详细
+- 2023/04/01
+1. 增加新模型
+ - 关键信息抽取[LayoutLMv3](configs/kie/layoutlmv3/)
+
- 2024/03/20
1. 增加新模型
- OCR大模型[Vary-toy](configs/llm/vary/vary_toy.yaml)，支持基于通义千问1.8B LLM的检测和OCR功能
@@ -294,8 +299,8 @@ MindOCR提供了[数据格式转换工具](tools/dataset_converters) ，以支
- 2023/12/14
1. 增加新模型
- - 关键信息抽取[LayoutXLM SER](configs/kie/vi_layoutxlm)
- - 关键信息抽取[VI-LayoutXLM SER](configs/kie/layoutlm_series)
+ - 关键信息抽取[LayoutXLM](configs/kie/layoutxlm)
+ - 关键信息抽取[VI-LayoutXLM](configs/kie/vi_layoutxlm)
- 文本检测[PP-OCRv3 DBNet](configs/det/dbnet/db_mobilenetv3_ppocrv3.yaml)和文本识别[PP-OCRv3 SVTR](configs/rec/svtr/svtr_ppocrv3_ch.yaml)，支持在线推理和微调训练
2. 添加更多基准数据集及其结果
- [XFUND](configs/kie/vi_layoutxlm/README_CN.md)
diff --git a/configs/kie/layoutlmv3/README.md b/configs/kie/layoutlmv3/README.md
new file mode 100644
index 000000000..cc18fd725
--- /dev/null
+++ b/configs/kie/layoutlmv3/README.md
@@ -0,0 +1,250 @@
+English | [中文](README_CN.md)
+
+# LayoutLMv3
+
+
+> [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://arxiv.org/abs/2204.08387)
+
+
+## 1. Introduction
+Unlike previous LayoutLM series models, LayoutLMv3 does not rely on complex CNN or Faster R-CNN networks to represent images in its model architecture. Instead, it directly utilizes image blocks of document images, thereby greatly reducing parameters and avoiding complex document preprocessing such as manual annotation of target region boxes and document object detection. Its simple unified architecture and training objectives make LayoutLMv3 a versatile pretraining model suitable for both text-centric and image-centric document AI tasks.
+
+The experimental results demonstrate that LayoutLMv3 achieves better performance with fewer parameters on the following datasets:
+
+- Text-centric datasets: Form Understanding FUNSD dataset, Receipt Understanding CORD dataset, and Document Visual Question Answering DocVQA dataset.
+- Image-centric datasets: Document Image Classification RVL-CDIP dataset and Document Layout Analysis PubLayNet dataset.
+
+LayoutLMv3 also employs a text-image multimodal Transformer architecture to learn cross-modal representations. Text vectors are obtained by adding word vectors, one-dimensional positional vectors, and two-dimensional positional vectors of words. Text from document images and their corresponding two-dimensional positional information (layout information) are extracted using optical character recognition (OCR) tools. As adjacent words in text often convey similar semantics, LayoutLMv3 shares the two-dimensional positional vectors of adjacent words, while each word in LayoutLM and LayoutLMv2 has different two-dimensional positional vectors.
+
+The representation of image vectors typically relies on CNN-extracted feature grid features or Faster R-CNN-extracted region features, which increase computational costs or depend on region annotations. Therefore, the authors obtain image features by linearly mapping image blocks, a representation method initially proposed in ViT, which incurs minimal computational cost and does not rely on region annotations, effectively addressing the aforementioned issues. Specifically, the image is first resized to a uniform size (e.g., 224x224), then divided into fixed-size blocks (e.g., 16x16), and image features are obtained through linear mapping to form an image feature sequence, followed by addition of a learnable one-dimensional positional vector to obtain the image vector.[[1](#references)]
+
+
+ 
+
+
+ Figure 1. LayoutLMv3 architecture [1]
+
+
+## 2. Results
+
+
+### Accuracy
+
+
+According to our experiments, the performance and accuracy evaluation（[Model Evaluation](#33-Model-Evaluation)） results of training ([Model Training](#32-Model-Training)) on the XFUND Chinese dataset are as follows:
+
+
+
+| **Model** | **Task** | **Context** | **Dateset** | **Model Params** | **Batch size** | **Graph train 1P (s/epoch)** | **Graph train 1P (ms/step)** | **Graph train 1P (FPS)** | **hmean** | **Config** | **Download** |
+| :----------: | :------: | :-------------: | :--------: | :--------: | :----------: | :--------------------------: | :--------------------------: | :----------------------: | :-------: | :----------------------------------------------------: | :------------------------------------------------------------------------------------------------: |
+| LayoutLMv3 | SER | D910x1-MS2.1-G | XFUND_zh | 265.8 M | 8 | 19.53 | 1094.86 | 7.37 | 91.88% | [yaml](../layoutlmv3/ser_layoutlmv3_xfund_zh.yaml) | ckpt(TODO) |
+
+
+
+
+
+## 3. Quick Start
+### 3.1 Preparation
+
+#### 3.1.1 Installation
+Please refer to the [installation instruction](https://github.com/mindspore-lab/mindocr#installation) in MindOCR.
+
+#### 3.1.2 Dataset Download
+
+[The XFUND dataset](https://github.com/doc-analysis/XFUND) is used as the experimental dataset. The XFUND dataset is a multilingual dataset proposed by Microsoft for the Knowledge-Intensive Extraction (KIE) task. It consists of seven datasets, each containing 149 training samples and 50 validation samples.
+
+Respectively: ZH (Chinese), JA (Japanese), ES (Spanish), FR (French), IT (Italian), DE (German), PT (Portuguese)
+
+a preprocessed [Chinese dataset](https://download.mindspore.cn/toolkits/mindocr/vi-layoutxlm/XFUND.tar) that can be directly used is provided for everyone to download.
+
+```bash
+mkdir train_data
+cd train_data
+wget https://download.mindspore.cn/toolkits/mindocr/vi-layoutxlm/XFUND.tar && tar -xf XFUND.tar
+cd ..
+```
+
+#### 3.1.3 Dataset Usage
+
+After decompression, the data folder structure is as follows:
+
+```bash
+ └─ zh_train/ Training set
+ ├── image/ Folder for storing images
+ ├── train.json Annotation information
+ └─ zh_val/ Validation set
+ ├── image/ Folder for storing images
+ ├── val.json Annotation information
+
+```
+
+The annotation format of this dataset is:
+
+```bash
+{
+ "height": 3508, # Image height
+ "width": 2480, # Image width
+ "ocr_info": [
+ {
+ "text": "邮政地址:", # Single text content
+ "label": "question", # Category of the text
+ "bbox": [261, 802, 483, 859], # Single text box
+ "id": 54, # Text index
+ "linking": [[54, 60]], # Relationships between the current text and other texts [question, answer]
+ "words": []
+ },
+ {
+ "text": "湖南省怀化市市辖区",
+ "label": "answer",
+ "bbox": [487, 810, 862, 859],
+ "id": 60,
+ "linking": [[54, 60]],
+ "words": []
+ }
+ ]
+}
+```
+
+**The data configuration for model training.**
+
+If you want to reproduce the training of the model, it is recommended to modify the dataset-related fields in the configuration YAML file as follows:
+
+```yaml
+...
+train:
+ ...
+ dataset:
+ type: KieDataset
+ dataset_root: path/to/dataset/ # Root directory of the training dataset
+ data_dir: XFUND/zh_train/image/ # Directory of the training dataset, concatenated with `dataset_root` to form the complete directory of the training dataset
+ label_file: XFUND/zh_train/train.json # Path to the label file of the training dataset, concatenated with `dataset_root` to form the complete path of the label file of the training dataset
+...
+eval:
+ dataset:
+ type: KieDataset
+ dataset_root: path/to/dataset/ # Root directory of the validation dataset
+ data_dir: XFUND/zh_val/image/ # Directory of the validation dataset, concatenated with `dataset_root` to form the complete directory of the validation dataset
+ label_file: XFUND/zh_val/val.json # Path to the label file of the validation dataset, concatenated with `dataset_root` to form the complete path of the label file of the validation dataset
+ ...
+
+```
+
+#### 3.1.4 Check YAML Config Files
+Apart from the dataset setting, please also check the following important args: `system.distribute`, `system.val_while_train`, `common.batch_size`, `train.ckpt_save_dir`, `train.dataset.dataset_path`, `eval.ckpt_load_path`, `eval.dataset.dataset_path`, `eval.loader.batch_size`. Explanations of these important args:
+
+```yaml
+system:
+ mode:
+ distribute: False # `True` for distributed training, `False` for standalone training
+ amp_level: 'O0'
+ seed: 42
+ val_while_train: True # Validate while training
+ drop_overflow_update: False
+model:
+ type: kie
+ transform: null
+ backbone:
+ name: layoutlmv3
+ head:
+ name: TokenClassificationHead
+ num_classes: 7
+ use_visual_backbone: True
+ use_float16: True
+ pretrained:
+...
+train:
+ ckpt_save_dir: './tmp_kie_ser' # The training result (including checkpoints, per-epoch performance and curves) saving directory
+ dataset_sink_mode: False
+ dataset:
+ type: KieDataset
+ dataset_root: path/to/dataset/ # Path of training dataset
+ data_dir: XFUND/zh_train/image/ # Path of training dataset data dir
+ label_file: XFUND/zh_train/train.json # Path of training dataset label file
+...
+eval:
+ ckpt_load_path: './tmp_kie_ser/best.ckpt' # checkpoint file path
+ dataset_sink_mode: False
+ dataset:
+ type: KieDataset
+ dataset_root: path/to/dataset/ # Path of evaluation dataset
+ data_dir: XFUND/zh_val/image/ # Path of evaluation dataset data dir
+ label_file: XFUND/zh_val/val.json # Path of evaluation dataset label file
+...
+ ...
+...
+```
+
+**Notes:**
+- As the global batch size (batch_size x num_devices) is important for reproducing the result, please adjust `batch_size` accordingly to keep the global batch size unchanged for a different number of GPUs/NPUs, or adjust the learning rate linearly to a new global batch size.
+
+
+### 3.2 Model Training
+
+* Distributed Training
+
+It is easy to reproduce the reported results with the pre-defined training recipe. For distributed training on multiple Ascend 910 devices, please modify the configuration parameter `distribute` as True and run:
+
+```shell
+# distributed training on multiple GPU/Ascend devices
+mpirun --allow-run-as-root -n 8 python tools/train.py --config configs/kie/layoutlmv3/ser_layoutlmv3_xfund_zh.yaml
+```
+
+
+* Standalone Training
+
+If you want to train or finetune the model on a smaller dataset without distributed training, please modify the configuration parameter`distribute` as False and run:
+
+```shell
+# standalone training on a CPU/GPU/Ascend device
+python tools/train.py --config configs/kie/layoutlmv3/ser_layoutlmv3_xfund_zh.yaml
+```
+
+The training result (including checkpoints, per-epoch performance and curves) will be saved in the directory parsed by the arg `ckpt_save_dir`. The default directory is `./tmp_kie_ser`.
+
+### 3.3 Model Evaluation
+
+To evaluate the accuracy of the trained model, you can use `eval.py`. Please set the checkpoint path to the arg `ckpt_load_path` in the `eval` section of yaml config file, set `distribute` to be False, and then run:
+
+```
+python tools/eval.py --config configs/kie/layoutlmv3/ser_layoutlmv3_xfund_zh.yaml
+```
+
+### 3.4 Model Inference
+
+To perform inference using a pre-trained model, you can utilize `tools/infer/text/predict_ser.py` for inference and visualize the results.
+
+```
+python tools/infer/text/predict_ser.py --rec_algorithm CRNN_CH --image_dir {dir of images or path of image}
+```
+
+As an example of entity recognition in Chinese forms, use the script to recognize entities in the form of `configs/kie/vi_layoutxlm/example.jpg`. The results will be stored in the `./inference_results` folder by default, and you can also customize the result storage path through the `--draw_img_save_dir` command-line parameter.
