From 3fc4bb3a121c01386b65da26884437e90c0cfa14 Mon Sep 17 00:00:00 2001
From: jasperan <23caj23@gmail.com>
Date: Mon, 29 May 2023 11:33:06 +0200
Subject: [PATCH 01/10] feat: added neural_networks_hero lab
---
neural_networks_hero/aboutworkshop/about.md | 72 +++++
.../creatingmodel/creatingmodel.md | 250 +++++++++++++++++
.../creatingmodel/images/2nd_importances.PNG | Bin 0 -> 125469 bytes
.../creatingmodel/images/3rd_fit.PNG | Bin 0 -> 53965 bytes
.../creatingmodel/images/3rd_importances.PNG | Bin 0 -> 46551 bytes
.../creatingmodel/images/3rd_leaderboard.PNG | Bin 0 -> 84227 bytes
.../creatingmodel/images/3rd_manipulation.PNG | Bin 0 -> 44974 bytes
.../creatingmodel/images/analyticstab.png | Bin 0 -> 181350 bytes
.../images/binary_classification.PNG | Bin 0 -> 32402 bytes
.../images/classification_metrics.png | Bin 0 -> 24171 bytes
.../creatingmodel/images/download_models.PNG | Bin 0 -> 37953 bytes
.../creatingmodel/images/downloadkaggle.jpg | Bin 0 -> 153417 bytes
.../creatingmodel/images/dropping_columns.PNG | Bin 0 -> 53060 bytes
.../images/example_data_structure.png | Bin 0 -> 132276 bytes
.../creatingmodel/images/example_ensemble.png | Bin 0 -> 111660 bytes
.../images/histogram_example.PNG | Bin 0 -> 49897 bytes
.../creatingmodel/images/histograms.webp | Bin 0 -> 307750 bytes
.../creatingmodel/images/leaderboard.PNG | Bin 0 -> 71954 bytes
.../creatingmodel/images/metrics_1.PNG | Bin 0 -> 59009 bytes
.../creatingmodel/images/metrics_2.PNG | Bin 0 -> 76030 bytes
.../creatingmodel/images/minmax_f.PNG | Bin 0 -> 31415 bytes
.../creatingmodel/images/new_terminal.png | Bin 0 -> 257695 bytes
.../creatingmodel/images/notebook.png | Bin 0 -> 123963 bytes
.../creatingmodel/images/open-notebook.png | Bin 0 -> 18573 bytes
.../creatingmodel/images/open-notebook2.png | Bin 0 -> 4119 bytes
.../creatingmodel/images/raw.jpg | Bin 0 -> 104537 bytes
.../creatingmodel/images/read_dataset.png | Bin 0 -> 51785 bytes
.../creatingmodel/images/running.jpg | Bin 0 -> 109699 bytes
.../creatingmodel/images/savepage.jpg | Bin 0 -> 235286 bytes
.../images/select_data_science.jpg | Bin 0 -> 116318 bytes
.../creatingmodel/images/select_kernel.PNG | Bin 0 -> 37914 bytes
.../creatingmodel/images/structure_2023.webp | Bin 0 -> 867748 bytes
.../images/training_simplified.PNG | Bin 0 -> 41594 bytes
.../creatingmodel/images/unzip.png | Bin 0 -> 125286 bytes
.../creatingmodel/images/unzip_result.png | Bin 0 -> 149633 bytes
.../infra/images/analyticstab.png | Bin 0 -> 181350 bytes
.../infra/images/cloud-code-editor.png | Bin 0 -> 145241 bytes
.../infra/images/cloud-shell-button.png | Bin 0 -> 365404 bytes
.../infra/images/code-editor-go-to-file.png | Bin 0 -> 70930 bytes
.../infra/images/code-editor-open-menu.png | Bin 0 -> 78918 bytes
.../infra/images/code-editor-open-popup.png | Bin 0 -> 78069 bytes
.../infra/images/code-editor-open-tfvars.png | Bin 0 -> 59527 bytes
.../infra/images/code-editor-path.png | Bin 0 -> 27874 bytes
.../infra/images/code-editor-save.png | Bin 0 -> 148146 bytes
.../infra/images/file_explorer.png | Bin 0 -> 398714 bytes
.../infra/images/file_explorer_2.png | Bin 0 -> 717748 bytes
.../infra/images/git-clone.png | Bin 0 -> 124042 bytes
.../infra/images/lol_infra.drawio | 1 +
.../infra/images/lol_infra.png | Bin 0 -> 77225 bytes
.../infra/images/notebook.png | Bin 0 -> 123963 bytes
.../infra/images/open-notebook.png | Bin 0 -> 18573 bytes
.../infra/images/open-notebook2.png | Bin 0 -> 4119 bytes
.../infra/images/open_terminal.png | Bin 0 -> 641099 bytes
.../infra/images/paste-compartment-ocid.png | Bin 0 -> 420358 bytes
.../infra/images/paste-public-ssh-key.png | Bin 0 -> 525848 bytes
.../infra/images/paste-region.png | Bin 0 -> 316898 bytes
.../infra/images/paste-tenancy-ocid.png | Bin 0 -> 383413 bytes
neural_networks_hero/infra/images/proceed.png | Bin 0 -> 234084 bytes
.../infra/images/select_data_science.png | Bin 0 -> 136539 bytes
.../infra/images/start-sh-beginning.png | Bin 0 -> 87625 bytes
.../infra/images/start-sh-output.png | Bin 0 -> 120619 bytes
.../infra/images/start-sh-ssh.png | Bin 0 -> 114941 bytes
.../infra/images/start-sh-terraform.png | Bin 0 -> 181952 bytes
.../infra/images/unzip_result.png | Bin 0 -> 149633 bytes
neural_networks_hero/infra/infra.md | 265 ++++++++++++++++++
.../intro/images/bought_items.jpg | Bin 0 -> 111796 bytes
neural_networks_hero/intro/intro.md | 87 ++++++
.../the_problem/images/diagram.png | Bin 0 -> 646176 bytes
.../the_problem/images/last_year_detail.png | Bin 0 -> 442726 bytes
.../the_problem/images/last_year_live.png | Bin 0 -> 177514 bytes
.../the_problem/images/last_year_normal.png | Bin 0 -> 132276 bytes
.../the_problem/images/lcu_architecture.png | Bin 0 -> 12923 bytes
.../the_problem/images/learning_curves.png | Bin 0 -> 84829 bytes
.../the_problem/images/live_client_1.PNG | Bin 0 -> 248685 bytes
.../the_problem/images/live_client_2.PNG | Bin 0 -> 102735 bytes
.../the_problem/images/live_client_3.PNG | Bin 0 -> 199014 bytes
.../the_problem/images/mljar-output.PNG | Bin 0 -> 278835 bytes
.../the_problem/images/nn_importance.png | Bin 0 -> 74167 bytes
.../the_problem/images/result_live_client.PNG | Bin 0 -> 4023066 bytes
.../the_problem/images/structure_2023.webp | Bin 0 -> 867748 bytes
neural_networks_hero/the_problem/problem.md | 189 +++++++++++++
.../images/activation_functions.gif | Bin 0 -> 296334 bytes
.../understand_nn/images/feedforward.png | Bin 0 -> 106806 bytes
.../images/grey_white_matter.PNG | Bin 0 -> 331893 bytes
.../understand_nn/images/inverse_loss.png | Bin 0 -> 32209 bytes
.../understand_nn/images/myelin.PNG | Bin 0 -> 405974 bytes
.../images/neural_network_visualization_1.gif | Bin 0 -> 5714303 bytes
.../images/neural_network_visualization_2.gif | Bin 0 -> 3242467 bytes
.../images/neural_network_visualization_3.gif | Bin 0 -> 458688 bytes
.../understand_nn/images/optimizations.gif | Bin 0 -> 2676767 bytes
.../understand_nn/images/recurrent.png | Bin 0 -> 208098 bytes
.../understand_nn/understand_nns.md | 117 ++++++++
workshops/neural_networks_hero/index.html | 62 ++++
workshops/neural_networks_hero/manifest.json | 64 +++++
.../neural_networks_hero/manifest_full.json | 40 +++
95 files changed, 1147 insertions(+)
create mode 100644 neural_networks_hero/aboutworkshop/about.md
create mode 100644 neural_networks_hero/creatingmodel/creatingmodel.md
create mode 100644 neural_networks_hero/creatingmodel/images/2nd_importances.PNG
create mode 100644 neural_networks_hero/creatingmodel/images/3rd_fit.PNG
create mode 100644 neural_networks_hero/creatingmodel/images/3rd_importances.PNG
create mode 100644 neural_networks_hero/creatingmodel/images/3rd_leaderboard.PNG
create mode 100644 neural_networks_hero/creatingmodel/images/3rd_manipulation.PNG
create mode 100644 neural_networks_hero/creatingmodel/images/analyticstab.png
create mode 100644 neural_networks_hero/creatingmodel/images/binary_classification.PNG
create mode 100644 neural_networks_hero/creatingmodel/images/classification_metrics.png
create mode 100644 neural_networks_hero/creatingmodel/images/download_models.PNG
create mode 100644 neural_networks_hero/creatingmodel/images/downloadkaggle.jpg
create mode 100644 neural_networks_hero/creatingmodel/images/dropping_columns.PNG
create mode 100644 neural_networks_hero/creatingmodel/images/example_data_structure.png
create mode 100644 neural_networks_hero/creatingmodel/images/example_ensemble.png
create mode 100644 neural_networks_hero/creatingmodel/images/histogram_example.PNG
create mode 100644 neural_networks_hero/creatingmodel/images/histograms.webp
create mode 100644 neural_networks_hero/creatingmodel/images/leaderboard.PNG
create mode 100644 neural_networks_hero/creatingmodel/images/metrics_1.PNG
create mode 100644 neural_networks_hero/creatingmodel/images/metrics_2.PNG
create mode 100644 neural_networks_hero/creatingmodel/images/minmax_f.PNG
create mode 100644 neural_networks_hero/creatingmodel/images/new_terminal.png
create mode 100644 neural_networks_hero/creatingmodel/images/notebook.png
create mode 100644 neural_networks_hero/creatingmodel/images/open-notebook.png
create mode 100644 neural_networks_hero/creatingmodel/images/open-notebook2.png
create mode 100644 neural_networks_hero/creatingmodel/images/raw.jpg
create mode 100644 neural_networks_hero/creatingmodel/images/read_dataset.png
create mode 100644 neural_networks_hero/creatingmodel/images/running.jpg
create mode 100644 neural_networks_hero/creatingmodel/images/savepage.jpg
create mode 100644 neural_networks_hero/creatingmodel/images/select_data_science.jpg
create mode 100644 neural_networks_hero/creatingmodel/images/select_kernel.PNG
create mode 100644 neural_networks_hero/creatingmodel/images/structure_2023.webp
create mode 100644 neural_networks_hero/creatingmodel/images/training_simplified.PNG
create mode 100644 neural_networks_hero/creatingmodel/images/unzip.png
create mode 100644 neural_networks_hero/creatingmodel/images/unzip_result.png
create mode 100644 neural_networks_hero/infra/images/analyticstab.png
create mode 100644 neural_networks_hero/infra/images/cloud-code-editor.png
create mode 100644 neural_networks_hero/infra/images/cloud-shell-button.png
create mode 100644 neural_networks_hero/infra/images/code-editor-go-to-file.png
create mode 100644 neural_networks_hero/infra/images/code-editor-open-menu.png
create mode 100644 neural_networks_hero/infra/images/code-editor-open-popup.png
create mode 100644 neural_networks_hero/infra/images/code-editor-open-tfvars.png
create mode 100644 neural_networks_hero/infra/images/code-editor-path.png
create mode 100644 neural_networks_hero/infra/images/code-editor-save.png
create mode 100644 neural_networks_hero/infra/images/file_explorer.png
create mode 100644 neural_networks_hero/infra/images/file_explorer_2.png
create mode 100644 neural_networks_hero/infra/images/git-clone.png
create mode 100644 neural_networks_hero/infra/images/lol_infra.drawio
create mode 100644 neural_networks_hero/infra/images/lol_infra.png
create mode 100644 neural_networks_hero/infra/images/notebook.png
create mode 100644 neural_networks_hero/infra/images/open-notebook.png
create mode 100644 neural_networks_hero/infra/images/open-notebook2.png
create mode 100644 neural_networks_hero/infra/images/open_terminal.png
create mode 100644 neural_networks_hero/infra/images/paste-compartment-ocid.png
create mode 100644 neural_networks_hero/infra/images/paste-public-ssh-key.png
create mode 100644 neural_networks_hero/infra/images/paste-region.png
create mode 100644 neural_networks_hero/infra/images/paste-tenancy-ocid.png
create mode 100644 neural_networks_hero/infra/images/proceed.png
create mode 100644 neural_networks_hero/infra/images/select_data_science.png
create mode 100644 neural_networks_hero/infra/images/start-sh-beginning.png
create mode 100644 neural_networks_hero/infra/images/start-sh-output.png
create mode 100644 neural_networks_hero/infra/images/start-sh-ssh.png
create mode 100644 neural_networks_hero/infra/images/start-sh-terraform.png
create mode 100644 neural_networks_hero/infra/images/unzip_result.png
create mode 100644 neural_networks_hero/infra/infra.md
create mode 100644 neural_networks_hero/intro/images/bought_items.jpg
create mode 100644 neural_networks_hero/intro/intro.md
create mode 100644 neural_networks_hero/the_problem/images/diagram.png
create mode 100644 neural_networks_hero/the_problem/images/last_year_detail.png
create mode 100644 neural_networks_hero/the_problem/images/last_year_live.png
create mode 100644 neural_networks_hero/the_problem/images/last_year_normal.png
create mode 100644 neural_networks_hero/the_problem/images/lcu_architecture.png
create mode 100644 neural_networks_hero/the_problem/images/learning_curves.png
create mode 100644 neural_networks_hero/the_problem/images/live_client_1.PNG
create mode 100644 neural_networks_hero/the_problem/images/live_client_2.PNG
create mode 100644 neural_networks_hero/the_problem/images/live_client_3.PNG
create mode 100644 neural_networks_hero/the_problem/images/mljar-output.PNG
create mode 100644 neural_networks_hero/the_problem/images/nn_importance.png
create mode 100644 neural_networks_hero/the_problem/images/result_live_client.PNG
create mode 100644 neural_networks_hero/the_problem/images/structure_2023.webp
create mode 100644 neural_networks_hero/the_problem/problem.md
create mode 100644 neural_networks_hero/understand_nn/images/activation_functions.gif
create mode 100644 neural_networks_hero/understand_nn/images/feedforward.png
create mode 100644 neural_networks_hero/understand_nn/images/grey_white_matter.PNG
create mode 100644 neural_networks_hero/understand_nn/images/inverse_loss.png
create mode 100644 neural_networks_hero/understand_nn/images/myelin.PNG
create mode 100644 neural_networks_hero/understand_nn/images/neural_network_visualization_1.gif
create mode 100644 neural_networks_hero/understand_nn/images/neural_network_visualization_2.gif
create mode 100644 neural_networks_hero/understand_nn/images/neural_network_visualization_3.gif
create mode 100644 neural_networks_hero/understand_nn/images/optimizations.gif
create mode 100644 neural_networks_hero/understand_nn/images/recurrent.png
create mode 100644 neural_networks_hero/understand_nn/understand_nns.md
create mode 100644 workshops/neural_networks_hero/index.html
create mode 100644 workshops/neural_networks_hero/manifest.json
create mode 100644 workshops/neural_networks_hero/manifest_full.json
diff --git a/neural_networks_hero/aboutworkshop/about.md b/neural_networks_hero/aboutworkshop/about.md
new file mode 100644
index 0000000..15d79e7
--- /dev/null
+++ b/neural_networks_hero/aboutworkshop/about.md
@@ -0,0 +1,72 @@
+# About This Workshop
+
+Estimated Time: 5 minutes
+
+## Introduction
+
+League of Legends is one of the most played videogames in the entire world. In this workshop, we'll leverage the power of AI with League of Legends in a unique and innovative way. We'll dive deep into extractable data (accessible through the game's API), how to structure this data, and how to use it to train our own Machine Learning model to generate real-time predictions about any match.
