxXPn86O2KJYuq1Zye1t+nb0FKtN6NsMV5LY/8n7fAP8A7m//ANM1etH868msf+T9vgH/ANzf/wCmavp+Av8Akb0PV/kz1+G/9+p/10Z+0FFFFf1SfsgUUUUAFFFFAClmICk8DpX4cfBn/kpnxz/7Knr/AP6Ks6/cavw7+DIJ+Jfxzx/0VPX/AP0VZ1+beKv/ACK1/iX5M+U4y/3P5r9T33Fa76/rMmgp4YkuGNhFO1ysHG0SsoUv0zkgAdaySp+tJtPpX85xqSjfle5+WKTWxsWXiDWdN0q90OxuGjtdREYuYhjEgibcmcjPytzxiuo8LfFX4h+C9PbSfDGqzWtqzGTygFdQ5/iUOG2n3GK4Ag9qNprejja1NqVObTWis2tOxpCvOLTjJqxZu7261C9k1G/kaaedzJI7HLMzHLEnuSTk17p8Tvj94v8AGWoX9joeoXVrod2kaC0cqCFWNVdSVydrMCdobBzyOTXgRB7UbTWlDM69KnKnTk0pWvbra/4asuGLqRi4xdr7/j/ma+o6/rGq2dlYajcNLFp0ZhtlOAIkLFiowPU55rrda+LvxL8RaH/wjet61c3FkQFaNiPnC9A7AbnA/wBomvOyD2o2nGazhjq0bqM2r6PV6rsyViKivZvU9Btviv8AEaz8LnwZbaxcJphRovIyMCNvvIGI3BT3UHFT2vxh+Jtn4fHhe31mdbFYjCsfykrGRgorkFwuOMBsYrzcg9qNpxmtVmmKW1SW1t3t2328i/rlXpJ9t+h6Bc/FX4h3XhhfB0+rTNpqxrD5Pyj92n3ULAbyo7KTin698W/iR4nsYtN17V57iCF0kVTtGXj+4zFQC5XsWzXnhB7UbTjNKWZ4lpp1JWatu9l09BPF1duZ/edRqfjXxVrEF/banfSTJqk63V0rBcSzICFc4HUA9sCvUPh18ZZvD8t8niqa9drrT4tOt7yxdEurSKFtyrGWAG09D3x3rwcg9qNpxmrw2bYilUVWMnfz17/5v5u+5VLGVYS54vX+v8z2T4t/E2y8fw6Rp1it3MmkxSRm81B1ku5zIwb94yDG1cYUZPfmvGacQe1G04zWOOxtTEVXWqvV/orL8DPEV5VZ889zwL4wj/i6nwI/7Klof/pNfV+4dfh78YgR8VfgT/2VLQ//AEmv6/cKv6E8K/8AkVv/ABP8kfp3B3+5/N/oFfDP7R//ACd7+zgP+ox4n/8ATBcV9zV8NftH/wDJ3n7OH/YY8T/+mC4r9JPq2fctFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAV8Q/8ABRb/AJNM1T/sYfCH/qUaVX29XxD/AMFFv+TTdU/7GHwh/wCpRpVAmftNRRRQYBRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAf//X/v4ooooAKKKKACiiigAooooAKKKKACiiigD8pPjZ/wApSfCX/ZJ9f/8AT9otfSdfNnxs/wCUpPhL/sk+v/8Ap+0WvpOg2hsFflR+0T+wV8eviR+0v4h/aB+DnxB0Dw9B4l0nR9NutP1nQbrU3R9I+1BJI5rfVLEASC5O5WjbG0Ybmv1Xorty/MK2Fqqvh5csl1XnoY4rCU68HTqq6Z+Lf/Dvn9uE/wDNWvA3/hH6l/8ANBSf8O+P24P+iteBv/CP1L/5oK/aWivof9fM4/6CJfgeV/qzgP8An0j8Wz/wT5/bhP8AzVrwN/4R+pf/ADQUn/Dvj9uD/orXgf8A8I/Uv/mgr9paKP8AXzOP+giX4B/qzgP+fSPxbP8AwT5/bhP/ADVrwN/4R+pf/NBXlvxv/Za/am/Z6+Dnij46fE74y+BbHw/4R0y51W/mPhDUsiG2jLlVH/CQcu5ARF6szADrX77V8E/8FJP2IPD3/BQn9l7Uv2cPGni3WfCWhXVxFqF++iCDzrxbMNLFbSmdJB5PmhJGCgEsi84zk/18zj/oIl+APhnAf8+kfkn+wd4C/ab/AG/f2U/Cf7VPwx+K3gi0sfEcDi4sJPCeoSy2F7A5iubSVl19QWiccNtAZCrgAMK+vf8Ah3x+3B/0VrwP/wCEfqX/AM0FYH/BBL9grw3+xN+xZo3iHwZ4w1nXrH4paZpXie60vUhb/ZtP1Ce1Xzms/KRXCyqUVw7NkRoc5zn9xKP9fM4/6CJfgJcM4D/n0j8PfFX7EP7X3gfwvqXjXxh8ZvAWnaRo9pNfX11N4R1FY4La3jMssrk+IOFRFLH6V8U/8E41/aC/4KV/s5L+0R8IPir4LsLeLVL3SbzTrnwnqD3FrNayHy/N26+MefbtFOvHAk28lSa/fL9tL9lLTf23v2bvEX7Lev8AijVfCOk+LlitdRvtF8n7W9osivJbqZ0kQJNtCSfLkplehNfkp/wQE/4Js+Cf2Jfgzq/xd+HnjXXtVt/iBcahaX+i6h9mNjHNomr3thbXUPlxpIsxhiIkyxDbzx8q4P8AXzOP+giX4B/q1gL/AMJHtp/4J8/twn/mrXgb/wAI/Uv/AJoKT/h3x+3B/wBFa8D/APhH6l/80FftLRR/r5nH/QRL8B/6s4D/AJ9I/Fs/8E+f24T/AM1a8Df+EfqX/wA0FJ/w74/bg/6K14H/APCP1L/5oK/aWij/AF8zj/oIl+Af6s4D/n0j8Wz/AME+f24T/wA1a8Df+EfqX/zQUn/Dvj9uD/orXgf/AMI/Uv8A5oK/aWij/XzOP+giX4B/qzgP+fSPxbP/AAT5/bhP/NWvA3/hH6l/80FJ/wAO+P24P+iteB//AAj9S/8Amgr9paKP9fM4/wCgiX4B/qzgP+fSPwp+IP8AwS6/a8+Kfg69+H3jL4ueEF0rU/KS6Nn4Tv47jy0lSUiJ5NdkRWOzALIwGehr93LiXz53nxjexbH1OahorxM0znFY2anipuTWmvY9HBZfRw8XGhGyYUUUV5h2BRRRQAUUUUAFSQyNDKsy9UII/Co6KAPyo8Pf8EvdQ8EafL4f+H/x08c6Jo/2u8u7ewitPDc0dub25lu5ESS40aWZlEkz7fMdmxgZOK3P+HdPxC/6OI8e/wDgB4V/+Udfp3RXj1OHcBOTnKhFt/3V/kcEsrwzbbpr7kfmJ/w7q+IX/RxHj3/wA8K//KOj/h3T8Qv+jiPHv/gB4V/+Udfp3RUf6s5d/wA+If8AgK/yF/ZOF/59x+5H5if8O6viF/0cR49/8APCv/yjo/4d0/EL/o4jx7/4AeFf/lHX6d0Uf6s5d/z4h/4Cv8g/snC/8+4/cj8xP+HdXxC/6OI8e/8AgB4V/wDlHR/w7p+IX/RxHj3/AMAPCv8A8o6/Tuij/VnLv+fEP/AV/kH9k4X/AJ9x+5H5if8ADur4hf8ARxHj3/wA8K//ACjo/wCHdPxC/wCjiPHv/gB4V/8AlHX6d0Uf6s5d/wA+If8AgK/yD+ycL/z7j9yPzE/4d1fEL/o4jx7/AOAHhX/5R0f8O6fiF/0cR49/8APCv/yjr9O6KP8AVnLv+fEP/AV/kH9k4X/n3H7kfmJ/w7q+IX/RxHj3/wAAPCv/AMo6P+HdPxC/6OI8e/8AgB4V/wDlHX6d0Uf6s5d/z4h/4Cv8g/snC/8APuP3I/MT/h3V8Qv+jiPHv/gB4V/+UdH/AA7p+IX/AEcR49/8APCv/wAo6/Tuij/VnLv+fEP/AAFf5B/ZOF/59x+5H5if8O6viF/0cR49/wDADwr/APKOux+E37ALfDv43eG/jp4x+KvivxxfeE4tQj06y1aDRYLVG1K3+zTyN/Z2m2krN5f3QZCoPODX6FUVtQyLA0pqpSoxTXVRSf5F0stw8JKUIJP0QUUUV6p2hRRRQAUUUUAFfBPi7/gmv+y/4y8da/8AEW6bxZpmo+J9Ql1XUl0jxbr2mW0t5MFWWYWtnfRQIzhF3bEGcCvvaisq1CFRctRJrz1InTjJWkrn52/8Ou/2Yf8AoKeP/wDwvPFH/wAsqP8Ah13+zB/0FPH/AP4Xnij/AOWVfolRXN/ZmG/59x+5GP1Kj/KvuPzt/wCHXf7MP/QU8f8A/heeKP8A5ZUf8Ou/2YP+gp4//wDC88Uf/LKv0Soo/szDf8+4/cg+p0f5V9x+dv8Aw67/AGYf+gp4/wD/AAvPFH/yyo/4dd/swf8AQU8f/wDheeKP/llX6JUUf2Zhv+fcfuQfU6P8q+4/O3/h13+zD/0FPH//AIXnij/5ZUf8Ou/2YP8AoKeP/wDwvPFH/wAsq/RKij+zMN/z7j9yD6nR/lX3H52/8Ou/2Yf+gp4//wDC88Uf/LKj/h13+zB/0FPH/wD4Xnij/wCWVfolRR/ZmG/59x+5B9To/wAq+4/O3/h13+zB/wBBTx//AOF54o/+WVH/AA67/Zg/6Cnj/wD8LzxR/wDLKv0Soo/szDf8+4/cg+p0f5V9x+dv/Drv9mD/AKCnj/8A8LzxR/8ALKj/AIdd/swf9BTx/wD+F54o/wDllX6JUUf2Zhv+fcfuQfU6P8q+4/O3/h13+zB/0FPH/wD4Xnij/wCWVH/Drv8AZg/6Cnj/AP8AC88Uf/LKv0Soo/szDf8APuP3IPqdH+VfcfBfg/8A4Js/sweC/Hfh/wCI9o3ivUtS8L6hHqumrq/izXtTtobyJXSOY2t5fSwOyLIwXehxk196UUV00aEKa5aaSXlobU6cYK0VYK+Gf2j/APk739nAf9RjxP8A+mC4r7mr4a/aP/5O8/Zw/wCwx4n/APTBcVqWz7looooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACviH/got/yaZqn/AGMPhD/1KNKr7er4h/4KLf8AJpuqf9jD4Q/9SjSqBM/aaiiigwCiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooA//Q/v4ooooAKKKKACiiigAooooAKKKKACiiigD8pPjZ/wApSfCX/ZJ9f/8AT9otfSdfNnxs/wCUpPhL/sk+v/8Ap+0WvpOg2hsFFFFBQUUUUAFFFFABWR4g/wCRe1D/AK9Zv/RbVr1keIP+Rf1D/r1m/wDRbUAfI/8AwTk/5R//AAV/7EvRv/SSOvs+vjD/AIJyf8o//gr/ANiXo3/pJHX2fQBZs/8Aj8i/31/nXwP/AMEz/wDkzDw5/wBhfxT/AOpJqlffFn/x+Rf76/zr4H/4Jn/8mY+HP+wt4p/9STU6APu+iiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAor4j8b/ALWvxH0/44+KPgf8IvhJrHju48H22lz6lfWuqaVp8CNq0Mk8EarfTxSOQkTbiFwDgZqt/wANK/tZf9G26/8A+FJ4d/8AkygLn3LRXw1/w0r+1l/0bbr/AP4Unh3/AOTKP+Glf2sv+jbdf/8ACk8O/wDyZQB9y0V8Nf8ADSv7WX/Rtuv/APhSeHf/AJMo/wCGlf2sv+jbdf8A/Ck8O/8AyZQB9y0V8Nf8NK/tZf8ARtuv/wDhSeHf/kyj/hpX9rL/AKNt1/8A8KTw7/8AJlAH3LRXw1/w0r+1l/0bbr//AIUnh3/5Mo/4aV/ay/6Nt1//AMKTw7/8mUAfctFfDX/DSv7WX/Rtuv8A/hSeHf8A5Mo/4aV/ay/6Nt1//wAKTw7/APJlAH3LRXw1/wANK/tZf9G26/8A+FJ4d/8Akyj/AIaV/ay/6Nt1/wD8KTw7/wDJlAH3LRXw1/w0r+1l/wBG26//AOFJ4d/+S6P+Glf2sv8Ao23X/wDwpPDv/wAmUAfctFfDX/DSv7WX/Rtuv/8AhSeHf/kuj/hpX9rL/o23X/8AwpPDv/yZQB9y0V8Nf8NK/tZf9G26/wD+FJ4d/wDkuj/hpX9rL/o23X//AApPDv8A8mUAfctFfDX/AA0r+1l/0bbr/wD4Unh3/wCS6P8AhpX9rL/o23X/APwpPDv/AMmUAfctFfDX/DSv7WX/AEbbr/8A4Unh3/5Lo/4aV/ay/wCjbdf/APCk8O//ACZQB9y0V8Nf8NK/tZf9G26//wCFJ4d/+S6P+Glf2sv+jbdf/wDCk8O//JlAH3LRXw1/w0r+1l/0bbr/AP4Unh3/AOS6P+Glf2sv+jbdf/8ACk8O/wDyZQB9y0V8Nf8ADSv7WX/Rtuv/APhSeHf/AJLo/wCGlf2sv+jbdf8A/Ck8O/8AyZQB9y18M/tH/wDJ3v7OA/6jHif/ANMFxS/8NK/tZf8ARtuv/wDhSeHf/kuvOnn/AGm/jl+1F8JPGXi74R3/AIF0DwLda3e39/f6zpN7u+3aXLZwxxw2M8spYyOMkgACgTP0qooooGFFFFABRRRQAUUUUAFFFFABRRRQAUV8O/te/tU/En9n/wAZeAfh18J/B2n+LtZ8dHV2Uanq8mj29tFpEME0jGWKxv2dpPPAVfLUDBy1eBf8Nmft19R8G/Bn/hb3v/zN14uP4jwOFqeyxFVRlvZnBic0w9GXJVmkz9X6K/KD/hsz9uv/AKI54L/8Le9/+Zuj/hsz9uvqPg34M/8AC3vf/mbri/10yr/n/E5/7ewf/PxH6v0V+UH/AA2Z+3X/ANEc8F/+Fve//M3R/wANmft19R8G/Bn/AIW97/8AM3R/rplX/P8AiH9vYP8A5+I/V+ivyg/4bM/br/6I54L/APC3vf8A5m6P+GzP26+o+Dfgz/wt73/5m6P9dMq/5/xD+3sH/wA/Efq/RX5Qf8Nmft1/9Ec8F/8Ahb3v/wAzdH/DZn7dfUfBvwZ/4W97/wDM3R/rplX/AD/iH9vYP/n4j9X6K/KD/hsz9uv/AKI54L/8Le9/+Zuj/hsz9uvqPg34M/8AC3vf/mbo/wBdMq/5/wAQ/t7B/wDPxH6v0V+UH/DZn7df/RHPBf8A4W97/wDM3R/w2Z+3X1Hwb8Gf+Fve/wDzN0f66ZV/z/iH9vYP/n4j9X6K/KD/AIbM/br/AOiOeC//AAt73/5m6P8Ahsz9uvqPg34M/wDC3vf/AJm6P9dMq/5/xD+3sH/z8R+r9FflB/w2Z+3X/wBEc8F/+Fve/wDzN0f8Nmft19R8G/Bn/hb3v/zN0f66ZV/z/iH9vYP/AJ+I/V+ivyg/