+
+
+ 
+
+
+ example.jpg
+
+Recognition results are as shown in the image, and the image is saved as`inference_results/example_ser.jpg`：
+
+
+ 
+
+
+ example_ser.jpg
+
+
+
+
+## References
+
+
+[1] Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei. LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking. arXiv preprint arXiv:2204.08387, 2022.
diff --git a/configs/kie/layoutlmv3/README_CN.md b/configs/kie/layoutlmv3/README_CN.md
new file mode 100644
index 000000000..1f2628ad9
--- /dev/null
+++ b/configs/kie/layoutlmv3/README_CN.md
@@ -0,0 +1,246 @@
+[English](README.md) | 中文
+
+# LayoutLMv3
+
+
+> [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://arxiv.org/abs/2204.08387)
+
+## 1. 模型描述
+
+
+不同于以往的LayoutLM系列模型，在模型架构设计上，LayoutLMv3 不依赖复杂的 CNN 或 Faster R-CNN 网络来表征图像，而是直接利用文档图像的图像块，从而大大节省了参数并避免了复杂的文档预处理（如人工标注目标区域框和文档目标检测）。简单的统一架构和训练目标使 LayoutLMv3 成为通用的预训练模型，可适用于以文本为中心和以图像为中心的文档 AI 任务。
+
+实验结果表明，LayoutLMv3在以下数据集以更少的参数量达到了更优的性能：
+- 以文本为中心的数据集：表单理解FUNSD数据集、票据理解CORD数据集以及文档视觉问答DocVQA数据集。
+- 以图像为中心的数据集：文档图像分类RVL-CDIP数据集以及文档布局分析PubLayNet数据集。
+
+LayoutLMv3 还应用了文本——图像多模态 Transformer 架构来学习跨模态表征。文本向量由词向量、词的一维位置向量和二维位置向量相加得到。文档图像的文本和其相应的二维位置信息（布局信息）则利用光学字符识别（OCR）工具抽取。因为文本的邻接词通常表达了相似的语义，LayoutLMv3 共享了邻接词的二维位置向量，而 LayoutLM 和 LayoutLMv2 的每个词则用了不同的二维位置向量。
+
+图像向量的表示通常依赖于 CNN 抽取特征图网格特征或 Faster R-CNN 提取区域特征，这些方式增加了计算开销或依赖于区域标注。因此，作者将图像块经过线性映射获得图像特征，这种图像表示方式最早在 ViT 中被提出，计算开销极小且不依赖于区域标注，有效解决了以上问题。具体来说，首先将图像缩放为统一的大小（例如224x224），然后将图像切分成固定大小的块（例如16x16），并通过线性映射获得图像特征序列，再加上可学习的一维位置向量后得到图像向量。[1]
+
+
+
+
+
+
+
+ 图1. LayoutLMv3架构图 [1]
+
+
+## 2. 评估结果
+
+
+### 训练端
+
+根据我们的实验，在XFUND中文数据集上训练（[模型训练](#32-模型训练)）性能和精度评估（[模型评估](#33-模型评估)）结果如下：
+
+
+
+| **模型** | **任务** | **环境配置** | **训练集** | **参数量** | **单卡批量** | **图模式单卡训练 (s/epoch)** | **图模式单卡训练 (ms/step)** | **图模式单卡训练 (FPS)** | **hmean** | **配置文件** | **模型权重下载** |
+| :----------: | :------: | :-------------: | :--------: | :--------: | :----------: | :--------------------------: | :--------------------------: | :----------------------: | :-------: | :----------------------------------------------------: | :------------------------------------------------------------------------------------------------: |
+| LayoutLMv3 | SER | D910x1-MS2.1-G | XFUND_zh | 265.8 M | 8 | 19.53 | 1094.86 | 7.37 | 91.88% | [yaml](../layoutlmv3/ser_layoutlmv3_xfund_zh.yaml) | ckpt(TODO) |
+
+
+
+
+## 3. 快速开始
+### 3.1 环境及数据准备
+
+#### 3.1.1 安装
+环境安装教程请参考MindOCR的 [installation instruction](https://github.com/mindspore-lab/mindocr#installation).
+
+#### 3.1.2 数据集下载
+这里使用[XFUND数据集](https://github.com/doc-analysis/XFUND)做为实验数据集。 XFUN数据集是微软提出的一个用于KIE任务的多语言数据集，共包含七个数据集，每个数据集包含149张训练集和50张验证集
+
+分别为：ZH(中文)、JA(日语)、ES(西班牙)、FR(法语)、IT(意大利)、DE(德语)、PT(葡萄牙)
+
+这里提供了已经过预处理，可以直接用于训练的[中文数据集](https://download.mindspore.cn/toolkits/mindocr/vi-layoutxlm/XFUND.tar)下载。
+
+```bash
+mkdir train_data
+cd train_data
+wget https://download.mindspore.cn/toolkits/mindocr/vi-layoutxlm/XFUND.tar && tar -xf XFUND.tar
+cd ..
+```
+
+#### 3.1.3 数据集使用
+
+解压文件后，数据文件夹结构如下：
+
+```bash
+ └─ zh_train/ 训练集
+ ├── image/ 图片存放文件夹
+ ├── train.json 标注信息
+ └─ zh_val/ 验证集
+ ├── image/ 图片存放文件夹
+ ├── val.json 标注信息
+
+```
+
+该数据集的标注格式为
+
+```bash
+{
+ "height": 3508, # 图像高度
+ "width": 2480, # 图像宽度
+ "ocr_info": [
+ {
+ "text": "邮政地址:", # 单个文本内容
+ "label": "question", # 文本所属类别
+ "bbox": [261, 802, 483, 859], # 单个文本框
+ "id": 54, # 文本索引
+ "linking": [[54, 60]], # 当前文本和其他文本的关系 [question, answer]
+ "words": []
+ },
+ {
+ "text": "湖南省怀化市市辖区",
+ "label": "answer",
+ "bbox": [487, 810, 862, 859],
+ "id": 60,
+ "linking": [[54, 60]],
+ "words": []
+ }
+ ]
+}
+```
+
+**模型训练的数据配置**
+
+如欲重现模型的训练，建议修改配置yaml的数据集相关字段如下：
+
+```yaml
+...
+train:
+ ...
+ dataset:
+ type: KieDataset
+ dataset_root: path/to/dataset/ # 训练数据集根目录
+ data_dir: XFUND/zh_train/image/ # 训练数据集目录，将与`dataset_root`拼接形成完整训练数据集目录
+ label_file: XFUND/zh_train/train.json # 训练数据集的标签文件路径，将与`dataset_root`拼接形成完整的训练数据的标签文件路径。
+...
+eval:
+ dataset:
+ type: KieDataset
+ dataset_root: path/to/dataset/ # 验证数据集根目录
+ data_dir: XFUND/zh_val/image/ # 验证数据集目录，将与`dataset_root`拼接形成完整验证数据集目录
+ label_file: XFUND/zh_val/val.json # 验证数据集的标签文件路径，将与`dataset_root`拼接形成完整的验证或评估数据的标签文件路径。
+ ...
+```
+
+#### 3.1.4 检查配置文件
+除了数据集的设置，请同时重点关注以下配置项：`system.distribute`, `system.val_while_train`, `train.loader.batch_size`, `train.ckpt_save_dir`, `train.dataset.dataset_root`, `train.dataset.data_dir`, `train.dataset.label_file`,
+`eval.ckpt_load_path`, `eval.dataset.dataset_root`, `eval.dataset.data_dir`, `eval.dataset.label_file`, `eval.loader.batch_size`。说明如下：
+
+```yaml
+system:
+ mode:
+ distribute: False # 分布式训练为True，单卡训练为False
+ amp_level: 'O0'
+ seed: 42
+ val_while_train: True # 边训练边验证
+ drop_overflow_update: False
+model:
+ type: kie
+ transform: null
+ backbone:
+ name: layoutlmv3
+ pretrained: False
+ checkpoints: path/to/layoutlmv3.ckpt # 导入ckpt位置
+ num_classes: &num_classes 7
+ mode: vi
+...
+train:
+ ckpt_save_dir: './tmp_kie_ser' # 训练结果（包括checkpoint、每个epoch的性能和曲线图）保存目录
+ dataset_sink_mode: False
+ dataset:
+ type: KieDataset
+ dataset_root: path/to/dataset/ # 训练数据集根目录
+ data_dir: XFUND/zh_train/image/ # 训练数据集目录，将与`dataset_root`拼接形成完整训练数据集目录
+ label_file: XFUND/zh_train/train.json # 训练数据集的标签文件路径，将与`dataset_root`拼接形成完整的训练数据的标签文件路径。
+...
+eval:
+ ckpt_load_path: './tmp_kie_ser/best.ckpt' # checkpoint文件路径
+ dataset_sink_mode: False
+ dataset:
+ type: KieDataset
+ dataset_root: path/to/dataset/ # 验证数据集根目录
+ data_dir: XFUND/zh_val/image/ # 验证数据集目录，将与`dataset_root`拼接形成完整验证数据集目录
+ label_file: XFUND/zh_val/val.json # 验证数据集的标签文件路径，将与`dataset_root`拼接形成完整的验证或评估数据的标签文件路径。
+ ...
+...
+```
+
+**注意:**
+- 由于全局批大小 （batch_size x num_devices） 是对结果复现很重要，因此当GPU/NPU卡数发生变化时，调整`batch_size`以保持全局批大小不变，或根据新的全局批大小线性调整学习率。
+
+
+### 3.2 模型训练
+
+* 多卡数据并行训练
+
+使用预定义的训练配置可以轻松重现报告的结果。对于在多个昇腾910设备上的分布式训练，请将配置参数`distribute`修改为True，并运行：
+
+```shell
+# 在多个 GPU/Ascend 设备上进行分布式训练
+mpirun --allow-run-as-root -n 8 python tools/train.py --config configs/kie/layoutlmv3/ser_layoutlmv3_xfund_zh.yaml
+```
+
+
+* 单卡训练
+
+如果要在没有分布式训练的情况下在较小的数据集上训练或微调模型，请将配置参数`distribute`修改为False 并运行：
+
+```shell
+# CPU/GPU/Ascend 设备上的单卡训练
+python tools/train.py --config configs/kie/layoutlmv3/ser_layoutlmv3_xfund_zh.yaml
+```
+
+训练结果（包括checkpoint、每个epoch的性能和曲线图）将被保存在yaml配置文件的`ckpt_save_dir`参数配置的目录下，默认为`./tmp_kie_ser`。
+
+### 3.3 模型评估
+
+若要评估已训练模型的准确性，可以使用`eval.py`。请在yaml配置文件的`eval`部分将参数`ckpt_load_path`设置为模型checkpoint的文件路径，然后运行：
+
+```
+python tools/eval.py --config configs/kie/layoutlmv3/ser_layoutlmv3_xfund_zh.yaml
+```
+
+### 3.4 模型推理
+
+若要使用已训练的模型进行推理，可使用`tools/infer/text/predict_ser.py`进行推理并将结果进行可视化展示。
+
+```
+python tools/infer/text/predict_ser.py --rec_algorithm CRNN_CH --image_dir {dir of images or path of image}
+```
+
+以中文表单的实体识别为例，使用脚本识别`configs/kie/vi_layoutxlm/example.jpg`表单中的实体，结果将默认存放在`./inference_results`文件夹内，也可以通过`--draw_img_save_dir`命令行参数自定义结果存储路径。
+
+
+ 
+
+
+ example.jpg
+
+识别结果如图，图片保存为`inference_results/example_ser.jpg`：
+
+
+ 
+
+
+ example_ser.jpg
+
+
+
+
+## 参考文献
+
+
+[1] Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei. LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking. arXiv preprint arXiv:2204.08387, 2022.
