From b1e48da7cf34304ffc3ef721d3add254eed01cbb Mon Sep 17 00:00:00 2001
From: lhl616 <145307078+lhl616@users.noreply.github.com>
Date: Sun, 3 Dec 2023 23:01:02 -0500
Subject: [PATCH 01/12] Create week14.md
---
src/content/post/week14.md | 1 +
1 file changed, 1 insertion(+)
create mode 100644 src/content/post/week14.md
diff --git a/src/content/post/week14.md b/src/content/post/week14.md
new file mode 100644
index 0000000..8b13789
--- /dev/null
+++ b/src/content/post/week14.md
@@ -0,0 +1 @@
+
From 373022885fa90b5e5aa401d61abd8c682f1b163c Mon Sep 17 00:00:00 2001
From: lhl616 <145307078+lhl616@users.noreply.github.com>
Date: Sun, 3 Dec 2023 23:10:39 -0500
Subject: [PATCH 02/12] Update week14.md
---
src/content/post/week14.md | 497 +++++++++++++++++++++++++++++++++++++
1 file changed, 497 insertions(+)
diff --git a/src/content/post/week14.md b/src/content/post/week14.md
index 8b13789..409ec68 100644
--- a/src/content/post/week14.md
+++ b/src/content/post/week14.md
@@ -1 +1,498 @@
++++
+date = "03 Dec 2023"
+draft = True
+title = "Week 14: LLM Agents"
+slug = "week14"
++++
+Presenting Team: Aparna Kishore, Elena Long, Erzhen Hu, Jingping Wan
+
+Blogging Team: Haochen Liu, Haolin Liu, Ji Hyun Kim, Stephanie Schoch, Xueren Ge
+
+# Monday, November 27: Ethical AI
+
+
+  |
+
+ |
+
+
+**LLM agents** are the "next big thing", with the potential to directly impact important fields like healthcare and education. Essentially, they are LLM-based systems that have the ability to use external tools, such as Internet browsing access and calculators, to augment their abilities.
+
+
+  |
+
+ |
+
+
+## Toolformer
+
+> Timo Schick, Jane Dwivedi-Yu, Roberto Dessì, Roberta Raileanu, Maria Lomeli, Luke Zettlemoyer, Nicola Cancedda, Thomas Scialom. [_Toolformer: Language Models Can Teach Themselves to Use Tools_](https://arxiv.org/abs/2302.04761). arXiv 2023. [PDF](https://arxiv.org/abs/2302.04761)
+
+LLMs have limitations that can potentially be addressed with these "tools":
+
+
+  |
+
+ |
+
+
+- **Outdated information**: LLMs cannot access up-to-date information without access to external sources. Giving them the ability to access realtime information (via Internet queries) would lead to better responses, such as "who is the President of USA today?"
+- **Hallucination**: External knowledge sources can help ground generation in facts and work to supplement the model's knowledge, reducing the possibility of hallucinating.
+- **Lack of mathematical skills**: Access to a calculator can help model generate correct responses and computations involving math. Using zero-shot learning can help reduce hallucination, but providing access to a calculator (assuming it is used correctly) can guarantee correct responses.
+
+Other limitations include limited multi-language usability, having no concept of “time”, etc.
+
+
+  |
+
+ |
+
+
+### Key Contributions
+
+
+  |
+
+ |
+
+
+The main idea is to develop a system that has the ability to use external tools (translation, calendar, search engine, etc.).
+The key lies in knowing *when* to use a tool, *which* tool to use, and *how* to use it. Training is self-supervised, unlike other capability-enhancing techniques like RLHF.
+
+### Data Collection
+
+
+  |
+
+ |
+
+
+Key step: generating candidate API calls via in-context learning. The method starts with examples generated by humans, e.g. in-context examples for “Coca-Cola”, etc.
+
+
+  |
+
+ |
+
+
+$k$ positions are sampled at random from the text to serve as "candidates" for adding `` tags.
+
+
+  |
+
+ |
+
+
+Tokens up to the position with an “” tag are provided to get $m$ candidate API calls.