+
+Are you interested in learning machine learning (ML)? How about doing this in the context of the exciting world of gaming?! Get your ML skills bootstrapped here!
+
+Here's a short 3-minute introductory video into League of Legends:
+
+[League of Legends Introduction](youtube:OfYU4gbk13w)
+
+And here's a video to get you hyped about League and understand advanced strategies from one of the best teams in the world:
+
+[LCS Finals - 2023](youtube:0b0TXaJMUMU)
+
+## About Product/Technology
+
+OCI Data Science is a fully managed and serverless platform for data science teams to build, train, and manage machine learning models using Oracle Cloud Infrastructure.
+
+The Data Science Service:
+
+- Provides data scientists with a collaborative, project-driven workspace.
+- Enables self-service, serverless access to infrastructure for data science workloads.
+- Helps data scientists concentrate on methodology and domain expertise to deliver models to production.
+
+## Objectives
+
+In this lab, you will complete the following steps:
+
+- Data Collection - Download datasets
+- Data Preparation - Preparing datasets
+- Data Load - Searching for patterns in datasets
+- Implement ML Models - Develop, different ML models, evaluate best model and tune
+- Data Integration - Connecting to real-time data sources
+
+## OCI Elements
+
+This solution is designed to work with several OCI services, allowing you to quickly be up-and-running. You can read more about the services used in the lab here:
+
+- [OCI Data Science](https://www.oracle.com/artificial-intelligence/)
+- [OCI Cloud Shell](https://docs.oracle.com/en-us/iaas/Content/API/Concepts/cloudshellintro.htm)
+- [OCI Compute](https://www.oracle.com/cloud/compute/)
+- [OCI Autonomous JSON Database](https://www.oracle.com/autonomous-database/autonomous-json-database/)
+
+## Data Sets
+
+We'll access datasets uploaded to OCI Object Storage. You will get detailed instructions on how to download the full dataset in the Infrastructure lab.
+
+You don't need to download any datasets by yourself, everything will be **automated**.
+
+I have also published my datasets into Kaggle, should you be interested to look at them in more detail, or even use them yourself:
+
+- [Old Dataset](https://www.kaggle.com/jasperan/league-of-legends-1v1-matchups-results)
+- [New Dataset](https://www.kaggle.com/datasets/jasperan/league-of-legends-optimizer-dataset)
+
+Make sure to give them a thumbs up if you're enjoying the content and find them interesting.
+
+## Useful Sources
+
+- Data Integration with leagueoflegends-optimizer: [YouTube video](https://www.youtube.com/watch?v=SlG0q4oWGsk)
+
+You may now [proceed to the next lab](#next).
+
+## Acknowledgements
+
+- **Author** - Nacho Martinez, Data Science Advocate @ DevRel
+- **Editor** - Erin Dawson, DevRel Communications Manager
+- **Contributors** - Victor Martin, Product Strategy Director
+- **Last Updated By/Date** - May 29th, 2023
diff --git a/neural_networks_hero/creatingmodel/creatingmodel.md b/neural_networks_hero/creatingmodel/creatingmodel.md
new file mode 100644
index 0000000..10eec47
--- /dev/null
+++ b/neural_networks_hero/creatingmodel/creatingmodel.md
@@ -0,0 +1,250 @@
+# Creating the Model
+
+## Introduction
+
+In this lab, we'll be creating all Machine Learning Models. This model will attempt to use as many variables as possible, whilst taking advantage of the power of AutoML (remember, work smart, not hard!). It'll be useful to us as it introduces the most basic and fundamental ML concepts.
+
+Estimated Time: 45 minutes
+
+### Prerequisites
+
+* An Oracle Free Tier, Paid or LiveLabs Cloud Account
+* Active Oracle Cloud Account with available credits to use for Data Science service.
+* [Previously created](https://github.com/oracle-devrel/leagueoflegends-optimizer/blob/livelabs/hols/dataextraction/infra/infra.md) OCI Data Science Environment
+
+## Task 1: Set up OCI Data Science Environment
+
+[Having previously created our OCI Data Science environment](https://github.com/oracle-devrel/leagueoflegends-optimizer/blob/livelabs/hols/dataextraction/infra/infra.md), we need to install the necessary Python dependencies to execute our code. For that, we'll access our environment.
+
+1. We open the notebook that was provisioned:
+
+ 
+
+ > **Note**: You may find the Data Science section by also searching in the top left bar, or in the Analytics & AI tab, if it doesn't appear in "Recently visited" for you:
+
+ 
+
+ Now, we have access to a [list of our Data Science projects launched within OCI.](https://cloud.oracle.com/data-science/projects) We access our project, and inside our project we'll find the notebook.
+
+ > **Note**: The name of the notebook may be different than shown here in the screenshot.
+
+ 
+
+ 
+
+ You should now see the Jupyter environment
+
+ 
+
+2. We now need to load our notebook and datasets into our environment. For that, we open a new terminal inside our environment:
+
+ 
+
+ Then, we execute the following command, which will download all necessary datasets:
+
+ ```bash
+
+ wget https://objectstorage.eu-frankfurt-1.oraclecloud.com/p/FcwFW-_ycli9z8O_3Jf8gHbc1Fr8HkG9-vnL4I7A07mENI60L8WIMGtG5cc8Qmuu/n/axywji1aljc2/b/league-hol-ocw-datasets/o/league_ocw_2023.zip && unzip league_ocw_2023.zip -d /home/datascience/.
+
+ ```
+
+ This process should take about a minute.
+
+ 
+
+ Now, download the repository (if you haven't already):
+
+ ```bash
+
+ git clone --branch livelabs https://github.com/oracle-devrel/leagueoflegends-optimizer.git
+
+ ```
+
+ After this, we will open the notebook called _`models_2023.ipynb`_ located in _`leagueoflegends-optimizer/notebooks`_ by double clicking it.
+
+3. Now, with our Python dependencies installed and our repository and notebook ready, we're ready to run it from the first cell. Make sure to select the correct Kernel (the one that you have configured and has all Python dependencies installed within it) from the Kernel dropdown menu:
+
+
+
+## Task 2: The Data Structure
+
+From our dataset, we can observe an example of the data structure we're going to use to build our model:
+
+
+
+
+It is important to remember that structuring and manipulating data in the data science process takes about 80 to 90% of the time, according to expert sources (image courtesy of [“2020 State of Data Science: Moving From Hype Toward Maturity.”](https://www.anaconda.com/state-of-data-science-2020)), and we shouldn't be discouraged when spending most of our time processing and manipulating data structures. The ML algorithm is the easy part if you've correctly identified the correct data structure and adapted it to the structure ML algorithms and pipelines expect.
+
+
+
+## Task 3: Load Data / Generate Dataset
+
+First, we load the model and train-test split it.
+
+To perform ML properly, we need to take the dataset we're going to work with, and split it into two:
+
+* A **training** dataset, from which our ML model will learn to make predictions.
+* A **testing** dataset, from which our ML model will validate the predictions it makes, and check how accurate it was compared to the truth.
+
+In ML, it's very typical to find values around 80% train / 20% test proportions, as it provides enough data for the model to be trained, and enough data to check the accuracy of the model without having too much / too little data in either of the datasets.
+
+After this split, we divide the whole dataset into two separate files, one containing training data (85% of the original dataset) and testing data (15%).
+
+
+
+Then, we begin with simple data exploration of our initial dataset. Histograms are particularly useful to find the distribution of one (or many!) variables in our dataset and see if they follow any known statistical distribution.
+
+
+
+It's also good to look at our new variables `f1...f3` and their minimum, average, maximum:
+
+
+
+This will also help us determine what to return to the user when they're playing a game in the end: the closer they are to the maximum, the better they will have performed, and we also need to adjust that accordingly.
+
+We're also interested in other variables' histograms, especially the ones around people getting multiple kills in a row, number of wards, big jungle objectives, match durations... Basically statistics that I find personally interesting after finishing a League match myself.
+
+
+
+After getting a rough idea of what our dataset and some of our variables contain, it's time to tell the ML model which variables we want as input and which ones as output.
+
+For this example (in the notebook, we create several models), we'll first drop those columns we don't want to use as inputs or outputs - in this first model, we don't want to use any `f1...f5` variables in our dataset, as we're going to create a model with League's original data to begin with:
+
+
+
+After we create our `TabularDataset()` object (which extends a pandas dataframe and therefore has most of panda's functions available), we're ready to start training.
+
+## Task 4: Model Training
+
+Now that we've seen the shape of our dataset and we have the variable we want to predict (in this case, calculated_player_performance), we train as many models as possible for 10 minutes. We can instantiate a `TabularPredictor()` which takes most of the difficulties of usually writing this kind of code out of the equation:
+
+
+
+We need to specify that this problem is a regression problem (we're predicting numerical, continuous values (not integers)) and we specify which variable it is we're trying to predict through the `label` parameter.
+
+The preset is a pre-configuration that restraints the amount of iterations, models, and time dedicated to train each model, to achieve some "quality" defined as a preset.
+> **Note**: find all available presets [here](https://auto.gluon.ai/0.5.2/tutorials/tabular_prediction/tabular-quickstart.html#presets).
+
+After our training is done (about 10 minutes), we can display some results:
+
+First, we display a leaderboard of the best trained models ordered by decreasing RMSE. If you're not familiar with this concept, don't worry, we'll revisit all metrics right below, in our Model Testing task. This will help us see which models perform better against the target variable that we specified before:
+
+
+
+Note that our Level 2 Weighted Ensemble has the lowest RMSE of all: we'll probably want to use this model.
+
+
+> **Note**: this is an example of an weighted ensemble model, in which decisions are taken using a technique called **bagging**: every model makes a prediction, and the best models will weigh more upon the final decision.
+
+## Task 5: Model Testing
+
+After training is done, we need to check whether the training we did was actually useful or a waste of time. To achieve this, we make use of some metrics, which depend on the type of problem we're dealing with.
+
+For example, in a binary classification problem (where we're trying to predict if something is either 0 or 1), I typically use **accuracy, precision, recall and f1-score** as standard evaluation metrics:
+
+
+> **Note**: an example on how each one of these 4 metrics are calculated just by looking at the Confusion Matrix.
+
+However, as we're dealing with a regression problem, the most popular metrics are: the MSE, MAE, RMSE, R-Squared and variants of these coefficients.
+
+The MSE, MAE, RMSE, and R-Squared metrics are mainly used to evaluate the prediction error rates and model performance in regression analysis.
+
+* MAE (Mean absolute error) represents the difference between the original and predicted values extracted by averaged the absolute difference over the data set.
+* MSE (Mean Squared Error) represents the difference between the original and predicted values extracted by squared the average difference over the data set.
+* RMSE (Root Mean Squared Error) is the error rate by the square root of MSE.
+* R-squared (Coefficient of determination) represents the coefficient of how well the values fit compared to the original values.
+ The value from 0 to 1 interpreted as percentages. The higher the value is, the better the model is.
+* The Pearson correlation coefficient is a descriptive statistic, meaning that it summarizes the characteristics of a dataset
+ Specifically, it describes the strength and direction of the linear relationship between two quantitative variables.
+
+Note that, in our code, we need to use our **testing dataset** as the one to validate our metrics (we'll have the testing data to check against).
+
+
+
+We're also able to extract the feature importance of our model. This is an awesome calculation provided to us automatically by the AutoML library.
+Feature importance is an index that represents the measure of contribution of each feature in our ML model.
+> **Note**: it's important to note that feature importance depends on the model and dataset used, and different algorithms may assign different importance values to the same set of features.
+
+For more advanced Machine Learning practicioners, there's a caveat I need to make here about certain types of regularization (like *L1*/*Lasso* regularization) - a technique that's used to often prevent overfitting and improve the generalized performance of a model -: it can force some coefficients to become zero, rendering those coefficients' variables useless in a model.
+
+
+> **Note**: if I have two variables with importances N and M, the first variable will have an importance N/M times higher than the second variable, and viceversa.
+
+This means that our model takes `deaths, assists, kills` as the three most important variables, and the fourth most important variable is the game duration.
+
+After creating this function and invoking it, we will obtain a resulting CSV file or dataframe object. We'll use this new object to create our model.
+
+## Task 6: Creating Extra Models
+
+The rest of the notebook is similar to the process we've followed until now, with a few changes, which shall be mentioned here for clarity.
+
+### Win Prediction Model (2nd Model)
+
+The second model we create is a winning predictor (a binary classifier that tries to predict whether the player won or lost, based on all input variables). We specify that this is a binary classification problem this way:
+
+
+
+The results from this model are very promising, reaching up to 99.37% accuracy:
+
+
+> **Note**: if you're planning to run inference (deploy your model and make predictions) on a low-end computer, you might be better off with the Light Gradient Boosted Model, as it's prediction times are about 100 times faster than our L2 Weighted Ensemble.
+
+### Live Client API Compatible Model
+
+And now that we have one model for each, we attempt to create a model only using our `f1...f3` variables as discussed during the workshop. We call this model _Live Client API Compatible Model_ as it utilizes as much temporal data as possible from the API return object.
+
+These variables were calculated like this:
+
+* Kills + assists / gameTime ==> kills + assists ratio ==> `f2`
+* Deaths / gameTime ==> death ratio ==> `f1`
+* xp / gameTime ==> xp per min ==> `f3`
+
+In our dataset, we also had two other variables that I was hoping I could also calculate with Live Client API data, but these variables weren't possible to accurately calculate:
+
+* `f4`, which represented the total amount of damage per minute, wasn't present in the Live Client API in any field
+* `f5`, which represented the total amount of gold per minute, wasn't either. You can only extract the **current** amount of gold, which doesn't add any real value to the model.
+
+So, the idea now is to create a model that, given f1, f2 and f3, and the champion name, is **able to predict any player's performance**.
+
+
+
+
+
+
+> **Note**: the RMSE in this third experiment, compared to the first model (both models predict the same target variable `calculated_player_performance`) is higher, which I expected, since we're using only 4 input variables for this model instead of 100+. However, as our leaderboard indicates, all these models are able to properly **infer** a player's performance, even if the RMSE is a bit more elevated.