4bM/br/6I54L/wDC3vf/AJm6P+GzP26+o+Dfgz/wt73/AOZuj/XTKv8An/EP7ewf/PxH6v0V+aPwU/bV+Pfi79o/w18AvjV8OdE8NReKtM1jULPUNI8Rz6sVbRxbNJHLBNpVhtEguV2ssjYKnK1+l1e5gcdRxNNVqEuaL6o9HD4iFWPPTd0FfEP/AAUW/wCTTNU/7GHwh/6lGlV9vV8Q/wDBRb/k03VP+xh8If8AqUaVXWas/aaiiigwCiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooA//9H+/iiiigAooooAKKKKACiiigAooooAKKKKAPyk+Nn/AClJ8Jf9kn1//wBP2i19J182fGz/AJSk+Ev+yT6//wCn7Ra+k6DaGwUUUUFBRRRQAUUUUAFZHiD/AJF7UP8Ar1m/9FtWvWR4g/5F/UP+vWb/ANFtQB8j/wDBOT/lH/8ABX/sS9G/9JI6+z6+MP8AgnJ/yj/+Cv8A2Jejf+kkdfZ9AFmz/wCPyL/fX+dfA/8AwTP/AOTMPDn/AGF/FP8A6kmqV98Wf/H5F/vr/Ovgf/gmf/yZj4c/7C3in/1JNToA+76KKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAPhr4F/wDJ9f7Q/wD16+B//TbeV9y18NfAsH/huz9oj/r18D/+m28r7loAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooA/K/9uzj9rL9n7/r38a/+kWn1Yqv+3bx+1l+z9/17+Nf/AEi0+rOOK/nHxV/5Gi/wr82flXGP++fJfqHFJxSnjgUY4r81PlWFJxSnjgUY4oBhScUp44FdB4T0aHxD4o03QbhzHHfXUUDMvUCRwpIz3ANXSpuclCO7KjByaijn6TivsKT4IfCLXvGmofC3wlrGoweIbMyrGLyKM28zxLuKhk5HHOSBx2PSvmPRvBni3xHeT6f4e025v5rYkSrbxtJsIOOdoOOR3616uOyLEUGotXu2vdd9VutOqOvEZdVptJ63utNdVujmqTitW50TWrLVP7Du7SeK93BPs7xsJdx6DYRuye3FdHc/Dbx/Y6jaaTqGi31vcXpxAjwSBpMcnaNuTgckDoK8+GEqyvyxemm3U51Qm9kziKTivVfin8JPE3ws12bTNSilns42REvRC6QSs6B9qs3BI5GAT0NeV44p4vB1cPUdGtG0lugr0JU5OE1ZoKTilPHAoxxXMZBScUp44FGOKACk4pTxwKMcUAeY+E/+Uh/wZ/7F7xv/AOitJr9la/Grwnj/AIeIfBnH/QveN/8A0VpNfsrX9PeG/wDyJ6X/AG9/6Uz9c4U/3GHz/NhXxD/wUW/5NM1T/sYfCH/qUaVX29XxD/wUW/5NN1T/ALGHwh/6lGlV90fRM/aaiiigwCiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooA/9L+/iiiigAooooAKKKKACiiigAooooAKKKKAPyk+Nn/AClJ8Jf9kn1//wBP2i19J182fGz/AJSk+Ev+yT6//wCn7Ra+k6DaGwUUUUFBRRRQAUUUUAFZHiD/AJF7UP8Ar1m/9FtWvWR4g/5F/UP+vWb/ANFtQB8j/wDBOT/lH/8ABX/sS9G/9JI6+z6+MP8AgnJ/yj/+Cv8A2Jejf+kkdfZ9AFmz/wCPyL/fX+dfA/8AwTP/AOTMPDn/AGF/FP8A6kmqV98Wf/H5F/vr/Ovgf/gmf/yZj4c/7C3in/1JNUoA+76KKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigApwRiM4ptfzP+KPDuv8AxT/aN+NepeJ/G3jaAaV4+1HS7K203xZrmm2lvaQWtm0cUNrZXsMCKrSOflQZJOa97hzh2tmmI+rYdpOzet7aW7J9zzM2zWng6XtqqbV7aH9MWxvQ0bG9DX80P/Cj7bv45+Iv/hdeJ/8A5ZUf8KPtv+h5+Iv/AIXXif8A+WVfef8AEG8z/wCfkPvl/wDInzP+v+D/AJZfcv8AM/pe2N6GjY3oa/mh/wCFH23fxz8Rf/C68T//ACyo/wCFH23/AEPPxF/8LrxP/wDLKj/iDeZ/8/IffL/5EP8AX/B/yy+5f5n9L2xvQ0bG9DX80P8Awo+27+OfiL/4XXif/wCWVH/Cj7b/AKHn4i/+F14n/wDllR/xBvM/+fkPvl/8iH+v+D/ll9y/zP6XtjehrP1fSl1rSbvR5pZ7dLuGSBpbdzHNGJFKl4nHKOucqw5Bwa/mx/4Ufbd/HPxF/wDC68T/APyyo/4Ufbf9Dz8Rf/C68T//ACyo/wCIN5n/AM/IffL/AORD/X/B/wAsvuX+Z8I/8Ep/2bv+Cndj/wAFqPir4B/aH+Kvi7VPB/wplt7nXLu5vZvL8SI0cieHIpzwJUe3kadl6Ksbxn71f2mbG9DX8y0P7P2jW15PqNv4x+IEdxdBBNKvjjxKry+WCE3sNRy2wEhcngHirf8Awo+2/wCh5+Iv/hdeJ/8A5ZUf8QbzP/n5D75f/IiXH2DX2Zfcv8z+l7Y3oaNjehr+aH/hR9t38c/EX/wuvE//AMsqP+FH23/Q8/EX/wALrxP/APLKj/iDeZ/8/IffL/5Ef+v+D/ll9y/zP6Xtjeho2N6Gv5of+FH23fxz8Rf/AAuvE/8A8sqP+FH23/Q8/EX/AMLrxP8A/LKj/iDeZ/8APyH3y/8AkQ/1/wAH/LL7l/mf0vbG9DRsb0NfzQ/8KPtu/jn4i/8AhdeJ/wD5ZUf8KPtv+h5+Iv8A4XXif/5ZUf8AEG8z/wCfkPvl/wDIh/r/AIP+WX3L/M/pe2N6GkII68V/NF/wo+27+OfiL/4XXif/AOWVfav/AASzm8RaV8Tfjh8P9Q8Ra9rul6LdeGpLBNe1e+1iS2N3p88k4im1CaeVVkdQxUNtyOleBxH4e43K8P8AWcRKLV0tG7637pdj1Mp4pw+Mq+xpJp2vrb/M/YeiiivhD6UKKKKACiiigAqodQ08Eg3EQI7b1/xrWsP+P+H/AK6L/Ov48f2Zv2Vf2ZPHXwS0jxf41+Hvh3V9W1CfUpbq9vNNtpp5n/tG5G6SR0LM2ABkknivq+EuFKmbVpUac1FxV9fWx4meZ3DA041Jxvd2P69/7R07/n5i/wC+1/xo/tHTv+fmL/vtf8a/mA/4Yq/Y8/6JZ4U/8FFp/wDG6P8Ahir9jz/olnhT/wAFFp/8br77/iC+K/5/x+5nzH/EQqH/AD7f4H9P/wDaOnf8/MX/AH2v+NH9o6d/z8xf99r/AI1/MB/wxV+x5/0Szwp/4KLT/wCN0f8ADFX7Hn/RLPCn/gotP/jdH/EF8V/z/j9zD/iIVD/n2/wP6f8A+0dO/wCfmL/vtf8AGj+0dO/5+Yv++1/xr+YD/hir9jz/AKJZ4U/8FFp/8bo/4Yq/Y8/6JZ4U/wDBRaf/ABuj/iC+K/5/x+5h/wARCof8+3+B/T//AGjp3/PzF/32v+NH9o6d/wA/MX/fa/41/MB/wxV+x5/0Szwp/wCCi0/+N0f8MVfsef8ARLPCn/gotP8A43R/xBfFf8/4/cw/4iFQ/wCfb/A/p/8A7R07/n5i/wC+1/xo/tHTv+fmL/vtf8a/mA/4Yq/Y8/6JZ4U/8FFp/wDG6P8Ahir9jz/olnhT/wAFFp/8bo/4gviv+f8AH7mH/EQqH/Pt/gf0/wD9o6d/z8xf99r/AI0f2jp3/PzF/wB9r/jX8wH/AAxV+x5/0Szwp/4KLT/43R/wxV+x5/0Szwp/4KLT/wCN0f8AEF8V/wA/4/cw/wCIhUP+fb/A/p//ALR07/n5i/77X/Gj+0dO/wCfmL/vtf8AGv5gP+GKv2PP+iWeFP8AwUWn/wAbo/4Yq/Y8/wCiWeFP/BRaf/G6P+IL4r/n/H7mH/EQqH/Pt/gf0/8A9o6d/wA/MX/fa/40f2jp3/PzF/32v+NfzAf8MVfsef8ARLPCn/gotP8A43R/wxV+x5/0Szwp/wCCi0/+N0f8QXxX/P8Aj9zD/iIVD/n2/wAD+n/+0dO/5+Yv++1/xqzHLFMnmQsrqe6kEfmK/l3/AOGKv2PP+iWeFP8AwUWn/wAbr9Nf+CPXh3QfCP7M3i7wt4VsodN0zTviP4rt7W0tkWKGCJLtdsccagKir2AAAr5Ti3gKtlNGNapUUuZ20T7XPbyPienjqjpwi1ZXP1Wooor4E+nCiiigAooooAKKKKAPyv8A27OP2sv2fv8Ar38a/wDpFp9WKr/t28ftZfs/f9e/jX/0i0+rOOK/nHxV/wCRov8ACvzZ+VcY/wC+fJfqHFJxSnjgUY4r81PlWFJxSnjgUY4oBhXb/DL/AJKPoH/YRtv/AEatcQeOBV7TNQvNH1G31fTn8u4tZFlibAOHQhlOCMHBHet8LVUKsZvo0zSjNRmpdj7K8f8AxU8CfDX4q+IdX8K6FNJ4mE00X226n3Qxu4wzxxAenQE/pXNeI9c8ReFf2dfCl34LuZrSPU7q7l1K4t2KO9wHIRXdcEcA4GecD0r5i1/XdU8Ta1deINak867vJDLM+Au5j1OFAA/ACu18E/F34g/Dyyl03wvf+XazNveCVEmiLcfMFkVgDwMkYzivrP8AWdVK1VVG4wlzW5Uk43kneyau2lZ63t1PZ/tbmnNSdou9rJaXad+m9rPU+oPEHji78KSfDP4l+K7N9Q1hLK5N4uP372/3I5W/2grFgTjnPIrgPH9vfR6BpnxR8B+KdSv9Gm1No0ivHdLi1uXUlgGzggpkErjj1rxlvi38RG8aD4gyanI+qqpRZmVWAQggoEK7AmD93GPxqLxr8UfG/wAQYYLbxNeCS3tSWhhijSGJGPVgkaqN3uea2xnElCrTqRvLW3LpZ3SiuZyUt3y3as9dn1LrZrTlGS18vWyV279baqzPof4+6rot58cNW0H4kajqkeiRJDJbw2ZEgWUxJgiORggBBbJHOfxr5f8AF8XgyLV9ngSW7msfLX5r1USXfzu4QkbemO/Wug8XfFvx7450S28PeKL4XdtaMrJmONXJVSqlnVQzYBPU9/WvN8cV4+f5tTxNac6aupNu7XvK+tt2rI4syxkatSUoLd3139NwpOKU8cCjHFfPHmBScUp44FGOKACk4pTxwKMcUAeY+E/+Uh/wZ/7F7xv/AOitJr9la/Grwnj/AIeIfBnH/QveN/8A0VpNfsrX9PeG/wDyJ6X/AG9/6Uz9c4U/3GHz/NhXxD/wUW/5NM1T/sYfCH/qUaVX29XxD/wUW/5NN1T/ALGHwh/6lGlV90fRM/aaiiigwCiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooA/9P+/iiiigAooooAKKKKACiiigAooooAKKKKAPyk+Nn/AClJ8Jf9kn1//wBP2i19J182fGz/AJSleEv+yT6//wCn7Ra+k6DaGwUUUUFBRRRQAUUUUAFZHiD/AJF7UP8Ar1m/9FtWvWR4h/5F/UP+vWb/ANFtQB8j/wDBOT/lH/8ABX/sS9G/9JI6+z6+MP8AgnH/AMmAfBX/ALEvRv8A0kjr7PoAs2f/AB+Rf76/zr4H/wCCZ/8AyZh4c/7C/in/ANSTVK++LL/j8i/31/nXwP8A8E0P+TMPDn/YW8U/+pJqlAH3fRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFfzk6B/yXv49f9lN1b/0jsK/o2r+cnQf+S9/HrH/RTdW/9I7Cv1Hwh/5Gz/wP80fGcd/7j81+p6NRRQa/p8/GgpMilr9ItH1Txd4d/Z28A3PgvxPofhgz2+oPcLqaReZcst020oXglJ2jjqOorxs5zV4VQ5Y3cnbVtdG+ik+nY9DL8Cqzld2sr9+qXdd+5+btFfVkXwd8Kz/BWL4wa2+r6ld6p9skkl0uKCS0sZ43IRLtSfMUSN8xYBVVTxnjPoXirwx8HbG0+FU/gcajp+saolpJBcGK22uDe7HluOu6ZTnZwVIADVyVOJqKkoQi3q47aJxTb/Lt57G8cnnbmk0tE/k7Hwj7GivtrxF8Kfhoj+M/iV8W9b1QjTvFdzpRFlDAZbpjuffghY42OGZsDbgYAyRVzxF+zV8I9O17W/h1oniLUbvxHZaXLrNsz28aWnkJEJ0gkOd5laIgllwoz0zxUR4swuilf7nbpfW2yurvzKlkdbpb71frbS/WzPhmj2r7H+Cv7PPhD4u6DZwwReI0v72KUnUks0Gk28yBisbSN87g4ALAjBOMelfWvC3wcH7Mfh3WILe/Gv3t/eW0cyJB+9u0WMGOV87jbqSPLx83JyK1nxNh1W9hFNvmUXptfm19PdZEcnquHtG1a1/y/wA0fIFFfbD/ALOfwrk8ZXfwOstd1F/G9pavKZTDENNe5ih897defNGFyN54yDx2Pn3xE+GXwl+H/wAN9E1O51PU5/E3iDSrfUoLZY4vssPmNh/Nc4YhsNsCgkFeeDVUOJsNUnGnC95Wto9nez9NHr5E1MnrQi5StZb6rddPXyPmckDrS19RfsjeKrLQ/jHo2iyaRZX0+rX0FuLq6QyPbxEt5ghUnYGfgbyCQBgdTXgPjRVTxlrCIMAX9yAB0A81q76eOk8VPDONrJNO+97rb5HNPDJUY1VLdtW7WOboooNeicgV7Z/wTLI/4aB/aE/67+Ef/TZcV4nWT+xr+1D8Hv2fv2k/jppPxNn1WKfU28KTW407RdU1VSkenTq299PtbhYzk8K5UkcgEc1+W+L/APyKV/jX5M+04E/35/4X+h/QNRXwp/w8g/ZY/hn8Ut9PB/iU/wDuMo/4eO/svn7p8Wt9PBnic/8AuMr+YT9kPuuivhUf8FG/2Zj92Lxifp4J8UH/ANxlL/w8Y/ZtP3bTxs308DeKT/7i6AufdNFfC/8Aw8V/Z16rp/jpvp4E8VH/ANxdOT/goj+z2zgDSfH7An+HwD4sP/uKoA+8bD/j+h/66L/MV/Lb+x5/ybh4d/66al/6crquQ/ZI/wCDnjwn43/bG179kn9oP4c6yksPim/0jQdX8NafeXNzLbwXkkVuL/RZE+3RTCNV83y1ZlbO6BMHHV/sbyLN+zV4amXIDtqLDcCpAOo3R5BwQfUHkV+x+DH+/Vf8H6o+B8QHfDQ/xfoz6booor+jT8lCm7loflTX6WftRa7418PpqOlaH4o0O00c6ZZodE2Q/b2WW3jEgUeQWyxYsD5mdvIxwK8XMs2lQr0qEYpud921azS6J9/JeZ6OEwSqU51G7ctune/mux+atKATwBmvr3xH8Efhn4F8GaD4n8RNr99DqlvaXT6jYRW76eTOR5lsrE7kkjGRlzywxj09j8TeD/Ael/tQ654a+Fk15oVzb6XqLXaLBbG2RVslcR26kNhXGd5YBgT8pHWvPqcWUNfZxbVpO/T3bX7vr2OqGR1PttLVK3XW/wDl3Pzeor668CfBb4NXfgbwZ4j8f6xqlteeMbmeygt7KKJkjeOcQiV3fpGu5cjBYk8cCjxt8CPhnpPhnxdF4R1u/vdc8CyQrqRuIEitZhJN5DiAAlwY26lj82OODmun/WbC+19lre9vhdvi5d+3Np/WmP8AY9bk59Nr7q+1/wAtT5For7O1f9nfwS/wx1fx1ocXiazXRIILs3Gq2cdvb30LyKkn2cD5lYA7k3EgjrXVfEv4O/Czxn478C/Dz4Yrd6bqWuaVp8zSSxQi3+ytC8j3EojO5rkquXx8pPesFxbhXKyvb3ru23Kk3f5P+maPI6yV3a+lvO7t+Z8D0V9P638K/hF4h8CeIPFfwh1XU55/Crwm8i1KKJVuLeaTyhNAYjlcNztfnFVv2g/hn8JfhJqT+C/C+panqGvW7QyXBnjiW1jimi8zYGU72kGVOcbcHHUV20M/oVKioxT5nfSz0tZ69laS+856mV1IwdRtW733vfb7mfNBI5z260tfcH7Jnj/x3eeIrHwSkdhF4O0qOe61xprSEo1qQzO9zMylixJCx8joBjANfGmvy6Xca9fXGiIYrKS4la3Q9ViLkoPwXFa4TMZzxNTDzjblSd0773tfRWel7a6PcithIxoxqxlv5W2tt3WvkZVFFFescIV9Y/8ABJv/AJIL47/7Kb4t/wDSxa+Tq0f+CeX7Xfws+Dfw18f+BvFmk+MLy9i+JHimZpdF8I+INZtCst0rKFu9OsLi3ZgPvIJCynhgDX474z/7jS/xfoz77w+/3mf+H9Ufu/RXwwf+Ch3wI/6F/wCIv/hvfF3/AMqqT/h4d8B/+hf+Iv8A4b3xd/8AKqv5xP1o+6KK+GD/AMFDvgR/0L/xF/8ADe+Lv/lVSf8ADw74D/8AQv8AxF/8N74u/wDlVQB90UV8MH/god8CP+hf+Iv/AIb3xd/8qqT/AIeHfAf/AKF/4i/+G98Xf/KqgD7oor4YP/BQ74ED/mX/AIi/+G98Xf8Ayqr4O/4KT/8ABZvR/wBk/wDZL1j46fAvwr4kv/EGk6hpiLbeKfB/iXSNMmt7i7jiuEkvbuytYYXMTN5RMoJfACuflIJs+iP27P8Ak7L9n7/r28a/+kWn1Yr8i/gV/wAFc/hZ/wAFWvjb8E/E3gnwlrnhHWPDdr4uXVbbUITLp/mTWViALPUUCx3GCpLIVjlQEFkwQT+uuOK/nHxV/wCRov8ACvzZ+V8Y64z5L9Q4pKUmtrw3aW9/4j06wuxuinuoY3XJGVeRVYccjg1+cU4OUlFdT5aEbtJGLxSV9yXnh/4AWvxgk+C6eHJ2F1Obc6gbmQPBNIuUWJOQUXIGWJJPJyK4n4ffD74eaRpfju/+I9pJqK+FbmKOPyZGjZyJZE28EDEhVQSQcDOMV9PLhOr7T2aqResk3d2TirtPTt2uj1nk0+blUl1T30aV30PlLikr6r07S/hWvhHUvjdq3h8/YXvI9P0/R1uX8sSiMNI7y4DlepA/DHPCan8NPBnxG8O+HPGXw9tG0T+1dVXSLu1aRpo45G+YSRs3zY284PqPQ5x/1ZquN6c4ydrpK93G9k9UlrvZu9uhH9kza92Sb3trte19v+D5HytxSV9LfES/+CHhy/1j4fad4bn83TxJbw6mLpzM1zGMbnjPybC4wQO3IA7eoaJ8ErHw94U0K5Pg2bxZcatbpdXlx9qMC26S4KxxKGXLheSW7/XAujwtVqVZU6U1Ll3tzOzva2kbv5JrrcqGTzlNxhJO29r6fhf7j4a4pK+0tK+CfgrS/jB4k8Hyw/2u1hZrdaXp81x5P2h5AG8p5AQSUzjAPPU96xPCfgjwX4q+LEug+IfDT6DPb6fJImkS3Lol1eKfkVZXwyo6nPBPQkHFV/qliE4xlJJuTj13Ts9Umvle7WyH/YtW6Tau3br0+Vv18j5n8MeGdY8Y69beGfD8QmvbtisSFgoJALHJYgDgGszULG60u/n0y+XZNbSNFIuQcOhKsMjIPI7V9N6f4X0TU/jh4c8J6z4Ul8NGaQpd2vnyskvDFXjckMBxglWIPrS+BPB/w0Gi+O/EfjbT5ruDw/dxC3SGVkfDSyKI92ejkKGYgkDJHNTT4clP3FJXvJXd0rRipPRxuv606sjlTl7qet3rqtlfa1z5YpK+mPGPhX4f+Kfh9oXxF8IaedAN5qn9l3cHmtNEpIyJFZ+eByfr7ZPt1t8OPhfYeMtb8Eal4Kl+w6FYvdLqk9xcKJ2jQNl2UhAr5ONvTB49NKHCNapPljUjbSz96z5k2vs3Wid7pDp5LOUrKStprrrdXXT8z8+sE8AEn0FeifEv4fS/DbVrTQb27W4u5rOK5uI1Ur5DygnyiSTkgYJPHWum+B/hix8a/Fe0a5iEGmWTvqNypJZY4IP3gUk8kZ2rk8muG+IPiy48d+NdT8W3BOb6d5EB/hj6Iv4IAK8z6pThgnWnrKUrR9Evef4pL5nL7CMaDm927L5b/ocd9aSlJoxxXjnAeY+E/wDlIf8ABn/sXvG//orSa/ZWvwd8XfFrwx8Fv24vgz428XWur3ln/YvjO38vRNJv9aut8kWl7T9l02C4uNg2nc/l7F43EZGfvr/h4d8B/wDoX/iL/wCG98Xf/Kqv6e8N/wDkT0v+3v8A0pn67wp/uMPn+bPuiviH/got/wAmmap/2MPhD/1KNKqsf+Ch3wIH/Mv/ABF/8N74u/8AlVXx5+3n+3d8FvFP7Meo6Lpuh+PYpn13wtKGuvAvii1i2weI9MmYGWfTEjDFUIRS253KogLsqn7o+iex/TlRX5+f8PLP2fv+hb+Jn/htvGf/AMp6P+Hln7P3/QtfEz/w23jP/wCU9BgfoHRX5+f8PLP2fv8AoW/iZ/4bbxn/APKej/h5Z+z9/wBC18TP/DbeM/8A5T0AfoHRX5+f8PLP2fv+hb+Jn/htvGf/AMp6P+Hln7P3/QtfEz/w23jP/wCU9AH6B0V8+fs9/tPfCT9p/Sdb1b4VS6mD4c1H+ytTtdY0nUdFvLW78iG6EclpqdvbXC7oZ4pFby9pVhgmvoOgAooooAKKKKACiiigAooooAKKKKAP/9T+/iiiigAooooAKKKKACiiigAooooAKKKKAPyk+Nn/AClJ8Jf9kn1//wBP2i19J182fGz/AJSleEv+yT6//wCn7Ra+k6DaGwUUUUFBRRRQAUUUUAFZHiD/AJF7UP8Ar1m/9FtWvWR4h/5F/UP+vWb/ANFtQB8j/wDBOT/lH/8ABX/sS9G/9JI6+z6+MP8AgnH/AMmAfBX/ALEvRv8A0kjr7PoAs2f/AB+Rf76/zr4H/wCCZ/8AyZh4c/7C/in/ANSTVK++LL/j8i/31/nXwP8A8E0P+TMPDn/YW8U/+pJqlAH3fRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFfzk6B/yXv49f9lN1b/0jsK/o2r+cnQP+S9/Hr/spurf+kdhX6j4Q/wDI2f8Agf5o+M47/wBx+a/U9GoooNf0+fjQV9K2Pxm+GmofDzw94E+IXg+bWH8OR3EUFxDqT2m5biUytlFhfoSB1PT3r5qoNcWNy+niFH2l/dd1ZtNOzW6aezZ0YfFTpX5evdJ+fU+mfA/xz8GfDnRLz/hEvDE0Gs3dnc2LztqMrWrx3G4B5bUrh5EQhR8wUkbsDpWdD8bfD7+EvCGm6n4fM2seDZYhaXyXTIj20dwLgxvDsI3H7obPGc47V870GuN5BhXJzad2735pX2a77WbVtjoWaVrct9LW2Xk+2+i13PdvHPxq/wCE08Ka/wCGDpv2f+2/EkniDzPN3+V5kbp5ONo3Y3535HTpXQT/ALQ5n+K2ofE7+ycfbtFbSPs/nfd3Wi2vm79nP3d23A9M96+aaDTWQ4Tl5eXSzW72dr9fJfcJ5nXvfm7Pp0vb82fZfgf9qvRPC1n4VutW8Lvf6r4TtDY2syX8kFs0J3De1sEKmbDEbySD1I4GPJLD4s6H/wAKhPwy1/RWurmyvZ7/AEu/juWiNtNOFDb4whEoBXIGRz16V4fQainw7hIScoxd2095bpt6a6L3notNXoXLNq7Si3pa2y8v8lrvofX97+1DoMusXfxJsPCUVv44vrRrWXVPtTm3VpIvJe4jtdoCylOOXIHXnPPhfxG+Ih+IFv4dgNp9l/sDSINKzv3+b5LO3mdBt3b/ALvOMda80oNXg8hwtCSnTjqttW7K1rK72Sei2XQivmVaonGb0fkv6v3Z7L8FPiN4Q+FniiLxj4g0GXWr6xminsWS8NqsLpuzvURuJA2RwcYx71yfxF8Q+FfFXiifXfCWkyaNb3JMkkEtybomZ2ZncOUTAOeFxxjrXDUGuqOXUlXeJV+Zq27tb0vb8DF4qbpKi7W32X52v+IUUUGu45gr3D/gmTLJH+0B+0II2K/v/CPQkf8AMMuK8Pr23/gmZ/ycD+0H/wBd/CP/AKbLivy3xf8A+RSv8a/Jn2nAf+/P/C/0P2J+0T/32/M0nnz/AN9vzqKiv5hP2Ql86Y9XP50nmy/3z+f/ANeo6KAH+ZJ/eP50olkUhgx4qOigD5y/Z6/Y3/Zg/Zo8Z6743+CXgvTdD1/xfqlxqms6skfmahe3N7O08xlupS0uwyOxESsI16Kor8Bf2PP+TcPDv+/qX/pyuq/qSsP+P6H/AK6L/Ov5bf2POP2b/DvP/LTUv/TldV+x+DH+/Vf8H6o+B8QF/s0P8X6M+l8DpRRRX9Gn5KIRkYr6e+Jnxp+FnxPe51zWPBk8evT2cdst6mqP5aPDCIo5PI8kAgbQSu7n1r5iorhxWXUq041J3vG9rNre19mr7LRnTRxU6cXCOz7pP8/U+kn+OXhbSfhnqvgLwL4cl0uXXoYIb6SW/luLYGFg5kgt2UBHdhkkscdBXSat+0t4e1D4jt8VLLww1tq19p11Z6iReM0c8lxbLbrJGpj/AHYQDcV5LE9RivkmiuD/AFbwd23F3d7+9LW6Sd9ddl6WVjp/tavok9rdF0u1083957PH8XjH4d8CaB/Z+f8AhCrqa53+Z/x8+dcJPtxt+TGzbnLZzmto/HaQan4+1OLTF3eN5kmCtJuW2KXf2oAjb+9B+6fu+tfP1FdMslwz3j57v+bn/wDStfw2M1mNZbP+rcv5aH2L44/ak0Dxdpfin7L4YltdU8X2aW17dSajJPHGUZCPIhZAqRnbyuc5xggDnnD+0hb2914O8XaXoSw+J/CcFtZ/bWuGaC5tLZHjEbW+0bWdXwzhiRjj2+XaK5KXDGChDkjDT1l25bb7W0tt5G085xEnzN6+i737b367n0b4o+Nng8+ENX8I/C7wsPDqeIpYpNTlku3u2dIX8xYYQyKI49/OOSeleefF/wCIZ+K3xDv/AB4bT7D9uEI8jf5m3yoUi+9tXOdmegxnFea0V2YTJ8PQl7SmtddW23ra+rb/AJV9xz18fVqR5JPTToltfsvN/efU/hH48/DvRPhBF8JNd8HzXcEkxuL6e11JrRr2QMSnnBYWJVBgKm4qCM4zXzXrdxpd5rF1daHatZWUkrNBbvIZWijJ+VDIQC5A43EDPpWZRVYPK6NCc6lK95O71bV/Rtr7vQVfG1KsYxnbTRaL/IMDpRRRXonIGB0r6z/4JNMw+AnjsA4/4ub4t/8ASxa+TK+sv+CTg/4sL46/7Kb4t/8ASta/HfGf/caX+L9GffeH3+8z/wAP6o/UDe3qaN7epptFfzifrQ7e3qaN7epptFADt7epo3t6mm0UAO3t6mvEf2hv2c/gv+1d8Mpvg1+0HoUXibwtdXdpeXGm3DyLDNLZTLcQiXy2QsgkRSyE7WAwwIJFe2UUAfkR+1z4J8GfDf8AaI/Zx8DfDzSLLQdF02z8aRWmn6dBHbW0CCy08BY4YgqKPoK6+q/7dp/4yz/Z+/69vGv/AKRafVnHFfzj4q/8jRf4V+bPyrjFf7Z8l+ocV0Xg7/kb9J/6/bf/ANGrXOk81JDLLbypcW7FJIyGVlOCGByCCOhBr85oz5ZqXY+XhK0kz7w+Ifjj4K+DfjXqPjG903UZfEOlzHbCjx/ZJpggCTMxO9SAeQARkZwa+dtJ+KFovgvxrpGtpI+oeKJoJ0aMDy1ZJWkfdkggfNxgGvIdS1LUdYvpdT1WeS5uJjukllYu7HpkseSapY4r6LMOJ61WtKdNKKbk7WX2002+7sepic3nOblFJLX8dPvse6/D/wCIng2DwJffC74lW11LpdxcreQXFkV86CcAKTtfhgQP58HPGp4g+MHh/RLDQfC/wptZ4dN0K+XUjLelTNc3IPVwmQqgZGB2+lfOxNGOK5IcQYiNJUo2Vla9vesne1+1/wDLYwjmVVQ5F6X623tc+kfHPjD4B+KJ9T8YW+m6oNa1NHb7M7xraxXEg5lDqd7AN823oT2APD5PiJ8LPHvhjR7D4mpqlnqWiWy2aXGneWyTwJ9wOshG1gO49/oPmsmjHFaS4irOcpOMfe3XKrPW9353+4t5nNybstd9Nz2DS7z4KS+ItR/tSLWLWwbyzp88Mkb3MRQfOZQdqtvPIwfl6V33i74m/Cnx/r1jaeJ4tWOm6Vp62dreq8bXrSq24yygnawI4xnI5Pfj5hJoxxWdLPakYOmoxs3d6b63s/JdOuxEcxmouKSs/Lc+sb346eE08a+DLmxS/udM8J+YGubvY95P5gwcgNjC8AAt0rz2w+JWh2vg