+
+
+  |
+
+ |
+
+
+An additional weighted loss term is introduced, corresponding to the utility of information added after using candidate API calls. This loss term is meant to provide feedback for which API calls were useful for some given context.
+
+
+  |
+
+ |
+
+
+Given the loss term and general strategy for inserting `` tokens, the model is fine-tuned with the augmented dataset. At prediction time, the model uses a variant of greedy decoding, making API calls if the `` tag is in the top-k predictions at any token position.
+
+> Professor Evans talked about how the method could benefit from having some “feedback” from the API’s quality of response, and not having an implicit bias in the design that considers API calls as “costly”.
+
+
+  |
+
+ |
+
+
+Interestingly, performance for some cases (ASDiv, Table 4) is better for the version with disabled API calls (so no agent-like behavior) than the variant equipped with API-returned information.
+
+### Scaling-law Experiments
+
+
+  |
+
+ |
+
+
+- For small model sizes, performance does not change much with the inclusion of external knowledge.
+- The utility of API calls is clearer for larger models, where performance drops significantly when API calls are disabled.
+
+In terms of limitations, these tools cannot be used “in chain” (an in iterative-refinement approach, where multiple API calls are made) and require sampling a lot of data.
+
+## ReAct
+
+> Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik Narasimhan, Yuan Cao. [_ReAct: Synergizing Reasoning and Acting in Language Models_](https://openreview.net/forum?id=WE_vluYUL-X). ICLR, 2023. [PDF](https://arxiv.org/abs/2210.03629)
+
+
+  |
+
+ |
+
+
+Research on reasoning and acting has been detached from each other. This work allows LLMs to generate both reasoning traces and actions.
+
+
+  |
+
+ |
+
+
+Learning based on fine-tuning and prompting (ReACT prompting strategy, uses reasoning & action steps together as prompt). The new few slides (below) talk about different parts of ReACT via secific examples, showing how just actions or reasoning in isolation are not sufficient for good agents.
+
+
+  |
+  |
+
+ |
+ |
+
+
+Only when these two are combined together do we get powerful LLM agents:
+
+
+  |
+
+ |
+
+
+
+  |
+
+ |
+
+
+ Reasoning and acting together create an augmented action space, which is key to unlocking these models' capabilities.
+
+## A Survey on Large Language Model based Autonomous Agents
+
+> Lei Wang, Chen Ma, Xueyang Feng, Zeyu Zhang, Hao Yang, Jingsen Zhang, Zhiyuan Chen, Jiakai Tang, Xu Chen, Yankai Lin, Wayne Xin Zhao, Zhewei Wei, Ji-Rong Wen. [_A Survey on Large Language Model based Autonomous Agents_](https://arxiv.org/abs/2308.11432). arXiv, 2023. [PDF](https://arxiv.org/pdf/2308.11432.pdf).
+
+
+  |
+
+ |
+
+
+The survey breaks down the agent construction pipeline into four components/modules: profiling, memory, planning, and action.
+
+### Profiling
+
+
+
+- **Handcrafted**: captures the role of agent properly and allows for flexibility, but labor-intensive.
+- **Using LLMs**: starts with profile generation rules (can specify via few-shot examples), controllable seeding for profiles.
+- **Dataset Alignment Method**: foundation of agent design, and has significant influence on the following 3 modules.
+
+### Memory
+
+
+  |
+
+ |
+
+
+**Structures:** Unified memory is short-term and simulates our "working memory" (added via context), while hybrid combined short-term and long-term memory tries to model human recollection better.
+
+
+  |
+  |
+
+ |
+ |
+
+
+**Formats:** natural language is interpretable and flexible. Embeddings compromise on this flexibility, with the added benefit of being very efficient. Databases allow efficient manipulation of "memories", and structured lists can also be used.
+
+
+  |
+  |
+
+ |
+ |
+
+
+**Operations:** Memory reading allows for weighted retrieval of information, with operations for reading (memory reflection) and updating (memory writing) information.