diff --git a/configs/kie/layoutlmv3/layoutlmv3_arch.jpg b/configs/kie/layoutlmv3/layoutlmv3_arch.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..9f23f3fc9d985e640aa7fe1c71386ad092f6e7e2
GIT binary patch
literal 380147
zcmeFZ2UJwemMDChCg&W528n`zl5>MdQV~T-g0$o$K{A5fhzdvs1q2ijR8RyY3nER<
zk~5M8mDmje4RrXA-ub-qzL|CJeRJo{x@-MYT77m`SM5`KSK6r{O_1gR=5zW+`Tzt1
z07>9KfV2Sk>jk?x1AwtHAPE2fH9!vG1t>rWtO963_>=(-`mgC+ec7d<`kfE#>klbCq^*+
zf<=FU%p?49lZik%`S);|C+oN~NYWyMppmXF;*zO}zR?A}UxY9)+xht1Bx42uPcMI8
zQ-d>tmR8pUslI`|VF5E0plavf=c9A!(uJQm|9Jkw|NCQa>{s7`Vdr6)xpQt9{`}aV7YLRzt2y&6@=M+!487(#824eFYx+L*!~y%
ziw*3)S50*RfSes15_&s3Ul#zN7X<4F20FNac<8l3_>`BUn->WCfUuUQr?(>rzXM?j
zunt&8{}aE8{lCNG^6y|fJNw@_?d)8B!+)U#Yze;jtec~&L8D=z995(cmh1kewVvC8U4n8
z+x4o&A9VuV%zv(kU+?wvxbnL{w_SCAm-_~o{_ek%m%*7o>O1@D|EAs34=nvRJg#m=
zzuWq|UN!uq+`;az{de-x`72I8%in!-bULf|M|~$Rv)}ak_@DXxE}xsf*2b?sJKZ$+
zUG5in<#$_ua6S-gz!l~sXyoM}b5c<5l$-(p{9La;>j1zG=dX1L;q3lv**zfu&!TQ*bM-k=?($Dfq%jGYb{X#v;Y&p4)6d1fG{8dNCR?!5}*!f
z1A4$Yzy!DoSOKAo1Ixe~fC2V^BM1aS1z~`&L3kiAh$ut~A_q}{XhHNK7a*n(E65Fq3&aZ&011QK
zgTz8oAeoTokk^nhNG+rZ(ghiWOhOhQ-yz$OLoz5CJsCS0KbaVrESU_pAh|TTI=Laa8My=bEpjAz9C;>rA$c`f-DJo4W6Dm6@f2v5T463(OXsRKqWhxvs9W{(vj#{7k
z8nqX71a%tq8|nt?A?mNxM>NbdqBLqWCNz#TAvB3JFKFs$`e~MFj%ZnE#c4HZuhP2H
z-lcs)TSD7L`22r(=o9D*=o{%L
z>30|y7{nN~8Llz-GbAv)W@urUWxz4AG0HF+F*-7$7_%6w8HX4*nP`|qnRJ+JnQk+s
zF;y`2Gp#ezFpDwkGTSjDnX{N{na7!TS=d-kvzV}Wvc$2xVd-M|#!AI1#;V8a$awN5d>U<7-F?{8GQ^&}TNggvj7J4lISnsi2egS?1ejok}
z{ucgq0ZsvJ0e6AN0u2J+VQerhm>cXd3=LZqi5QE7i4=)UiPDIwh`NfVi*|_ai;0U_h((Fj
zh>VEnX!4S%N_VE^$jDPh#W*$kfPe%8JNdlTDWGkR_f{JazNbi&N95
zSx=uk9dWwu^tPO&oP%7J+=x7byuLh2{=Ga_K}x|%AzNWWkxkK9F-oynk)WikeA{S>TlG)X^3k$X}r)_
z(mbwdr}uAFi=r33F4mX;CPpTYOlB^LU%GYa!)40L
zmo8^rUNJpw8fMyYh3yLB%9|^DSI=BcygF?rVHRN4Y|dnEWB%HF-{P!Aip9L8tmPfc
z9xGleH>>y8sIHk`dv$HsTF*Mwdf7(NCemi?y7=|r>s_|Iwl{4*BA5_%h>9B&H!N6ORH9(oKt-rJmHDH#}>-SiL;FTD=9lgS`iRqZqyNCZ3xSP0Y#%nHN>nFmz_vjlqu_l8J@JPi4ITmSaUQ1VcR
z(B?3qu)ARkcXaMNM?#SH$Yzu<>ON{Y+#tO0F6~{9yPqOXMWjUhxMy*%?mqwh@cT;-
z3?39kGDiAFjz_6S(n2JSo9Bp)XM>
zF*}Je$t!6r8J=92!jf`3W$BUeqw2?kk7FMHNJXS}r75K4q|>Gcq|ZOO@Z^1lNJdfy
zKGQXGH0w-O>C?=ZnSHXOtX<_SY<=3y@9D9>kL{=0~^!=^P+kxV<#qUd`OP-fS~LIN>iuBtx!UGB^}4DLG9L=+
zh3d0Da(ql|pl!I{KtkU}?>720t~a?fEj1&Wr&}yrMp`eo_O+dF>ulF=Z|TtKKzC|(
z)_19O)po0N*Yqg$RQD?OR((?ZRMn@{SKY7N|9(Jqpl(oOuwe*3)HHl{xP8QEq-XTv
z=-`;y*u=Qa`22+9#J9h0Gb$bhZlsfWZm?*gzSA*7`5<^sgF_zvuuU4EeYGv-nT&Uu5o|
zKY%g>Py#-FibcAO0PqgXV&Nce@cC0Dx~dJ3$;A2!EhiT@53i`0xWoxb1w|!g6;(CevwHdlhDPVETs1Sd0L3#0
zM<-_&S2uS*|A4@t;E>z*?mvi(ihdZA^5}7DTKbcW%)I;;FJBcDzJ60)QCU@8^S-vO
zxuvzOy`!_MduVuMbZmTLa%y35>C5uU*KglfvD-U8cK7yi2ZulDf&kFJkoB9gKhebk
z(nUs24kf4hNf(4H2+U9xatfi-l&m_JsqAjC9hbXH&3-2Nd07*Uu)HaT!`^q0mQzGw
zK@|Iwv|p6{V}wQgmni#9*k9?I21RGGUkOS^2Bm;Pp%j!9V4|c3B`-=!YFg@FiT2+T
z{jbFMGco;6BoGM%#6eC@P6hsFrlX-_{=ZJ78BlS*OPT`cp%8E~L0JGeK)~n4NC5wv
z|9{N`C~4q-bRIyrna&XE(C#E423AScii$vy6WLwsexR6k@8~y&u6_9KX>TYk^QHRX
zN}!N#&+?b!5f{1dJ`nLk0)z$d@Fj+-a0Pt{$&IJQd0_5c#}+&cqkC1A{Sphknr=BJ7LskAy=f6b!`C%&_l0la
zd+4pk(hjDH4@A>8>=E%j?hy&NIOY%8MXXOAdE$L=kB%;{!760ibD{zT7)C3NwN^Hl
z*7sSBkJpQ_2@IXr$9AS{nqP^a793yY4}st~Fe7*m#1xDFm=A#-m<#n@*8V~s
zm0j4pw{Rf9)5x**1Gucm{x`G5it>ijkq6!xCwpXwGuR8nY0lHZcvl-_hCEA)(?Y()
zT+!H9zEtjXF0cKe_s3_^w91@cI_}xc=of
z*Da}}<9(mL_9uC_`?R{SLk>c*=l6~X?;05K=9s9a96r1hrd2=;TU&WuO(nqSRdm+$
zBRA$0y#>^5YE2*2VstdD#(TpLbAGYrt`t48dIPrNb11{4OKgH2T8Wwi3i0;+
zE;&viZ?0K?Zd(ydR&EoCS*ztf)yZQ@JsC{{ja`7WowUIg@dC^B_Q>v&$y%9-+No<*;vkZ1<
zQqwDW_;RX^_vObIC&*EflAA(64|D#>5Y0~kCxYBWYo>)Wc;_hU8yjB@j%9(R^rkf)*Rd
zgR_Jz*)x!nW9mF*UCf^XSK(ys;c$fF$535NU%FP!qOtz8(Qt9N+~=9zBpq&Xpgx7U
zsH;P)@S<=vU8*pW5Rv8@^_tr3Dm+#dSN8s=pwPNfq_t{#bJY&#kM~Wd1=Y_
z#5-lRYfv=Cmjg7w4k&ozp5V=TYN@Zhlq99BUcjs;J
ztDVL}gNHvpY_WXrbgUZ|{1{>Vjm6M(Xu2qanNpCon=&|;u{$PqgPG@3VRG2~hi*r&
zsf)@ECGzdAM)N5~LJ4|q$J=+wS!=&I26pDsxaK`SpB3ViGOK!dIf`%K)XCZ}Ui|u@
z&f#LEIYbTO@EjG&`J>#p8LsB_jN7|8DkY+RPEeqTqaeqQ;KRlEhjw>`Uxiu>aA2bF
z>~F1RiLE4nef`;xkx9i8Wfb5WdJ>pQwhA>F;=7f`eh
z1V#_%wS}pYT%lhPlUsW)a7ITaQ^EI4aRxyFD2IifC&e!B-&cn^CNG~^wYNI8?4$bm
zjls*%JEaf9RPPkq27C3GRlXnn@xw^w=2EtD?Olt`dj
z&qlnGC*ydQE>JmN?P@&wpsj_@%iK(U9VBF?ZzK*tK4-3k87!eFaO%2R;@CiR!MEFL
zNemmiBj#g`Pv?1@oG$v@`UC{9n|nr9yl}6(z`K`zhhilofcCR)0DX}N>zSTfM1JAcr51yFfsSM2;~cmn?A~nOS0ZiKkp`mCROi;hbm=ogq5GFt#CQ;*muD1eyS|n#WV3!*bSDjk?P;RvZ})vhi(KljO9;V}|p>;p})0aO-0t
z6ook*8KGLFNkDTMW-McgyHuuF{OCGi$XahtB|Ai9hsCCxJE@%Gj3r0o3i_Islskge
z9|du*D`+;BHyI1N>>evxoL!)hwm1CXhNG^?x!z;1HZmaRZ=8mTAcc@~mtd1cL|O*S
zqrIFeM1`BJpvRTMZc1gvSm{(u|Hu1-h5a9qdBhn6&!c(Ndg)RQ?=>6hD)k=um*4Nh
zp3J(#-y4`*VXZL!p{os3LbkxG`#VU$6bWck9_O-!3|7LSt|A?;-HrTwT;#jhDo;%d
zla*`D7mFFv@$?0;GaTk1c#(}La`b{Ou}6ymQ?-nq+%Z(LgE
zGi1jc{-I$khWqhm
zveLb(n26`^f^NV%V<)24OThuXS6v5P&xYa3b%@_k8}pZThKXmAy$+E9d*gTx<;B4K
z1=wT-k)t#cYP^Vu$&9awTF*@+c(e&Wlu=4pXnreBtV=8g{88^pH*~SXdnJQauxKrL
zjCtxpBDwC$)TGVy=ZpImL4$Q(lgGkKlVWV-)2Moz3svr*b4R}*03!Q$G_-*p2Vc-i
zuk2X(sK)1ZoszspW$u;|{QKM9P1J1Z^<1(TiGM0Hkp}-qH431#3HB%9J@qJOF&%p*
z*xM9E-sYm!ApPZt)?LnsBg@8N-#+MpfL>zIti+3@+|qXRU33}K1{xE!motiXvwr5+
zcsqVSro`LYs*3A&X8YdZiTVvM>;=IiG-adGcTpTZqb)8e*-H}$u$7CIw%9`T4`Em(iS
zw5~^RUx2LAa>Fk?(B!;}thy^#+iUlD7sZ3=8S9UoM|6>?&ng7Xr&KOCb`kDjtZ|?4Qm>R-
z4J&bBhITM-V@V^YGPa;?0e>Uf+uoN#LkFl<$_TGXfR}O5{L0HCeODNDy=q=C+_?Jv
z2)ta;^9*LK!t&atTyM;~Hr9p9PpN2o^CgQ@`=?M9u#Qh$dl@+C+^`#iCq;&?n6nvZ
zJ}wdU@Q4lJJ>jqeTw1I3Ev|Gj6oDpbT_lttm4b^Fy()r^SdgCvONy8JIi@NFj*rM8
zI~F#M`)~%INoQZx*t;~ir)V3n>sYNfU8s0eFKP9{8
zi%o1U7CRejF(R0-bkU7*a+GxjYh)7BHuzQtVD2c6M&G+n*s!Z|tIybSkcM1dQ(336;#1BV0?$^@e
z-fsH0^P0!Wy1X0u@xdbPOty`ZH%(el-+}0p+CFxGne7J?p1ozUG`ybSNHTc5Sxb1V
z`0ef%rzM}KWYgz6uQu*u9@5`ReDc<=d}DyI;H3+)F^3l9^j_$(w
zJHemSItWh}tvIv>_$bK3ooE)&Q3{ZGIL2^NF^91qJH0@k8Mj}@br)~IQT^;#pN(GK
z`S(teYBD6CBL5UXHmTC3xO$BQG-hSZO{sU}aHBj4sFuw1W>@zn0YVG_Knskjl`6UmzQ^==a7d(ge(${9q^LrNFWUjVS!@I`bZNqLB
zdz{se-fkP)eGO^;cC$~&j)nFskWJ|k)C!0Bqd|eINv2BEWd7x7K}^*Kv?)_RxyZ7;
z$WQccCby?-{ZSzKibVZazH~QgL@H&t1_@|^mlrh23ZR-a659y&!y7rwn|S%4#f-4)
z3fo3SGR5Sk*S$i^+s`D|h1|?&AGoNDADu!q$q{Kl%L00R?Rl8vSUo>OPTa1#G@@nv
zc2#{1=P<&usyDg*MVsVZb=l3S1??d*TNuu15j`c}ej6ScAa&kl
zEKI}{Q9(Cm7W!UeLZQn!Valkc!hoZ_{qx05V`NVl3l4oGYm5n>!Lwqu;<>`>nOs@D