+
+
+
+Just as an interesting observation, our model's importance is mostly based around `f1` and `f2`, being `f3` about 5-8 times less important than the other two.
+
+## Task 7: Downloading Models
+
+If you want to use these models in your computer, while you play League, you will need to *zip* all generated models into a file and download it to your computer.
+
+In a terminal, you can run the following command to bundle all directories into one file:
+
+```bash
+
+zip -r all_models.zip /home/datascience/leagueoflegends-optimizer/notebooks/live_model_1/ /home/datascience/leagueoflegends-optimizer/notebooks/player_performance_models /home/datascience/leagueoflegends-optimizer/notebooks/winner_models/
+
+```
+
+Then go to the file explorer and download the selected file:
+
+
+
+Now, we have successfully completed the Data Science process! We have:
+
+* Loaded and Explored the dataset
+* Created some useful visualizations through histograms and `.describe()`
+* Created some TabularPredictors to help us autotrain models
+* Evaluated each model's performance against a **test** dataset
+* Saved these models for future deployment -
+
+You may now [proceed to the next lab](#next).
+
+## Acknowledgements
+
+* **Author** - Nacho Martinez, Data Science Advocate @ DevRel
+* **Contributors** - Victor Martin, Product Strategy Director
+* **Last Updated By/Date** - May 28th, 2023
diff --git a/neural_networks_hero/creatingmodel/images/2nd_importances.PNG b/neural_networks_hero/creatingmodel/images/2nd_importances.PNG
new file mode 100644
index 0000000000000000000000000000000000000000..e13920c4ffa23646eef8efb9399e2846f85dc20d
GIT binary patch
literal 125469
zcmdSAXEdCD_cuxeNeF3(UP2HpNOVRCLiFByFBu}bQG+C+MARrl^eCfuW)Kl|lIV5x
z&ghJ3XLA34&vWN~)_vZb^Xl1a%~-S7Wqr$TpIxqZI$EkXZ!p{-A|kq}_WYS15fK?b
z5fMq)H8KK6v*#d&@Q2t(PgRkqYJ_Qv@Z+kpf~EoyQB5KR#+H=u`?}Y2Qy(HCs-8a|
z;yz%h0}&CvNbQ+|L4egR7WBblHh7Z4%JqqH!LRWl)&l9@&yUF~Z$EhQoQ`vUo$4_&
zb)j%FV?^si5^tZEJYIR+WS_Iu^d+OpNSqnmHGICj-VWNl_b_)k+|c_~o)i;LEJ<&?
zTJ*P+PZFu$Qh;7xKS|t!J}Kjj5wP|CDZ0~eHWLhKIG-7BYyCQh-~u`Lk5_d}S_#3A
zD>h7p@TBbjFC-wzZWHn5zf%3ln{d)Q|CN6KzI%oC`G2Pu+Sg11itFk`30L$N@n5^j
zFLPI@;xQ%hUtA)hYNA`!2O8H*{)0&rM?y>iCoKOT+-iyc+a5)_!jqQRh<^ErSPtobv4eSoydmET1`h1=W#4*A*p%>6F#0UHe4)EpB!L8%0_9Ar)_>)xv!%mEc?tK6f5A;TElj=T_FHHm)D=F7Izi
z@xi#R@-O^C#eoa)lm2o=h<*ral@~O6j^UI=H{c69{R<)3CsQ91Mt9x8p^|#vD799hBPuJT~W0CHqlNJOutz#mBQY3
z?zDBjsQ6xON%Gf~H8ZJSGr>q;Q!b@*LZd>r(V-G720ufNKH#>1rGv>^0PRwrZiucz
z!kqj!_e(0s&bl^1OkX^jKV?9ssy(~vz8)-+UZLGlwlc;37Q)`az5^U`Zwldhq0gb0
z<*G)p=
zTQ&WLmlE~P-6dyrP4%o`%ShZj^aXU|9K*82=GioR5*&s^(j=e(m#TQ4n`BW_4I#6f
zp#uKfiGB8$XF$B)r2urT=_e5F3Dxu%K5>Vhx!3Hd7JQU&kF@SVjfP!n#|1xZYFP?9
zhC|oPJC*0+0C*@GiIYBL6t&5Zx{nFgG?HII;L%~)_*F3G?x7$QC60IU2IR7ak`dt-
z{GgnepirlnPHun7WB>8BZ_dq=ZVl72Pn8GVSdOmHLN9lKThgJ6&Rb#UCAf$wx%c?s
zjb9X(Gm!61w=Svi?1ij>DBvdkY!ZC}y)1?bRbe5(oZov;$-Fu>DX@@Ln1J72*eL=J
zy0i{GIM>q(Ky2WrTomzH0B?1`!R`ku>~ClR@SV<`4aXzQ4jKF(hZvUZLh
zMPN1B#o>Hm;&C)+&j1&srxEHQe5Lp$UhEkX!+VLs58(twZEUjxYXYT=STDK3;l0QUW@uFB
zAtO|Htztt91x96pGcC(=4#1(8TSK%-!OB
z&WoM3j?Z3`x@FUrCcUc3jiJ67W-77kIr&-Zf&)IO5ROF#%2B+`z{GPu@Xi_82s`!!
zJcpv-MaP3-ct)V(;}7JVax<9ug+#MFUU3VwQ0N(9|{H
zX977ZZi(R4YT$K?5w$rv%(t)BfvEqwDHE9&1Y_sEM6s?y(Y8Sckl|FSKjtTjLnZ2{
zLc&dc+=ye%M`DY5E2uc}$7ppyxv0PoPuSV7sT?
zLKyQcFZ|lEbqN+p|1&J`*vAcjJV}u6IpRnh&n1i9vZjq7!%c6};hC{G{5k$Pe%2fy
z56724-oXQsmS*~7Ys?mb*iZO(#ZW;A2s@~M3B|2h4PCchMM+(QF{7Y6LOA`)LooXD
zwTB0o19)I~6eRaC37?G3VH5xtXGcQ{#oNC|lGE07euagPRP)Ap5FT
zkZ1n5zA%A(R2JUCxcvb%Z38uH2GA-2_=yE`=Njkwv&>znMp^S`PjWaGEs^)*aRl$6
zJSz-985y!=G
zA;NnRDCsa&iLuiS;6nRt(I(BlR=@6J_-j~10m;3fr;(i_jqewM_&unm^|`OLDga#v
z;KS{|mg=6yC*ltQ36|+}Nk1-g$!}gXO36?IPW1t)mkTOZut*@O@6LokO~=?DzxD8p
zS%`w=%@S$}wG2k3rQ$+ZEQl%gfX
z+&KXE&ibCh+a6@1QL@*`4br_gLYE=TSD#n4KVL8==MV6T1ojm|Ru0-fcSR@80=SMY
zSNjTVqGC3{wtJ@r>2p^3TOQ=~yzC+d2VvMTK;pP|3==DB-$GAQg-3`8s$xMRIqLuJwA!)BqxN{Da!H_5-nN6Do@pn`vQ`y-5
zMa7BjyUZGo!Ra#Y-QQ@de+-rbO*0XP;k_~uv4LUae+i!__63CrrqB9<*E^%}nmVVy
zgUCqwkJYdXS1Dt_JMOFe`}1waL9k^6ZZ_;iTU%Q?OZ8*pgbICFB&%W$rGd*(c<&*d
zP_Hxc1uPOujYnIzXPdwaeh%HOJ`mtLmSzdKco4RlIMm2^907SRB_*W@yYlbV!?|8v
zCyv8E1I7Zh5T#Kx>o;!XEUbk05&#xnJzJrx$w)nSFlpmW7!K1S@dMreiF*R6dC@UM
z6cYnh=%QZE<+kZ+s$Mo1o^SS!vVtazsQ*wT=f9veY()Cr`qs#G574)q1x01`o@1nN
z(;_6an8mI671n(Gp!N2zRYY&jvRc5;aB}JZf}?K8gI6zlm9*1W|5S*8qSnf%pEB(o
zO#X4W8x~J&x{Tk3?DTA$SYyw_w3%KH1JGeOeCS*&*FESNL?&%k`4AS}nFi$U1#E0+
zZ@Pa{xRn630u2se62P`yQC5GD20dDaf+Y#v2t5968X{QJc
z18tu5gkAdB&6eQLW}$`Fn9nlSQRxJi2c>`}AU?6ehN;JM#cgy8E1pDkg-5dGv0g;5
zz6Is>!m)GkcXdu$T3Qu|zrh?WoB$smB_(zFw;QC-=w%X9*wgf_Z<&BL!678HVi!iD
zLB6CVZ|_<`lB(w?=f1wfqnBb?i>sEvpiyyB2IjUdetJ-qV-pc}JJht3DindfHTZp3jAQ;bZ8o(jwdox=4)1+WHi9oe32k{Cx!a|7x5LkGUcqcb
zphe#S7~_*LTpZpZ%@`*H;76DALG3zl_*v#(_$By&$O>w0*7}g7ad;=dYegF0H8?Fv
zi$AT@ujw5F1e{TDSRb#Z>JyX*3_7F7?jx{m%I42Q)89b$rxAz}Aa(+%*#P*r*CK+g
zdAr8G6cR&SkkpAe>7=xwSpIXQWY!Zj9yRcEUcH0VEZ=OFPvY4;8(
zLMtA!h+05m7p60L)M9$_B_6D|5fZM=z1ht?um4z0UlF!k)|IHUVF7B{bjVf{CM54f
z;gRHQ0c?@Lo2OBLi>R_ZQoudjpHxwe6?AgDG4d}jYeOmz>lS6VA%y
zrwQtV>!MaaX@7(EBmY{CC>HoGzS*rg&DiIOPP9ApU`hO+Yt?R@Tod1;yiwFTT>AzE2?yYKh#s5lXL4*tZuWMcZ
z|FIa4Bz6Oj0dJKeWiw4W$&s>4NuR|@W}1x(_8AAy%fDp_j#qi=c-5)({5+C1(bj<&
z=k|YM@O*i-K3JrXzoNTQm?QX&d^Hr(%;3dV;%G@062UIdo?d5
zTJ&p+eQ(c$0qN>$uCC&9re!=^#)*y{+G&>}UG-APk)<|nuo^y3vnbxR=6X2RQaQLN
z;u%3RO7cnWDYc&~-;V)!?SLCg)mKoBO^_BKz{t?bE~d
z*~+Pfe8^UBY(*TijBA#>Pzl8hxT!~xKb^srnpcq&B(XGqJ?o7;9G`Zq^CONUJtbzn
zRSlB>vd}tE1SXPCyz~8)nSSL8Ez+=-Mv%xRBE_X%ZH1f+jwaiF`sG77=|CONfa|Fo
zDSx*d^O66InTth1o`cA9;#U}TT!;Za0?K4_o78BwL6=$m(*G$@t`;D53!bu~;
zNq-QZG7yR}A_3GSv~gY!*=)R@%ERn<*=%Y&(#pvPCTFji{CQ4U?3&3pShK^s_L8%y
zna!aW#KdvukEU|r*E?2zzErIoJUQi$xz$v60epz19S3^q
z!pMCXvkY1+OD5`EcNzLqRJ0ik+R7j|U3OaMgoGXIJF;}}g(@^SjWg`JRy>2Y%9E!M
zwdyBP&c910a<)AZpXE+EA&M>x={G?fQ378V_I_@9E=%EX)H?=izA#9eQGSQC07Eu+
zm3~YN4|FSOhMzTSyAw?vyLEI2Eorp+w7rUoVGQJ2&PN@iIpcIb9s0OF1W@dkD!Zcc
zDW(oy{A9nT@s&9MJdW2ig_xSlScG;kB)s$K^nEuVBsBU-;j4x(cu57;JedbFbj}*@
zF|fXU8Wto`KDii_x=oFKQZtOx@noCPfBor{4%a4`ZRF;aqC8|`xHo|6uww}n22JNH
z2?zzCo%dP+@2ppbi|V_|>8lTB6I4<)X`^pCIGW7vKYbubUHHK(Oemv&Ac>6ILaVlw
z?nPVQ82SzH#hst~ScP`xq5RkZj#oOd!kJxb>Klf3v~%4j^29kggJ?5|{x-`2daO)b
zGmP~t@!9lv?z1&%(7e4^&%!RM>(ufqv;U}HxB@S6oaDfOI7B3@Hndo1#sD(lcgF%z
zG7BN9Gi^veBVy%K!l2Vir2W4l{mGH)S5pnXQ@8mE3dcE
z)^7*Wjgfq1{)~VWg4lN-PHc>PAH%*y_!NtR_p
zPIZVKP#IH$>WzD#?ap_TmE$kR0?rHp(~o>3nkS{5jG2&ph-;s6{M()gIu=E-qCDL)
zr9`_2W5w#`r|`0#A)_y6n;W@yx{4OCCt_IxuGk
zD;^0F3lP~aOEJOLH9lo6{{0LVj4!m6(VsPU7u-|WQ*s)$=oP|GT*KFw**^vK*rgZ6<93-*Uh
zm-a=wa&kpoIA(j8Y7vEArl<|W^~9U?tkWKf$Jprx>*ie*wL*7(xc?8^X``yF}Dr8^AN*S?4)Rhh@O>#RXR2opR^h)5SN}lPmS;c5~~?YzI7*D=Je)q_#BP
zHBoE!>kBdBX?D^l{e=&nJv}CJY0*aFIC~m(c?YMgexsVAW+@adrBD!k%<#>JBPNiq(hIqUdaA
z)9jPueHjr*+gQ+Tdb=gm+aT`vW+hiY0Z}sG)6%fwsxZ3W7lB!3-y@<{`hExMpqhII
z)Uk5eI`~k-6k9I{tEAzF2yioGA?&C^0D#wFr7F57sjuj1WPEu(YO6{`w(+FU!Kz6=
zDg+hyrOK+l^sHL7z!ZDD!M7$o==_PLran3kdNNVx?{vz*$~kaE;c~9OVm4uGO{c1
z!u=)BjZOMd<$;
zOsPi3G`8VAEmk<8DzM%0egF)B3S=jv$9#v#&bOli@9~#&1U57o3&yT(gorhDmb?xh
zFGHI(@D-B>cbkfb!Pu%|$K5i$G(4AJHfv6lo!r_CV-n2IT2pt}C#p~|*=8g|fOR7a
z&vuCLVSjmRSa*cC?fdNrO4I7sXuD$3qB>tp%gdEo*_+zU@y=rk#AuEMt9ic_j~H~0
zV7wvt27<@@$pf1tUZLbm#nY9o(FcbW@!G?FqWSyYQBv@}d)ub>1UG&{Yw`&$?80s*RytzV7~q3>+d5XU7aS9%dQt3Ox1