+
+### Planning
+
+
+  |
+  |
+
+ |
+ |
+
+
+Without feedback, planning may proceed via single reasoning (dependent, connected steps), multi-path reasoning (tree-like structure, kind-of approximates human thinking?), or using external planners (using domain-specific planners).
+
+Similarly, planning with feedback may rely on information from humans (e.g. RLHF), environmental feedback (e.g. RL for game simulation), or model feedback (using other pre-trained models).
+
+### Action
+
+
+  |
+
+ |
+
+
+- Agents can have different targets: task completion, communication (communicate with other agents/humans), or exploration (explore vs. exploit tradeoff).
+- These actions may be produced via memory recollection (using short-term or hybrid memory), or following generated plans.
+- Their exploration space may include API calls, or internal knowledge.
+
+
+  |
+
+ |
+
+
+**Impact:** These agents can directly change the environment (e.g. starting a calculator service), their own states (e.g. recollection), or trigger actions in other agents (e.g. a chatbot agent calling a legal-information agent)
+
+> Subbarao Kambhampati. [_Can LLMs Really Reason and Plan?_](https://cacm.acm.org/blogs/blog-cacm/276268-can-llms-really-reason-and-plan/fulltext). Communications of the ACM Blogpost, 2023.
+
+
+  |
+
+ |
+
+
+The blogpost discussions use Blocksworld as a benchmark. Blocksworld defines rules, goals, and allowed actions etc. via natural language, expecting a set of instructions in response.
+
+
+  |
+  |
+
+ |
+ |
+
+
+Performance seems pretty good with GPT-4 (Left, ~35%) but when names are obfuscated (Right), plan generation results drop to 0-2%.
+
+> Professor Evans talked about how the benchmarks are not a comparison with human performance, which would also understandably go down when names are obfuscated. It is thus unclear whether these drops in performance are expected (given that humans are bad at the modified task as well), or a result of the model not really "knowing" how to solve the given task. An alternate explanation for these results, would just be that the model has a hard time identifying entities that are labeled with non-sensical, multi-token strings that don't revaal them to be blocks. That said, there is tons of data about Blocksworld in the training data, so a difficult domain to test what the model is really learning (if anything).
+
+## In-class Discussion
+
+
+  |
+
+ |
+
+
+- *What are your thoughts on LLM reasonig/planning?* We talked how in psychology, reasoning is divided into 3 domains (knowledge acquisition, reasoning, decision making). Even for the literature in this field, it is unclear how these three domains interact with each other, and thus even more complicated for LLMs.
+- *How should we proceed with this line of research?* We acknowledged how it is difficult to define “planning” for both humans, and even more so for LLMs. Professor Evans mentioned that for this line of work to advance, we need to come up with a good benchmark (but this is very labor-intensive). Students recalled work on performing activities in Minecraft as a useful benchmark for planning and agents. The “granularity” of planning is much more nuanced - humans can infer steps in between (or use domain knowledge), but harder if talking about agents or "just LLMs". At the same time, we do not have a good answer for "should we expect our model to behave more like a hard-coded program or like a human (performance changes due to new factors, ex. Semantic changes, etc)?"
+
+
+# Wednesday, November 15: Applications of LLM Agents
+
+
+  |
+
+ |
+
+
+The experimental setup comprises two restaurants, serving as competitive agents, and fourteen customers, acting as judge agents. To confine the action space of the Large Language Model (LLM), a management system is employed. This system functions as a question provider, formulating precise inquiries for the LLM to ensure that its output remains within the defined action space. The customers exercise their judgment when evaluating offers from both restaurants, ultimately selecting based on their individual constraints and requirements.
+
+
+
+
+
+
+  |
+
+ |
+
+
+
+
+
+  |
+  |
+
+ |
+ |
+
+
+**Data inputs**: Daybook provides data regarding the previous day's patronage, menu evaluation, and related insights. Likewise, Rival offers comparable information concerning the competitor's restaurant, encompassing visitor statistics and menu alterations. Customer feedback is used to make decisions about the next day.