zk9&)0%1wzn&-gy_Ogs6&SyqCgZ$&f#JlHgepWqewxY)ffI07ce5bIS2D!k<1j-I0v
z`ZoGr-f-Bhz8^!20!OJZ+ut#chVdODiTLK%D8)7=H
z)Ns{s(W&tj=05%K*+{r1i+xTmO{!KG;1eq*w!vTx?5l9hC3He&OmM=Ox@ggehsyVpfU;>@vKx*?T^rzywX$_EXUW$7(*yumJui4VLzUk>kT!do&w
z6}R=Om#Ng(W(vB9gdM&3+!pzH!?PlOr|mrwz~&%Dq@_CIM)o0T2+3gT;8
zu{YCwy0_is>ZliQ7NiyOWIS2+4X*-__mCYn#4Zv*gH21ZYzQZ0@9D6U?{I7&BbdT?$4}x77UC(}`L(2`c{mMMGH0|i
zroCs9zlhFUq;^&XRgh2&-6CBWIY{ucRWcNVhbTE30od|Ki#tG9=4Ow6XNu7JSKwo
z?%5cun_;%j;ex=!Lk0JB!IRe`;|GE6OeUX6wT)4boL+!mK_Bxcd5y#xbscBvgcnaU
zK|{>^BUq()Gn~c^&hZ0r4ELs`B|S5{Wa?9$J6my0mis%y9X0tRrpIMzK8)I5x!;n2
z$(+4RoIEHL^en*=72zD$0JkuT6*4BX1Q>I6A!>*~*?eQNI?gL*^UlLR*?dZrM&3iAjt@
zR3tW`$v_?(XdS~J8^oXREDWYv)KArHf4)$$!gtnB2z
z-xTOZ*1;R#4_kIS(f;~M5-Rjzh|vdQkYeSLj1d1Yl{IRwfF@1#
z4B`PtJF3aPK@vyVv3DnNL4Gap)8dB7mEkS63ldnAe1
zCS+~}Ooy7Vd(_|U5hK&
zLkFvu*tqsYWK=^b;D~P8b=}$OIif;)+?`hH?GVF@G&|a|uea1;5ShxY>7Aw0HWWAh
zy1^Lo7+!ruc(7i1p;px4mHLWV+TkjPF`;{pG5Z~*?5m&-
zNB85p?iI|%3@^IiSFV{u%_~33NkHC|f!d4VjWbruX4EerkKXg*vk2+P<8CO731SMd
zn9~HCg7#vz?&F<82>H)FOk0F1l7&(bCu#(}(=1=mwmKS!gmTBF0BSH+JUz}EBTuAl
z!I*a}AkzgNy!1FU-uRKm>~*A}409UE^P^2U>$&Up)`ZMKgTW?M3UrdJxUk>Vitls?+YUrqA3
zt(Q>c!H#J`kz3I4_p(}FjFpU{J1H+|#1`(;x%#$1C|ocG82lIWS3|eAh_p+X93m&v
zdVav-v`W|~gGRLn4TUfDZKk@BJ-+#bBvC8tDK%Wd1sr%(FwDFmcI8xjOW)Yz<&VZv
z#f>~$?a$|`pi?1oKC$i^uNo47l+L1~1?YzwgoD{yZGmyFz??;6t+k8qmqhde*v__w
zhEEFY9u$F+*=NH2dY2=Kx#0db&}{!@cr#3|_ys4Yw4Kz&WXtvo6;98D1$)HPQ#%AK
zcXOHE!re>xkVCdCnEj}eaAy$<<2aKj*2oe?WKOxpq@-VS7LbMO}+&GOMr0e9~gfqR+&h5
z8nZ{7YlbtTT;U8W3|E?_UYDf>cl151=OeyQ5%y-!C{)%z$X+P
z5lBYdwH`z=RX^+6?YEcv%Dz@68)6*;7~JP~+K{QpQ5dDk2J}b^H@%=hDcUf$ZoMvh
zT8bB8UEc40JzUhHFCNNFLu*3yG7Oye=?TrRIeC(Z%RrSfvm!+U~zDFe#WgP0d_)0y}Co;XI$R4ho^);9Svgl26XpEzBn5fZ@d)^L0*O!|$REz6ieR6tQth;YS-ak(v>(GxpS
zrELzQ!0f=&z>iPJV_^n^L~&#%ysDrD9fzvid+mlDzKFfvIaImf9xTwjrSn)VGW5EZ
zDsP`g$aEuZTXmWF!-E@4vvoh6J^w_8lZO6h&4}Q{f__3up2tSQ7~1E^J^lI6%zO_R
z8;)nM&h}d0ywh9k#y-RYU2jfeFIj&l2t7zYf+KqoH64xchiKLYvB~Rrm9b6ai8%%X
zvzqLEix`znPPdgSUZx$Axt}|gsyCH(k{=xs^nl}=Q_2Bb2^lZ{<
zXHTE!NT~6~kbrgsxoklZ?t(hUtK$;{IP;t!ZNzP!-r0xQ`
zJuwnS_iBKsh0BxW3T{6#YZ#G{_1tu+4Hq4bG+dN#qdP9E67O*J!h*ik_y0U{2&#U0
zNa2~iXq0(*#GZnBs!P1nr90X;LW+`fDBhm%_T+72cc6_XD-;N4q&Tk!=!jNTmWvFv
zCW?ei_2#Db+prB~?C~8he_fM!b7F_GDXPQxvmb8sjRe{I$}0K|dvdB+KDM?LR)j42`pw+3#l3m+{OS!xfFH=E
zowF||%pGc=>QqgV!d?UkF448>BY91N3_Y=-u@mq`{OJV6UVI0?>e}UoX9b;B*<)g|@vV7Jmy)wpJKSsa
z0JD_#Zp3ai7Ts;JqK&-LbS&Y%sp5m!_A<@Nn-BM|<|S#}%EpCnG4~^REzs@5{O8YbH6~C
zYWf0KODWxVTQS_=Ta%la3^`H7^PcBM>z8+&Yg0ZmiH+#h!?Vw@-3zz4jeLyP!ND-z
zi}eFG_zM`Vcu=c1n`WPizt1JtUfVa4a#rSz$t6YxQ;gOljQT=Hj0}<^Ol8R5gl^5#
zPM}yR{pFO7lkj9^G1+IeGtoR@Z71G;`MRdf)xZ!ed>I34Mv(`rw^oSd7vnT&0JhncmC^J
zLy`8GCjIr>E+vw58_E|)g*XqNy|uU{MM)!fGKSk<<%Yh2gN17@7@c{yk;8P~e1+E*=CGU|#_8rEguocZTe@9>lGmFsA44Q5`%DYR}Amh3B_H2cx
z`n}Uv!Y}b2ShUSyE}BfiJF=P57xrV(?w_t)!hC8rV9yL0;8YKaVSOolLLfKaK9$?{
zDWgj1xUhnh+;l}@>S)Fl;ag^cr=hQbcU>FFYo?dTZ=FqQO=?woaYHguTZKdP?>l^c
zI)tb|e{AK}I(iYFKorN_!9p9+%q2*H;G%{4{dC@d&*nxb_V~~*QC4;#$LZfkxlzQV
zI@7;6X7Q~5hzqy0nAnP@#v2d{8^lq~Ayju=j%*+0xlLWTt(Bc_mwJCVSz+6xTWIQB
z;$!-alf{RC*n?Ld0ReJ(MRJZ_w5<pwf|{#F@j_o7I*R4T+?4QHLXM1v=gXcqYg7Dv^)<6siRTQwPe07Rp})Nz(tnN!MzOgdpm=SZ
z3&y1d#)K4GQ<@bT*HCvqX-p_gHVqA?vG#4&`Ea|HgZt^G)CW}s$cF#lah(5jtaBWI
zMkNX#LZK{@Elg&pfe|NlIaHP%H>OwnK8@;xVo8RZpnX$1{i~42$CF1*FC$7*@!O~-
zW0ZYPbPXMrX|DzAk|-TOXRNtyUpcr`k<3fSBlEnx=7E!3YBx<-htR1$n|ZHXrt#8=
zrM-LQgjaAEP#Fq4&zLh!2`}?t5u}>Fs0qaM)1Y-<@ev3t%{+_l6+c
z88j&MSdCVOmI}8a#b@%rOrdq)YnU$hx>Ahnl3h($6WM-JN)
z--cP;F%KL7J!&e_vc~l+53~Fe-<>#^qvfgSG~+LXZlj3QpN@?9q^)fT85$-*N5uxc
zV+yd-br)l_CU4df4{vuI$a~^oO?=Tc-1vpjt(Os@
z9>eT{6SiIK+8vb!4&)+HG(E$Wlb0+1>#YE;A(S-CD*2W=0
zw}hJY$f+HZCk`cFi=k7brS>+kHn191LXlTo%Z0iThPPupT>`M1+S1kcD(lXW-z$6S
zy@nXgX+uP6ao~F9rL7lAQk`NrIA)yfs0ZrskZ|~^?xzuYNLee;
z{-p(Xra^nHk$}4%S
zKC}{4o1zlLd38rbH0D)cq4@G8b{~o$|Gybm`}g8*|KI<93(l$`UlQEk2wH6KhbfmR
zM+QhccY7EHs4VP!sgysVd-B#+qRbm3-S>9yLfWaRcG|G9MA{P}FG1GoR?bD==3nc_
zY&C8hEo2;zPD@xbNz<;2?=D<;m~l3lrG;8(zx2CU`(gjpvYL^~??Wp?n+lmc4mY|)
ze1Y!OmI!WU9ezEOynqm>d|P}wgl%Xpm`pAsg-sN+LuDaNqm-FO)x<
zad&nBd;gI7R~#DZR$$d%=*}iuRs7cTWO-+ksmg$tv$F;Oh&&oF%fzCl*cQwyupbtf
zq8r%1&5EI@b6A_DWn>p?j;_YKuWR5^>64k}#2oCcH!E(XPy={2++}braef#30>*57cPkzE``hg=C3apGCn<}6QqVg0DxL%s
z#7Z9RHGuZ@h&<3`5ff&7=sI^8T%OBRuh4%$IfWXRAOSyEEZoyTbD7|@$>B1L7$Ldw
zH$%Vv2n|EtMm5QxzVwADlK?wuZ!nNwe*D|@|Ge}cLdbEjRzL!}uT&w9L~emJV?BOI
zqBZ?Vz`?md&;sX}04Hck1IP#e?{|&AeOurkct7M|ZE5Sk$h*1!h@~orAfb53fyizE
zNAqzS{v_z_aF@{c&t-mUPe!hN{ux6&RHoCTT>^=)3mfdquPq3{n3{+yq`}08B
zShRwX3LS7x4On3?bP?8$=3GO=8dK#1a^Uyu
zE~>7J+ITQe$s6?mvb~N2nWZK)vjX>+u}MqKda=f>(^E3iuS`-_N3$MeYT@YL
z7*c(A-x1|}mNA>@{!Jn;DJeI_hw8R2XNE{}Yow9o@r0KNR9_KU0`i1hqBwCtZRPm_
zYI4pDL5+tGi>D0jwA-Guv32)ropz4ZPtdlvyc7{1xNz5d_gvtnRv>u9n=~1lBWiG4
zCLgh5Z$5XS7NzyKOly-rnq-Nmq0j0Gj85x9Rg2c+?espHjDb?5G7$+L=jd3X0?O?>
zdS{o5fJb0UyJ0woxyeN0!#?m<9Wo2=N&+G^2~WULefk}~{~g_4L+sA+C6@G$ZvB^0
zB(A^SXDX`L4Lve3CILBx=&eN(5D%NW?TshskpL7I-eWt8$i2S@ZZCO$FTlO`6$#*l
zFC?Oz;fLaCCWJ$8Z2pntotD-0U2mM|#_nTy+1JRW%tZye*B)W=;UOa~>&D3Y|5s%F
zH};8OOB^0XeO;A3!l9Q8_OijaU-)Iz9O&`m#XtWuI}`uJZjie^hacb;=*k~+2Zyml
zqAYHZPyo{3KZ4lDqT10CsQ;4>|B+AS1)sJ4RmSIkdGXh|`hV~Ok$re)Hyy7jel=@@
zA2~@ZsFOn9SKQ&8ViFf-9~(52~nvxbiPD;6!x;ynIywfjuW0jYXNCN0L
z!1{XPi%dk0sT>l(pZ{<4=n~9Z_Yo*Obk{2e4<22SfuK~B_XtY`>IoYrBp_k%f7|{q
zul|X-F`~gj0#3;N{G$iu1Z6xc%q%Jh55_{8m1&O8kpO^T{f{l=|Dm?|dm#z%
z?D0iZt9l5PfeDwpb=3S2%Rz+1f7bPDd1%?20p3&j^F<6MP1(|@*ale5iYHa;?1gLj
z`l{Y#s{4RPlPViasm4;=U6tF)d6s^oT$Qw4LZsS;maw|!Lm#X;`yeUST`j*>w=j)`0UH@8+{q%p!2);n?@i;9xP
zEBf8J^GdJKV7aO~X9c|Zd?KaiqIw+XtsV)TIIihRRyln=;5K^WwF9nM*iJhBwY3*lT
zZGc5xiRrisMx{Iqpb7Y3U^P0c6u%kU8$C^{!s<=&IA2!ruohFaNSTF2EaW^uin*~U
ze$0vS?=e27@-zZVSD0oZ9X(&pId!XD4{3`WUs~eY7|#n1nm1`GjXKzwHgSk(KA^Q1neMFYiA!NHIHu|L`OLUgKPVZP_~ydS{ehBl`uEqFQE;VFF4r
z77nCEhi}w=OjAT0SAS5Kb-T#!l3cW}k<2jx3FxkU4E`C;qzKidr*+C5P1T@a6@1TN
zP>XM^$nDk}*AzWA8xr6wk$9(dNP*%@=H98Hb|FX0SV2LC_L{dSH#q3M0x#u2mJkm}
z0Cl~mCuV%1zeU;F*uvJb*h^8<^iV$-pQRI4N`6{OPj-vkYeIW&d+sjA
zy6dD)!p7HWYstO%Lr5iRm^J+*acSzH5XouE|CbDY0CM
z0TQBhm6+MApx2++As5OGT<9)*7g#K+jbN@7ba~COgeXtnpkvz5#URHQH%jgk1**~Y
zAFCtBjM`cb0?VJIPBshDPjp?PzS6!l_TGESyHyFziE>6nr=uv4qHdVB7vFiG*2T9!