z|8q%Su)Y3ZGU`Wjbs9NCQT+XcCl2+(Vnqwgf}E-u@~y}>%F?7szPBGQ5}K0*gKxVUQ80(zVfQ=RL%)^
zp5AJ@f1VkoqkSXY`8%xcZTsTdZHd+1MSZ{d_L$x7*}O#dHm7^99|knErSJ8qYX61H#$Ax@RYJ8PnUA?iJbn($`IvI5s
z7vAWKkp6Std|sM75lRn5xuSzqogr9fx|8mj0j>S^yUa|?be2KKXrWeEwqMK}kksnr
zDwSVd>iU6#ok}_~Pd>Gd5m2Je&=A(-dhyki#I?Y1k49=%e1@%us2LMv=({*>1v)N4
z>ygd4ru;I(ko}mxCkX!_Mhi7=Z2uWk&xmF@K=HTUyu4rQ9s`-v322ahfoOhAdfe~b
zh+OS32>lc`h=D_52fAfp@#lPprdxj;MsC!3jaBOKUC-NJ8l!3@qKk#v#w4b4a)N1u
zOGtjFD>`x=4wr}FVni1GR+55((iS}$w_{#uafkKvr}fl=3c~jRVezQ7pMZE?Tt%11Kg#
zae{0|jZbCgGirTXf_@j8WVMFmWM)D$sN=DFyxQ|3JUQo%V|;(|Q1hiLUa4;>bUep6
zaYk3*s#62g@7==P?jDtO#ou{PUtt7Oywgdqv++%aQ`-CfV!f&_F0iT^ZgfM$#&$^W
zkolekENSS0WSqee<@4boUu62DACHLyp;m^>0S(3o49$JEW~V$65%T=pe@l?ZiHXL(
z#~F85Bd)dhMJ<(
zM_&8v)AM_s*N8ZT)zm&po8&;nHBP7{HE@R}Fs1E1z+zqlUV&{nyarHf+Yxp_v)m9`
zNjW#T`O}}&J{oo{)e6A
z&6KKPDzE0g&YiuPboih%TQjUAH2t&defr3cUWTI5<9Bv9XF}fdmh=jP7E;i6zAq?9
zhKrmXq-%nn`RJa;rYzX9;Y`I<%kcf~E
zJ?i!7|1@`IFWv}mvBoKGBOE6|Nyd1lQ_Qz;_E+^6lF=4FFtqpEo$G5jjub~QZ__OU
z`D<^c*_T${Bh9&Toj9vIAuemgR1
zW#Rc!x3MT&X>97O*Qidejj48((>L^JXwc0sSl~>HZI@dgCqL#fFzP+p1_PW{c0F{w
z+`(=^(LD_}#uM6}49tCgU`{o7_A#_`DE*gJ>I1fmB=E4ERTlR~dO3Oa9^$>Wl_dxh
zSt_<#NiO0r`Fk(UB0RP?zlc>uG;Uhr^2iC~3A$pXkEyF8nU&*PijnA<9T)GKQw_3=$>Q!i3
zc8;8!=L?6dm3`ev!Hp}3D_IE?{2Na`vw-m}Wn;`KWS$Zp9^BIJBu)NQJuQAH-L%`#uCi^WmAo@GT!@Nv0ZujsdYP!bmGR*cJhIn?JpwZ}WqOyUHyro$vrage
zt5L3N5X?8#ZoyrZD#c%)A$MZUWAshLP8J|Aj~9UC3=OdU{6lwa^F_gfkkOLw0o$=t<5HYVfH}SftUsj;UiHeSf{E!9IU9Ap*+gCHS;x;2ZB>87m
zFHyYY2|;~(+GtTMUGB|NpLT9%d+9aItJ82u4m$zPy)qNA47_lZVe@7GgXyLlUZXkK
zrIS5*C)US;dCUl9sY-gO@1Cs{Y6jPGC
z7nN0^nNpZuJClJA;N+P6-};0X4*V#`208tyM8{TMkquYd4*JyIY4*IaY9ye{B+D#v
zgnv=#IIX0P$UW*(5w8;Ka(tQCU(WR(j{PE=_{!ZkBSp1cg8CUTx4%*LpCqe0jmdgk
zV`tK@0%qFWL)XnbzDdj_Z^=n^ElxfyJ)VtuX(RxWac7`e&~b?N{ym46NUvYDc2Q&7
z-vyZ$9KR>~ejj*iHsODRXK|0?4C6TI{{1)#w9T33iDp!uHrLGnv4z-{mFKBfl830g
zvA?w)mr;}6EZ*axG_&cOe{7~<6{5T{3cCHeZvMEN>}95>4*j`c$mOj1CqD7lmaWwd
z{*pdYzI)M^eAElmmmg3e+JPTKHdcuQx8x5dl;YUSsf(!$EZ+R_Cltq>6XmZWLqeX%
zDsYq*-j}VP)gW$@!(jq1ZKy~G2LKRBD!sexC8Ou7uL5?Q;y=SF+tk+jqNf*%1l0k6
zHP{l#r9s;|kF3ph{uQ|OlNP`4Oo40WQ&T60y_bHSEGCR-j$f<^Sq
zR@`Np(nz;$*0y2byiK>%u)xt{TA>)zOfJ+#0^sTA?1S4Jm;w>1w9tirpPuPJ50W}XwikCY`X`Vd`lS8h3NJmz?Is^
z*#D&iT3F_tSgRlx;X-uY>&>G|`Q^GZdWzp9(Xah<7e<3p%@t~bNfYF{J&L>DwR8-!
zFw3#LHT62&fKQXx@mLrUrkD1{OE4E~pg7CH+2faAaFVmTD7U<-qm^zCq%xE8QaKf~
z8@7ZA+A;aSbfK8#(E&!e?%DKX^$MFUYo}DwL6*
zlT5*PYwd0e~I#
z)DAk9S>A9O;#0GfhHEC9mC7L|1w_??!K|3HO`!x|LN`H3ThEV?ldkkYSr0Gy>CH84~t2vf7Z$
z&P0lOg?EDUtPq6G3yRG~SMN{F|33)x?`!&te-ZJ&Ht`pe=z!#(^!jfm(c-_FMCMCY
zmWm9yyL#0%Y7N{HCx6eA5XsB^PhTY1mtIE+>=M?dba&@zZu{0x7AX7>hQt2O(An_2
zj%*^MObE2x*WHN627l_`%B?u34~3Haz1W}Ru1+!Mq%2S;UvsH~P`#gE_qEcpkM6WE
zc+CEQh%IO2CWA@mu@a=ZRHwtgi-y+cU;P(XU`b#E25m!a@_eH+DwU%3qQ;6ijw7Ow
zH$L}CX=5GlaZsgV!-`=1!c3K=+-N$pRC`13C5_Oe=2xaCWq=y%onchCng7>%nuB1G
zK;Jr*;m!`Y7Pz6b3e4*9t7-=|dfPYA-Js<=vPFSV={Bv>KbCq>y2bI!`XTn>xZL0D
ziG(=P_0IbpcX1Lwp{v7!S}GvRCobmCooOMHhg+T7qKsn!FI!?{p}x+*2`fuaX@OY|
zy>rczi2uQVJWn7k{3BBJLYLw_#Zm894O3Ftgh)~9W&Q2UdA
zqdSGpR~8;Yp{pB86VP)97Li?Fpt(l>gr%hsW@ffr?BUfoPN$~#m(Eh7w)S*Ale9;h
z5IEykP;1SMR<;aUQFyCq>%hi%nu;kdG2a1wr5#=|c>!Ac)!C?l`EIfJXsIk#{r=qt
z1O>c9>h@^=K27m%kHL~qEGE47;+Npyu`-x$JWS!5qQ%EOL&@0yKnG4>pVIJ3S@b^jQy{A1s=
z#`^A986Ab0hU=XBZ`_aqEaIcDH|@RZgNaO^2|`uu)(=Dx**vf{VMCPA^8fKJq@|d2g)H`a8uT-x9@G+sr|L!&d@?JsF
zt0Y~xXRcPle+{Qck_HYMBois(mTuSDp8i%E-O&7cbcwIdg3Njla%?&zq#ao8
z4t(6!XV4o9kHIPqr`J4Qpt6Dk@~1-WDEplQVE{aj6?^Kuo^6c`&~6d?9GAnNk2G
z*dg?Sav@;)r4e=6@xT`k`xlOmRypS98Ean~N8Sk-8#@j#cSqfoE9cLvOl&@sDPG_X
z;S+MQSV)|y2`PEN(4O^s(bjS;6&pjJzAq4nxkB3#DDVLY>(;tB(MbUswRL#R1^y<%
zN}bxx3Ismh)6JRpL4Uh5?*w``+4oBn)x;DqyEeI)(L_2Sb_ZM%;J4bY%CuH`LZ8Qz
z$;iriv+Pv2Ubd&bODxHO&<9qOstf2CAsZFyp4Pw_z?S8J(5KrE*QUzTeq>4J5p3J!
zdRf?c^6Zz-<9aQC!fKsGC}9M83>R)~{xZpmOaMfdJCJX8ze9P=aAu*VYxR!50n6*r
zvxrBt9qlEM)O1sK{#0J!#L}AuBBpPVf6~v}xX604fp;F&q9#g?@7pK`drM+r-
zXf$n=-Z?{6LFV>*(c*5rL?RTfbZ0!moU;ZYtZ4yJL~Z6SBs}wYeDcI$(ptGE&7rv<$0*l@nDVcn
zK|i-GZa4YP8vO#JFb&%ux%hTXUzR5?euxDo#*y)_UlSaO{=5Vs2{RKfA-gF?2!N9J
zUZ%7}*319=y{yCg2=dH)V4Tj)>uKF5egwTE-<1bNY1fGcou1!(3FOSl-cZ_p`%ZAT
zO>%Wb@l-yUpSgf)<#bB3Npk2TTN1k&1Cho7e!GC?5%xx{go&YtCAU(cS6N9coRuF#TV;%$m}
z_T72q9;y)FCudXI+ea^kn6!7TndYqH>7Ulxx++Cqzsc@Fm8$!NsZ+gqldC8-Se9X@
z=YwL$RjHnu(l;rfy!^A4~+pCa82_|=XY&eaLoWE#f53$x7oG?3@_S5JR0>c5+r9sxpiKg@?y+_k7
ze!;Kz1RFlc0(*{cpK382DkNG}jWl&jj7}a!SSww%a5A?Y8Za0pqNzim5VN~29y+Im
z4{;Ny_=SFxTKr4Jj$pkSCb@*
z;p4s}IE*8wVJ)&_*!Xd*DcFp)ziUXtN|>GvhpX1?nNHKtXMj_gsRK{%H2~GtUc&8T
zo)N2Z*<;?4Pd(bA|3KE8%2QqNVURw{0F|c_KiV}XYAyIhv~I`nW^pFv&%NhOs#%B3
zROat+-J5wXdkmcPr9^(uZf0h>hdk?46vVJ@_|l
zMFW<1NeF2neV42?T5)wxB`OHBcsSchzRk!F>m)+*l}m(L+shC}kUvkQ`8d11Q1+yj
zJ}pa*i-9=340~WA)$1Yro#}r{?_@g{o!ojBrJ+?@ARMY9#;1nXmG*TiGTr77It&{4T=MYb@Ws5M_Fl9rvUBI=a$KHlrR`3^y|X=d|@2
zYdk4EZht_guRHyyzLzYm*w*IECS
z1AUgXf$SAgD;d?DZ1_A%(_Vd^Mkav>eLl2|50|DWy7f`fMVFYO*d~|xm(h_`cH!*}
zfn5%VK3;Ac_VY`SOZ_!!V(&_2aX>q0^kq
zCeC!a>c!bO@0CnO>GD!aM-(^N(T$GFhb>&A(Ke^>oEW+Dd!iieMm--&d(Xf4j>Y*I
ztV?$G7{7vQ2-_S@z5ikTF0w^uw!dXHgjSUPaqP(JaraJ(k3TSO7ns~P})#qoS5ZyjS2|pzG;j;6Cadkn5eMr*dMhWdN)k{
zHAV566xDb>k5SZvSPk1^$}c&xU$+RoM!$2IBiq`IwrkVgwM+}!Y(g8mR3u#W`UwWI
zqT+`ksNLSH{2ImZEIXj~G4rb1EwpWc@IR&|N&3<4xN7>#5%b=aA!$f7FKzXG@Wi(d
zO@_Id?C5gux8ompeW?THhZyH-@*DMSNiy5TD$w#a^iw7M_wxdz9+kjX`5C~eGzU!R!)!{K*vg$2qYmE9E08
zN5}6?il*VBUE(g8TRQz;-nL62_2Sb;Ig5T>hUY8MGv`Ra!Q+hGUTAB42EI%CK6;oYC7RAuIE^e*;|!xA-XC&0HU(<_o?ML8MWG%gpjDP@ATUW
zc0tQ^s7b7)myDu9qn7EX9bfk7gN6#(;8`Xrs1-2UP8y~y^~h6;XocZv+(RVO2hLEc
zn>ra@>IGIK6J$$CW+B=~HgRs&Q<0wuRe~?4Zs|>KKZhJl3gEdAU(#Ul4q4lY8*v97
z@}9Fs&UN&WYkT@2TZ^Fj=TKZ3ZIsEHb~UTrK3F~k|7X8x@7K3C2g>=H}8GX1|3Op
zIE`4w8j_w@D8x`&fK*5y@DT$UtV%KdP^q7d)$eFN8y{Wb6-|)zlx&zaI+DZTR7TuVRvG>fn8LizFNJy
zrH-5GY8Efs#$p$9^-q$N)eiee-$rP~&AQR(C+$KxcOG_C(mq9ne`BcjvT0ZP7#@wz
zb&XYDR~WcNsVV;}5<1EPI*mgXk~;&~`uHHx{UbfQ%)V`}mXVYll-Z7JQb~emp$SBv
z*~vMYKMCxIteDIavDQ<68c;obL2+gim00i_q^&uYcvNLs8Dz;=;TiqWF+2n6ef1XB
zcfEIeoC|mwE(Zdr47Kc~@%8)Vl#m}sl~KU*RpVBl>Nk1wi_Bs)HSiA)uA8Z2m>$uk
z8>dB7D>)m;?H8??NxYt+YwxfK!jkCR*gp*SN!!so4F|XXNR5i|IBx7^TDhf0>TF^`
zoW1v`=jFg;rt+nH7tZWSbBXs9-Q3GdZdEt*1O?ghIH~fh8dZ42I;U
zejS{JIN3PLsWQD(ZNhNpRUA&G+5>Z{WBNa!vsJ!CNc#1C>6Bj*DV~NI*U{hf+@CUX
zaP1;AOYzlO!IxfEB>Lg2PY0$Nz=L({V*Y-v#^8RNq*weUA85u+nQqF4N;oa&elC;A
z5XE8&YZ>+N0pP}OLth1E0BLo~gjP#HXz?${_9(&f6e1&aj#d{dn4Q6hMieox1zH=!rUzUtq%$m1bW41N
z+ySAv|H6H_rdx4bH1@pb54Hy$k?B_K14?yQUd!Jd+^tTzbJz1UX{NuUj3zW;K&IE_
ztbij*axwmwkwlm8f>rC2q0iCj(sf45C89z`5Us$$jLu!X6}DuB%lTyxAFpu=rNf@o
zbkl&yhTq@9ZM?_pDrTKLXtn$SqkY>2MoknNM@vDbv8~FYFA2sq?~Fty8X(2jZiCO
z=bWVhCm93OW@x>Se4V=F7U!)uEG)ac6GrLNHYQnBV#D0m3BZHK@g@Ows{VnU;|c>1
zrkHawCqwa8%Md6br&DKX;u
z_oEdz|9LlJ?gyyH-x+Q8>KiVXI{3-R6ymN0*R8+bwOM@h|DpZxpYOB4sQ;M*`+EyS
z;jI7cLB8yblG_5M%Zgci>;J>sTZTp1wQsz@2&1HmpfpH#cZ*1OOXo-oAl)gTlypc-
zhop3;ba!_Q9Ydr0Mjzkzc|Yu9fB7H#tH(GxYu)!+*ZN)Oc`h4<|FHwM;QwvF#Ppt`
zK&`AEp3768TM8s5AAGvuNS$l`cW1+=59PikqLnvh7>{4gz490(5olo?Wr%|MoYl`z
zLsIzQOf_v^U0RI(acIb3?_N-;-~$?qJEs6HaUaU`Z%LdM0(c{;`lmT!?yLO%1|j#)
zzbh-}0cXx&YJ&t@iQ%u0NXOb*fh!f{vu55l-hAo9HS$MQ3PFhVPPwA(wy#p=V2OA-
z95kw2e>xZ0i>}dhU^$q=-jtiYQ2JyF9rUhpYAgEA^86b4h(hiqC|r;$a!9ster2?*iT)pt>l
z&E4R->@TuD>O<{nequX&DT?d9&G48BTD)FbP4$4Iyr~@-^>Z_ADHsm+p_7`39xc@!