+
+
+  |
+
+ |
+
+
+**Discussion Notes:**
+1. LLM scores can act as a baseline, but there is always a possibility of bias. For instance, changing the order of options presented to the model may sometimes result in a different score being outputted.
+2. Designing a model based solely off of customer/restaurant data fails to capture other experiences of dining (i.e. customer service, environment/ambience, etc.) and thus indicates the simulation’s low fidelity. Capturing decision-making factors in customers is especially difficult, as they are difficult to define and quantify. The current simulation does not account for customers’ risk-aversion for trying new dishes, and it also does not consider the influence of star ratings or reviews on customers’ decisions to choose between the two restaurants. There may also be issues with prompt-based tasks, such as over-generalization.
+3. Utilizing simulations has the potential for real-world social trends and phenomena to be reproduced without requiring a large number of real people or complex variables; it is not necessary to recreate an entire town in order to gain insights into real-world trends.
+
+
+  |
+  |
+
+ |
+ |
+
+
+Agents are able to learn from each other while maintaining differentiation. This is visible in two ways:
+1. Agents **imitate** observed strategies that provide a high reward. For example, a restaurant may copy the popular dishes of another restaurant to compete for their clinetele.
+2. Conversely, **differentiation** is used to attract patrons that the competing agents don't specifically cater to; one restaurant may provide inexpensive food for customers on a budget while another provides fine-dining options.
+
+
+
+
+
+  |
+
+ |
+
+
+The agents are shown to adapt to various customer needs in an effort to retain or attract further patronage.
+
+
+
+
+
+
+
+
+
+
+  |
+
+ |
+
+
+A number of sociological and economic princples were demonstrated in the experiment.
+
+
+
+  |
+
+ |
+
+
+1. Is competition among agents the best mechanism to take advantage of their capabilities? What are the limitations of this approach?
+2. What other interactions are feasible?
+3. What are the benefits and risks and/or pros and cons of these interactions as compared to competition among agents?
+
+**Collaborative Approach Limitations:** One potential drawback of adopting a collaborative approach is the propensity for bias in a single agent to propagate through multiple agents, thus amplifying its impact.
+
+**Employing Negotiation-Based Tasks and Games:** In the context of collaborative endeavors, employing negotiation-based tasks and games is a valuable strategy. These involve the participation of diverse agents, such as a managerial figure, a chef, and a waiter, each representing distinct stakeholders. The amalgamation of their inputs contributes to a holistic decision-making process.
+
+**The Feasibility of Restaurant Collaboration:** We explored the possibility of restaurants engaging in collaborative efforts, including the exchange of information regarding signature dishes, the potential collusion to elevate pricing structures collectively, and the coordination of operational hours. However, it is essential to consider potential drawbacks, particularly the willingness of competitors to engage in such cooperative ventures.
+
+1. Limitations of having collaborative approach: bias in one agent might cascade into bias in multiple agents.
+2. Discussed negotiation-based tasks and negotiation games to collaborate with each other. For instance, one could have an ensemble of different agents (i.e. manager agent makes final decision, chef has a say, waiter has a say, etc.)
+Each agent represents different stakeholder
+3. Discussed how restaurants could collaborate together, e.g. communicate signature dishes, collude to raise prices of everything, coordinate times they are open. Noted potential downsides, including willingess to collaborate and power dynamics between agents.
+
+
+
+  |
+
+ |
+
+
+
+  |
+
+ |
+
+
+This work explored learning through collaboration via multiple types of interaction as shown in the next slide.
+
+
+  |
+
+ |
+
+
+
+
+LTC Communication Patters:
+1. **Monologue:** Letting an agent train by providing feedback to itself. The agent will play the role of the actor and the instructor.
+2. **Dialogue:** As opposed to the previous approach, training is conducted with separate agents acting as the actor and the instructor.
+3. **Analogue:** Similar to the former approach, but raining rewards and examples are provided by the instructor agent rather than by the environment.