/wAceH5IpvO8S3EMtsQF2osczSHzPmyDgjGM814gTRjirq8SYqbcm116d48n/pKt+JUs1qyd/X8Vb8j6I8CeLdN1bwl4c+FMIMeoDxJDemWdQbfy2AQBvm3E56jHTvX1f4ptoviZ4+1P4b6zpviaysZJZd159pYWKhAWEhjcbBESPlG4jkYA7fmTuIO5Tgiuvu/iF491DTP7Gvtbv5rTbt8l7iRkI9CpbBHtXrZVxXGlRdKvG602tqkmuV329VqduDzhQhyVFfb5pdP+Ce16DHD8OPgd4h8QxSq954juzpFpIvBNtCSZpB7Pgj8q+Y6eXcqEJO0dBngZ9Kbjivm8fjvbKEIqyirW+bbfzb/I8vE4j2ijGKskrBSUpNGOK885TzLwlkf8FD/gyR/0L3jf/wBFaTX7Lb29TX40eEz/AMbEPgz/ANi943/9FaTX7K1/T3hv/wAiel/29/6Uz9c4U/3GHz/Njt7epr4h/wCCi7Mf2TNU5/5mHwh/6lGlV9uV8Q/8FFv+TTdU/wCxh8If+pRpVfdH0T2P2mooooMAooooAKKKKAPz5/Y2z/w0R+02T/0UGy/9RfRK/Qavz4/Y2/5OI/ab/wCyg2X/AKi+iV+g9ABRRRQAUUUUAFFFFABRRRQAUUUUAf/V/v4ooooAKKKKACiiigAooooAKKKKACiiigD8pPjZ/wApSfCX/ZJ9f/8AT9otfSdfNnxs/wCUpXhL/sk+v/8Ap+0WvpOg2hsFFFFBQUUUUAFFFFABWR4g/wCRe1D/AK9Zv/RbVr1keIf+Rf1D/r1m/wDRbUAfI/8AwTk/5R//AAV/7EvRv/SSOvs+vjD/AIJx/wDJgHwV/wCxL0b/ANJI6+z6ALNn/wAfkX++v86+B/8Agmf/AMmYeHP+wv4p/wDUk1Svviy/4/Iv99f518D/APBND/kzDw5/2FvFP/qSapQB930UUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABX47+Mv+Cb37RE3xd8efEH4W/FfQdH0vxv4huPEJsNS8LT6hPbS3MMMTx/aYtZtFkUeQCD5Knk5zX7EUV35dmmIwlT2uGm4y2ujlxeCpV48laN0fjD/w7t/bPxj/AIXP4S/8Iq7/APmho/4d2/tnf9Fo8Jf+EVd//NDX7PUV7v8Ar1m//QRI83/VrAf8+kfjD/w7t/bPxj/hc/hL/wAIq7/+aGj/AId2/tnf9Fo8Jf8AhFXf/wA0Nfs9RR/r1m//AEESD/VrAf8APpH4w/8ADu39s/GP+Fz+Ev8Awirv/wCaGj/h3b+2d/0Wjwl/4RV3/wDNDX7PUUf69Zv/ANBEg/1awH/PpH4w/wDDu39s/GP+Fz+Ev/CKu/8A5oaP+Hdv7Z3/AEWjwl/4RV3/APNDX7PUUf69Zv8A9BEg/wBWsB/z6R+MP/Du39s/GP8Ahc/hL/wirv8A+aGj/h3b+2d/0Wjwl/4RV3/80Nfs9RR/r1m//QRIP9WsB/z6R+MP/Du39s/GP+Fz+Ev/AAirv/5oaP8Ah3b+2d/0Wjwl/wCEVd//ADQ1+z1FH+vWb/8AQRIP9WsB/wA+kfjD/wAO7f2z8Y/4XP4S/wDCKu//AJoaP+Hdv7Z3/RaPCX/hFXf/AM0Nfs9RR/r1m/8A0ESD/VrAf8+kfjD/AMO7f2z8Y/4XP4S/8Iq7/wDmho/4d2/tnf8ARaPCX/hFXf8A80Nfs9RR/r1m/wD0ESD/AFawH/PpH4w/8O7f2z8Y/wCFz+Ev/CKu/wD5oa+q/wBin9kHx9+zFrvj7xh8TPGVl4w1fx1caXI76dpL6Rb20el28luiiKS9vmdnEm5m8wAYwBX3nRXBmPE2PxdP2WJquUd7PudOEyfC0J89GCTCiiivCPTCiiigAooooAmt5fInSfGdjBsfQ5r8MfBX/BLL9qX4aeG4vBPgj4zeG10izmuXtVvPB1zPOsdxcSXG2SVNeiV2UyEbhGgOOgr9yKK9LLM4xWCm6mFm4t6aHHjMBRxEVGtG6R+MP/Du39s7GD8aPCX/AIRN3/8ANDR/w7t/bPP/ADWjwl/4RV5/80Vfs9RXtf695v8A9BEjz/8AVrAf8+kfjD/w7t/bOxg/Gjwl/wCETd//ADQ0f8O7f2zz/wA1o8Jf+EVef/NFX7PUUf695v8A9BEg/wBWsB/z6R+MP/Du39s7GP8AhdHhL/wibv8A+aGj/h3b+2ef+a0eEv8Awirz/wCaKv2eoo/16zf/AKCJB/q1gP8An0j8Yf8Ah3b+2djH/C6PCX/hE3f/AM0NH/Du39s8/wDNaPCX/hFXn/zRV+z1FH+vWb/9BEg/1awH/PpH4w/8O7f2zsY/4XR4S/8ACJu//mho/wCHdv7Z5/5rR4S/8Iq8/wDmir9nqKP9es3/AOgiQf6tYD/n0j8Yf+Hdv7Z2Mf8AC6PCX/hE3f8A80NH/Du39s8/81o8Jf8AhFXn/wA0Vfs9RR/r1m//AEESD/VrAf8APpH4w/8ADu39s7GP+F0eEv8Awibv/wCaGj/h3b+2ef8AmtHhL/wirz/5oq/Z6ij/AF6zf/oIkH+rWA/59I/GH/h3b+2djH/C6PCX/hE3f/zQ0f8ADu39s8/81o8Jf+EVef8AzRV+z1FH+vWb/wDQRIP9WsB/z6R+MP8Aw7t/bOxj/hdHhL/wibv/AOaGvuT9in9mfxD+yp8IdQ+H3i3xHB4q1TV/EOq+Ibu/tbFtOgMuqSiVo47Zri6ZFjxgEzMT7V9dUV5uZ8RY7GQUMVVcktdTrweU4bDycqMEmwooorxT0QooooAKKKKACiiigD8r/wBuz/k7L9n7/r38a/8ApFp9a2mWY1HU7bTmcRC4lSIuei72C7j7DOayf27T/wAZZ/s/f9e3jX/0i0+tjStMvda1K30fTU8y5upFiiXIG53IVRk4AyTjk4r+dPFBXzaKtf3V+bPyzi9XxqVuiPqj4o6T8Gvhlrd18O9S8K3rvFAPK1U3TrJLIyZEixkCMoGODj0PHavmq58J+ILPwtaeMrmHbp17K8EMu5Tukj+8NoO4Y9SMV9v/AAuT4oeInuPhl8ddFkudAtraUve38W17Ty1JVkuDw3TAIJPfOAa8ouPDOteLv2bdFtfB9rNqTWGtXSSJAheRRJnYxVecEEc+9cua5QsQnWpxsuWTjHlUZJpxunZe8kno93Z32McZglUTnFWVnZWs9GtH3tfc8Sj+GPjWXUdG0lLPNx4giE1gvmJ+9Q5wc7sL06Ng11l1+zv8Y7TSP7Yl0WVkDKpiRkeZdx2qTErFwCfbpzwK+mrvRtR8O/Fv4S6Fq8flXVrp6xSpkHa6hwRkccGuI+EWva2fib8RNQ+1ymf+zNSk8zcd26KT922c8FP4fSlDhnCRqKjV5ruXLutPcjLVW11draaAspoKXJO927f+Sp9u7PAPG3wd+Inw802HV/Fen+RbTN5YkSRJFV8Z2PsY7W4PB9K6DSP2dfi/rdtDeWGlZhuII7iGRpolWRJQWQKS3LEDJXqO+K6PwtcT3X7Mvi5Ll2kCanZuu4k4Ziu489z39a6/4sapqdv4k+GdvDcSJHDpemSxorEKrs4BYD1IABPoK5IZNgeT6xJS5XGLSur6ycXry+V9vIxjgMPb2jTtZO1+7a3t5HkkP7P3xfm0afW/7FlSO33lo3ZFmKxkhmWIsHYAjqBz2zXP+CfhV45+IVtNfeGbRZLa3YJJPLLHDErEZC7pGUE4I4Ga+tdP1TUp/wBt6SKad3VZJYACxIEX2UnYB0255x689a8o8F/DrQL7wDqnjnWrHUdeVNWe0i0vTpCgVsZ82TarHkEKCF9PXjapw3hnUtSUmoynF3a+xbXSLtvtZsuWVUua0E9HJPX+W2uz77WZ4/e/Crx5p3jSD4e31g0erXRHkxF02yAgkFZM7CDg857etegaF8F/G3hbxRoM/jXw99ut9SuGijsWuEieZgjNtYhspwN3zYzjHevpHxBatZfHT4W2r2klg0enon2eaTzZIgokwjPxuK9Ca8F+Hep6hq37U1pealM88r6vPlnOT8okUfgAMD0FayyDC4fERg+Zt1FFbWStF6px1fvWtZea6Onl1GlUS1vzJL8Hrda7nE6T8F/H/jm5v9Q8IaWi20V5Nb+U1zCpiZW/1f7xwzbQQN3Q+tZ9j8FviTqHie+8H2+n/wCm6YFa7DSxrHCHG5S0hbZyDxgmvdvhj4Ju7LXvEvxnudPl1NdHvriPTrSJGkae9Mh2sVUE7Isgk+vPauZ8EeAtQ8W/8Jd4w+JkGqXd3YeVNPpdtuhubmSdiQzqQSEUcjC8Dp0weZcP0pKneEuaTk90lypPX4dL2dl72iemqM/7NhJR913d+ullfXbT8dEcNbfAb4gweN9M8G61ZeS2okSLIksRR4FI8xo5N2wlVOduc9OOa6bxX8FNU+HXxSsNMutPOq6Td6ilvaQvcRLJdDK/IxVsx7s4ywWvV/iNZ2+k+DPhki6XPocMequy2t3KZZYozKjZZmCkAj5sEDA4rm/Fnh3X9L/aytdV1GzmhtbvXbcwTOhEcnKH5GIweOuDXdiMiwtBcsINvnp632Uo3s046r1SvpddH0VMupU9Ixd7x17XV9dPzseKyfDfxT40+Imr+HfBuitbvazy77USKyWqKxG15mIXAxgEnntmsTxz8MPHHw58l/FtkYIbnPkzI6yxPjqA6FhnvjrX1dqGn6rr/hX4p6B4PRptWbXfNmhh/wBdJah+igcsAQ2QOvI78+dazputeFP2Wn0bxrFJa3F7rCy6fbzgrKsap+8cI3Kr19OvvzxY3h6jGE5WlfllLm05dJNctrb6d92tLHPXy2moyet7N36aNq23l/wDjLT9mz4y3is0WkYwiuu6aJfMDLvAjy/zHB5A6dDzXiV3aXVhdS2N9G0U0LtG6OMMrKcEEdiDwa+x/i3q2qR/tBeEoVuJAltFpflLuOE3su7aO27v614T8dkVfjJ4lC8f6dKfzwTXBn2UYahGboXvGXK7tO+l76JW28znzHBUqcW6V9HbX/hj5Z8J/wDKQ/4M/wDYveN//RWk1+ytfjV4TP8AxsQ+DP8A2L3jf/0VpNfsrX7n4b/8iel/29/6Uz9D4U/3GHz/ADYV8Q/8FFv+TTNU/wCxh8If+pRpVfb1fEP/AAUW/wCTTdU/7GHwh/6lGlV90fRPY/aaiiigwCiiigAooooA/Pj9jb/k4j9pv/soNl/6i+iV+g9fnx+xt/ycR+03/wBlBsv/AFF9Er9B6ACiiigAooooAKKKKACiiigAooooA//W/v4ooooAKKKKACiiigAooooAKKKKACiiigD8pPjZ/wApSfCX/ZJ9f/8AT9otfSdfNnxs/wCUpXhL/sk+v/8Ap+0WvpOg2hsFFFFBQUUUUAFFFFABWR4g/wCRe1D/AK9Zv/RbVr1keIf+Rf1D/r1m/wDRbUAfI/8AwTk/5R//AAV/7EvRv/SSOvs+vjD/AIJx/wDJgHwV/wCxL0b/ANJI6+z6ALNn/wAfkX++v86+B/8Agmf/AMmYeHP+wv4p/wDUk1Svviy/4/Iv99f518D/APBND/kzDw5/2FvFP/qSapQB930UUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAH5X/ALdn/J2X7P3/AF7eNf8A0i0+rSs6MHQlSDkEcEH1Feff8FMPiX8PfhH+0X8AfG3xS1uy8O6NGnjGB77UZ0t7dZZbOwEaGRyFDPg7QTk4OK8L/wCG9v2Icf8AJW/CX/g2tf8A4uv578TsDXqZkpU4NrlWyb6s/MOLsPUli7wi3ov1PuPUviD471jTP7F1bWr65tMAeTLPIyYHQFS2D+NUdC8XeKvCwlHhvUrmwE4xILeV4w312kZr4r/4b2/Yi/6K54R/8G1r/wDHKT/hvb9iHH/JW/CX/g2tf/i6+DeGx7kpuM7rrZ3PnXRxLfNZ3+Z9p/8ACVeJvtdrqB1G5NxYjFtJ5r74RkkhGzleSelVbTXtb0+4uLqxvJoZbtHjndHZWkSTl1cg5YN3B618b/8ADe37EX/RXPCP/g2tf/jlJ/w3t+xDj/krfhL/AMG1r/8AF1H1LHXvyS+5ieHxG9n+J9iw6vqlvpk2iwXMqWlwyvLArkRuyfdLL0JHbPSn3mua1qEttPf3c0z2aLHAzuzGJE5VUJJ2hewHSvjj/hvb9iL/AKK54R/8G1r/APHKT/hvb9iHH/JW/CX/AINrX/4up+oY21uSX3Mn6rXt8L+5n2gvifxHHrh8TJf3A1IksboSN52SNpO/O7OOOvSptD8Y+LPDMk0vh3U7qxa4/wBaYJXj3/720jP1NfFX/De37EX/AEVzwj/4NrX/AOOUn/De37EOP+St+Ev/AAbWv/xdXHB45PmUZX9H1KVDEp3Sf4n2o3izxQ19bam+pXTXNkMW8plcvECSSEbOV5JPHqaz7XV9VsdTXWrK5lhvEcyLOjkSBznLBhzk5OTXx5/w3t+xF/0Vzwj/AODa1/8AjlJ/w3t+xDj/AJK34S/8G1r/APF1LwONbvyS+5ieGxG9n+J9waZ498c6JC9vo2s31okkjSusNxJGGdvvMQrDLHuepplv458a2msv4jttXvE1CVQr3InfzWUYADPnJAwODXxH/wAN7fsRf9Fc8I/+Da1/+OUn/De37EOP+St+Ev8AwbWv/wAXWiw2YWSUZ6baMv2WJ0snp6n2fqvibxHrkaxa3f3F4qO0qiaRnAd/vMNxPJwMnvV698deNdSazfUdWvJ209g1qZJnYwsMYKEn5TwORzxXxH/w3t+xF/0Vzwj/AODa1/8AjlJ/w3t+xDj/AJK34S/8G1r/APF0vqmPu/dlr5PoT7DE66P8T7Qs/E3iPTtXfX7DULmC/kZme4jlZZWLcsS4O45PXJ5qPXfEfiDxPd/b/Ed7PfzgbQ9xI0jAegLE4HtXxp/w3t+xF/0Vzwj/AODa1/8AjlJ/w3t+xDj/AJK34S/8G1r/APF1LwWOceTklbtZg8NiLctnb5n2Xea/rmoahHq1/eTzXUIQRzPIzOgj+5tYnI29sdO1Ur/UL7Vb2XUdTme4uJmLySSMWdmPUsTyT9a