zIk|g3drF;V>b$u6q@7%Z)~%gFn5@k>sN7@DZ$-hmH_wCV5PX|Ki%vXa7FE&VaLd0r#D+5&e!Ez}7pbkSxC?747T?>%p_aCy1x^rh@p`X<
zzc!`SDb!pui%j%BCjY6hoz*OIO!1+U8Ldd{e#!DgKK)bBYefd`XGks&M0L~#55}Yo
zcFbGDyAUrnzL<%}p;~TzOFPqP;CeStSsDfYF_cB;8J;_xx$)gmd?xk87SJVc5{Gg}
zPsVHbz4=(3y2R&J>K_|hIKLy^Fij?<=eKufEy>h!W@$FJdj93rIKUR?gz4COJ3s>5
z=lF4;ywjjrGm2z$ukOp#iYr;isB1Vg3vG`IoqE$*Di|o|MILeS%7z*~iSSrU0XZyi
zobUlD9>xILI1R%L@^LV|+EZ;Vb`%dPddL4i^4>fe%J}aaSCq=GWH&-7OOiF)LlNLI`0nj3xU{DBF}>Vyr{R*d-x5V;w~H8Os=E>34n4?|JU~d!OgI&+q=8=RVIr
zzkhHX=W?C-T-WFGe!t$Y#mx2rp3Cxf_cbf2;g*~CAeO(u&WVX23`ZMknge)7(Pzmi
z^If*W#B$uM`AajB^0AA4`D;RX+Sqce$ECZ1h69B+o1(QtZi8w}BNqdV0NG53e_dh(
z8frAQ?JAitu=Y#8(z^MTSX$EZ4ZTZIJ*r(ncY`KHL}@KZNB8hb`Y$i4{qVLx&m&9%
zr6V&a#9btZ3F=xa;j1Fz+PRLdJByT!!7P|Yf&<`LU0@!%IQ?g)UMK3U|5(Ob1f=Yx
zS>|rtY$E9`lhJchI_Au(v8PTyTStssf^Bh7l;KU?ST4k4($0mM
zjpO)SRG?C}mT9JJT506PUuU?6Z{=hcL4A!nZlSYUSgF3WW`sTE3Br_ylBF6J!XtJ%
zjK6($u1`0%So41(KO;(U##t3-Y3z!hieb(U?o9WLddYZrE{5ttbnHUoR!Prjm2_^Z
zGC7&328zz*Bo}(HWhb}9I~5*r_o+>*q+{?oh@upg$16##NgGK5ExEKNq%#K}m0^g=
zwt$Y{v;^#X9*sq$7R}mzO2PHs(G{v^mE<*nQC~H)v{PfX8D|DbIhqR??TcoTx@PlLA0C?TYEc0mIS;%T
zj|}r>T$GCunCKFOam}(le%LtP>k?~tYw59&AkKh=xW|3&3ZzfKD~Y4%<$rV8M^4D^NF!@$s8ClZ=#=CK7*2odLl=Pgvv=p1lQhDo?onX;ZxHSLi7I(5KVd5Uy
zbb;c|vgUT^E7R{)0{iJ}z0SKQJi+VB?%6MaPBD;VJg0iH5jia%zY6eTr*?EuE45jj
znK?_B{=BFj+RnUkTxuWlM|y4doOHUW+9rbTZMt{$%?xly$LBl#&$_68=e}wPE)~D^
z?tRlF7Yko&w`b@vc2shj(VPV}XL0h(V7{f8D`<$(Q29S5gH*A`n#4Llh=
z^yrjGdy73d?_Ivq=&?{{ne(GSjye|r5t~l!!p-Fn$_&YNc&qx#@gS*@vw47&dk12+oi3!sgdP$6AHzmArqT+DippVp+
z*kL3>GanAo9-+$Pr=L`fC^m>=Fro6{30L1ql^mZK-Sy9Pc@ZPM!L2F!e>egCx9KQa
ziI#)T!h^TKi|*kUi+c$@o&CzN3u}Zo3G)`!DXK~mQ9hKro*n8wGv#rPsQYLU)NkOO
zI;lpm1YHCM-fz~3RiMl;GL+qpX-#FayYUb*l6f_+rh9?#GrXWq*D_t
zJR@>|;NGrorRiu~2VURN&@PE^Vfx^I({j4>gWln?O>409el_;Gmau#yc^=SEDFM4
zS|L0^qoO+?VV33DSb_yAM)$?z)wRdd#o{X|g6hIQH1&6<#V3!UTaZ|Ud=Yu1BZGc%
z>~~|tph@D7w9Sg5>IRl1q_2k?v5$Z0^9|LdPw042px||(_h2%=6O~YJ-M0?fq#?9?
zP?SaR?;&(P$Rt-1xQEq)_cX9aCbci
z>0hNaQK6buu(v1z1Qm%rQPFaXYC%JRolxhTEYS%IV>$aZa`<>*wG|bKJW=bC$c|8qpbr|L`lg+HOXt=(Tg8~9rVE!x&Oldgac!}6
zdxazdItU7XP(M)5T(9Dd4w8iHNca0QaTOK1XpmF_pxZksQ%ie3OTYIRx3*#!%JL90
zw1gIR#ACuhJbwNx9M%EjRPz2BD57q~_t@AWL2_?7E|2%_?%_4hCn8XZY^cE^Y)YKu
zKupIX7g0P2`Cr~PuqSE7jg1Xe@xL>nl!48Ah9WDqTI}?n^Gq#_)-Y!fD^)pNAh#%1
zc
zY2AA|$1O>w1}Nr#G);;MW}Q0?d3p+tq@HK#3c5H9w%dtFKdvZuvXkv1mD&l+K!omD
z_0}0?BrvhxTeGpa|FKVqAiZy8He5UFd;i<7gRD2rm7{hkbBkwBFQ@ar%g<+D|Di8F
z++s#np`4@l&7Ww#ukoC)(^h)Ao-3)%+tG{^)-6
zb3XI2_YR+V?V5x2t+S!g`0y}n}3_x-tP81K`&
zQ9tjCF_cn8p~DwJtj(7k3B;C*wpRUL$evafUy8y^zoJZ10{jl$6Pzk?4IG%{83VFe
zPXwQ40AFpn_|jIARsenhj3{trq(X9UneL}~+&a|wkDACQTKK0at7!uXX|};BQlEN8
zB7;U&%hQ4;FPlUv*GF6rMRMIq==1Z`F0~MECO0;c)$J1^O}D}Kj~}}ox7xXH3+|JGBrCE@g^R5OK_&dk
zP@SwBOP0N6OTK@KrLIrJt(NZgr>-B;qB!{3Kc<<|y0B1Y+&06g0@5#m|uO0^8kV3|Ss==*036?>T`BbwJ
zB@Rv}x-cdIH;%J$G)WK>6Ew3hlYX~X)N|TNML(eG3mp3l8*B!kiPH;PbrDq@VVO%?
z%nu=)>x<%)8m?Hd%j`0b{d=#<`kzpZU(DaZgnQ~G#k6-yQCuMb?mYAP2@6?<%KTJ;
zN~FCN#VV0}b$&}v=;4>2KTdN{i4ErSyY^u_uonn0Egb0W!r4kt5&rke@bAXTi|QH^
zdO8J~J-(-QKYJnN_I;1XU`q5TX}5rrO?yFZQkwLPY_}8uixy!b9MDr2t(=-k-;BIW
z{jT#hB!{=Atff2nFf^)4{blH6V?Y2izEdY#YGX+=wB%>$x_XJW@mZ1|GmNXB4337L
zaz(b8+6`+!N)XIs*P@4+kVGXwI2ZBjAT}D2WpvB_rTvE$b*tA-mxDhiczzfp)Q|sV
z;KDuyCYX!!@z+}}QFRGhr20jcQl?n%>-C>M1xL)3xh2m^>BRUw(LGoEk%x%da>YX;
zQGl6*6EgTOLnSimRY79M>nTWu(dU_&evi+!@*O@^Q`&x>!v>UC*8_Ot@zE|}K#7T*@cLJe}H_{nr
zPSJd_UZjhWFBLN|v#afX-^b9q8x~=kUj;m#mhg>5R3i9A^p2LKbG?YDje>Pc<BO!k&pTq><7s~F`4cs&}R>M9SP?cXqWzl;*N0vG|fUc1aEEu^sI8gYUDvZ&k$6d
zVu$Q?q~%~YmC(qw^u-3X}%
ztj=xpX6LI4xcg3K&MGtpy1XxDeAbb63ImpdR
zi|#Z<&7ffDDuC#C|7EzmK?1&3Mgmv*b+AWvU(_2tRz>Kq|D%gY9r|ypxQH95BMne6
zavP5R%V1P$+hMEe^_M|1jKDz~1Rl1vI_Xpp^&0;;+2Yts_X4zE1pZ&3g#A~UUjNeH
zL4VhG$Q{YRYME&XgMS%{`UGeO=sx!d*seMDB?mz<+gAPh4#Yd-2P!^WyUu%bETfpRd#ph?iKdyx2F26PVa<)KgAR0gv7
zF<>-H9nk2Z%@yZU6UPDKD%Kn<{;wHlco29>qNpSRtU8*EeW!hPU9~bMcT7KDbmC
ze3XAzLQDGt4Q`*}`K>n5Blk8)i2R`dAvmF7`rL8J$VXf=JwrTMV~9Gb!S>n8<&~Kz
zhW*!Tm$8z6>&^0Fc2SWap%yNx?I=RKfT}<=&4{X?%BoE}8{=--yNkQAWC!z`=QC`i
zUDjt7U0@t5`5)+j^lvOql604BN-?l~34CM^8WTbm(X2ND*f}^)OCiK=_C~1
ziSPSrpkts@7!5Wtz6}CFD56eunW{)VavxEy8^pE$vX$G(GW_6MI_fp=c5GDbbby5<
z!>*ZyjlnuwAoSBqHOcUO7vM8?YiE-)9LY16EqBc~w~LVAdvfJZyNpuozxyI&x~7-eGx2|KKK_hc5q)45cCOBF4R_rfYON
z;N`^l{WFW7ISFA;?kPg>{Nn9W+lf5LalEgayj?u(n*Y^Gb!mEUhB^J_XW!SYH~7>m
z<#RZ~0OTGaOzdbKM{y!JB+83sdu6PybLwfB==`>jxc&3QpN`Zt(ZdQxQqw5x0WCA=
zk|9XWt4K31(X9D-JtOhvvld=jxRMQI!Mz6^g{;nqag-_YZ62{`F_yuspM4IqiZw~0
zoY@F1y7H*vBMFIueBUaLG|5}8hh^*!PBJi?S{TjV@MS
zzsQQE40BvJX07g(e(*jK6BhLcr)4
z?n$1%IvTEeFmv_ML+zM&irQ?Sl{FFE?d)of1<_UNRFE?z||(7?;hsh-3sn
zT~Ned29Q#qZwH+_%y^{EoqbcqC$q3|`e9Q}a^>fpLt&o=RVjVpjXyYg5M2bGWk%)#
ztTg17R*G-ISgFu&aWb?pOZENW5Xy6}^@3|rL1
zcY28Q3KFO@>fAfr!?aWYF@=kde^Je%rl;yMrh3>lWyNf-K>lulVls8$hfYba>apV=
zN?)Ii+CZSrv1up{r<`sfgg54Ndm&lL;4rbUosX>8K-Qt!L5XvsPTX?QxtA>7Z1|nl
zh`%maviT~6IbXv;A9ANb$RYdv=7Z9L{3c{@CEQzi2BWu)3IA+R{%72>8Ka_RKX*4J
z%c|>1!`p^|UQ_;2wS7@}{vB(*E>Lt5Rpu8GFWyYpu3BVK?TiPLG^?$@HG5Q$Ql9vU
zt1}2gJxD!DYiSwft@UcMv>LW{E5Ao!GdHkLOIi}DDHOQv12T)U9=#Az@dP-pkMzHyJ#x=2(ms_J8cLZ5Z>
z)YY8Fj0Hp+_55g-S7q(I)}9TWI5+~BZA=BxdL2lihs-~ZYtKOtG7O0;F8fxCl3j_B
zsB<#1bs*4&v6DgI)d<{8v{F}RG5Jy`kDr4^0vG#Dh0(`tP|v*2PXZ#Y$HYU=i2wS7
z<{|KIu@EvA^`a^?gz?{==M#DI{8H93Nc94Ne;zo#j#yXy5HWlf$9|%>d?yOT2N;yd
zrF2hpXSU;D&JMBaSD&SClUChfQsYw+A6mC{p8AS4n!Z%(4`lbf5B-fPE;6>&CCo3?