+
+
+
+  |
+
+ |
+
+
+The agent model is optimized with the trajectory data collected in the exploration phase. This relies on a multi-objective loss function composed of a standard loss function for unsupervised language model training and a reinforcement objective to maximize the expected reward from previous communication data. Beta acts as a balancing hyper-parameter.
+
+
+  |
+
+ |
+
+
+
+# Readings and Discussion Questions
+
+## Monday 13 November: Introduction to LLM Agents
+### Readings
+- **`Required`**: Timo Schick, Jane Dwivedi-Yu, Roberto Dessì, Roberta Raileanu, Maria Lomeli, Luke Zettlemoyer, Nicola Cancedda, Thomas Scialom. [Toolformer: Language Models Can Teach Themselves to Use Tools](https://arxiv.org/abs/2302.04761). arXiv 2023. [[PDF]](https://arxiv.org/pdf/2302.04761.pdf)
+- **`Required`**: Subbarao Kambhampati. [Can LLMs Really Reason and Plan?](https://cacm.acm.org/blogs/blog-cacm/276268-can-llms-really-reason-and-plan/fulltext). Blog@CACM. 2023.
+- **`Optional`**: Lilian Wang. [LLM Powered Autonomous Agents](https://lilianweng.github.io/posts/2023-06-23-agent/). Blog. 2023.
+- **`Optional`**: Lei Wang, Chen Ma, Xueyang Feng, Zeyu Zhang, Hao Yang, Jingsen Zhang, Zhiyuan Chen, Jiakai Tang, Xu Chen, Yankai Lin, Wayne Xin Zhao, Zhewei Wei, Ji-Rong Wen. [A Survey on Large Language Model based Autonomous Agents](https://arxiv.org/abs/2308.11432). arXiv 2023. [[PDF]](https://arxiv.org/pdf/2308.11432.pdf)
+- **`Optional`**: Karthik Valmeekam, Matthew Marquez, Sarath Sreedharan, Subbarao Kambhampati. [On the Planning Abilities of Large Language Models : A Critical Investigation](https://arxiv.org/abs/2305.15771). NeurIPS 2023. [[PDF]](https://arxiv.org/pdf/2305.15771.pdf)
+- **`Optional`**: Lin Guan, Karthik Valmeekam, Sarath Sreedharan, Subbarao Kambhampati. [Leveraging Pre-trained Large Language Models to Construct and Utilize World Models for Model-based Task Planning](https://arxiv.org/abs/2305.14909). NeurIPS 2023. [[PDF]](https://arxiv.org/pdf/2305.14909.pdf)
+### Questions
+**(Post response by Sunday, 12 November)**
+
+1. What are the key methodologies or techniques used in the Toolformer paper, and how does the tool use of LLM differ from the existing use of LLM, e.g., prompting, demonstration, etc.?
+2. Which potential applications or industries could benefit (or suffer) the most from the LLM Agent concept? How might it revolutionize or impact these areas?
+3. Regarding [Can LLMs Really Reason and Plan?](https://cacm.acm.org/blogs/blog-cacm/276268-can-llms-really-reason-and-plan/fulltext), do you agree with the opinion that what LLMs really do is a form of universal approximate retrieval, which was sometimes mistakenly interpreted as reasoning capabilities? What is your perspective on this question?