+Qv8Ahvb9iL/ornhH/wAG1r/8cpP+G9v2Icf8lb8Jf+Da1/8Ai6mWX4yW8JfcyXha73i/uZ7N4T/5SH/Bn/sXvG//AKK0mv2Vr8DfgF+0B8Dfjh/wUV+FEHwa8X6R4qfS/DfjN7waVdxXRt1mj0tYzL5TNsDlWC5xkg46Gv3yr+j/AA8ozp5TTjNWfvb/AOJn6pwvTlHBRUl3/NhXxD/wUW/5NM1T/sYfCH/qUaVX29XxD/wUW/5NN1T/ALGHwh/6lGlV9se+9j9pqKKKDAKKKKACiiigD8+P2Nh/xkP+02f+qg2X/qL6JX6D1+fH7G3/ACcR+03/ANlBsv8A1F9Er9B6ACiiigAooooAKKKKACiiigAooooA/9f+/iiiigAooooAKKKKACiiigAooooAKKKKAPyk+Nn/AClJ8Jf9kn1//wBP2i19J182fGz/AJSleEv+yT6//wCn7Ra+k6DaGwUUUUFBRRRQAUUUUAFZHiD/AJF7UP8Ar1m/9FtWvWR4h/5F/UP+vWb/ANFtQB8j/wDBOT/lH/8ABX/sS9G/9JI6+z6+MP8AgnH/AMmAfBX/ALEvRv8A0kjr7PoAs2f/AB+Rf76/zr4H/wCCZ/8AyZh4c/7C/in/ANSTVK++LL/j8i/31/nXwP8A8E0P+TMPDn/YW8U/+pJqlAH3fRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUANZEfh1DfUZpnkW/8AzzX/AL5FS0UAReRb/wDPNf8AvkUeRb/881/75FS0UAReRb/881/75FHkW/8AzzX/AL5FS0UAReRb/wDPNf8AvkUeRb/881/75FS0UAReRb/881/75FHkW/8AzzX/AL5FS0UAReRb/wDPNf8AvkUeRb/881/75FS0UAReRb/881/75FHkW/8AzzX/AL5FS0UAReRb/wDPNf8AvkUeRb/881/75FS0UAReRb/881/75FHkW/8AzzX/AL5FS0UAReRb/wDPNf8AvkUeRb/881/75FS0UAMWKJDlEAPsAKfRRQAV8Q/8FFv+TTNU/wCxh8If+pRpVfb1fEP/AAUW/wCTTdU/7GHwh/6lGlUCex+01FFFBgFFFFABRRRQB+fH7Gw/4yH/AGmz/wBVBsv/AFF9Er9B6/Pj9jb/AJOI/ab/AOyg2X/qL6JX6D0AFFFFABRRRQAUUUUAFFFFABRRRQB//9D+/iiiigAooooAKKKKACiiigAooooAKKKKAPym+Nas3/BUrwkFBP8AxafX+n/Ye0WvpXyZv7p/KtP4/wD7Ef7On7TfjDSfiD8WdO1R9b0Syn06zvtI13WNEnW0upI5poXfSby0MsbSQxvtk3AMoIxXi/8Aw6o/Y8/ueNf/AAv/ABj/APLmg0jOx6r5U39w/l/9ajyZv7p/KvKv+HVH7Hn9zxr/AOF/4x/+XNH/AA6o/Y8/ueNf/C/8Y/8Ay5oH7RHqvlTf3D+X/wBajyZv7p/KvKv+HVH7Hn9zxr/4X/jH/wCXNH/Dqj9jz+541/8AC/8AGP8A8uaA9oj1Xypv7h/L/wCtR5M390/lXlX/AA6o/Y8/ueNf/C/8Y/8Ay5o/4dUfsef3PGv/AIX/AIx/+XNAe0R6r5U39w/l/wDWrJ8QQzf8I/qHyH/j1m7f9M2rgP8Ah1R+x5/c8a/+F/4x/wDlzWN4i/4JW/sfweHr+aNPGm5LeVhnx94wIyEPY6xg/jQHtEePf8E44pT/AME//gqQp/5EvRe3/TpH7V9n+TN/dP5V+cP/AATu/wCCaH7KXjX9g34OeL9eTxeb3U/B2jXM/wBn8ceK7WLzJbONm8uC31aOGJcnhI0VFHCgDivsn/h1R+x5/c8a/wDhf+Mf/lzQHtD16zim+2RfIfvr29/pXwN/wTOilP7GHhwhT/yF/FPb/qZNUr6cH/BKr9j1SGVPGoI7/wDCf+Mf/lzXwj/wTW/4Jsfsr+O/2P8ARPEniJfFxupNY8TRH7N438V2se2DxBqMKYht9VjiU7EG4hQXbLsS7MxA9ofqF5U39w/l/wDWo8mb+6fyryr/AIdUfsef3PGv/hf+Mf8A5c0f8OqP2PP7njX/AML/AMY//LmgPaI9V8qb+4fy/wDrUeTN/dP5V5V/w6o/Y8/ueNf/AAv/ABj/APLmj/h1R+x5/c8a/wDhf+Mf/lzQHtEeq+VN/cP5f/Wo8mb+6fyryr/h1R+x5/c8a/8Ahf8AjH/5c0f8OqP2PP7njX/wv/GP/wAuaA9oj1Xypv7h/L/61Hkzf3T+VeVf8OqP2PP7njX/AML/AMY//Lmj/h1R+x5/c8a/+F/4x/8AlzQHtEeq+VN/cP5f/Wo8mb+6fyryr/h1R+x5/c8a/wDhf+Mf/lzR/wAOqP2PP7njX/wv/GP/AMuaA9oj1Xypv7h/L/61Hkzf3T+VeVf8OqP2PP7njX/wv/GP/wAuaP8Ah1R+x5/c8a/+F/4x/wDlzQHtEeq+VN/cP5f/AFqPJm/un8q8q/4dUfsef3PGv/hf+Mf/AJc0f8OqP2PP7njX/wAL/wAY/wDy5oD2iPVfKl/uH8v/AK1Hkzf3T+VeVf8ADqj9jz+541/8L/xj/wDLmj/h1R+x5/c8a/8Ahf8AjH/5c0B7RHqvlS/3D+X/ANajyZv7p/KvKv8Ah1R+x5/c8a/+F/4x/wDlzR/w6o/Y8/ueNf8Awv8Axj/8uaA9oj1Xypf7h/L/AOtR5M390/lXlX/Dqj9jz+541/8AC/8AGP8A8uaP+HVH7Hn9zxr/AOF/4x/+XNAe0R6r5Uv9w/l/9ajyZv7p/KvKv+HVH7Hn9zxr/wCF/wCMf/lzR/w6o/Y8/ueNf/C/8Y//AC5oD2iPVfKl/uH8v/rUeTN/dP5V5V/w6o/Y8/ueNf8Awv8Axj/8uaP+HVH7Hn9zxr/4X/jH/wCXNAe0R6r5Uv8AcP5f/Wo8mb+6fyryr/h1R+x5/c8a/wDhf+Mf/lzR/wAOqP2PP7njX/wv/GP/AMuaA9oj1Xypf7h/L/61Hkzf3T+VeVf8OqP2PP7njX/wv/GP/wAuaP8Ah1R+x5/c8a/+F/4x/wDlzQHtEeq+VL/cP5f/AFqPJm/un8q8q/4dUfsef3PGv/hf+Mf/AJc0f8OqP2PP7njX/wAL/wAY/wDy5oD2iPVfKl/uH8v/AK1Hkzf3T+VeVf8ADqj9jz+541/8L/xj/wDLmj/h1R+x5/c8a/8Ahf8AjH/5c0B7RHqvlS/3D+X/ANajyZv7p/KvKv8Ah1R+x5/c8a/+F/4x/wDlzR/w6o/Y8/ueNf8Awv8Axj/8uaA9oj1Xypf7h/L/AOtR5M390/lXlX/Dqj9jz+541/8AC/8AGP8A8uaP+HVH7Hn9zxr/AOF/4x/+XNAe0R6r5Uv9w/l/9ajyZv7p/KvKv+HVH7Hn9zxr/wCF/wCMf/lzR/w6o/Y8/ueNf/C/8Y//AC5oD2iPVfKl/uH8v/rUeTN/dP5V5V/w6o/Y8/ueNf8Awv8Axj/8uaP+HVH7Hn9zxr/4X/jH/wCXNAe0R6r5Uv8AcP5f/Wo8mb+6fyryr/h1R+x5/c8a/wDhf+Mf/lzR/wAOqP2PP7njX/wv/GP/AMuaA9oj1Xypf7h/L/61Hkzf3T+VeVf8OqP2PP7njX/wv/GP/wAuaP8Ah1R+x5/c8a/+F/4x/wDlzQHtEeq+VL/cP5f/AFqPJm/un8q8q/4dUfsef3PGv/hf+Mf/AJc0f8OqP2PP7njX/wAL/wAY/wDy5oD2iPVfKl/uH8v/AK1Hkzf3T+VeVf8ADqj9jz+541/8L/xj/wDLmj/h1R+x5/c8a/8Ahf8AjH/5c0B7RHqvlS/3D+X/ANajyZv7p/KvKv8Ah1R+x5/c8a/+F/4x/wDlzR/w6o/Y8/ueNf8Awv8Axj/8uaA9oj1Xypf7h/L/AOtR5M390/lXlX/Dqj9jz+541/8AC/8AGP8A8uaP+HVH7Hn9zxr/AOF/4x/+XNAe0R6r5Uv9w/l/9ajyZv7p/KvKv+HVH7Hn9zxr/wCF/wCMf/lzR/w6o/Y8/ueNf/C/8Y//AC5oD2iPVfKl/uH8v/rUeTN/dP5V5V/w6o/Y8/ueNf8Awv8Axj/8uaP+HVH7Hn9zxr/4X/jH/wCXNAe0R6r5Uv8AcP5f/Wo8mb+6fyryr/h1R+x5/c8a/wDhf+Mf/lzR/wAOqP2PP7njX/wv/GP/AMuaA9oj1Xypf7h/L/61Hkzf3T+VeVf8OqP2PP7njX/wv/GP/wAuaP8Ah1R+x5/c8a/+F/4x/wDlzQHtEeq+VL/cP5f/AFqPJm/un8q8q/4dUfsef3PGv/hf+Mf/AJc0f8OqP2PP7njX/wAL/wAY/wDy5oD2iPVfKl/uH8v/AK1Hkzf3T+VeVf8ADqj9jz+541/8L/xj/wDLmj/h1R+x5/c8a/8Ahf8AjH/5c0B7RHqvlS/3D+X/ANajyZv7p/KvKv8Ah1R+x5/c8a/+F/4x/wDlzR/w6o/Y8/ueNf8Awv8Axj/8uaA9oj1Xypf7h/L/AOtR5M390/lXlX/Dqj9jz+541/8AC/8AGP8A8uaP+HVH7Hn9zxr/AOF/4x/+XNAe0R6r5Uv9w/l/9ajyZv7p/KvKv+HVH7Hn9zxr/wCF/wCMf/lzR/w6o/Y8/ueNf/C/8Y//AC5oD2iPVfKl/uH8v/rUeTN/dP5V5V/w6o/Y8/ueNf8Awv8Axj/8uaP+HVH7Hn9zxr/4X/jH/wCXNAe0R6r5Uv8AcP5f/Wo8mb+6fyryr/h1R+x5/c8a/wDhf+Mf/lzR/wAOqP2PP7njX/wv/GP/AMuaA9oj1Xypf7h/L/61Hkzf3T+VeVf8OqP2PP7njX/wv/GP/wAuaP8Ah1R+x5/c8a/+F/4x/wDlzQHtEeq+VL/cP5f/AFqPJm/un8q8q/4dUfsef3PGv/hf+Mf/AJc0f8OqP2PP7njX/wAL/wAY/wDy5oD2iPVfKl/uH8v/AK1Hkzf3T+VeVf8ADqj9jz+541/8L/xj/wDLmj/h1R+x5/c8a/8Ahf8AjH/5c0B7RHqvlS/3D+X/ANajyZv7p/KvKv8Ah1R+x5/c8a/+F/4x/wDlzR/w6o/Y8/ueNf8Awv8Axj/8uaA9oj1Xypf7h/L/AOtR5M390/lXlX/Dqj9jz+541/8AC/8AGP8A8uaP+HVH7Hn9zxr/AOF/4x/+XNAe0R6r5Uv9w/l/9ajyZv7p/KvKv+HVH7Hn9zxr/wCF/wCMf/lzR/w6o/Y8/ueNf/C/8Y//AC5oD2iPVfKl/uH8v/rUeVL/AHT+X/1q8q/4dUfsef3PGv8A4X/jH/5c0f8ADqj9jz+541/8L/xj/wDLmgPaI9V8qX+4fy/+tR5Uv90/l/8AWryr/h1R+x5/c8a/+F/4x/8AlzR/w6o/Y8/ueNf/AAv/ABj/APLmgPaI9V8qX+4fy/8ArUeVL/dP5f8A1q8q/wCHVH7Hn9zxr/4X/jH/AOXNH/Dqj9jz+541/wDC/wDGP/y5oD2iPVfKl/uH8v8A61HlS/3T+X/1q8q/4dUfsef3PGv/AIX/AIx/+XNH/Dqj9jz+541/8L/xj/8ALmgPaI9V8qX+4fy/+tR5Uv8AdP5f/Wryr/h1R+x5/c8a/wDhf+Mf/lzR/wAOqP2PP7njX/wv/GP/AMuaA9oj1Xypf7h/L/61HlS/3T+X/wBavKv+HVH7Hn9zxr/4X/jH/wCXNH/Dqj9jz+541/8AC/8AGP8A8uaA9oj1Xypf7h/L/wCtXxB/wUXjkX9kzVCykY8Q+EO3/U0aVX0V/wAOqP2PP7njX/wv/GP/AMuaaP8AglF+xXLdWlzqmm+KNTSyu7W+jt9R8aeKr61aeynS5t2ltrnVpIJRHNGkgWRGXcoyDigTmj9G6KKKDMKKKKACiiigD8+P2Nh/xkP+02f+qg2X/qL6JX6D1+fH7G3/ACcR+03/ANlBsv8A1F9Er9B6ACiiigAooooAKKKKACiiigAooooA/9H+/iiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigArC8Uf8izqP/XrN/6Aa3awvFH/ACLOo/8AXrN/6AaAPjP/AIJhf8o5vgb/ANiPof8A6RRV9018Lf8ABML/AJRz/A3/ALEfQ/8A0iir7poAK/Or/glL/wAmQaB/2HfFn/qS6nX6K1+dX/BKX/kyHQP+w74s/wDUl1OgD9FaKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigD8+P2Nh/wAZD/tNn/qoNl/6i+iV+g9fnx+xt/ycR+03/wBlBsv/AFF9Er9B6ACiiigAooooAKKKKACiiigAooooA//S/v4oor8YdP8A+CwGpeJ5tSuvh98BPG2u6TY6pqelRahFf+GoI7l9LvZ7CaSOO51iKZUM1u+3zI1YrgkDNcOPzPDYWCqYqooJ6Xk0tfmehl2U4rGTdPCU5TaV7RTbt30P2eor8cv+HrvxF/6Nu8df+DTwn/8ALyj/AIeu/EX/AKNu8df+DTwn/wDLyvJ/1xyj/oKp/wDgcf8AM9n/AFHzr/oEqf8AgEv8j9jaK/HL/h678Rf+jbvHX/g08J//AC8o/wCHrvxF/wCjbfHX/g08J/8Ay8o/1xyj/oKp/wDgcf8AMP8AUfOv+gSp/wCAS/yP2Nor8cv+HrvxF/6Nu8df+DTwn/8ALyj/AIeu/EX/AKNt8df+DTwn/wDLyj/XHKP+gqn/AOBx/wAw/wBR86/6BKn/AIBL/I/Y2ivxy/4eu/EX/o27x1/4NPCf/wAvKP8Ah678Rf8Ao23x1/4NPCf/AMvKP9cco/6Cqf8A4HH/ADD/AFHzr/oEqf8AgEv8j9jaK/HL/h678Rf+jbvHX/g08J//AC8o/wCHrvxF/wCjbfHX/g08J/8Ay8o/1xyj/oKp/wDgcf8AMP8AUfOv+gSp/wCAS/yP2Nor8cv+HrvxF/6Nu8df+DTwn/8ALyj/AIeu/EX/AKNt8df+DTwn/wDLyj/XHKP+gqn/AOBx/wAw/wBR86/6BKn/AIBL/I/Y2ivxy/4eu/EX/o27x1/4NPCf/wAvKP8Ah678Rf8Ao23x1/4NPCf/AMvKP9cco/6Cqf