zx6j84L=@(crOg%AkHOTSwL|5-DckA2+Kl?I`pc~nCtEqsGGr&)B*tra17};LCF`6d
zu^a84@AR6pT&ZL0R92qo@0{uC5}5!-qbIS}y?-Ph8tm4U?oMoZQ4A4xXj%JKHpqJf
z(~7Emn4xfif8613CSOmP1o42ov3@pHDMI&|-PjdK?5W;Lq%_BX*PkvoMBEaFesp}K
z&%b&0ABLJ&^?K$bK%6ZsQ)F<*L`G~`j_vsFRk=5P8;nZL9uLZ8jJ%od|6mDe8x)VB
zxeZ6m-t5njB8^Z)M1X#yyBH2*Rz
zPwXQpY;d4EhWOU;7y>JL_?ICS22eOQm4K>Jb8ceLgSms_!?r_K-*O-t22@IQ6Vwd!
z)C9d<Be
zXmaB1=-4{x9C23b`^5czH%T0#M(UHC3G+;|k80ebyfornX88nR*b3Y)&Hv&qS+v
zB_Fxe=`XX+Im8QeR1rtKR34m2J&=ph5{Ekij`jzj7+;g}?r@jFmJ&Yu^S7VvJ-Y7y
z;lv3&&)ewlgJem`XzzJo#}(8tBzSd^A`MczGF>rcuFk95KdNL$@8?SjZaK;X#0RU0
zN0+IaCN#<$;Ph!CfC^RuVNBL0?5-|CgU{b-7};0F8Qo{Iwm$0sg)hCeQoLT)jLg?75iuThj5%Rl)w>$f
z(!y20*
z^s|nioh*IZL424`ery#=-+KC!AD@%WTizB=^k1dR1%{C_XbD;bayBPDkCWMnn7b?nyx|&N?hPz-mq1fSl->=Vjp(t
z9rR54}DsswjF@t)YBRKawF}!omCr?&4=SWhP)CZaQ(f`K*
zCvmKb&XdmLHZjN>X(|2(=$VlaBu|wm+{-7P$JYu7ld?WqYL3w=iZZU?$Ii(7(zq$c
zR+@5$-J@;G)}3+%;X+=cXumW@=n|DOB5M=RDRp&~gpR9Yuxa>5DrL>x61SHblRAUp
zLINZh&J(OcdP9C+Tp#Iu4^U0GX8&clPj)1or{w__HH@UdVd~|@9A_&dOWZcez|KhP
z_;sE|?4`X+A-{gVhTA7lCuCNgOH~TvrVs12`f2$HM`9Kp8+srY1!KGPYo~UHW6IDE
z>GE06(@og!DCc7eKhNn=rmg3z!U52Xpk9<VY4Zv@3a2j2vmwj?JOCQ2~C=vBr#UzWH
z?7ZUg>}0KC@Rm5|PCf6fL+{Hp0CknDGyi1}=hD+$FDNVwJxq_I{D?Wa3C0=1h==h?
zJEy282r=+B4Ve`-L+Cdwy??~MktxibCCbrn{Htx#~Z^!{@=e`(ynl_DbZG!rsM3BUTb|JrxDjj+C@}`
z>F3ERceA>+)W`5D21RW*t?szKEQn}m`ue($6(MA25#7r2eQ?4Gf`I^tLpx*h~p5QlICT0+Jgc)Ha
z#b$<@7kKZ?bn1xw{I!Qco9?L-s3~;pJl7jQH$2Zu)g_)cIk7Y(EAny)yo>YN)#f}*
zF?#~C+^^AKwX=`{vtp>_hq*7H3Ge>YTP-I1Wyql~t-#I)P4PNWZrcP60_Z;b+>Hqt
zx`i8ip>TZT_hi43YmN~Q&P4v`xiy={s5)pvxf?W;t3pF577=|u5eP217vswpQ;z4&
z^p_Kyx|P2*-8R39e5d-1gP&39XqBMedOXn%nZ~wef@6yPnza|p*_TYlmXm&GJsOpD
zt0;av{X1FNW5Ufplac{c5PZlfm98(y-`U6Z$mM#NuwO5CZH9gvqpG39)b&en|Kp1Sl#Zl#vOy_UwK3A~Hw5RP#7gOzyK9s4oymXPr7gZifz`35A?
z5H0$b62*Boz|~S`fZ>ljQ2Y+WyCHWzLJrXV9Mi?E{ylL)mk{^KWuJ_sY^l;B7lZvA
zev39G_!W92`07a{-OzkjRnl%(GTQm@UH{fGI_s_p&sZej-A-IjC123?;Y(BAztZV&W96~4xvJOFJah61Tx*2{Bn28_Y!%!$ye@ofyvFlYJX@LMOSEgspd#3KVdmJ1GT
zLlPdNK+Y}K$7O2Oxr*_QEe#1caT$z{!&N^Es0Y)6qvUDq1CAa@=xsMJCqjc6O{yyf
z*w<)7`Hp+nMSchdwDN~|ARJs+@`cy
z+Tk>=sVbX2?MRkTV@WuS0WTD{+?YVfmEhKo9gc1PxP4hMjFOPHdzo0R-**j%f44x;
zW=@6`fG@Tvl35m7lATK71M7D@N`K1MQdso+yAG(!!D5(@AjemgOoS7n+-LbM5(K1P
zhhMd*McGX5Za&C%RI6z;aVhHib>J`kI~In+wx^4z{Gu`vLG8zf!AczGF|$%OFV37Q
zA4%b{KW_z5H8}hH6iCqU>NAyd2Z^zO3*adNwBt^fh!$2keeZsy)g5mB3b$hSrVCth
zKR1!0wbw2(1DQmW-kY^uPS`vuhXexN2|0JbLO9~rnIQ(YvXFab1
zW^zfp+u%SVm|-~u_>`rb?N%?QT5YYEYTT8`g1Ax>ZvrHmMNsq{2id+H-hsDwg5d<%
z3j3+V^~(O6hVHH`KQ?$%CE)M2j(nd7fGiiW_Jh&%o8%sXxE?ud>FZ*{Se;c}-=v23V?yA!7d0wnDi?G!KDyFVn9lsiTg@8fl&ugEqe)TkAD*UIM(Bt>H
zcsH}zSjH8_+ox`IOT*}A7v(8e;J1(aO97A`aqtJiy7bn~c;bWq=Dx({~;Z(e*ZL0P@5aP;&S`nHT+k5h+0bBejEZ0ZhKeyd_x{)IVR7
zvt5iAjHo5TVk&s`hgTurUJXkdXtoaC5z=qtFc!bGdb$xk`IzMK3+%WP;xB0MDiv^|
zH2bxlD>|Ab=jG-7MguSSW9nHF^JIRozGL_5vmgNN3SF3Tu7)$>U*w;0-<(v^ymu`M
zB!`lbSG=XwWS;?w4;mtK*p4)tE+9GJDE%ptFzjJ%Q~mD1;~%*^$;!#2u6Dpce1r6t
zftC(zqjwS9T`(}5@0I`ZWfo^jri)+PuF?YBpv`C&AnuOJbe-T|S39t(26*|4GGjj#W
z;molAmw|klx62)1{GLNP7;el*2VEaVNV6TrDseG7XH-rdb-BI@;fl5}Oz!qs72Y7J
zcjQVECgR?>Ai3arX0;QFWt9(9d=z@T+^;e-6;vsRKaCR-sU!CQut6AEoRHB4s;Q6C
za9zZnU27!Ol(^LUa6?LPT|Q>Rf&reDAJfdGE+)sGIE+53W@w{GOm&78`D|u
zT*AMm^Et%G{(3HiO8S1MAz{#@?u8nq+Lowi_LZy
zeV*l0TJ7h)-S-1H)P`9fE?%`W+1}Rq{5dYgV1eP5#00))%YFfdV_Z^h0xkeM0d@?u
zZ_QO9hGh9xlb|QS^Bqi{uqHb%c{^KOna+1#BXUJ5p@9*#
zISutLYYW`etWQroE8p;Ra_+`@`QR;b6a5?`VB)F<>OxI1cR(@exR$~aa?HqsG4E?z
zc@(oRnBg_E=^a6)zHbKKJzE_T-kqQ@;&);IQ(}aT9r82zakrXs-Sfg%6bF)Jc->F4
zxfecugFP`w_lM$|fA7RkNv$bSmSVZ+CbGX5w6Rxo-;IvMCkWgYce(r+cV}=U}xHc@nNdN}+vi*2I6ax>G%-2o5>c9`pY2
zygmU^>x}C@ontTm)KEhHo?B>YZOVz$!CJJeuH1_l!Z3MY)uJZ0W6LeW-Z&s?{O9y{
zMu|B;ce%c&+H6fsoZ1!Qtw2o!q)t2-{XG;UOr~fPmGc>%k#IhRuHHAEtj}3L{P_e4
zXAPKGhHiNiS)k-k6hGi*JW=iSbQJ<`2SvZmWrxKH!sJ?Nt5nqxH;}>I!IZ2cz%g6Z
z3c11Ho2>V*zbTv?I1oR4uMNar`vMRiP_>K6F8A-9ajTOqDHola<^A~~%~z!3mS$NN
zfwy}w411v{Wg1_fpa{7DIma*Yj}*@B!n;f-2ZbIEOAomF&3&
zaH&JoW*f-~VD|pJ_2b)^iy83Tm9zO@zea66BOcw8lfRQ6!T9rt;UmYdnE#9!`Jep$
zziCw12gWec%J2_S_>hUt76}3=llic@|NhxC)Vtn9z;&DP_Lvqo)tu}_F(p6bzEas3
z^Mfk)Ljry_g>xA>_!2Ss{x5^~6X+1QPJrq~gv20t3IM*jyBFF?gq_Hqv&d4`jjh5Y
z>}@`+Zadj`Bgu1B=Esd6H~744vBZ4*R?pTc3fRS#Yjpp^8&@lfs=e~^>$ofF<8;tm
zYXwoghhC*q1r>MF*wyueX9kVrT>f15&G^qRRL!!xV*qvV#@C5jEKYFgFihqRfA~&V
z3&?4lF5T$y!?1MS1EBAGRu;~@YB-{$Mm}SLqn*1mxBUiXIurBZUUj
zd7UqgXWFbzuY7I6j$E2s;f{$4-D=r+KOHTe+$x@V=za!aNGoa9T+jsRj%hJj&uyqZ
zf2TaG29GpII9m{Y(l7X^@ST>5E6!(?U$u8Q#}JyG>ErYUx$LOsn}4PVjeZhn3b?vi
zzO=kpSK_y}x@M+es}T@!a98j7vv0RApJD#gwvX=HiAFL(62J@DbF5#h!+wryNES@P
zLuW$vYE3?6u6a$I<`;P2>tKns|3sfq$yDi<<0j|hn^NoQROC+{#-
zg?-z;#qjWcNXkF~W@e8xvALfs=l)hrF0@4guI$b-M@BxZ9ljjuw^i}6$uPkCZSQ%j
zUeU=mKPW@o(u&ZCD-=;33x(tZqq36=8CBLpPV0?F$Tlrm;6xSHTk6?FFUW;AYHlFg
z?U$z8@S68@Z>0+wUVx6e4OYCuq#tITtoqB~Zlv&#pwbDOYTvjQ?kc0GsauYZS2h1s
zliS*vKlOxDD{T#14cZJTMX;0WmjHhdc<2cdpCgbR`pUtA#eAS0njLHay($))t$Zip
z?fHSAStOg*EriFohC)77o{aSM28C@GH_(SO4Fj#VBY|r3?v$P!s$&GpQ6^CBks*rWf
z5BcdJ{`2r=&+alHnY8cGwjC|{>X&#~d(WI&0UZ5ah;w8^q9HIgj8PgCA^U5>9#xNB
zoV#XgkyLPnDX<|y7X(jDQGk5{T!N~0HmAH?q=5W_(?-SOTfBGbOWs3&pepM-b`pUK
zaB+R^965|IM|?BSWjj5P8UBr)BW1}MaZBLk`104=%BUoT*SAePU;%*I4-$`1!1rjK
zBm~aKDOq)GQulyg1ZeQRHI|1uP8SR*5v^@s|Ka$9&JOXr-xFQ87oFb`+v-|ttRFl`RF`sG=S
zX>k8p^B0$6{XQ%p%Oy2W`OH@icI@DoNxo{>FuEv0saef2TF&*k6~%?7W?C!Cx9peC
zs~%^k|Ln2XQ=x0BJWWLP1&$tYIIK3XXG@ZzHmSX6?!5ptAG!?m|9TcDkMOo;rqbeK
z-wWg&Qpx=A3g_3T%HsT?jiRExcY(DBmgP#nhKLz$95|~*-=0B0kal1r-I03{$!y(z
zJ$KNgp&@tW`vVd4UgciY><2Druy1w}8-U93NRZDH@=3YXEZxi
z0j?}7_$8d!oF{9wQ26gBLX3MERB2>9=rs0~+(oMM2!cbUQ;wb7GW99K#W2gEx#s#M
zK=`2Mdd^<{pe>QvN89pGwu$wS(QLUT^_AUO<*h4e>AiGbAh@cXetir9Oq5TF5=dCd
zaBq%Nn3|72>n~;|#b2)wOBfGK`H|EV-Wzj+=Y4V|=>9%PU_);a2>61vBRQIN5HyZf
zoI}Pq#*<@KiHhk)sx$B0+>Q&I?3SO`sAyV1;wB)pQVlT0z#5@eJHfs9!%)7!d_5`O
ziO}qMH`&_y$H@<;|1!uI=)UZ|tSton7z7C0P6jD*#Qr0bna;9?Ic$*~cEv|Rhm3|l
z|KQ=b)>UjV`5vM2GE1E=_)A%KNKD+3ZQ^3fGm}%{AzV
zPPvB*8W&X{zM?I`d!khC*P5q-Vw@b+o_{)s0MnSDE4*7!JPeme${<@5uVuP-
zbQ1b=D)L=#jX1W1u9`Hy&+l1r)2ZxWvKsJhAxTPH6`}`IFN3)wLpDXpX#3(x7t)_(
zwY7Qvh?CVl(gjDJ7!{IC2(1YMAve4v#8;#LDnm6n2ioqtaC
zY;mPk1o0!=CuzA<7(N!>JgdX<80o;!7g*GH_`PgL^f73tKk%~dgT;c(XU^+Ai#=_q
zXjh6hm5mIc$VJCZc)4)6qf0a@LO!K6J(?YRZWh5e{psao0Mwh}J!M3&U)pI;#5vx3
zTllS+IJtiEQgEO1G$z|O6K7yA(hD=D55THVhfu(XyF_fF40@PKj=2|W6mGZMA0GST
z#Y;ms^L?{g4t@ZZvi2JI$>Hbb)f!Nxf$d2*$5{CImJRx
zpAQl*5DrVKy&I-MGH}?UT%3mLF*6XAWeK;~F?s1$;r2>^4v~5?O-YgrH|{p*b2~n_
z5$cin;j@|E3b3?Rc)j+WC@>4vxCFk@ZV8QY+dTM~Y=nHy_A?qZwdfjBV5aVsTCdux
z0o_Ol!K2v7CVdoz_G8ncM8}*tqin9}o@{${@pB4^&g1}604+I4I;eld8V6Td3{3x&
zx9-b181UThlQO99$h=EzZd=dc$m{!```OT0|4z#<$GV#WJe_8l?~?Q}H4y&HtHn}x
z-H2tEyv^FAgbjooyuVbwn|YdQ<}ebUUP3M-HKrqDv?K_i&9Z`--V?;Gc~Znz38u56
zLvNN=uo9d_7Q9?3XJ+1ut}>Um(2^9{AiAiI#>h^Ti1k{=BPF0oB!?PYN@2E%7Zkr|
z`mVIin1V*RB9^mLv-e=1v_N4tn4wn-M5P&C+!v1!