+
+## Wednesday 15 November: Applications of LLM Agents
+### Readings
+- **`Required`**: Qinlin Zhao, Jindong Wang, Yixuan Zhang, Yiqiao Jin, Kaijie Zhu, Hao Chen, Xing Xie. [CompeteAI: Understanding the Competition Behaviors in Large Language Model-based Agents](https://arxiv.org/abs/2310.17512). arXiv 2023. [[PDF](https://arxiv.org/pdf/2310.17512.pdf)]
+- **`Optional`**: Yilun Du, Shuang Li, Antonio Torralba, Joshua B. Tenenbaum, Igor Mordatch. [Improving Factuality and Reasoning in Language Models through Multiagent Debate](https://arxiv.org/abs/2305.14325). arXiv 2023. [[PDF](https://arxiv.org/pdf/2305.14325.pdf)]
+- **`Optional`**: Kuan Wang, Yadong Lu, Michael Santacroce, Yeyun Gong, Chao Zhang, Yelong Shen. [Adapting LLM Agents Through Communication](https://arxiv.org/abs/2310.01444). arXiv 2023. [[PDF](https://arxiv.org/pdf/2310.01444.pdf)]
+- **`Optional`**: Daniil A. Boiko, Robert MacKnight, Gabe Gomes. [Emergent autonomous scientific research capabilities of large language models](https://arxiv.org/abs/2304.05332). arXiv 2023. [[PDF](https://arxiv.org/pdf/2304.05332.pdf)]
+- **`Optional`**: Yuzhuang Xu, Shuo Wang, Peng Li, Fuwen Luo, Xiaolong Wang, Weidong Liu, Yang Liu. [Exploring Large Language Models for Communication Games: An Empirical Study on Werewolf](https://arxiv.org/abs/2309.04658). arXiv 2023. [[PDF](https://arxiv.org/pdf/2309.04658.pdf)]
+### Questions
+**(Post response by Tuesday, 14 November)**
+
+1. The [CompeteAI: Understanding the Competition Behaviors in Large Language Model-based Agents](https://arxiv.org/abs/2310.17512) paper shows that LLM agents can be used for simulating the competition environment. How might the competition behaviors observed in LLM-based agents translate to other real-world applications where strategic competition is critical? Essentially, are there specific characteristics unique to the restaurant setting that might not directly apply to other sectors?
+2. What are some considerations (ethical or otherwise) that may arise as a result of programming LLMs to compete with each other, especially considering the possibility of this being implemented in real world scenarios? If there are valid concerns, how could the models be calibrated to ensure that the competition remains ethical, preventing the agents from learning and adopting potentially harmful or deceptive strategies?
+3. Agents can be used in various ways. One way is to make them compete (like in the CompeteAI paper). Instead of competing, how can agents be used in other ways (e.g. by collaborating/communicating with each other), and how might this impact their behavior?
+4. Given the adaptive nature of LLM-based agents in a competitive environment, how can we ensure transparency and interpretability in the decision-making processes of these agents, so that stakeholders can understand and trust the outcomes of such simulations?
From 4002e2cea7816c2bb936d5a43088b1058de0ec06 Mon Sep 17 00:00:00 2001
From: lhl616 <145307078+lhl616@users.noreply.github.com>
Date: Sun, 3 Dec 2023 23:14:14 -0500
Subject: [PATCH 03/12] Create week14
---
src/content/images/week14 | 1 +
1 file changed, 1 insertion(+)
create mode 100644 src/content/images/week14
diff --git a/src/content/images/week14 b/src/content/images/week14
new file mode 100644
index 0000000..8b13789
--- /dev/null
+++ b/src/content/images/week14
@@ -0,0 +1 @@
+
From 7f370d8483f8e74b143e8113eea68c22a0d05074 Mon Sep 17 00:00:00 2001