8A4HH/ADD/AFHzr/oEqf8AgEv8j9jaK/HL/h678Rf+jbvHX/g08J//AC8o/wCHrvxF/wCjbfHX/g08J/8Ay8o/1xyj/oKp/wDgcf8AMP8AUfOv+gSp/wCAS/yP2Nor8cv+HrvxF/6Nt8df+DTwn/8ALyj/AIeu/EX/AKNt8df+DTwn/wDLyj/XHKP+gqn/AOBx/wAw/wBR86/6BKn/AIBL/I/Y2sLxR/yLOo/9es3/AKAa/Mz4L/8ABTyT4n/H3wj8AfG3wg8V+Brvxq1/Hpuoand6HdWpl060kvZY3GnaldzKTFG20mPbnAJFfpn4o/5FnUf+vWb/ANANezgsfQxNNVcPNSj3TuvvR4WPy/EYWo6OJg4SXRpp/cz4z/4Jhf8AKOb4G/8AYj6H/wCkUVfdNfC3/BML/lHP8Df+xH0P/wBIoq+6a6zjCvzq/wCCUv8AyZBoH/Yd8Wf+pLqdforX51f8Epf+TIdA/wCw74s/9SXU6AP0VooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAor5I/ag/ba+BX7Id/4b0f4utrM1/wCLTef2XZ6Ho2oa1czLYLE9y5h0+Cd0SMTR5ZwB83Br5p/4fBfso/8AQD+I3/hBeJv/AJX1xV8yw1KXJVqRi+zaX5nfhsqxVaPPRpSku6Tf5I/U2ivyy/4fBfsof9AP4jf+EF4n/wDldR/w+C/ZR/6AfxG/8ILxN/8AK+sf7bwX/P6P/gS/zOj+wMf/AM+J/wDgL/yP1Nor8sv+HwX7KH/QD+I3/hBeJ/8A5XUf8Pgv2UP+gH8Rv/CC8Tf/ACvo/tvBf8/o/wDgS/zD+wMf/wA+Jf8AgL/yP1Nor8sv+HwX7KH/AEA/iN/4QXif/wCV1H/D4L9lD/oB/Eb/AMILxN/8r6P7bwX/AD+j/wCBL/MP7Ax//PiX/gL/AMj9Ta5/xX4s8MeBPDN/408a6jbaTo+lQSXV7e3kqw29vBEpaSWWVyFREUEszEAAZJxX5pf8Pgv2UP8AoB/Eb/wgvE//AMrqxfEn/BV/9jPxh4ev/CXinwv8Qb/TNUt5bS7tZ/AHiZ4poJ0McsbqdOwyujFWB6g0f23gv+f0f/Al/mH9gY//AJ8S/wDAX/kcX+wP+2Z+yL8Qf2rPj/4O8BfE/wAK61q/ifx5az6PZWOr2lxcX8UXhjSFkktY45WaZEaGUM0YYAxuCflOP2br+AP/AIIlfs4fsm/8E0v26PjP+0p470Hx/f6al3NpHw0c+BvEc0sWi3hE89zMBYFo7gIUs/mCsQkzY2SKT/Vf/wAPgv2UP+gH8Rv/AAgvE3/yvo/tvBf8/o/+BL/MP7Ax/wDz4n/4C/8AI/U2ivyy/wCHwX7KH/QD+I3/AIQXif8A+V1H/D4L9lD/AKAfxG/8ILxN/wDK+j+28F/z+j/4Ev8AMP7Ax/8Az4l/4C/8j9TaK/LL/h8F+yh/0A/iN/4QXif/AOV1H/D4L9lD/oB/Eb/wgvE3/wAr6P7bwX/P6P8A4Ev8w/sDH/8APiX/AIC/8j9TaK/LL/h8F+yh/wBAP4jf+EF4n/8AldR/w+C/ZQ/6AfxG/wDCC8Tf/K+j+28F/wA/o/8AgS/zD+wMf/z4l/4C/wDI/U2ivyy/4fBfsof9AP4jf+EF4n/+V1KP+CwP7KTHauhfEYk/9SF4m/8AlfR/beC/5/R/8CX+Yf2Bj/8AnxL/AMBf+R+plFeW/BH4y/D/APaI+EXhz45/Cm7kvvDfiuwh1LTbiWGW3eS3nXcjNDMqSRtg8q6hgeCBXqVemeQ0f//T/v4r+W79kf8A5JBef9jd4z/9SjVa/qRr+W79kf8A5JBef9jd4z/9SjVa/FPHT/kWUv8AGv8A0mR+7/R+/wCRtW/69v8A9KifTNFFFfy0f1ygooooBBRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQBw3hH/AJSEfs4/9hLxT/6jt3X9Cnij/kWdR/69Zv8A0A1/PX4R/wCUhH7OP/YS8U/+o7d1/Qp4o/5FnUf+vWb/ANANf2B4Pf8AIip+svzZ/E/jX/yUFT0j/wCko+M/+CYX/KOb4G/9iPof/pFFX3TXwt/wTC/5Rz/A3/sR9D/9Ioq+6a/UD8nCvzq/4JS/8mQaB/2HfFn/AKkup1+itfnV/wAEpf8AkyHQP+w74s/9SXU6AP0VooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigD8TP+Cmv/ACeN+zx/2DfHH/ojS686r0X/AIKa/wDJ437PH/YN8cf+k+l151X8leNP/I6/7cj+bP7L8Cf+RH/2/L8kFFFFfkh+yhRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAVNb/8fEf+8P51DU1v/wAfEf8AvD+dJ7Aj7V/4JB/8owfgX/2J+nf+ihX6O1+cX/BIP/lGD8C/+xP07/0UK/R2v9DaXwI/zTrfGz//1P7+K/lu/ZH/AOSQXn/Y3eM//Uo1Wv6ka/lu/ZH/AOSQXn/Y3eM//Uo1WvxTx0/5FlL/ABr/ANJkfu/0fv8AkbVv+vb/APSon0zRRRX8tH9coKKKKAR9AWvwFvr/AODX/C2LG/EkgSSZrLy/m8qKTY7Bs87R8x+XGK87+GvgS9+JHi+28LWcot1kDSSzsMrFEgyzkcfQcjkivq7wX4tt/CXgj4bHU8HT9Sk1Kwu1b7piuJFXJ9g2Cfauav8AwhP8AvB/iGS6JXUNcvv7JsG6N9iVg8kg9nUhT74r9Tr8L4N+wxMI2pwjF1dXvyRmvTnvyq2zR+Q4bi3HL2+FnK9Sc5Kk7LRc8oP15Lc7vumfPfxQ8CN8OPHF34NW5+2/ZxGVlCbN3mIrj5cn1x1rh57K9tXWK6hkiZ+VDqVJ+gIGa+0PG+n+L9Q/ai1A+CZoLS7t7OOVrm5UNHbxC3QPIQQ3K544zk07xfrL6z8CtT1abxMvie80jU7WS3vFgaF7d2YAhWcAsD1BHT8q4cfwhRdbFODcVCVTlWlmoa2Tc+Zu3aLtpd66ejl/GtZUcJGolN1I0+Z63Tnom0ocqV+8lfWy01+KRbzmX7OI28zptwd35da6rUNG8OW3g/T9Ys9QebU7iWRLi0MTKsSKTtYSHhieOB0z7V9169Jc2Xw7m+Pdnp3k+L7rSoo51BXMEcjFDeCP7wLIOD2A9jXzvrZJ+Avghm5J1a7Jz3/eVljuEY4RTUp8zdPnV4te65xUXurSd3dO6W2remmA41njHCUYcqVTkdpJ3koTcls7xVlZqzlvolZ/O32K92PJ5Mm2LhztOF/3jjj8aY1tcpAt08biJzhXKnaSPQ9DX6Nap8TPE/8Aw1Avw4VohoszeTPaiNds3mW+9nkJGWbOB16DGK858NeO9d+IH/Ce+C/EPktpNlpl1JaWkcaJHbm1bbH5eBkYH681tX4KwsakqUK7clKcF7iS5oK/82z2vvfpYxocd4x041p4dKLjCb9+75Zuy+x8S3ttbrc+Lbezu7ssLSJ5SoydilsD3wDiq5BBIPUcGvrTxb408R/CL4c+DdO+Hk405NTsft91cxopeediMqzMDkJnp6YHQVy/7RsFtcan4e8Utbpa32taRDdXsaLtBmOQXK9i39K8XMOG6dGhOUajc4KDkrWXvpWs762ur3S6tbHu5bxRUr4iEJU0oVHNRfNd+42nzKyteztZu2ie5850UUV8mfZBRRRQBw3hH/lIR+zj/wBhLxT/AOo7d1/Qp4o/5FnUf+vWb/0A1/ND4t0L4q+I/wBs/wDZ60n4L+IrHwr4gk1bxIYNR1HTW1a3jRdAujIrWi3VkXLr8oPnrtPOG6V+tV98Dv8AgpPf2U1jN8efBwSZGjbHw/nzhhg4z4kx3r+wPB7/AJEVP1l+bP4o8bP+Sgqekf8A0lHUf8Ewv+Uc3wN/7EfQ/wD0iir7prwz9mP4Kp+zf+zr4G+AEepNrK+C9DsNFF+0QgNz9igSHzjEGcR79u7aHbbnGT1r3Ov1A/Jgr86v+CUv/JkGgf8AYd8Wf+pLqdforX5W/CX9jP8Abi+APg5vhj8H/jl4btvDcGo6pfWUGo+B5by6hXU7+41BopLhNft1lKPcMoYRJlQOM0AfqlRX58f8KZ/4KWf9F48G/wDhv5//AJpaP+FM/wDBSz/ovHg3/wAN/P8A/NLQB+g9Ffnx/wAKZ/4KWf8ARePBv/hv5/8A5paP+FM/8FLP+i8eDf8Aw38//wA0tAH6D0V+fH/Cmf8AgpZ/0Xjwb/4b+f8A+aWj/hTP/BSz/ovHg3/w38//AM0tAH6D0V+fH/Cmf+Cln/RePBv/AIb+f/5paP8AhTP/AAUs/wCi8eDf/Dfz/wDzS0AfoPRX58f8KZ/4KWf9F48G/wDhv5//AJpaP+FM/wDBSz/ovHg3/wAN/P8A/NLQB+g9Ffnx/wAKZ/4KWf8ARePBv/hv5/8A5paP+FM/8FLP+i8eDf8Aw38//wA0tAH6D0V+fH/Cmf8AgpZ/0Xjwb/4b+f8A+aWj/hTP/BSz/ovHg3/w38//AM0tAH6D0V+fH/Cmf+Cln/RePBv/AIb+f/5paP8AhTP/AAUs/wCi8eDf/Dfz/wDzS0AfoPRX58f8KZ/4KWf9F48G/wDhv5//AJpaP+FM/wDBSz/ovHg3/wAN/P8A/NLQB+g9Ffnx/wAKZ/4KWf8ARePBv/hv5/8A5paP+FM/8FLP+i8eDf8Aw38//wA0tAH6D0V+fH/Cmf8AgpZ/0Xjwb/4b+f8A+aWj/hTP/BSz/ovHg3/w38//AM0tAH6D0V+fH/Cmf+Cln/RePBv/AIb+f/5paP8AhTP/AAUs/wCi8eDf/Dfz/wDzS0AfoPRX58f8KZ/4KWf9F48G/wDhv5//AJpaP+FM/wDBSz/ovHg3/wAN/P8A/NLQB+g9Ffnx/wAKZ/4KWf8ARePBv/hv5/8A5paP+FM/8FLP+i8eDf8Aw38//wA0tAH6D0V+fH/Cmf8AgpZ/0Xjwb/4b+f8A+aWj/hTP/BSz/ovHg3/w38//AM0tAH6D0V+fH/Cmf+Cln/RePBv/AIb+f/5paP8AhTP/AAUs/wCi8eDf/Dfz/wDzS0AfoPRX58f8KZ/4KWf9F48G/wDhv5//AJpaP+FM/wDBSz/ovHg3/wAN/P8A/NLQB+g9Ffnx/wAKZ/4KWf8ARePBv/hv5/8A5paP+FM/8FLP+i8eDf8Aw38//wA0tAH6D0V+fH/Cmf8AgpZ/0Xjwb/4b+f8A+aWj/hTP/BSz/ovHg3/w38//AM0tAH6D0V+fH/Cmf+Cln/RePBv/AIb+f/5paP8AhTP/AAUs/wCi8eDf/Dfz/wDzS0AfoPRX58f8KZ/4KWf9F48G/wDhv5//AJpaP+FM/wDBSz/ovHg3/wAN/P8A/NLQB+g9Ffnx/wAKZ/4KWf8ARePBv/hv5/8A5paP+FM/8FLP+i8eDf8Aw38//wA0tAH6D0V+fH/Cmf8AgpZ/0Xjwb/4b+f8A+aWj/hTP/BSz/ovHg3/w38//AM0tAH6D0V+fH/Cmf+Cln/RePBv/AIb+f/5paP8AhTP/AAUs/wCi8eDv/Dfz/wDzS0AfoPRX58f8KZ/4KWf9F48G/wDhv5//AJpaP+FM/wDBSz/ovHg7/wAN/P8A/NLQB+g9Ffnx/wAKZ/4KWf8ARePBv/hv5/8A5paP+FM/8FLP+i8eDv8Aw38//wA0tAH6D0V+fH/Cmf8AgpZ/0Xjwb/4b+f8A+aWj/hTP/BSz/ovHg7/w38//AM0tAH6D0V+fH/Cmf+Cln/RePBv/AIb+f/5paP8AhTP/AAUs/wCi8eDv/Dfz/wDzS0AfoPRX58f8KZ/4KWf9F48G/wDhv5//AJpaP+FM/wDBSz/ovHg7/wAN/P8A/NLQB+g9Ffnx/wAKZ/4KWf8ARePBv/hv5/8A5paP+FM/8FLP+i8eDv8Aw38//wA0tAH6D0V+fH/Cmf8AgpZ/0Xjwb/4b+f8A+aWj/hTP/BSz/ovHg7/w38//AM0tAH6D0V+fH/Cmf+Cln/RePBv/AIb+f/5paP8AhTP/AAUs/wCi8eDv/Dfz/wDzS0AfoPRX58f8KZ/4KWf9F48G/wDhv5//AJpaP+FM/wDBSz/ovHg7/wAN/P8A/NLQB+g9Ffnx/wAKZ/4KWf8ARePBv/hv5/8A5paP+FM/8FLP+i8eDv8Aw38//wA0tAH6D0V+fH/Cmf8AgpZ/0Xjwb/4b+f8A+aWj/hTP/BSz/ovHg7/w38//AM0tAH6D0V+fH/Cmf+Cln/RePBv/AIb+f/5paP8AhTP/AAUs/wCi8eDv/Dfz/wDzS0AfoPRX58f8KZ/4KWf9F48G/wDhv5//AJpaP+FM/wDBSz/ovHg7/wAN/P8A/NLQB+g9Ffnx/wAKZ/4KWf8ARePBv/hv5/8A5paP+FM/8FLP+i8eDv8Aw38//wA0tAH6D0V+fH/Cmf8AgpZ/0Xjwb/4b+f8A+aWj/hTP/BSz/ovHg7/w38//AM0tAH6D0V+fH/Cmf+Cln/RePBv/AIb+f/5paP8AhTP/AAUs/wCi8eDv/Dfz/wDzS0AfoPRX58f8KZ/4KWf9F48G/wDhv5//AJpaP+FM/wDBSz/ovHg7/wAN/P8A/NLQB+g9Ffnx/wAKZ/4KWf8ARePBv/hv5/8A5paP+FM/8FLP+i8eDv8Aw38//wA0tAH6D0V+fH/Cmf8AgpZ/0Xjwb/4b+f8A+aWj/hTP/BSz/ovHg7/w38//AM0tAH6D0V+fH/Cmf+Cln/RePBv/AIb+f/5paP8AhTP/AAUs/wCi8eDv/Dfz/wDzS0AfoPRX58f8KZ/4KWf9F48G/wDhv5//AJpaP+FM/wDBSz/ovHg7/wAN/P8A/NLQB+g9Ffnx/wAKZ/4KWf8ARePBv/hv5/8A5paP+FM/8FLP+i8eDv8Aw38//wA0tAH6D0V+fH/Cmf8AgpZ/0Xjwb/4b+f8A+aWj/hTP/BSz/ovHg7/w38//AM0tAH6D0V+fH/Cmf+Cln/RePBv/AIb+f/5paP8AhTP/AAUs/wCi8eDv/Dfz/wDzS0AfKf8AwU1/5PG/Z4/