z7ZLFuU2Tul?y=NTM7)mPQn5-OnKq&(NjkXcn(o|q#W^MN?*hhFcaX8Tc
znT9k-l#44(7Bz-7;qxoEZUCZFQtBl{Z^@Er#3E^!@Nu-n=GeA;eH{#Dh=~@qX@NUQ
z4roUnC!3r%$Zlw6bcHA`*o$DCYdwDOj1cSV7K>Q6w}P-lJlV}2)35&9##@i09j+UX
zr0~?^p($pACqg!gol!r+4AS@@=S%B$lNon2%dHfDS`+rAB;@d0po`%!{!2bOuuHJn
zZOl<1t@VI^?!3sL<<&VoTF;h0QmK%GPFbAgkzr8Rm46YIF!v1$b=Jt*7dDqw
zo{fY$;dqkCU5`1^cI{?`ALINp+$fu&YL2SBectzdi(^Z)7&TUpT{W2BI%l$~UgMk^
zqHCus|JQTT1`NrkZu9v|NCn=BD<6;zvTI>(2GVPK(dQW8vawGU@@@#1pp-``kK$wz
zdC`G>AQD3;k{b=^{sq#Pb60(i^H>q?m4@=UYy-1<^A1f@Hs4tS|8eU#6u6!oggv|s
zAW6iNdGJ@dg|V442c
zK>N#Ad^i5}o;T<;jIzdvkvK0+uIgJ)ek@jxq;)&HaA*9yD|qZPRYu_ty8Ne1-
zhEVe|=16EBU?m)T_1rk>SNlFH)OeDd>@1QG3c@K!!!)OZg$ykMjxj(BRH&hd%x!un
zAI4Xmi89HL0rN16cV2{+`gp}(tY4L=Y*Fqi-2N(r<+=5l^a(K!NutQ!(D7LGB6M}$
z6cYV%BU0Xt=E4|!iTAV#rmWlD6*!_E{GHz0m9iMe_t3{lTOCp=`-b_{geA}an>xV2{k_jFe
zD4?`cH6$?$RIo>i+tvu@`=}uu#hFU603oD78O4A}lAT5@yxTlZmL9*{g_HA`AW2E$HKg
z&TagW&u&Uhts=jg^u6UR*n@kn{Itk+X{5qCYAwL)XBscw(==tOXLe3iRlb0mGE{RX
z)3>0vl=~7t1|5F>y6=&Naaj6eN)HRcJDoj)y>&xL1k?Yj8%>iYg}EsWGv70pPQwFP
z`};537w{D%17^fMt=X4WquHKt^aH{r+`b>v%gcHq
zmDK&y&%?J(TxtA;rgPgrU!5Z(?t4c%oF^3lhyc^=spu!?qM`)l=Ql429GAoBel1Ag
zLYl9cX4Gd(u%`x^`naFL4XC(8R1ky0Qt&IUjZxbbt?E^_5UIFc5b#05-jBnAJ3_e#
zRuYpD-7*;Ii#fzBj-28MIZ}b9(5E<(D?5fq@oGE1EKAWukI#9N8C79=ufUO3Q&$2_fRY8!CbDCyB^`oEJ3?m
ztlAS*ofrta2xlTsy(RDfha&gy^;`?^0wMq>GWgK!Jy(J6jpm=x%i^MXJ&6cUuO0j=x++Y8jR%)T+o(c-b9Tnu0)I$S4-_y(J*tb1v(
z6xFLc*~KbmEOh21S?7JU-z|Us^oIISrFxkxime3?uKPZ(;hn;W=KxnXcsz`G4Xk-G
z!yS~{<6qy!u9bb`87GvQi*dc{oTSvW+OR)H7L8H$ZFBRWgb8(mUSbzdE&#JRw6E#a
zes8d@-Xl-XOi@8E766y~OJQFh?q1}>+nk6HEP+`-eZb&@@Ovr2c))<`6TyyWOxPZT
zJ@@k7w=*?Yg9GSY7^&&+O*Vx;UNzoTZkcG7eqGbyFZEUefbNl*>Ys9G6y7D|#5OqG#!e
zEomGKrY1-*wT|F=G{baLe`z4v+sx;v4-eyNL8-WrgEb$Fkp0qzOKE$4NdP;C2}7n(
zj@y^n;m3qfmQyd*$pIytUcp@K4Q}J&pRgRsM(L$1-8jv?v3JRmy$cc|(~9{>Bii_^{6X=r}bI_1X9Y>$M{GjMGmnVX#P&TR#NaF0(Z
z%qU{hLAT4WwYOGW8h3`Dudym-
zGm}|ccgO02_O&FKC{006-ydB#JcBZ}qIm1ZALnamOjlKuKR#ik@0c{GY9;J8bz1&z
z7JY8!ZokXex0#|vunsK->GI0`I?Rm=bYp0aFT4UHwVlOcnGn{v0AsITX
z-&_)%Nte~?`ST#0@dbeW@LZ+}kCQ$Q*D%?6Ju}yknm>`=(
z9qW%1sn}&3Qb95h_a&AD+{S8uc1U1H#1$EZM8CeWwWF4C12b4*?myt^+Z1dq4m!C%
z{=_h7dIzD@6vU8}DpvGHy-KaX{`{7IE>~H(6NJ0xd5OqZJaaC!v88I%?;^C)Re^sQ
zyi|6f#vZ>el;3qT7Xu?OGq5Lj&K13*4l=SWy)&Q8uqh-g9p0y4{ixUtDm$$*p?gTL
z(sVnzm1^|n^>8Wi7YDfQ@J^ck06G=eGdw1
zCCvNTI{Vs<9;Eg;oGT(APLyEkJa0y!J1Ik;MIXFs!e5=1`O-8s=23y?X%0?%73{jv1i
ziA+->9UKMNvSWP*KMX!7L_oqtYuVIh6>UM%iTOa>dw>1|cC$Mo{N(~=uNEW8ffNi>
z>wSre%@Vadr}=CpqX6FuPddt2TvF*oljLn?4D)l3xuV3Z6M6v0;Ju{H?hP{$8dJb}
zdbXqmuBPoY@9CpU?{lmwFzEthC=D1BTm0jL!LvVUkKV;woNF4;Rv(+o?P(;XI%X@G
zT3agAL8npJEgZ#enhB(R`#PEbfxhO9tclV?g+Ka
z&St)&k1`4`;00=F;Cb@Rgc0o$fOYg;y1RLM>9B{~Ek&N^i%eaG4t=L(}5M9d>~xuM3(kgk{IJg=WcWdukGqDzTC;L
z13fM|3>{`;5^5i4+B}K{WCDRzLKMj*Ez9(DMxxl1&#r<)G92^2<7-kpgZ0`jX4Yxq
z>Sk}+b4t2$XSag99^<%AxmTG^V8cFTZzZV;_Po{o>cXFWe2Pi5u-ckJY9Eph2q0T`
zBHtFI=u8riZa-NH52^IM(Hc?%#f8yv0`&mgcX}lzaP^{U2Eun~LV!@aauploRExEP
zGINxk_|CHW3&*hO(1n@$ey7`uw>(+6OIWWTr<>ue3g2&k50
znhl;dIGH{W=xA!JvP7D#)5^>#v#NDzd5K@tbx714^znC=6m@ES6k8K!r)l1!~tL;d>4KFIB^BIlTe-g8%kWhaggI#qRl7~rz$O6?cI
zxS&;vqMy-V&91@hPrCIo1J7r=zR4_JRktEtFU*#pt9P4c_jP{aO`w_8$IF?w#jtAF
zozN=Y@teH5G%nHmB6B@Jk@%xe|b$}=ITCl(fi=g$K=BA#1j{XIL+Jd0>I5Q^E8
zg!SU(9Cn`Hp}U^HD+qTgP~G_XC4tys?l~buuf;+|1k3au(Ey<*iZP=gTv=B1@ak!Z
zfJhlXKPj(~yV+$eiuK#l=~%_p2s3dQv5Kq6>4D=n7p{D}dijB|);*s$Nh1V0OZT1k
zz&E9$o9ykE3A6+H=_{VDr6-G9GZmMk5UV1ue<3|`D1h|Z6kjIsm+yhEOV1xDX`4D;
ziH6}fFhSWeGW|H@xeG{-;=$WfnR%~*HlRVG>`$X@A`{bP+x2yce(7qSJ6gniVrA4$A12Y_tTM=UvI#R(Yy&&%x+Nv<^euL7gYjjh=mQ3xpLc3e!rH2<
zY9*iSaE6-ZoFgi}otUS|yvhcZRMl(ze0NtUqn!ZOql`GP_hz}Tw`F@wJGeFsD
z8?#rsBi-76vC3NI#s771p&-UrNBz-Kz-%RXKDf5QMKf2BaSlD%6|qoGRB#>XA9-z!
z?7CGdln<7B!IdJUv?Vh?^d;qguv_0>AujMq1{bfcj)a`9J?Ew%@559bPsLwsME(E9
zUOs9F|M}wNn=f>J+E9EguYY4Kg!^UvkHKz6_L{L^X)*iTOzYH13-tY_h|!SZzrRfr
zV0FJ;<^TRe1O%`DJ9V)pm`cIod+n3qgPw?2egDqrM6kvC`_B{(ec-F+erN1ImhKD3
z*GCy&zNlTu_yl+V_kYXw9_2|r-_BYB#Qcn?HEUecUtuY@7=vTTSS_zy79Y0;b=mhh
zV*;cO)*ez^nm}%FTqgsc!?CztcOK0>U%=nj!!q&NcpGr8`3t7-)e_~0z|m8PJ>)YC
zAeUYQy4*iK6g-|2Vwt5q0^(ATPn^YjkJtcOJe(272*i!>zUx}_WCENBbt>76$AgH_
zg(9U2;W9mbPuzHD{os~@`B|1|Ws00qctfhi4_$n`4PpNY^W8yWwc;MO>P3yp9R?#u
zP9mj?8m@YXt*3^Nw}lR9(`RCLhfFQ(rw0Gd*m}zTEK(`gus|LQUdL`oWvh13xGQg^
zexFpYugM*O>@}26U|MZI$@oz(?ly>hkxi>Q^B^(GSsHUzv(@0>40vVCt|~5U00KE~
z_ieA_jn+be%C6QMU7U>~v75T$sB$fA`Mh5d6T;1dN~zH*v`bKmxnMbn_`shz22$6e4-n@@>fvi=U(FXBP-UP*lo?fjt$(;nwe8xU;L`
z3E%%X3#K!`C11&-8wS5r)v~>6Es^V~ceGQjJ$>uS5~{l!BU)4h(~nzGRV){aw;a7<
zBT_juS+eD0i`(!}+NUyb-I$^}ti3EpGQbnTdcEFF4}>$T{Dua0iK8`BS)SyY7?5k^
z`zi{B*NqhIp3O9{{K(~;72X+<1XH(lyJ~puvAH4AJlU9MeEFc)SQIhDPzA3Xr$Cta
zH9c@<7uV+hCz7UDqQBqile$uov@+6Iym{3Vrbt=_4
zzpKxC3y}MMKPAnQI+Prju`RJF>VITz>z#(FB`9asd{{WM4XBC9x7v?dtK5Djyb|PR
zPaZvrL=WFf+3dGAIUKHUU*8Qw!8ogU@mnQ4U$Zs$#PRS5f;}8kZB^I!Mp?H(LY?Y&z*lEJ;fQ_oK`4e
zRw>wAp-krg6~`jEYQ2Sa#Ut{>&$CG-cTh%MVzNA5ax26Xw-Px64VCwCtZwyK`XFSq
zCwdeIjfoy14)ILnb!x*~Wb@qHF`a(KCjtxoqi1GsRE>klARST@yOI88s8ww^
z4c22!&x>e{aea`NznL&pd?=sYc6gmq@0M@<6uMsEZU?@0klg22<3(ed{(2YfyP{ex
zz~9sVV$#3I2h;t^-%z{7k38)7b&Fr);WAG;RFGcU3Je6FC=PP)c;I75#!21lYNgzk
z3&GcpFr0;hce9y%^;_ps9&ZwSD#?T-dEBUerFv;WhFC!SPa1ByxWetgyj9LG7Pp#B
zH&S#w&kHC`%!*INkjoqVT8F}u4-rlt=6g>qmY7x8Z9e)!yHGQ5Z8#s<0XaVLtuwoo{VFP^|H%^dogh|FBvC{r$(S8Q)4woV+~BR72#^_d
zZ=*qihEm4`Bb7b|P}{?x5wCsq@2vSaieVPlvPk~jWmc|EPrr4H8;m*!h7H&PVpZAN
zcQSd89EDA$jxRPq%ADPg=R}Dd62B6COq48S8;%LT`>S~k5j5A*3wh}2%l6WVL@W6Y
z#((!yxol7Q#&NZ7#`!AnE2d_L-($Cpe)+JscdP+qDn>(ds#{QKuBg~nzf|~J-Iav9
zcTZD$@Y0{pmaj^;Uklqnq#!`L5MSCJYS5qnsJG|&&7N})=~GM49o0VV73;2lf!uZf
zax8R|{G&!XmgIQo!g9oQo7N^<&Bm7HPTp$a3zF(J?dgfp_6-wI}BSu^WHfz?S%1WVFmV*-B{aT&|Oj0j0j~5C&U`^7
zK0XQIu8oIjS*1RV)U7AA5yTPyY{QV-yk5F&zr@C6yCY1^1e(qj7N-q#cdB=nbdPX0
zQZcimf^8(S4%%zgGtU+2ijEw$>52qH+Isp*t{2fS%|*G}65BjXwYL{
z-UO@RjN%Pxph;5JOM~PN61EN>V^H+}k(9cc9(IR?8hEQK%xfm3!t>#g4nhwu)DlnA
zF##C^7Y1Z*L2iarj>IXhdg5tXVvj%CPK8ICA^kCv$4#L8PsADCLi#9Zmfcj5x8g%0
zj`0&;8wscK_-QeZv^r!gpZGbP*Z%1LrE4i~!LiM>*Z(X53j_hp1dA1
zYHIlqdc4(8Qtp&K)J8jS^|TRN*v0!$KE$nL;5wedv1=1kIICIAU)5b*{^hysw;!l@
z3zRW92r?iyk!9wRNx9$HdB73?&uo$VtfOS<4^BMT-)^!PX}?aT_5Pen`j*rnqdK{0
zepg)EJIL*3_cCtOAf+aqSnb`>YDq;a8SMH&otD_uwt