From: lhl616 <145307078+lhl616@users.noreply.github.com>
Date: Sun, 3 Dec 2023 23:17:05 -0500
Subject: [PATCH 04/12] Delete src/content/images/week14
---
src/content/images/week14 | 1 -
1 file changed, 1 deletion(-)
delete mode 100644 src/content/images/week14
diff --git a/src/content/images/week14 b/src/content/images/week14
deleted file mode 100644
index 8b13789..0000000
--- a/src/content/images/week14
+++ /dev/null
@@ -1 +0,0 @@
-
From 8ab6dc25dcb03c387e883f0870a6bf60d3a9efc6 Mon Sep 17 00:00:00 2001
From: lhl616 <145307078+lhl616@users.noreply.github.com>
Date: Sun, 3 Dec 2023 23:17:24 -0500
Subject: [PATCH 05/12] Create test.md
---
src/content/images/week14/test.md | 1 +
1 file changed, 1 insertion(+)
create mode 100644 src/content/images/week14/test.md
diff --git a/src/content/images/week14/test.md b/src/content/images/week14/test.md
new file mode 100644
index 0000000..8b13789
--- /dev/null
+++ b/src/content/images/week14/test.md
@@ -0,0 +1 @@
+
From b77c2e475078224d12dd6048b0100d9993de60d6 Mon Sep 17 00:00:00 2001
From: lhl616 <145307078+lhl616@users.noreply.github.com>
Date: Sun, 3 Dec 2023 23:18:05 -0500
Subject: [PATCH 06/12] Create test.md
---
src/content/images/week14/day2/test.md | 1 +
1 file changed, 1 insertion(+)
create mode 100644 src/content/images/week14/day2/test.md
diff --git a/src/content/images/week14/day2/test.md b/src/content/images/week14/day2/test.md
new file mode 100644
index 0000000..8b13789
--- /dev/null
+++ b/src/content/images/week14/day2/test.md
@@ -0,0 +1 @@
+
From ce3cd96af15272a4aeadbe1cfb371e2281af599e Mon Sep 17 00:00:00 2001
From: lhl616 <145307078+lhl616@users.noreply.github.com>
Date: Sun, 3 Dec 2023 23:19:18 -0500
Subject: [PATCH 07/12] Add files via upload
---
src/content/images/week14/day2/A.jpg | Bin 0 -> 37963 bytes
src/content/images/week14/day2/B.JPG | Bin 0 -> 43233 bytes
src/content/images/week14/day2/C.JPG | Bin 0 -> 46538 bytes
src/content/images/week14/day2/D.JPG | Bin 0 -> 49003 bytes
src/content/images/week14/day2/E.JPG | Bin 0 -> 77558 bytes
src/content/images/week14/day2/F.JPG | Bin 0 -> 47445 bytes
src/content/images/week14/day2/G.JPG | Bin 0 -> 64355 bytes
src/content/images/week14/day2/H.JPG | Bin 0 -> 70581 bytes
src/content/images/week14/day2/I.JPG | Bin 0 -> 89438 bytes
src/content/images/week14/day2/J.JPG | Bin 0 -> 44617 bytes
src/content/images/week14/day2/K.JPG | Bin 0 -> 90196 bytes
src/content/images/week14/day2/L.JPG | Bin 0 -> 88754 bytes
src/content/images/week14/day2/M.JPG | Bin 0 -> 62051 bytes
src/content/images/week14/day2/N.JPG | Bin 0 -> 49248 bytes
src/content/images/week14/day2/O.JPG | Bin 0 -> 62500 bytes
src/content/images/week14/day2/P.JPG | Bin 0 -> 61773 bytes
src/content/images/week14/day2/Q.JPG | Bin 0 -> 59360 bytes
src/content/images/week14/day2/R.JPG | Bin 0 -> 57838 bytes
src/content/images/week14/day2/S.JPG | Bin 0 -> 84263 bytes
src/content/images/week14/day2/T.JPG | Bin 0 -> 79363 bytes
src/content/images/week14/day2/U.JPG | Bin 0 -> 69286 bytes
src/content/images/week14/day2/V.JPG | Bin 0 -> 86322 bytes
src/content/images/week14/day2/W.JPG | Bin 0 -> 137557 bytes
.../images/week14/day2/Wednesday blog.zip | Bin 0 -> 2008889 bytes
src/content/images/week14/day2/X.JPG | Bin 0 -> 107989 bytes
src/content/images/week14/day2/XXX.JPG | Bin 0 -> 37966 bytes
src/content/images/week14/day2/XXXX.JPG | Bin 0 -> 65382 bytes
src/content/images/week14/day2/XXXXX.JPG | Bin 0 -> 81051 bytes