zFU?w6ch(g4FUbnuAAxkX4hrAK<%?LzV$L}gzVHJqvE))Ix2O-i<~tfQ0n;h7ZR~h6
zh<>(96gg#o2eu7!2FOBJ`2I3jl8`tSJfzb-E{L~@0PxrBobay9P>f69ouKy<^(GOo
zl7&z&qmA#!7rc_J;Der}CUO=@l_+w@x%h9y5T;l5X#rC>2RA?
zdDi&c+ufeb`~jxYPETuAzOG|yvzP!MSIP4LM`@q}tbR)u;Y;D2UpE6)`qTyHIa2No
zaBv#9hlQ=?=?^U===H}OH!U3kk1N19q{{m)xU<09N7@@j(9WdB$PYzI2cu<|j^26(
z=3Ul(exyGMkz|z~K)ygtH7$ZSHym0xAtFG)OxrL-^C!mh&z8HVM5ZRTG25oWbdKl!
zO;vHxr`e$%sbv99E{u@(|41975SDC+#ck>v*mDTTxQ5iT@#`0NJ0LN=Ne`aIzA7vZ
zHMmLK0tJBNgu-BM)XTKATDfo}j}ki06OJ|~%0(+}sMIv2K5_|65pOnX_;RCXkK;1K
zDuZLpt~7WuzYD}d7hShiqh->?upwJ@p;}`FHs-Rl52s>zY=*DmzG|6Yweh^2bW%6)
zGNEsQUXKTJ+LAXGnd96M?u$*6r#^m7SE{bT)G^7&;JPo$Us++O$r+dwO7OaHO)=%2
z&)yZTx>UVB97mT6Ru~w!cc)S=!(-oEz)j`57E$N&-6c{(Ay>X-wg1DG
ztrPvWMT_Vg6~s)v7j}C{!$|?+{XDQxcF*#=&8+3jQAt71$_=9qK2cQ!qXj_qj6t{)
z_LfdfbnZlAFQPXizMe@mciG^RLs!J!G?3@+qZHm$+kX7}g;v<>#rzcCSQp1-X-!3q
zLz4R#eTsrac-Z&_kJ-+dm)`-s$4w8OylTBi|8n=WI*0Qgu=_PLa1630**;j~$(Nwe
zl^@^imq44T-b)-YB#;OYoYx3BNnYmrt>To}LxKNzIuaqKp%$3WU0
z;2oEt+zp`ASe0D+XOi&=*e&{m<<}wb3-_qQ;Dx^Y86eSi!Tchbiuef~O50W0asjw{
z8QqZ6282F*Bstm`!@e%;bEm-tremy#haQMb-An3f|X`8jO6}zbWl~oif6dbQ}B2zS*rJNN)BLu}^5}ZwBMAjLl|<
zEtd{YBactFwZYD#mbx`u_lhsS9r&gH8Qb=)YII@*T&vRVLv{oJ-Tm=qL7>NaIrb%Q
zWT%e5cV=W=MIi3H&3CJ^=d-PnI+p=T#6%Uiv(QtvJT+BoHedI7qY+^H{1mJQ9
z{Y0@}_sYq;m+TzQvcMCkL;~bX7
z{R1Wv;|#_#q=)M5r`mcce0$Q0X@fsNEH86~i4f|A!yX(m`c0pWhrPHeli8
z)OqYay^{hKGw(DM=q26MzFiivD)FWbJsf|3qnitEw7p-hHn7wgI?L~Rxe>nE-^=J9
z_)%xDQV8GtC5icmLXVH5$u_DcXX`?3pmAIJ2%xzuARQ2>C{_1~b`{|FT3W+CzA
zvJaEI4Skzx*vE}DzEnyibFcpNP99brL$1;)@}TGw8_wu<%=@3R;0
z*>l}XmtF0X`7ZoPOrP*2MpY!N^Rel(neocsEhL|HzYR9i;gm=Ikj!Zt*rzmN8OfNu
zN4dR*dzeaOTznYMB%m}gEjtYvl8AYd;Wo45f>CW7P+exkHKC^2|1yyLM-)+KY^YvU
zO%n%(KwGGAXGC#$g+$s|W{>5*dGhV%?APFdUuPw*VT;MWr0>Mpy4-Lgixpuf^i5rI
z(5xT6!%Ml$GdVjgq~b)xj6tj1p6Vv^oe;`n`T&@-MS{H43gT@!?|e}npvig3#JZGb
zTs)LYmFV3QeZ|=p#>&c&mOq==rFw1p1(0$ka6LEEG
zVo5G8SApOjasU<}p9C>Fh1o_;`r%@1ogvb9m;I7;LN1x)%M?z9F=)C>1Bt7a2cufN
zWUqy_-|$+a0Grv8v}Yy6X(n7^=uI%(oc+Ti{}40mOVrdLDW8`AhE3gc{J@4$yM=Ht
zZ0&T+Q`WE4gqfc+(%k4ins5-&q?T;Z09~C{&~mEyZIJxjnt*%!mRr!Y
zy;p(mt!hm2!+4UF7DfWi357eZs8~o_XVq1J1g9R(dTkLG8Ws2x(1n=kkk9
zv@3mYE=a}7vWKA3g9*KGaScCf%oK31HFV%Muloc0IOrgt3p~wCG9pCnXFUcTX=-};
zU|GSh1}{*rd;|D0gPd^R2eme8u9^n{j2>Pk>s)@oXz8;4VwC=kl%Myg4sTg~q0cFT
zK-=6w5Nz!QLJUO2M$r48R{GtGEv=diTr{?Yu_OFRYJr=@DM2s3BUnYwwsUseN2$4g
z0Jr77mAtiR|2Vl2Fnmg^{*xC&9DMp8N|^tv^nm|2jgN5^kal~K{0ZnSO+tV<*?6&(
ztr$m|F`vte43PfgwLI|Iqk%Qe`}<5>UvhTh;NTQ;QLblavR)T9wL?lLu-BK94RzCX
ze^H_M>1zjBck6uRep|~nHuv&fj`PzN3#;s@y@sAt0}==XNs(kM(U+7-0FbKI`6>H0
zfn|9zM!9MX0Y$GJT3-C{y!eN_N-F95Qm$hfi#StMn#t?`hrRC(YckvRj-sNVf)u4m
zR6rCEEFemYiZoFX1pyJFBGLpzKw3xu1*AqmKtYLs^b+YkQL0Ms2vQQ7^n?-uDSVqb
zGjq@AoO@@^z2}zinLqL**>BtL-fOSD`Y*HEc2%4JeUI6}{0z&T8LYD3?pB&;C1B1N
zovC)PxerJGM3Z
zm+O-^s3k`v$ogpIH|{^LEzoZ(h*%KfOy0ugd*}U$NO+qy3Dqk}@h8Ei>xQ=CMCs@9
zIpRQB8-Ask&(N=uefOmP8L9W29F5T4M~=X4UPSRM@E>!{KgI7g
zE#P?#_NLJa`31Ter;Iqn$Sglbdw_tDI+&-lEf3^(EqcgVNxu@kWKFvhEQ=N@oaQ+$
z`C{bE=!Ic!gj73>kcl5k2%=qcJ;@$~^&RAJIyLRoXzF2QJJexsUKF|g(UgMv%&2ll
zL!JF+UB~5aSWd?V(&b3n9akDLE6Fd21RKH!CXjjQX+qs9QtcRxR;TwIb8%zij#a(2jLA-*YVg(1W{&
z_dRH3Dd}YDhWRnBCwRdWWU+lgMAqsZTL!BzJ5)ANYoP)1sX5BYWyv^9!bFiMWZo41
z@yZJk(RXYjb+hM_GZvod
zZ&qe?GC;O)Qpl^UC+cR|rYqa%%|H^oPes>3cOqpFhK1B5(r^uvA5F|Iutz1`?5Cox
zXzM%5^_AYdG&1cU(7vva#qvRbD+g*K`d|=Sx|m0i%>l$_mK!?G!!UpOoGAgz
ze07hMp#9Q`PY+35?$3jiN3eM}y3b8(0E$%EK|Q?&*@mgncO>s2>W|>W$n4tCNQs-K
zko`r+hWM&?PYJjgZV-F`G3ar+!U}3wg^`ShyMZmPz6x5z`krHl&hV2$AIVGh8hO{l
zCo0c(VG9pcX_D*hKByij2|d44FZK*tr<4+BT3v|*6tm*4%gT!t)OIk3MI
z$33O6`?XkdMvA3NOp+fab^lAdgy+3yMUFQmxp>Ye>DjVXqo3pNp}iw);6>w4@s@*6
zaBLk&C|$g$J*nud=bhDa*3r?NnO6mk`LCYd-T%t%Yl7Fuof~WuDenIkCH40#@!$JR
zWu*Yp2ydn-qIzshK{m~i8YlfGW)FPc2+*fJZHLX^s(EpCAg*XXlYNP?Yr&|QY>=`7
zf~ZWtjQmI8Q-ApQ|Cr79@9n>T_IEkR^3Y)&g3{t7!MIr)yz54TdWWBe=pdZSu?Dqj
zdtn5GC9$bH<8~uofmH4Vjky18mi|9`J@1Mwsh+95X9j5YW_cjyF1Hqc)n>kr5KWH)
z2}sO3ph~oxhqx+fF((Z
z?jKXqADIu%@c_R!{sj8i8pI{-8%T#($_6Q1T^t`P|HK?|e2T
zk8+JlAOupk@lV|FKRTDg3xEH?ZE6Wk39ePW4Lfx-YFgM_kd4)q$uZ
zjM!`9^1d*cR)gVKQqS1Gp`6z!;NAIB&CS4zZci8t5aw(vcOYno
zIVVk$q*{u|(yn(M=+eP#rAya02}dsRZ`S{G$ho39t|<*?*&N>mXl#-7=|4*&-(SZo
zPv`F0&4{GK9>O~Nt#P6;CnnKZ1Cg+Er-eI-7$G_eHDtf30_#hZA2>#q&3NnTUaRC*
zr0{z1(@xgq1I@7_LYvi2iv-g0hFq%HWq^TTtI~c_cI{aOSBpzUdEvK?9(=v$g!be1
zhUa-qxrNgVy|1^5R33{Zw>?MeW5HNHQ9yw}tEuWD*pXlbCN@Pe&c{<#^VJ#rb<2;u
z%QJ{2+7-Gmj@n0he3gXv&*rMD5xC#Ait~Q726;4S(EgU&_=NE1>Ft-@Dis|}Hy@t=
zq@6Qc^X2ROXUTdMF2t2wDuz&07WHKj%{g&A_K3YEX0)W;nGa|5CQ|3r*KO+GrhF8=
zC!LoZ!mYn8MilF+@v8nNA><##wETzX%x~RQ0PB+;sB9n>F7HPpS~)u5av`4tSyaJfcP
z(#R&kr4Pu?`3gF|Gt!)INv@B+G$at!PDylwVq8}IyD|~4x-l<0FJNS@-!{D-g2sdG
zU}fRyzqcv;E00Oe4-ki$|{=z?|(*4x0Jl}Yo*3kkrM9^+^l!2Jm*%dh;r5EWPuBcbIZcD
zs!v0d^llcqdyZ#-30nzU_4=QrsQqhOU*!knK=A@eXA;R?pu8v8^6hOGq5oGB~gKONX1kX
z*y}7nC9jbS46B3+{kSgeyKY$Vq;UoU{mGtc4flgCj1qO|2c6N(^NCD>Xo5d|#Wey*