zbrglU_wA)L+!h=TEQ^koa^dKWMua~E!UPV2-n97fj~ya^zi@JrkQs3ZlmF8OS|~3nvN2n-}=1
zmop|-QasGQFqSp@(D*bx1P}i-9euF3-B+`qPy=qI6D|L>6;i)X_)!i`w%{p
z+8Iz9`Lx^+aWMWP>KG@Wc~*f?oF&*kr4%~lu9Kkwmh?wwBBHL9*1wk<^F~~tdxAJa
z4Cs?^hJ3`|#?#K^ZOEp2gkZSOo2d2r93Raa6SJF9qDWLvY#
zeS89R%9};q({?~lGI4MTND{O;YtfCM^Ft)1rA$)_bZ-Lh*BO*?zGaXhQHDle=)_I%
zw{$-dD7e#`e94$ICR5MwNtrAEZ2f8Bb33@$F%E8W#C=-boRYjBfZ#mdOsKf}=rcTQ
z!zvak?lj|=Bsm8PG&UonvDPfBIHPlrqnPG>dFG~;V^J0g-Xgw^WBlKS7WO@bhP1qNF3)mqd9}NT9yO=rH?j4~jTH0#
z^irYNu3=G6Xsv3n!(1ba`?&Oyj&OKniuzd)sU&)x4$3hy60Sw;Xct3*!RqtU$H_o_
zb3hJmVoP&RqesL)?__9_Jl>4?L753*?*vqmwYnOMak9vUj&1S!*Yhg%IFNSD8%eNThuUOr@m)xi6ieDL#}0gaD)IGU$HfVV&o~(>`jb&W%d%&
z>w@J-1JUuQk@I)F5q&WeOm-NQC{P6780FN;yB;c}_DvqiJ`}br*WR
z&CZJ1j6&q87|T`n)@U#L;mDP-rXyUK`Nn-a|UP{Y6ZL*fQv59G9&Xh`zG!>!tK3Bze>b
z5ngYuvp92<0{sC~Njk{*xl|of5VgnTx3Y
zjgLiPt;t0qko{I;-87*XABy@c*JM2=7I`VI7#-4<&95KBbo{;m=Jf4+C8ckhsiFu$
zlTSkjOGi>qyeRfBX3zHV=|s}BZFy-@mz<`6-9z3Km2wA7F+rHbdBC8~a@;n~9$cou
zr-={rEvUxtB0Q_7az-4a$NC#z>!Q4W)3WNQ-7h^`vk;8wAkTvBVuZ4okYm&%yhcRe
zLX|ws)lND4RMO@V<>?M?M6cv82Tmkh{gjJ5rYiXY<d!r#w*a}wpt
zBY4NTbkb_6gZpsH@_tO$5Wj;%2nMbuuPH`Or}FU_c|ouMWwkY``PoNia_%AYvW#RD
zz#0OaSK!j$ZKcJI-RDrz4p>&MDtO7}k8XR#>gR(T{H;ueqbe)4hmm4jr|$`4#t+T}S}O4z~$EM~RyJw_e~t
zYVOL(U&EM*0oc}_Du59z$1+oUAuF+>@nHg1Rb;)SPbh>A`}Spv%YJU{$PNd=z*BeL
zYxv6@#|-YsOSq-`m5hY$zcd8lKYMUKn18pH-kIYC-fcCoTZ>FiCk3ZfIzLcx;O0V)i~mS~l9
ztgQFl{P9+%SF*%UMaLmBxkf&>
zjjahia0Nzc0t{%=m0T->ft_h>T3@2r!M4jdeCSggs!*c6lJwFlg=8L*yI3
zTYdhzAo=f@RE-XGvD+`p*Gv-!RrwcuUq+(2+_>+V-M7@#+8^C&-wOdbx$O18Wk!U6
z=K4HpuW>-Y2nv_8rWUdDyQ40li16On*ex)pMKb0)Bg{)qx8-#TYe{*1r~LO3d1GtE
z8;u?s8Wig+f03uY{l&6(|WAP#z!Um!RicW~?a?IqJVu42zI_S6j4?)O}H
zJ)%e6A=X1p2G;vDMYzN_pPB`E64D{?S0n8-PZ%Fc+`p;QGTJ{dt@M05k^eUI(330+D+Q5Ie;qFrLSmgv(vQiKfZjF@Atm^pB--dXIde!>#J7WeZI7k
z*`>i8E08SaBoNEwW?Abr3K?3;?B)FF@xwEXd)*n2qyIW7JkMPxX@R|_i9Su3vl!97
z1qMH3$yu_X_WaUNp}Q+ZmXa7_N$|b4LLw-a{0miy-1mQM`Fm_1w^ll=nAL6saegZR
zP3OWZ>+!gF^J*4)BGOEi)3NMc>0tpeQ@qQ)Ye|^OU^{ljEsGAP8nIZ3sB>zX91o
z>dmGEU*KP?*UVXNxI88pS4geiDJ@9G3!XsNe75aB#tN4|S8Gvz-!u69_xPktscWVL
zlv&=bciz6K_GtD&iX=h5Y~iz?fVaeSHB#;Gwgdxc8~(eSTZARP&T@o=h*tU-X#Ui8
zpxK{fi4H7&%*mizww3meF35|Bb(r+KK<7%hbM$J@Jgh>a3k*xsC}K!N>@J47*Kcx0C~!LCb5Z#6yLf}8^Xki4ct<*I%@jAvSrNwKQf~VX(3gI%
zrCgL|Y62fZc&6{r5u#1u?G?~&+moeY5Bh9kYjVh;U?!jM9E4ZIjn6SWWP~ovpLK4(
zmI?%2HnykXD6WY+5FG_!kyg9vve%COhG%LIBtQk;yh|`fn-pHaP%ZomI1dkHay?7U
zKes?JUMl}tJ~SsXUYzU6*}^@?DaPMRN8&u~IZ?^XuloF=NYOs8SZ!$Q@NMH#ZmaQC
zeUXp2*NyHo@f*?`k6ihb_#%d@U%Ga0G2ThX#)_LHIX7B>YG1*}$m9YDmn+(KPi8-W
zU~MNx)6Xs{C%K2%>75Vy_dR+~KOmB5>K8W<8+fKPH^(rN!%WY2OnYK#Kw=^ytLo48?-#I%cWNPbji@4Yl_V*y2pp_2j{FDLp#X}_St0i
zCG?j8RN!~!YYRks%YD3%IrN?V%v7!GVBs+g@F>w4QLGafbBU}^I
zIsVZ6JH%g5f#uGT@Y?+uh7MYIc;^0-?>agIJ^@69gm_ixGuu&1!N@%NOiXhO@ZK}h
zc#^qKIBadSj{PZg#C9#(dU@&QZHRO$LxrV>>lUoM^f>+&rNtsx9`K3ZJnYcqVrOFf
z?~WGyEVA8z3An&5$13T!8m&4Ct?J#c6_#?)cWA_0rOl|6V@B$$$xwLE^Lao#MXKf2
z^84tdy%Gpmq&@m$N=iU@K4|GF?<%Qz!}WZ*tQlWam@^TU0-{dM{tgA3M;r5IAN2|5
zqG*h3o$MZQxR(^K&33nF+>bO8?r2%siw}VJKZ4e=%$uptW(6_wKWt1_uN>oFY>e2X
zGZ~9+x%3~y_lsbtb4Sq|;PFqC?$@6QfY_mIi&quZ%%bPF&S{O}b;9SU>kF_5*tMB`
zv>^r8X5|@fCG=)ffm`ITPg(&&a|1h~bt9j@pVF6czPij&kaK(!AcI+C-_l0(u)$-i
zuhUe!PLEoDucA3R9J^`ief0Tw>YSFNilm;&1G#pH#P@;O%@q}1{^!8(E{b3!EZILU
zb!q;ic)*(NK{`U$92d6Ku8vB&${2U_LPWzUs3|uv$GieITY3KXK<+@|G&th=wrv`?OI8
z!IFD?eOf|fOH2&sOo;nlqf&&65Hg7wN)R-*Ua)^S;dvub6c^7GYEHDEBTJ@PTHPc_
zSr7#woo?F(TE9TYu4082BPXmTyFGt=
zPI<75mH&azipPrYU#PFKM-i!Os&6O`RtGL5~A{9juOd6HGT^S4vkXL1r0?x$f1
zb9=iB1?r7IpxK#@U&J{KI(BFEK*FtXFZ%HN^!DZAUVwrMOh@-=*o$(;D)DtKZLgn(
z8oeFBvfuc%%&tVjSWwAubN7?Y0C+MHT`=Na`VGtsRjBnyq74<~5u&0{HVrg4st!n2
z!rm))f75Gr`!3$oT3@IL^o$M5t+ol8FMLzndq{$JSRoi*iv-8x*`nRZ&g{J-c+2q-
z<0V(VTeVo`mfcO3tJ!g_MvdfGydXS!od+|0E%2iJb24%WT2i)LGIq
zNJkDchbIZF3k)#W)K;AsvqM=|FvMoEULRw>A92ZpSm5)ls>%jBQ#|$>M
zQr{gH*dLqbG!|B%WhN5Lq%%5o`tH1?!MqL&TBT>;v0DBGG)9RyMIKe{pX^w-v7HniK|XwkHLJm;nQ`dPb}!=dCYz2jeL`VsWDP
z?Raa37AK^bUe3+K`S_oyJ(&+3Bp#V-Ho@??T@T=Y*IWIhYy%BgU?xSdI=p|-jOWG|
z{ha{M`w0NFfAc+2i->3&#;a+yJ|u9#mT%ybe`^0veCUH4$;jyx>jlU|z7bBi)Hr@$
zzxxCjU1xP25I~96u}dS~R4cT!u}NZ5bMQdBFwcrT4d2_Ac(lBVS2KE?5dk*Zg7&CB
zR;Ti~${rHz41M^+!0S3DP4^|()0|uktD$OVTgxMY?OL|DRO*11aio{YGO;P_N#&(9
z(ULDEi=t;l<);9+lx!es0DEJGlwRcuCyT~+~6w{ZKuXTGq2gIPoH*h
z*G`}Dp7UbQj&;ODFi!^?H5>PD&}jSpML2LHCSLS%pmoaN2fX6{q>ovk
zq@4@~`XFnvw2k!7zO78DpV#5L2F-eNWaaaU*wfKw%NF&)Ijq+N@k%3xoUI+*aRZ$Oj6U$4QDo0
zO6m^eM;)n8VeA#0GI}OpxfKt@L+W9W)aLsb-KzhLhA}ZOy4`Q0#U_z#%oqLqwx9I;
z0(|w8zO>k6*Y~#=cLw#Muo@dbwRhj#2n+(1K0Zstv`9SCGD!}EfGoU6zhHD{`7PyS
z=JsymXmdN}Wt=n1(E~#^ZG+jIY{H*mjAAJ{W*7wXz-w@A31^Go2C7@`K8R%WV&r_1
z3pKugw0|PT%Y#lVBbX;$HC8ZhiT?-S}TVH&E^Y%8b3Y9=aENcQLz>AZK^V>(F
zwYdX0+&YH3fVPC!r2WJU1%SiO2K<+DIi59F^FE3xui%wsTfUl@lc~YH*Qtq?|JA91
zdK(Jr-h5f-48kch78ifJ&!G?z0=lg#IS`c<{>H?+3cU)NZ+K)@vwYit1YN
z?ze>P8)`^5*CF8VP#M<`fGH$cJxOD|JAb{wNu8cX&zx|udQryfT;2lgb^ODi_o(n4
znW?{(Tyc5|gHJgU?%OwuJjp3?ygT%Q-NM`^xvA4Yfu0$evDE6`mh2{SFK?l>NBaeYsUH#mYlJy
z>p10duDz}9cVdFRhY0IYw!{%BJXzL(sS~6hVS<5`MG0j`q>WhLIUZskA&#*as&=GW
zlC#n2?n@c)x3|g6)XtG&E;Kzf-fG<)ocU;WQ+}TNl5cE80cUd&2ob^1NpOAN8@Fzl
zbk!FNu^XoJ7a)XWE2d+a^~m@tsSnyUZrT>H$kwjlPha*{y@dB(kUkHOqv!!SqxFSc
z`Q;Fz7qngj2R$EcJxKwRUDKWW2Ic*sfgH8lkcC8gP*=d)JBVV>z8O;^1;?(r4N2DBWN&GR1OpdXR-*17`zMD@5I<4b%^0cxt;>${cVc^*+-@aErmGibZAW5`{TTDBODrRs4SUZ^c>z)|;T{x7{gP
z+KP~qE#aaeleng)*u1${;dGe7dw163zqp+(c07}T`=N%R*Q4sTx>2X5_7H_$gCs)5
z^+8;<%kHYLa6Z$zy*j-YeFEV!f7RE)&^Kk<`=Bii?;3YO(%@ApPG3O8yVp`u~Aq&7*^^|Xmj4!qI@(;+@TqecA1(u^df9;Dej|&9t7AN#Ws$u)7r)_;!e1=
zT9-%8OoK3ii>gRZYNTTQso;R;w2_xYaaTUJ3dw88r^wxs#>?r;=1}mG}u1=)vB|%Ff4}
z$z5NPA{_-RfX;n{!T%&chA8C$uEAIn0QFdcC*7NW--i`zdGcl1jkEdN`p(79$2#Bu
zK4IE?Z)%%p8|A6e`$9I-)Hba-ya_*#Mb~bdFMHB2vYp`(zBGX{>goxPb*LP*e+Ke&_!jF8nANn(U1`i
zO*j=9>Sz0Ot}KX|J)=@j%sbo*%pws-!gbF*PrQg1kM!w0GSNP-pMuVsdLHMcaq{%j
z{8A$|xmqg}_X9NH*t{;A6^Mr#oup(nV@bvvE%y?Afqx}Mf@jRM`R|iOh6bO&7q{)u
z%}4p4fTXAW;HmQ#zsT+_7!!b@3=(Zpd?;e21t^9Iv~y!3w)
zD7=pERql?^d*jsB$Xy#dTrYJW+_Az3O^V+IjHiwS38ik0onf)O;23dOrDFYPUhA2>
zKVRC+;WyA~>}UR>#_pj@>72S6n}9Fvm+)kPJE6OBrjzf#-bt#YOgWRWADDP+G4i(e
z!d0l~wY;2#&BvClV77yD0CkK1WC0qQv$M5daXv1Y3aJ+TiMclm1ZHnT@Ve8j0`PMg
zf7D;a;j9&jRo+p-^>zfxu7Bwudty7G--FtFZJ6dCZP>U7#4nqaisIy44~Xsj=$3Z!