7Bvjj/wBEaXXnVec/tf8Agn9p/wAJfto/AqT9ob4gaL42gn0nxmLFNJ8OyaE1u6w6Z5jSM+p6h5wcEAKBHtxnLZ49Gr+SvGn/AJHX/bkfzZ/ZXgUv+EP/ALfl+SCiiivyQ/ZjpfBvhm58Z+KrDwpZyLDLfzLCsjglVJ7kDk17QP2ek1PULnQPCninS9S1a2Z1ayy8MjNGcMq7xgkEfT3rivgV/wAli8OZ/wCf1P5GpLpdVf48TLom77Ydek8rZ97d9pP+T7V9jlGGwiwkKuIpc7lU5d2mlZbW0vd9Uz4jOsVjHjZ0cPW5FGnz6pNN3a1ur2sujR5bqFjeaVfTabqUbQXFu5jkjcYZXU4II9QagjRpJVhX7zEKAeOScCvtnxHF4km/aD8W6j4BGnQi1ija6vtQVWhtMIm9wWBAckEdCetSePddv7PwT4Z+I8Wo6dq2t22qyWZ1C0gXy3jdG+R1ZFDMv+7gdRzXZV4MhH20nUaVNy+yrNRlytp8ybfXRW6cyOGlx1OfsIqmr1IxfxO6lKHMk/daS6au/VRaPkLxd4U1jwP4iuPC3iBUS8tdvmKjBwNyhhgj2IrnM9Pfp719pfG3xT438QfHhPAOlW9tfpaXlvLaWs0MZV5PJDYlfAZk5JIJxj6V1N9/a3iLwL4w0Tx3rOja1Pptg9xFBYRgSWc0R7OsaDAPynBJ7dzWtfgyjPFYilh5ytCUkrxVm4puzfN5W0TfVpIxw/HNaGEw1bE04t1IwbtJ3Sk0rpcr731aXRNs+D9PtGv7+DT4yFM8iRgnoC7Bc/rXTfEDwdcfD/xjfeD7udbmSxZVMiAqrbkV+AeR97FZHhv/AJGPTv8Ar6h/9DWvVP2kv+S2eIM/89Y//RMdfMUsFTeXTxDXvKcV8nGbf4pH1lXHVFmlPDJ+64Sl81KCX4NniAIwSOg4NeneGfAllrvw48SeN5riSObRGtVjiUDY/nuVO4kZ4HTFfXPw+k8b6brOg+FvGl/oOnWN7FFH/YfkqZ5onTAZsI2JH6/M4yc1zfwh8HeHNY0rx/4I1a5+waa2r2ttv7hEuJAiAngFsBQT0Jr7HLuCE6sE3fnVRWklG0lTck/iel2t7PTVHxGZ8fP2U2ly8jpu8W5Xi6ijJfCruyfw3WujufDmR2PSjIPFfeugw2+vfFvVtK8SaZa6TY+CtOmfT7KVN0KFSoWeXaCZQR85PPbGTknifHPifwx4i+GeqWnjDxFpeuazE8UumPZW7Qyp8wEkZPloChXpn+grza3BkYUZ1ZVkrc1rpK/JvvJSV3dR91t9eW56lDjmdStCkqDd+S9m5W59touLsrOXvJLW3Na58g0UUV8KfoQVNb/8fEf+8P51DU1v/wAfEf8AvD+dJ7Afav8AwSD/AOUYPwL/AOxP07/0UK/R2vzi/wCCQf8AyjB+Bf8A2J+nf+ihX6O1/obS+BH+adb42f/V/v4r+W79kf8A5JBef9jd4z/9SjVa/qRr+W79kf8A5JBef9jd4z/9SjVa/FPHT/kWUv8AGv8A0mR+7/R+/wCRtW/69v8A9KifTNFFFfy0f1yFFFFAHX6v421nWvCek+DbsRi00ZpmtyikPmY7m3HJzg9MAVr+PPin4t+I0+nXHiWRHbTIRFCEUqDggl2BJyzYGTx06V5zRXoSzXEuEoObtJRTXdR+G/p0PNhk+FU41FTV4uTTts5fFb16nrn/AAuvxqvxFl+Jsf2db+eIQSxiPMEkQQIUZCxJBCjPzda0NV+OviXVNDfwv/ZumW2ltPDcfZLe28uPdC4fnDZIcgB85yBgYrxOiutcR45KUfau0m2/Pm3+/r3ON8L5e3CXsVeKSXly/D93TsexQ/HLx3F44u/HcjwzT30BtZreRCbZoMYEfl7vujqOeueuTXJXnjzWbzwvpvhB0hSz0q5kuoNqkNvlbcQxJOVB6DA471xVFYVc7xdROM6jd7317tN/e0n6o3pZDgoOMoUkrWtp2TS+5Nr0Z6dP8WfFFx8TF+K0iQf2mrrIFCHysrH5Y+Xdn7v+11rM8P8AxD1/w1e6xqGnLCZNbt5ra43qSAk53OUAYYOemc1wlFL+2cVz+09o73cvnLd+rKWR4Tk9n7NWso/9ux1S9F0PYvC3xt8UeGvD0Hha5tLDV7G0YvbR6hAJvJJOTsOQQM84Oa4Txh4w1/x3r0viPxLN51zKAvACqiL91EUcBR2H9a5iijEZziqtFYerUbirWXpt9y0XboGFyPB0a0sTSppTd7v11f3vV9+oUUUV5h6oUUUUAcN4R/5SEfs4/wDYS8U/+o7d1/RjX853hH/lIR+zj/2EvFP/AKjt3X9GNf2B4Pf8iKn6y/Nn8UeNf/JQVPSP/pKCiiiv1A/JgooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooA/Ez/gpr/yeN+zx/wBg3xx/6I0uvOq9F/4Ka/8AJ437PH/YN8cf+k+l151X8leNP/I6/wC3I/mz+yvApf8ACH/2/L8kFFFFfkh+zHReEfEt54N8T2PinT0SWawlWZEkztYr2OMHH0r2Z/2itUs7u51fwx4e0jSdSui5e9hhLTgucsQXYgEk88GvneivXwGfYvCw9nh58qvfpvtdPdP0PFzHh3BYuoquJp8ztbrZrezV7Netz0zwT8UdY8H3upzXdvDq1vrUZjvoLvcyzZJbcWBBDZJOfetHxl8Xr7xb4WtPB1vpdlpVjYXX2m3S0Vl2nYVwdxO7JJYseSa8ioohn2LjQeGU/dd102bu1e17X1tffUJ8O4KWIWKdP31Z313Ssna9r20va9tNj3rWvj7rWra7p3jCLSbC11uwlSR76NW8yfYhj2uCcbWU8ge2CMVLefHuc6XrGk6J4d0zTE1yGSK7eBX8xml6vuLcYycLjbk14BRXW+LMwbk3U+LfRdVZvbdrRvd9bnGuDsttFKlpG1tXsndLfZPVJ6LpYtWV3JYXsN9CAXgkWRQemUIIz+Ve7+J/jvZ+LZbu+1nwjo0l5eIVe5KOZQSu0MGLfeUYx9BXz9RXBgs4xGHhKlSlaMt1ZPa/dPuz0cfkmGxNSNWtG8o6Jptb27NdkfRtp+0fq0GoWPiK50HTLnWrKOOH+0JEcytHHx0DbQxGQWAyM8Yrz1fifqy6N4l0UW0QTxPcpczsC26JkkaQCPnpk45zXmlFdVfibHVFadS+/RfaXK76atrRt3fmceH4Uy+k706aW3V/ZfNG2uiT1SWnke8J+0J4yS/0nW2gtpNQ0y3a0luXVma8t2x+7uFzhgMcEc55zXN+J/iPo2u6NNpOkeF9L0lrllaSe3V2l+U7sIXJCAnqB24ryuilW4kxtSDp1Kl0+6T6JbtXV0ldqzfUqhwvgKU41KVOzXZtdW1onZpNuyd0ulgooorwz3gqa3/4+I/94fzqGprf/j4j/wB4fzpPYZ9q/wDBIP8A5Rg/Av8A7E/Tv/RQr9Ha/OL/AIJB/wDKMH4F/wDYn6d/6KFfo7X+htL4Ef5p1vjZ/9b+/iv5R/2VPG/gnTfhVf2Opa1p9tPH4u8Zh45bqGN1P/CUaocFWYEceor+rivkzWv2B/2FfEus3fiLxF8FvAeoahfzPcXN1c+HNMlmmmlYvJJJI9uWd3YlmZiSSSSc18Xxxwes6w0MM6nJyy5r2v0atuu591wBxq8ixc8UqfPzR5bXt1Tvs+x+RX/Cxfh5/wBDBpf/AIGwf/F0f8LF+Hf/AEMGl/8AgbB/8XX6xf8ADuz/AIJ+f9EK+Hv/AITGlf8AyNR/w7s/4J+f9EK+Hv8A4TGlf/I1fmH/ABASH/QV/wCSf/bH61/xMRL/AKBP/J//ALU/J3/hYvw8/wChg0v/AMDYP/i6P+Fi/Dv/AKGDS/8AwNg/+Lr9Yv8Ah3Z/wT8/6IV8Pf8AwmNK/wDkaj/h3Z/wT8/6IV8Pf/CY0r/5Go/4gJD/AKCv/JP/ALYP+JiJf9An/k//ANqfk7/wsX4ef9DBpf8A4Gwf/F0f8LF+Hf8A0MGl/wDgbB/8XX6xf8O7P+Cfn/RCvh7/AOExpX/yNR/w7s/4J+f9EK+Hv/hMaV/8jUv+ICQ/6Cv/ACT/AO2D/iYif/QJ/wCT/wD2p+Tv/Cxfh5/0MGl/+BsH/wAXR/wsX4d/9DBpf/gbB/8AF1+sX/Duz/gn5/0Qr4e/+ExpX/yNR/w7s/4J+f8ARCvh7/4TGlf/ACNR/wAQEh/0Ff8Akn/2wf8AExE/+gT/AMn/APtT8nf+Fi/Dz/oYNL/8DYP/AIuj/hYvw7/6GDS//A2D/wCLr9Yv+Hdn/BPz/ohXw9/8JjSv/kaj/h3Z/wAE/P8AohXw9/8ACY0r/wCRqP8AiAkP+gr/AMk/+2D/AImIn/0Cf+T/AP2p+Tv/AAsX4ef9DBpf/gbB/wDF0f8ACxfh3/0MGl/+BsH/AMXX6xf8O7P+Cfn/AEQr4e/+ExpX/wAjUf8ADuz/AIJ+f9EK+Hv/AITGlf8AyNR/xASH/QV/5J/9sH/ExE/+gT/yf/7U/J3/AIWL8PP+hg0v/wADYP8A4uj/AIWL8O/+hg0v/wADYP8A4uv1i/4d2f8ABPz/AKIV8Pf/AAmNK/8Akaj/AId2f8E/P+iFfD3/AMJjSv8A5Go/4gJD/oK/8k/+2D/iYif/AECf+T//AGp+Tv8AwsX4ef8AQwaX/wCBsH/xdH/Cxfh3/wBDBpf/AIGwf/F1+sX/AA7s/wCCfn/RCvh7/wCExpX/AMjUf8O7P+Cfn/RCvh7/AOExpX/yNR/xASH/AEFf+Sf/AGwf8TET/wCgT/yf/wC1Pyd/4WL8PP8AoYNL/wDA2D/4uj/hYvw7/wChg0v/AMDYP/i6/WL/AId2f8E/P+iFfD3/AMJjSv8A5Go/4d2f8E/P+iFfD3/wmNK/+RqP+ICQ/wCgr/yT/wC2D/iYif8A0Cf+T/8A2p+O/wAPPFPhfXf+Ch37OdvoWp2d9Imo+KGZLeeOVgP+EeuxkhGJAr+k6vnP4c/se/skfB7xTF45+Efwt8IeFtbgR44tQ0jRLCxukSRSrqs1vCkgDKSGAbBHBr6Mr9d4Q4bWU4GOCU+azbva27vtd/mfi3GvFDzjMJY5w5LpK177K29kFFFFfTnygUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAfid/wUzVn/AGx/2eAgJP8AZvjjp/1w0uvPvs9x/cb8jX6n/tJ/sX/s3/tdSaDcfH7QJtXn8Mtctpk9tqOoabPb/bFRbgLLp9xbSFZFjQMrMVO0cV8z/wDDnP8A4J9f9Crrf/hXeJ//AJa1+Occ+F9XN8d9bhWUVZKzTe1/PzP27w/8WaOS4D6nUoub5m7ppb28vI+SPs0/9xvyNH2e4/uN+Rr63/4c5/8ABPv/AKFXW/8AwrvE/wD8taP+HOf/AAT6/wChV1v/AMK7xP8A/LWvjv8AiA+I/wCghf8AgL/zPuH9IbC/9Asv/Al/kfJH2af+435Gj7Pcf3G/I19b/wDDnP8A4J9/9Crrf/hXeJ//AJa0f8Oc/wDgn1/0Kut/+Fd4n/8AlrR/xAfEf9BC/wDAX/mD+kNhf+gWX/gS/wAj5I+zT/3G/I0fZ7j+435Gvrf/AIc5/wDBPv8A6FXW/wDwrvE//wAtaP8Ahzn/AME+v+hV1v8A8K7xP/8ALWj/AIgPiP8AoIX/AIC/8wf0hsL/ANAsv/Al/kfJH2af+435Gj7Pcf3G/I19b/8ADnP/AIJ9/wDQq63/AOFd4n/+WtH/AA5z/wCCfX/Qq63/AOFd4n/+WtH/ABAfEf8AQQv/AAF/5g/pDYX/AKBZf+BL/I+SPs0/9xvyNH2e4/uN+Rr63/4c5/8ABPv/AKFXW/8AwrvE/wD8taP+HOf/AAT6/wChV1v/AMK7xP8A/LWj/iA+I/6CF/4C/wDMH9IbC/8AQLL/AMCX+R8kfZp/7jfkaPs9x/cb8jX1v/w5z/4J9/8AQq63/wCFd4n/APlrR/w5z/4J9f8AQq63/wCFd4n/APlrR/xAfEf9BC/8Bf8AmD+kNhf+gWX/AIEv8j5I+zT/ANxvyNH2e4/uN+Rr63/4c5/8E+/+hV1v/wAK7xP/APLWj/hzn/wT6/6FXW//AArvE/8A8taP+ID4j/oIX/gL/wAwf0hsL/0Cy/8AAl/kfJH2af8AuN+Ro+z3H9xvyNfW/wDw5z/4J9/9Crrf/hXeJ/8A5a0f8Oc/+CfX/Qq63/4V3if/AOWtH/EB8R/0EL/wF/5g/pDYX/oFl/4Ev8j5I+zXH/PNvyNTW9vOLhPkb7w7H1r6x/4c5/8ABPv/AKFXW/8AwrvE/wD8taB/wR0/4J9qcjwrrYI/6m7xP/8ALWk/AfEW/wB4X/gL/wAwf0hsL/0Cy/8AAl/kb3/BIP8A5Rg/Av8A7E/Tv/RQr9Ha87+Efwm+HvwI+GOhfBv4Taamj+GvDNlFp+mWKPJIsFtAu2OMPKzyNtA6uzMe5NeiV/ScI2SR/Lc5XbZ//9f+/iiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAP/Z"}}, "metadata": {}}, {"execution_count": null, "cell_type": "code", "metadata": {}, "outputs": [], "source": "from pmlib.anomaly_detection import AnomalyDetectionEstimator, AnomalyDetectionAssetGroupPipeline\n\nclass CustomAnomalyDetectionEstimator(AnomalyDetectionEstimator):\n def get_stages(self):\n from sklearn.covariance import (EmpiricalCovariance, EllipticEnvelope, LedoitWolf, MinCovDet, OAS, ShrunkCovariance)\n from sklearn.preprocessing import StandardScaler, RobustScaler, MinMaxScaler\n from srom.anomaly_detection.generalized_anomaly_model