zfv?ThAYqbA$${9J$^PhJ=4}RAyNrp`7%8I5$y9Ny%#+2L+zMb{KA;PbHnmArOhmaB
z$Q)_{AvGT(SKHN~eVt`8b@VElaJ0)Gep)dm0LH0wnw0TgWgj(+IKgaAq(1y?(-$@D
zps80_&6-hCnXr5ZGNnOTva2VrL5?!(lKR&maFO`~3zSK!yaidHjz;+}
zffZ@42YNMI^DH-4t%S{aPR5?m?H0FBKvi9(9|g4wWVld-6Pc)aF_fb43F-=&wH7Lj
z#aq6VxzS$ZIK`f6{PD@V!tgyi>;){+&pi!GRn|Wv$B0BWqxkcoy{tonzPKY)l(j1M
zhEzml%E8_lj%ddXbF>ou=iD8T{;hBGdi$f77C5HeO;Oz+>Ga>56&N86$%kr=4uFwf
zcEMJ-@HZ?|7!}63bW{T@XToLux~M;RT_VlB$Qex7pI`~0FCd7
z(_rcn56b@fG#4<>_4^Dr^|cVmwJZXIB|Vh%9}vc4c;W6++(GW^sCQu~KtgMRm9
zc$Td0ADOs?K@cKe#9BG%kK`;Nb4{#v*7WA8+1fiNP71i|NES(mmllN{J+}q1`xBc(
zBeJu5z7@DW?lTgkkx~dpxNqsEY$hMo7T5X3akMeL9$j!+nIqEI%OX+g#VKnYd+h3boBV{Vo`8K4rRzI?>BV(fir>1fU5Vn70
z7elb2P)xpBb0BW7^S@bW3|JzT`XycF{#E>1Y2OgjNe~_lu39UEuaB(t?w7)
zA15Kqf8yHTCBL0rf5-PKY{(u&)=)#MHOLkumuq6R{6kW7Ojg7n72EF0OJKbbe!6s@qvgB;5tL0_;m8iX0i3c
zm!5(6_Jyd3w?$k?a`$LuwyI~C8t=$uJDnJvYi+?i`cWU$c=Vnll0cpMYoz7sxn1Jb
z&6}oNU>rD+vs{&O6)LikGW&JPb|zXViYMItZiEUV7i+;
zFSE@pY3k~?+0P2HQOi;DzGh#gGL<>388Qh)Ndd29gQVpIdgz~;XTsJX=cU&*r70Lw
zJrP!@!Qdq|SnLf{!5b8NSh>KCJAQWzP4ylg?DZg^de=z{rjBoH(PzyJ2B-;FW0-TE*C
ziG4t6n?>For+ft1DSN=S>FW$iYl!G#`b3ydP0&o_P1UeobyCVn9Te*u_629Kx
z=5>4F1EYpIlZDVE0pf=7Si@NtHOL$xn`3L<_KH*q+l?`LFElE?_1ND(_D)KEv&88G
z_%NDI&jJM>G<=q{Bjt(x9-{K=a4P2W9tz4z+dltk*Fv+rpuA@yv<|N{EY$WhsWQF4
zTSGrkN-1-v4N{rBr*duwC}Pi(uTGi^8p{o~E|el`jSgtcU-;Zomi&p|<<0g7nHxrS
znMUhOEqVACMmsKu|8^k<|90&%P9~tSowLIKxn-R;k;3eCPxK70Cv5Rg_(*4ety-fc
zyNj0psL)n~BSmi2LQUr9r+6dw(*&mmWV$lTjevDGj)Q-%ie7O(sp%!l4eG-Ir_)9U
zwTzN+<{}an`&nUSa?797(|vWlPC=YDMg@+(`ucjLq0jz0(2+HL{tHb8|B*@IH^)oR
z0%a0
zA#Onq|6MY*G8X#;MAqW%ct8&G!^cSW$bDW
zfKe})>Vib4IeoxQPLsufD6ajPVUSR=2I+hyj~Q@6W$|HUis`@^?7N%}P*{c5!F8S_
zjY3JdCHSfch6_b{UXJZ_2Ljvm2LmYAKYBh9De5}VPAKwn208BV8#<|K@kWWc@6Re+k>yTvaMx>v@78H|VXTWAEb3BHHCR
zVZS?DX8c&et7AQR+xWEj&DtG~2)C#^0@3t0JsmfzlSe_7EXe6J;>n`;sgq2FuDOve
zqF!#X_VX~Y0qiDo`6ph!di!pl^X|pJ!+c30;u6W&p=e*+zNyl?tI|at_eDCUvu}C0
z?Z-V{cJvnVI;!q015H2Pj*#qLDET?Y@YygL7z0XLYe;KBd&
zQn01O^|K>?Gn$FaJp&nFb9E%@Fu9^5SFvi)$tJ<(IgY0X*9{W_j1NiwK`
zaW0x1_p~u;pD~i{le9D?YRx69n3TR425=X*+Bi?CnQ5%%`7O|@7?w3|o^_+U!)R_)
z7XxZ*DqXCQBpxtcR#dcY$14HjPZ_b)iyS|Myr-JgM&w`0I&Vp*nd8N#Getn;3WSnmCPZggL7O&#MPN)=quRYFNYsV
zO=iggfXJbLek%>SOoajS>3B7UE+zw(_-}atGNbjkgYcWv4{Owg*s11nJce=Ib!;xE~hE}CA{(FY%{AN$Z8gz~NEE`**
zb*)m2%fnQ~ZM(VX73a9NAjhflB#ZZh8_V}__$Y0&Jz%_oc64!e+#f?>&YRvVEqg(4
zYuwL&Picei6&{H4jliXMFWB^W)70z2g6lf7=dK$7TVH9@(LZIb`IlpM|7sfg$wZWz
z;=}$ThW$mCVPAUj_7tIc4&Ea?8&*CI?-02FC4mZT{*=nkyVw!R9X%;gW!miRpZIER
z7OxI;^|rwdw(riSN!i+ulwUj!N#xAg}^d*>gK@Bi=(|IrAp!?@A^!~*-pF8b%b^PEF0G)c3O
z&e>DTEmkqM!M1WWjSWHDIcpzF&1;TGr0&()$42cPXR>V|ZL}KEVXYy;1dPws2+W-&
zQQ|$VN-l3Ts+I^7}F+NI7Gzhv|6muMDYs|LzK+s?j(7MZ)dQ4KUbwafX={ha>r)$vGQGyC
z9e7*#HiTmVNj#LpHy^ci;9K`j#4+T-J&s>rUXenYUYMEeT*Cj(>r7jG@&E7WJS{Mg
z>hi--NND1R-oQ85n!sziOb?NKBRF@=<`F6`bhTEaS%buj)D2R#
zCZTk-n^vEu101y?BfGKxSrZUrmIF|7cw
zdr%Jgw&8#Rh9gaJy54~hM7;yM*Ks8K?$@#7th2=X?yuOCgK0-8IkU(Q6ZrX^1ps;;
zNTF{6%iwHA%mK{yTX{bGO1p+0(Ae2i)$I
zk)8I;tvnyzex=}3-CUOR?$a$6UNt8pT4DmuX*w!}+b20&yBY9xzb
zY1|%!G^t{)dQiwvqF^ZsBXzw{!H)dgm~qa&|IGAh!^-j}53e|2fRWzs83QZ2(oTT-
z-MKy@LXEp;P7!mWb1!DapL=K!c=&8Z*y?^YB#_>!G?3kci&}I*H@#B{
z>%RH&u7Jpz9eZhTRI4~(K0Mt}Iu>|M6e&2MtmzO?^uyV(6}U#H#2Vy~$LU96!1T2q
z$?%G5&#b&L$b4miZ7HQcWV|du^)^%NsA?$)@mwm1GktVKWsEXp@(J5$*I+L#Je2+6
z?vSlakzz@$D;BW4@86q<f2;@_R|UPksEM
zs6~~gi8xtIk`8e=a85}zc`&))?YY@=Y2}B7PV~l$N-qaI<2wHSZa|bkf`J2fidYK+
zl917B(mU8&2)sbe2mB8
zM=-T-*d{d96jnIM*v5na8)9J^Ft|CPD#;AM4Zi7+SeZKtbeO7%lg_B;%FoN735y2f
zK-Px#*)yMjvJqh6oP_qeM?c$d
z{K0M9o+dm)^ZV*lR{_12L64tK>R|*Y;D7J2KfWzGtrL*M)oElXpN!!YjjMm)Ac_)J{Op+pyKeHKe`X#oD@~5h~{3t
zThvnUg;2r%wN@tUs`x&qt5b0=d6!ThWYEGJsY`^u*Y+RUZr&J}aP*AUYr_(0CzFy*
z%v3IQYzhz7ka(YaYSeIOSmL=kM}Xl=wX9|f;PyR!?tjXv^y_m+5w^qVtq6Hi#Tq0;
z$t|PL@`mt>IP=E-_NT-UN3Y}Vd0ZP)>yPb&Xl=xuwC}@T?9%B)?xl;8@qtdsV@J*#
z_E~jxZ$(bMFIql&@uSAR7hQR2*EtPhi_(%5PQ>POBmD=Q)@#r8GNl05VI79iR^9?*
zw=?vI7PF;E=mOdE+u$Vc8f3hW_HsFimjtJu#0SB?c~hI2Oe+}uwZ`}VyHiokHOp^S
zeNihE=9uS%8Uk>~9av)X>OSO&H3&Wii0IzgHOb
zUgTX|gZztnhjG`1Zw7mY4K^B-*is&k;}sd&YfC%_ua4L6ai^)aS$$KhVzcLN8b?%N2SU%^5T?cd~B5>A;F)
zdjn`1Pe6I0DZTdQ)v}jaQu*XV)1P=dzdG8AJTu~^0!B_
zc1W&4nv0fChsBlV2`v#?MAh-kTDE;8)`)XjRoofCTQ;L65ASNVTRE_
zYx^UEkvP=>8z<6Q{$*pK&nS1YVrZSt|-AeEWkE<<}O#2gz%}`
zUpddRM(*?3u*pEnMEua7jPu#B24E5Y#fl}%{POUn4otrN40{$rw);gRsgD5b!F8tG
zT@k2zDoxyT!R`#vP%7S%yldrRrnVgmOV2GIK|`~`!jL
zW8H%EoSooHY3X|~+1xTNqZiBX-@h03R^wgWbSELRQ03rN+kRgjUA=w3H`iEpP~Ng>
z;ieBX0Jih4fK2t=whR-?n71PL(V}K9J+sHq3cW)tg0(9*Po8OpJ0;|0TsRRF9Oc`d
zN)S6V>&Z=0C3^bTVxE}pqGqJ1QY_0arJQwb8CN;qR2(CYq!f#l)
zB9;wKXd*4#L`6r!&%FP(**%iSpu6_4q)7V(;M>2(_j`*)-?&dw`9hPYqOP}emR+J{
za-cqkP*`#A=g$w>o;^4_6j-1Jxr3|{c)jdz3kvORC@l`3tRqAf@VF~#*7P2MUN184
zGrrRpb=$5{)LM(MF_oa^IS$8Qa&_phr3v&wkWJzAYXW>T{6~a^bW{A
z2hQQ8@+XXm=ow|Q1`%1pd=|%wgKznE3?^yLn^k^o@blw09sy%jr&bF0T^w>;jfF=J
zZb6MWQvnM!U*DozK_YatnnL$&`^~qqn9