zt36SVPS&Xg&Qfy%5Kq&YuzN$gSwKNm@I#@!(bxR1o|GmeX(@hP%7JxpVW&p2nqg^n
z#1kExF2I9!OeIFIQVw+7mW@!&uM!;`pxeYO}a
z2e33bQqOm2Bp+Z9qEc2lDwc{*@PE6U6offzb)>4$7_?>-)D9@F^>kb0{ch42
z=(y9D=`_95L$>43TafUG^||$)&rda}1mOY{?T5R3Onb9@WByewxK}FTZDqTm>EPVp
zpS(sJZScOo@#0Xpoj*dA^TQj7J>H!OLi6eL9UhD4RYH8m03h3S4*>93>_k|oRzi!Y
ztoAnUSa|Ri2
z*TQR^lhnR-&Z$DJijU;?0gU3o#TUpu{SJ+B!zhWq14IE6KX6?$%?v_tG>e_-71f=oxz^h#^q|@dj*bGvu@ws?8V?^zGaE|Q>I0?bzL#7_Dz7@Z
zUdXOAIP(SLH`y{Ga)pk4b=~;bZNs7XOdd|!w>&`e3&+4EjsD=T0PV0G5yTA2Ra$v9
zk&h0WZXVm9n>(|QDT
z`W>^tEQ=&d778Fc%Ut!4Oi^%(m2J56VXN2MpPr5N!N_qXC@wgMHdbR=QxZz7Bnfd&
z2*KVj?B_2H0lA4AII26_kJ5ArM$gm~2B{3~YLdn@n6R~&>>miE_(#}DPVE8$s3^3!
ziqZ~rE8@$E3Lc_EE@!QQ;c%q~a;QoH1+NsPkArFRe?X8aZUR*5Aquq@t$3}GLqPXHD`Qe^u6aZ0;oF*cqLR0LQI5T^h
z4agvcFAgJBY20>6n;eG{y)?W?<^2m}4W`xcy6w83;J#g%mfXA&53vBG-)vQDiI}mk
zuq&c6EjKE5$HSlt?CIy!jD_!Eu^L@Rq5`FvkQ08KLuugV@pQ7Ei_u;HacqtFvKsDk
zKG+~XeIJVmBIHUJy7!6Zzc$;CtzoXTN9874tas&{&$h
ztd}BR@q_V9nIqPc{@@FT0iov`ibgGM!N46z*S-%K^6n0eA~nEzu4+&UmQ$Sborn$2
zbFI*ZYb~mRy`~ItP}9Ve)lt3FL_HC{pv?%=4SzF6PNX*Fk^9UiGs1G;=G_Qk-^Ky+
ztW541Z`H(lNnVndMe;!}GBlM7G}0uw^3SJBQhXm2{L&!CT-|W{MpwoE!lMhTs`=7@
ztXXeudM*}UmKxK)dQrP)=sH@d-)B|(?l;0SEs|b*HgRafC|_d9X>o%-%8b#Nog5OX
z<~u0b-B~ud8!(%mN?V}O_2MJm^f{z+X&=dyQgLvMXfid*6~TbWhnR+x`I%+LUvM-5
zT$YMRaVI0Cnx;_Vp?IhFIV9PRSI8-=Q+IFGv1VGV~&a1#LeftIRD^!)b;f|HXKHz;YWtvIS~
zSlVivt;IY7EP_dGWx6=+lM>`#DYJiLx+~PocjCd*Hr#N6h)=0xto|fX#QU^#)eQB3
zn`h7|TRnp^MPt&vFtM`YVig*Rr>HC;n|DL@D-`THu_vy7S3FpofZXGfl*-$-t_K?t
zjQcd12pe{99F?=;)`*{jFE}ROtr33EL5DQzWnsV)ztXz0Fg=DN3-
zShcyLw!D460T+4V+Nw$LR|Q^(0%)y#dSYXWIg~twrkK}b+J(&lI9-RHuzL-1@HR4Ix$N5n^;kVFAt$jAbG$bf;jJ4D)48-iX!d-RxqJU2fr_m
zN8N0W*B-wRT3l?|1dp`wp^hPoHOLWoc9*ncx&XVM;}JPb
zILwE
zGAX?~hUa)OESP0qGO8ZZPdqB|vzGN;?=7eZ=Sk|Qn)rXHd&{UOyY~%LL>QD(k#3c4
z5s>cg2I*E(S{g(^8tD${?(S|z>F%z9fuZ9(gYWzM$64o`Pv@+4_zbL>XYXh4`@Zfg
zCNHWeOPJZH%W-ZtzpkIu8+o@T^{f~$G38ryUJ<-?@l9(=9}iLQ-Zn-;^Nlb
zk7IFxYc!F@d;zgB;>0AZBC+CXQ|qGr`0*lYNgWCGTc%0rhXeSM*BR5KXL
zTN`wksf$^=fffQrM@~{}Tf|&1wH%NVv}F9oSkUFdy95v2_5Xyzf+NJjFJ|x%4i>?l
zhefS?cp_HI?s-=k*WShS0sd&S<$Parg@uKYznk%Ly_{o>#-Fd}`j4bltoKsDYG*UA
z@QXz8oV(^d&P0m@h%Cx@!s4fsNv9)Z1(D_iaI<@p0BewT{G&EL>DB|YwY<{7Uy$LS
z;5W0As@-!Y4|TI+H^!H19#wc3)q3Ao!H{&&ChKGLJn}73O_!5pyQ68*N8;y5>!K^{
zeds24DZ>VlK6?v6(G#3wf11#B?3+0r?SW3ckqVnMSRalU8Nj0_76mwC(g8C<5~_s{
zS*f#UW}`L|_Xzj%c>LJMK!NnS5Oc(Y7E_r)d6WB}P*TvHy1dKi}lo_wybWQ-#%%Pc>Dg?DF
ze~W(!Rg-yR9*Cg(5_T8p?dNuxw#HnO+%(|r-{nhm0K;TB)l88r?zo
z-o_FI81B0~0wPI#byJ@E!kKlVSZ*cgnVuaPvum8gn+Qum;!7M}`Jcu?{^JhwQNFD7
zo81grDXcd`XC@OqNW0a5d&08Ax8XB<`N+k4o>1Xd?)7Nni8ID6x2m~4wz`?D45um<
z2YjybJ{e_e4VZX_E5kIhD({a2tf(cEjFV_r())^v0qH30F{Vcs2!O9I1=FO|!|S39
z^k&i&W;dysAa4-H(iqyr*2%39ISis{+$I-O;BI0xL`l8ei71^TL7LoYUG)z+0YyWb
zNWILP^nH(+bymyZhf^+^{Ty&&m6wjwFL+ePwaihA^_cbQM|Bd*@1+<5J*T34I6pmq<1EI0MyYMWBsNcsM9>vHy7gl9I
z_GltyPu)gjVgv91Gz!Fm0-Hf6ja(eaQu*~ap(8uqK++}>VmWhiXVZgc9|}IivQs+<
zLK|M;U+MJKY^U;7Sbf&mt-fQfk8@M;m!p5j>Rj{bk1NdD*M8*H#x~E4gSVMx`{mt6jsz$+pJNL;hi)0rcL*XQ~P
zGI{oHz7v?af{1%;-9H@}A#-1?Tao|UAC`TE|KV}Rd*U?7WfV3RVsOx3D>)DBk0q@<
zj7<70>He2@ER#I??}!yn^nsWOnf;}tCIIIL*bN6^m9zO^>DGxLLe6_G-mG0m-JGI*xL3qzZa?>2
zb$2{Q<9Hpz-ng`Ss^cfUD*T?~HqTH)_lDA|khMjmtG19@8Ba*w*3#kcw~oWB0|h+v
zt{)_!3!5~BM995?W+TkkR>w`yJDY*R2c>T&M44F=h+)ap(J@1_a05ahUHeMbO!SN0
z{35K}R16wK=b6+^AD3YWU=H;Ut3bZp^a)VHKc((AbAC&h49hD?@uW$|lkM0-FPI*nna&+axaFvi
z_dNjaqO>rhsYKsfynI0S3*^$Yuz1@}WS^u=pX}-7_~MM^irckC)$uKk+@xU!J+Bb7
z>!u$2vD}FfTHM-i25piJn4rGs>Ys`++I8>SqT%fa7^X-8{&}M_I*V#-5T=2McZ9TQKDOF1
z=oi+lO*1D*^`}5pSyWctMW@e_mGewZh0l6ynf{rQpBdsV5r=-dxzMDVby>x;g7Iy6
zVB1c@jN%z~E-MtBTrt)!3gS|*IM?)%{
zL?_ZlUT1bXWe`L!ER?ygmU!=MN6sKlu3#zA$7JF%nLOe_aYQzHm3$@cq*Aiut2*
z7Ia`vBJ@n&_5EmAtU@)ClMc*wUhx$lcO=a5Bw`la``)ps7ZA!`-1f+&a52reo!oOn
z8A>-y*m4xWIT*R`-QR9#Su>M*V?S{vdkwjS_&tnxbAYN26=AKZyYnP9~OG`NP-ep+GQu%O~WP)
z5rS{Fq-BJi;I?pNNAI?COjmmcY3g*duY`N;997j0vkyorT|!z|Je07~keoe|4`R2A
zGUvOlLFfd~o7C~~QA0J|>*4Xy%~Sg9iTy@6AFs8lY|2D8cX9k?9&zLCG#`FWtC9aR
zAkqT7zBIeRwxuB~cvyi(QAElT_J
zk_?^hzA?;h91yCaCBniIEIny9>Y|r8kt-6*HR}0|5xroR0rc^l(YoKy+;9O6{J5Gz
zdw>q5QH*xbPN8Vbgh|Kl1_s2&U)UtI0@^jMSp%soU!UmSV?TgG!>e&AZ3^})iIq9H
z@oLag%nGl6tavcuCfv!7$aqY=bn&iySwspzi*{-Ed=E!Q4mNOfJZB!!`XD1_vTKmt
z+jTxhUJX&n_M!!(y#bvPwVbd@X_8k*(mrg0lYl&gF7m5^l+jnVl+;2|xi+B>zQ4dQ
zvNy66n<6n&R?4tglg}0d9|#q-h8(1{seaJ$ab`|B5fhd%lhgYm@1|hvs&K?sA1;`D
zo2uj6EtAsVM%JYC1jnIt-wH;})@(F$B)(79(K4>T`rcrkBOAHx5%h$pRFRi|`
z(-3oOKJ`%@%3YhGF(09V5UinP0s}AFU;CIW$04H4yu%BE%5!Rr1dLxAtD`jyZgYPe
zX`F*<4B#zy0_O@O84{gt`kXI4fUO>vp?&59rYDwN$2V4kw4~O2r8kie>oop|Gj4r0&
zz;umF1&E~Ez7ax6r3GUw$D^pw;Zg>kq+tU~aeMxES-+rh(u1NRXA99+^Zi|X9A=dv
zq$dZtk7sYIa2L*XBr?TsNc>k(uN}Jl%Ik?SjkxynPH?Nnc(JW{{!=|Jin)O_9VO
ze-8fbkK*;6G@-7t<~-#n>Dorx&5b+JlTudxgBJ>|)#!5g$LB!Cis6Txnmj@$`Cc@S
z%SKfqf!$@f=vv8ENw+HnuI?5v!WN4LXOL61{6R*n{DZ&Fl>|nfnMZIn_VIau6~Rbs
zu>5Bp?tOZ5b;)Ohk(SiPG
zk$xAC_{4=hMUo?_6G6v7M%Q{C`1BwXGu!neZ8+#R2PKMK6dHGb6@lzoWxy;Ds{j@4}~
z3L`YA#2--~+neP1S6;6J7~JhLY6}6t^HD=i;PODBQxE&O&dQn_ZFwH2-K{Df>h-wC
z7T*i@9)e^jZyH3^M&M=Ssvkb*P?U;BuDE#c7jRFh`_PjZ7>A?^i1~XuEKN<3h7hZyP>+B*_Ju7vt;@tcYi~iM#p_C-u}QG(
zW`P!&UZwF0$V$<5+E#di@>Q23-!%>&3f{2$?5x@NFEFWE?$c`I7uwaL^C~x2wB3lEQ<6&T;vSpr8M9Qd6JBXa%ay}gBTYZ(Rd
zyHU~)RdBS*=o`R?&v0ux&>w`K$idZEx~lsK!L+OnhCayXX(_L`Ejrqse;e_Udk1Kk
zQ>C6NcmNNa;y&y}l4#8HQ5u7yyHs8v7>xLSYY?2ZI%3~6H^?v6@s#JTf)P#U4jvpk
zJjNHQQ>h#MZ2(=vJpu#I_T3U^?!2~taHSRX&43mDHa5i)|h!+ORu#)^s(N|ZlJ6jzr)Z^?x%OQJn8cF&q`;RgPG*$
z;}B`u!tQi4yOB+b^LwS{HYaVjlZM&bp%w}B`
zZy%k^;>IR$)d$I(KOID{2pMu7q2Y*IS+yj5)@rAP@y@0w)kX}YY`*R@eaMio`3;8q
zkK5f7{`Gli%FUuKDGM+PAHJ@$aGEZoi47#-t1ePW
z2)IB@;1`JgyP$LKscEhyaOQ7U){RQZ0l?CuglNDkYR>=FUoUYw#W~FO4{Q=#=lJAU
zy1$ZNR`N$zzN_(S5qQcijKT%tk~$Jg5xQlkxN-RMR>^(zgS}<{>1hOh3Kv&hytm&y
zaAtUhIrACC>uhPPEF4})aOC2
zLaGl1K=1L8JhDvmErLj8wKVsn!lLYbTIO1hNNai5!K!^%KXNygQ~lAG`r~FN(fSf4
zVVSfupy;7~KRqu_6IDK*rJEEOj(-cV7e#-YWS_T(uEb64RQgW`b+b!5_*XDlUOC=S
zTg>51P|6TeUZz;77F!QaE+GKB&KADwqN*q0Y_*|vXP{TWhi^O_qB4{OqLnRGvD=0x
zt*~YNC>zk`6(y^&`T#InMBFlMO*`B6N%bzMX6&eLf>+NF%_GiTi
zLU1V%|7PZk;(6o>D0AP!GYbJ_ZV)=-A&fyQUp9{_R@p`g_SQBhIJ8!ofFJVv=Qg5X
z&f^za3n?Uus_Qgv`A#|HE{k!}k4wu1JtkWki&}%)LR|01U!H~_J^?OC-`L^yVoLO?