import GeneralizedAnomalyModel\n from srom.utils.no_op import NoOp\n\n return [\n [\n ('skipscaling', NoOp()), \n ('standardscaler', StandardScaler()),\n ('robustscaler', RobustScaler()), \n ('minmaxscaling', MinMaxScaler())\n ],\n [\n # Covariance Structure based Anomaly Models\n ('empiricalcovariance', GeneralizedAnomalyModel(base_learner=EmpiricalCovariance(), fit_function='fit', predict_function='mahalanobis',score_sign=1)), \n ('ellipticenvelope', GeneralizedAnomalyModel(base_learner=EllipticEnvelope(), fit_function='fit', predict_function='mahalanobis',score_sign=1)),\n ('ledoitwolf', GeneralizedAnomalyModel(base_learner=LedoitWolf(), fit_function='fit', predict_function='mahalanobis',score_sign=1)), \n ('mincovdet', GeneralizedAnomalyModel(base_learner=MinCovDet(), fit_function='fit', predict_function='mahalanobis',score_sign=1)), \n ('oas', GeneralizedAnomalyModel(base_learner=OAS(), fit_function='fit', predict_function='mahalanobis',score_sign=1)), \n ('shrunkcovariance', GeneralizedAnomalyModel(base_learner=ShrunkCovariance(), fit_function='fit', predict_function='mahalanobis',score_sign=1)),\n ]\n ]\n\n \nclass CustomAnomalyDetectionAssetGroupPipeline(AnomalyDetectionAssetGroupPipeline):\n pass"}, {"source": "\n### Package Custom Code", "cell_type": "markdown", "metadata": {}}, {"source": "Notebook is not a good way for code reuse. Maximo APM PMI is not able to get the code out of this notebook. In order to have our new custom model work in APM PMI (doing prediction results generation with live data), we need to package it as a standard Python library.\n\nWe need to create a Python project on local machine to put our new code. Here's a sample project structure:\n\n```\n+-- setup.py\n+-- custom_models\n + __init__.py\n + anomaly.py\n```\n\nBelow is a sample ```setup.py``` you can use, which creates a library called ```my-pmi-custom-module``` and has dependency on PMI SDK. You do need to choose a meaningful name as the library name to avoid libary name conflict and also replace all of ``````, ``````, and `````` with your PMI instance values.", "cell_type": "markdown", "metadata": {}}, {"execution_count": null, "cell_type": "code", "metadata": {}, "outputs": [], "source": "#!/usr/bin/env python\n\nfrom setuptools import setup, find_packages\n\nsetup(\n name='my-pmi-custom-models',\n version='0.1',\n packages=find_packages(),\n install_requires=[\n 'pmlib@/ibm/pmi/service/rest/ds///lib/download?filename=pmlib-1.0.0-master.tar.gz'\n ]\n)"}, {"source": "Put the new code in file **```anomaly.py```** under the module directory **```custom_models```**. **```__init__.py```** can be just an empty content file (this is Python standard way denoting the directory **```custom_models```** is a module). You can then put this project on Github. Create a project named **```my-pmi-custom-models```** and push the project content to there.\n\nWith the project ready, we can register the new custom model to Maximo APM PMI. But let's do some testing first.", "cell_type": "markdown", "metadata": {}}, {"source": "\n## Setup the Model Training Pipeline", "cell_type": "markdown", "metadata": {}}, {"source": "This is exactly same as using the out-of-box anomaly detection model (as in the notebook \"PMI - Anomaly Detection\"), except we have switched the pipeline class name. But we also need to install the custom code from the remote project repository we just created on Github:", "cell_type": "markdown", "metadata": {}}, {"execution_count": null, "cell_type": "code", "metadata": {}, "outputs": [], "source": "!pip install git+https://github.com//my-pmi-custom-models.git\n\nfrom custom_models.anomaly import CustomAnomalyDetectionAssetGroupPipeline\n\ngroup = CustomAnomalyDetectionAssetGroupPipeline(\n asset_group_id='', \n model_pipeline={\n \"features\": [\"SampleAnomalySensor:x1\", \"SampleAnomalySensor:x2\", \"SampleAnomalySensor:x3\", \"SampleAnomalySensor:x4\", \"SampleAnomalySensor:x5\", \"SampleAnomalySensor:x6\", \"SampleAnomalySensor:x7\", \"SampleAnomalySensor:x8\", \"SampleAnomalySensor:x9\", \"SampleAnomalySensor:x10\", \"SampleAnomalySensor:x11\", \"SampleAnomalySensor:x12\", \"SampleAnomalySensor:x13\", \"SampleAnomalySensor:x14\", \"SampleAnomalySensor:x15\", \"SampleAnomalySensor:x16\", \"SampleAnomalySensor:x17\", \"SampleAnomalySensor:x18\", \"SampleAnomalySensor:x19\", \"SampleAnomalySensor:x20\", \"SampleAnomalySensor:x21\", \"SampleAnomalySensor:x22\", \"SampleAnomalySensor:x23\", \"SampleAnomalySensor:x24\", \"SampleAnomalySensor:x25\", \"SampleAnomalySensor:x26\", \"SampleAnomalySensor:x27\", \"SampleAnomalySensor:x28\", \"SampleAnomalySensor:x29\", \"SampleAnomalySensor:x30\", \"SampleAnomalySensor:x31\", \"SampleAnomalySensor:x32\", \"SampleAnomalySensor:x33\", \"SampleAnomalySensor:x34\", \"SampleAnomalySensor:x35\", \"SampleAnomalySensor:x36\", \"SampleAnomalySensor:x37\", \"SampleAnomalySensor:x38\", \"SampleAnomalySensor:x39\", \"SampleAnomalySensor:x40\", \"SampleAnomalySensor:x41\", \"SampleAnomalySensor:x42\"],\n \"features_for_training\": [\":faildate\"],\n \"predictions\": [\"anomaly_score\", \"anomaly_threshold\"],\n \"srom_training_options\": {\n \"exectype\": \"spark_node_random_search\" # spark_node_random_search or spark_node_complete_search\n }\n })"}, {"source": "Note the module here is our new code's ```custom_models```.", "cell_type": "markdown", "metadata": {}}, {"source": "\n## Train the Model Instance\nNow you can train the model instance.", "cell_type": "markdown", "metadata": {}}, {"execution_count": null, "cell_type": "code", "metadata": {}, "outputs": [], "source": "df = group.execute()"}, {"source": "Once this method completes successfully, you'll have a trained model instance reday (for next step, see below) and also with the prediction results returned as a dataframe for verification.", "cell_type": "markdown", "metadata": {}}, {"source": "\n## Register the Trained Model Instance", "cell_type": "markdown", "metadata": {}}, {"source": "If the trained model instance looks good, you can register it to Maximo APM PMI. Note the difference here from the out-of-box model, we have specified the URL to the Github repository storing our custom mode code:", "cell_type": "markdown", "metadata": {}}, {"execution_count": null, "cell_type": "code", "metadata": {}, "outputs": [], "source": "group.register(url='git+https://github.com//my-pmi-custom-models.git', df=df)"}, {"source": "Once registration succeeds, you can see this newly trained model instance available for the asset group on IBM Maximo APM - AHI.", "cell_type": "markdown", "metadata": {}}, {"execution_count": null, "cell_type": "code", "metadata": {}, "outputs": [], "source": ""}], "metadata": {"kernelspec": {"display_name": "Python 3.5", "name": "python3", "language": "python"}, "language_info": {"mimetype": "text/x-python", "nbconvert_exporter": "python", "version": "3.5.5", "name": "python", "file_extension": ".py", "pygments_lexer": "ipython3", "codemirror_mode": {"version": 3, "name": "ipython"}}}, "nbformat": 4} --------------------------------------------------------------------------------