zrWHZy3EiTA3_oG>0=Wwj|I(a(j59LWOefA)O8r&~KyDEYWXiOYS=|
z;VWMwo-K&Q%;5gPand%!Anj)mV$ph4^yW7r4iQ*S`rlCn9YT-G1B`fhK2MtEF3PI;
zGveW__As#X7+1DR7ic2|7$l=Z2E*L?-`o+G3k)zdP%{0%vmAuXX*$J7S|Q$#@9J43
zzF>$cA)uPRYWak+M(+fEh*9rx(mt-olw++G`KZwLZROlOuysvm7+Su_q3B<5V>yk<
zqgLsnG}U`zk{bN8C(k!z%YMk!5ub6VCIhe~oygNA8DWo@Ro7!dMX@D1X#ZF;q@j=K
zYwRr?Yw?&w0FN>>DA|r}x5rf;E7L~fy3iZuLzUCMe($dEq_#8JSeKtjl)_;(su-0Ls#7I8vesUByl;wUTZ?lyed&X2lz)qyGxbZ
z?J2aXF3YT@Sn$#w)}9TDZ~4S?jwwS9)EZKR3rQ;VI=O0Ik2~M{(di#!`xP0SY~*ap
zL>#FGV8H610%Apy_xb0pZy^R~FM$&M!&)x-Z0m@h5ny|P&Z#A$ZDITMT&uj2kQTIm
zHSO{^7**p7G>z$JPsx^)Pe|v`n#qS4Q18Mx*lYt(L55PjIgdO8n2ax_TPQuu7iNM>
zuwZTy2a6vXYNe-XN`(Fz>jj)I$R37T!j1K)0UNuAQfyh{1s3`*jk#;@iQ64-N=?un
zX98fioY3-q-MbRF^GRQQ9}JGVXx|IS#eYWzK!N6*K)9&@!NjCx#jj7A3|LDEzwDz`
z?=1g_cws}+V0#%RvD84Oipx8)1q`JMc-(t`j4!7v^$bQ1Q{(uDyuO(9kYz5WN^Aci
zWPNmRk{?m9o4?qg4A4o5KxgTCTyVu;I=P213=L-Bs<5x6y=IHU?r_ei@sG9WxIJ|7#lwt)mq?(MZ-{7dT!nx|F`a5V{lXtf_!RB?Z
z#))^ogTSz_PJzeIg-uXzjk7utW>M?0(SOt4?)CpmMI89_zjGvzZc_ic*x8iU?o{GT
zz)nQHhFv`+036#Kh_+q|IlBSe#i9rR13OpD67`)U2NC7LYE&0K>b7XH+yIs4k-_MR
z_UdO7a#!zN%o=O(9%QrGEtWjBdRPFX>iXC4JMD_s##VdeQHVDUbHuguN0z=y9C>Dx
zKejl9Ys@x`G{3lQ#K%OHs55F=lPYa(ymw@PC|B}PmlOio5KPZlca*P75-_E;s3X*Y
z#B91plfi|>pm`^Y(UE^9DIox&B8|A#Gzu8f@sg8P2;ln$v-}OPzD)iAK}c6PQ&LAC~W2VS^`vmP>zR;|noqdrhQd)Eg-;
zD1^%mdXF&CTE)$OgYx;bPAnuoYs3MWETzpD62=FdfOr2QdtRS=`T-A&NaV$9fedfI)-vc8
zFRI4(qT+dd2$o7d5Yo`nZuWZNZ)!A0FIpBnzwAej9yOtqZc4MyE#>HJt;3fejRJs-PX`jFN0M=H>0m7X7
zV;RHgfFEi5XFv##HE5pJ4H6$+p24RchqjL4EMT=9`=hf0hBopcuI)nB^7kNRYF(eL4
z&r{25TVu9#V4xY@+tuqQKd!A=8j#Cet6SZkdNUDdJG~-~of!_v$5wcPvA3^2HIVArHL+R=Cl=4mNrUdeGjG@<
zS8t=)(9P=4lj0AhJ2p^_eg>-1FB+|b(Fg_LU`Vd8`BJ{ry-JzrHDDKKY2S~W%R{{!
z_js>;uLfmQ&mo%#u50^q84vYWf{|0{|BQF}6`K3L2EPoRj;vW>J%WcDlE+}`M^S>+
zg-yy4-(~p05lo5u;!+eT3-N?5olpqbdi~C-v6{IHr9?NI
z!qJ@N&nPj8T|e;mRG)LUl5S;O
zmo_p~`tZfEOw>fn%Lz}vHHy$qM~8wVFPd+q60y0^}u58FNpA?2wW
z=r88->rQj7q1xdfFj8J_Mn4J^JY)Bg9yecpGhQRFy1MuF0zsXEMB@BT|Eb22
zAh=gafdc*JFA?v3>M$isjc+|E{)gGV(ZP?R39U&=di0_}rXlt{K(K`}AXc3XKD=hk
z`t!yUQ~21=gG&i{$&#C-$00ASKrQAC+y4gGs{c;_TO@wouZ5j=D0#<1y18b>Du?(G
zaP7v+IW#1PagLDB6E)7H7&PMt@%W3Gx={r3^u8>P9*Y7D0~Q3zk~sP5fp5rS+iHO&
zi*|MXO7cd$Eo&)@*O5BKlz+?nq`C2AB})v^|Fn$S;sq8XH~zex$Caq@@S9e(L#8WAQXqP*u9u-4^=_sU9nO`6jAaDRieMw7ZUQq@NNQ4!K_#MSGpg@*b{>FPBMbR|f$
z>4WuIe@g(Mu!VOA0)6;
z?ljmp#@nP_o@1X|-VF50C}?^`iNIT6^EY(W`w|nwLwHx_60)9@-l@cfWRMp$e9gpQ
z`j3{+&%(b8Qa}oqBtg$r?!IxszDKbo=)r2ymA&2g^*kxDYfaAN;DDM#Q6Y+PdNoBQ{zt|CioZQjG?+B{44k
zVagqv60jnw^Mt4j(pd}FsE)8`&%p6e)(hQteWsB)VkgJqrdAa
zqp~Cg%UXqoE=A#D$s-NabWj&eu8~c`8r1anNds|9YRK@HXE#cY|Htl04Kb(T9d53r
z#GtQ`Crue;-EH&B`;$QUNOxzX7_)tp_0MIbm6#IlpH*DfAoMJMbkQ%#W*3!L6
z;Gs(bJcCx{?^vcu^U8A?#l$(_HPfClAVa=VQ|CXYiJQxRjd
zx4^z-H+(6F8Tb()=EGo5&pa4}q)-u59k?LR^*#uRR#l+*dwW)88hBO;WAZLFx)V9O
zN-!tw0wIylPGkn59{G;qojY0W#N0IxV{dM=FwGh-GKxX?VEW}gAx-Ok6sPH&TwV*e
zjqX$>P(zvnQe{tEVJid+52tU~f5keB>UMCV84BMhQC5V^Zj_PWAS?vy!^lS(eDU
z$gO|Z_=3fPwYDa#RO&q4|m&oNpA;hLyjOyX^R`zne{O2?)rvzRIG6axR(YiKc^4*e^2%iUr6_8L!1+
zDrH+kNfWe&hSH{BmcZ#r@EcL?^xEx%s=E9`$pIL07We>Q;e!VEwt*lDL~)VbCiNO|
z3#ak8kfoZRc}{@+LFX~jnJzWF6>E~%t~gunW}@h?^6RHvFTHbH@7c17kp1;o3)icJ
zrwx?|j$soOYQiX2+%>SaLZQ%>_gBA)8!xu}SH}=vIDW8IDYYs%KKN`p=o$ryem^q4
zRCz;r@3aiel1j03F%mDxCwpEp;4*zwaqBBd!khBD`eHta^T>6um@r^`48)>&G+gM_
zY#zNjH=WbTL^YCa+5*<&+Ttw;rdU_1`2k*({e0w2DpD=j&O2w<0C?23R09DLlF7FQ
zYkxiJ-d+hraK__ktpJ)_lY^o=){riGVY>?f>r92qI%y*c!ps&V%WJ^}HzwH%tnRx|2`6kpElaWRBH$W4;D+_CqqxaQrKUwKUGC-x60
zLJa2w==~++CizK6RpzMN=Ykf0Y8NDJc&}?EuTOKVl!>~@$RkZ9loQM?>0}aN1J!=Z
zsMU`Jb(ZZq?{rV_2j$NH)Obn_CfPzWXdwr`cHruML6#Q?rb@eNoqxaHh6?>UG@
z?~jZc#kwFl09%r({7x7=Q`}{Sp%?It)S`0!ofG;q5sacpluK4qxujgt0)W_$PlRdJ
zHUuZ`8$mdhERUZ7%j2WcRc=*W_<)14$-r?iD)#?|gJJQx6&7pY7eFfw_+?T--!0F>
zwxl;F+LK`ge-QVkdGR-SZwAdTKg_jZ47S$4mdpe_f85LcaW4iR6?4?AUt=!_;Lu5C
z3RPQBi7h1_lqS7xW}6pE#eB9Li<{qh#0F?^F_d6v0vvKi32Ovz&HCtk5^y2x+ar6M
z2!#6J5~I3Q(jl+kn;rW8Sqr8sUFUU^b{|B0+!Agl*~${!S0b$TTV(CStsqS$%D}Tr
zsoQmJGtqle<;H(}_m58Fb2@8#_G)A&LC6VX$Zq3^XX%x@XWMIditKQiayBhyQ;+4h
z$G{CHc4Yr!A4UKWoPNKotNpj11|*K5ymxef;}_*w3TWU%1)8o9<=bwkypcumQH)~F
ze}lpBBt(hl{-R*qiU0~$lzL%_ZEN*$M)s{9kFMfUE5!Y}y<
zKOmuZ@Cz-bSxbHzcmS@?l|Rk+g!3}stu(-YYL7AA7x?Vw5-wy?{TS|IhcM$DbuE$f
zwWv&Cqj%;Xl3W-zi=s|>%{fBh^%#O|kyK7PIj70(g>_7gih`YeuGp|=10U|0`!Z%6
zF4eIwIMeq9!}>^(j*sBG-a>TBlY!Zl1Fq_6$NZLD|sz@`{#gDM32@?X?F?rL#D(>gxF|=Mc~n^Y3%o0b*{UJ0udC7
zN)_0Q{{wM)^#5Oh?*I5656Ib0;TAf;k?YGkp4WYvPLDiZ7Tvfhv;oZ6|0TFF{b(=0
z)iH#FTsn^LmJWm<@3Y%|2@X8Rvjbt*FypHV;egs2>gOg^|Cg%U+S=p4P#3`C9RAwk
ztV`u_cEeksy*6bsf4=bj3Ug1TnWS&;@7g*ptUJhiLX5+?Y*t$lT4X*uxZ)N1&Pt|H
zLM*3n078J$+F4FrX3e$O0$BS=q{Jy-!=#O`
zBAda2Ku|1SHrl7Z84cS3RMncANK=IlUw7+j9(h7v3%4#EW*OgRf!8OjZgqJ|WxTj&
zy=$9X>Go7x?>9$MuWwOn02mB$+o!bUnV)<9!ut7e8O@}`zn2&7GEIOWpAA=4?o~4(
zkeM3w+O_Q=GZipI^+7_Wk7i!t7H#`QkAqtludv157|>0or!GXPM_Xg}>@45j^j?;6
zQ=eZ<;5v70x*7<4Wn-4qcL5&Lk^?c=tjwa(-(Ve01a3y>rpFVGN*7E28pBlxKN!PN)J=uIlJjZ8Ahj&q3a1%ot?QscC++u%
znwg@MCm7!k{Zw$$RYI7o;@2^E{t@;*i_)v~ocf1j!V$tQdugi{`tBb4KcLtXm3zA|
z{LC?ui5cEYNZ|ve%~$rnC~fHb-npgiJjXYZJzbB=P#W&GVkr(q-yIX
zKpgq9YQB)*``kWfZy5I?Zucb{Nd7AouMcYCVJ*nN{WD(aHlb)>K>UTL=fnGBt^D`uC3b
z?yW)_`p@nPL>u=>>PNy^@kLH;BafU>;yZvu+5@DEmrD1HrQCzw0^or_^hl2cCCGp`uFEu~6k^>75zM~j~bM&f|_ayc!V
zO7~OmG7m&mrsk4ck3jpJ$U)i4?sUd_KlKi*jzK)gYZisi7jD^vI>`P=4PzT&rRa`&
z_LtafS*#iJBIw_~D9*1HScpgPk%T$9>&ucE(7m|R->?jR$B(zUG@E@i*zvOLB3;sI
zPU$qBCCL^-e8PnVOKS4=n>q=mKVT{=nyD}1u#%B@)5)PgCVkMD-l
zpb3|=eUJC(xmsoxHiahWrG#rX4_}$%&x=x`zcKA7b57y7}|2TE8Ub
z`jJJB9sQ?Er}_D3$>33G#-M(7y)D#62qMiG9f)B!&96w8>it)3Tt|lF{JmfU%G1alYAv&WS3sIV8Hkkq$8OGpubPUs?b8I9hyu~@-4_-uxo;*)
z3jaB&J|Y-7cAI1nw8oqYNI?dfGTmepJO|g?Z^PV3C}z
zyM)X_;P99SbxMbg)?2Je^)7wav{_+Z`#|(OAkhtAoNXhrOT0*{{0AVL9Emdz2cl7T
zDri%uzD!HGlm0%WA|^=iB@ca<{$A^vfnc^o+Is>UR|L?xhCVBwHMenM&NNPIz~#YX
z0V}fIC3}I9DO_Pu!_QOSlmZgU*HigUlbOPQIDTDQFX~QozKLo-c}DD+5J5E*yK-fi
zbD#63^Eg++GT^7YqRY>5NRrBBx>)F#kI9*;L}x&tE2a}xKoho03*(y>6Tewx-xqTJ
zGU0_=Z_h7NGl90a0toNybBr$jkE^&UxjKgL>Q`j?3$}7AT<|Q^exVjP#Dai6ut7l^
zv+lt_gi*U>Iw6CANQvJWnVC8OAl7OUvxrVzCEFmy#FHB-QR^!j2=S3?oP2MkJmH!+
zwnxOd1mkL5YN)?J+y0@$VXwAlHXCc7uS%PP_z^P4jY=OsMfKw`^5>!RDa7e=%+O00;B0SD%(RtpVAu!`z<_
zvnxO02+uPk{9Wi8s5f}!nB(5P8-$u^thccuKtC~2xea)0ck4Tm)u7b@`qnW;3S(Iz
zzB#lN7wE?SJb-;!?nrb^TC4QU6MOaa&pmLJUfA#5kVW1^7GQme!pDdUO06tii1
zt&d?VYQ-;isqLQRVw|ki)AvH%nwV~t_T)1AQm->m=r4GeOf9M6pGFjsMZ9fUPrzR_
z+(5by!!LJE95?8xm&Ln`jj?}J_{zLN4mud9oL3U-pZe$3i=>Qmp%PDKU$MIfSZp?o
zfb5A@Q42gDRjSsJ{pgTH+$_G&gTigxt8Y(?FbL_S;4EXavQA`uA=xLk)}&*ud@Z8=
zgR5irit^A