src/content/images/week14/day2/XXXXXX.JPG | Bin 0 -> 88972 bytes
src/content/images/week14/day2/XXXXXXX.JPG | Bin 0 -> 108108 bytes
src/content/images/week14/day2/XXXXXXXX.JPG | Bin 0 -> 77493 bytes
src/content/images/week14/day2/XXXXXXXXX.JPG | Bin 0 -> 94340 bytes
src/content/images/week14/day2/XXXXXXXXXX.JPG | Bin 0 -> 76312 bytes
src/content/images/week14/day2/Y.JPG | Bin 0 -> 42034 bytes
src/content/images/week14/day2/YY.JPG | Bin 0 -> 98356 bytes
src/content/images/week14/day2/YYY.JPG | Bin 0 -> 82531 bytes
src/content/images/week14/day2/YYYY.JPG | Bin 0 -> 85390 bytes
src/content/images/week14/day2/Z.JPG | Bin 0 -> 80020 bytes
src/content/images/week14/day2/code.html | 182 ++++++++++++++++++
39 files changed, 182 insertions(+)
create mode 100644 src/content/images/week14/day2/A.jpg
create mode 100644 src/content/images/week14/day2/B.JPG
create mode 100644 src/content/images/week14/day2/C.JPG
create mode 100644 src/content/images/week14/day2/D.JPG
create mode 100644 src/content/images/week14/day2/E.JPG
create mode 100644 src/content/images/week14/day2/F.JPG
create mode 100644 src/content/images/week14/day2/G.JPG
create mode 100644 src/content/images/week14/day2/H.JPG
create mode 100644 src/content/images/week14/day2/I.JPG
create mode 100644 src/content/images/week14/day2/J.JPG
create mode 100644 src/content/images/week14/day2/K.JPG
create mode 100644 src/content/images/week14/day2/L.JPG
create mode 100644 src/content/images/week14/day2/M.JPG
create mode 100644 src/content/images/week14/day2/N.JPG
create mode 100644 src/content/images/week14/day2/O.JPG
create mode 100644 src/content/images/week14/day2/P.JPG
create mode 100644 src/content/images/week14/day2/Q.JPG
create mode 100644 src/content/images/week14/day2/R.JPG
create mode 100644 src/content/images/week14/day2/S.JPG
create mode 100644 src/content/images/week14/day2/T.JPG
create mode 100644 src/content/images/week14/day2/U.JPG
create mode 100644 src/content/images/week14/day2/V.JPG
create mode 100644 src/content/images/week14/day2/W.JPG
create mode 100644 src/content/images/week14/day2/Wednesday blog.zip
create mode 100644 src/content/images/week14/day2/X.JPG
create mode 100644 src/content/images/week14/day2/XXX.JPG
create mode 100644 src/content/images/week14/day2/XXXX.JPG
create mode 100644 src/content/images/week14/day2/XXXXX.JPG
create mode 100644 src/content/images/week14/day2/XXXXXX.JPG
create mode 100644 src/content/images/week14/day2/XXXXXXX.JPG
create mode 100644 src/content/images/week14/day2/XXXXXXXX.JPG
create mode 100644 src/content/images/week14/day2/XXXXXXXXX.JPG
create mode 100644 src/content/images/week14/day2/XXXXXXXXXX.JPG
create mode 100644 src/content/images/week14/day2/Y.JPG
create mode 100644 src/content/images/week14/day2/YY.JPG
create mode 100644 src/content/images/week14/day2/YYY.JPG
create mode 100644 src/content/images/week14/day2/YYYY.JPG
create mode 100644 src/content/images/week14/day2/Z.JPG
create mode 100644 src/content/images/week14/day2/code.html
diff --git a/src/content/images/week14/day2/A.jpg b/src/content/images/week14/day2/A.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..aed5a753581be220c6193d00ec1e586fbcaea3c6
GIT binary patch
literal 37963
zcmeFZ2Ut_g
zdX?TwXbClh5PsWp?wxzi%-m`Jng2X9&xB_~cFD^6_V?{~t@W<8I2%8k16@{CQc(h(
zI}ZY#1Aahflc4*cix)1CULd(hN=iybc9HxNE#)N&ic3sau2R#oF@xFJm|0n`aS8KW
z;}GOzW#yIN6TEp#OiT>SBPAz!TUJ<9?Dj7q=g7#&E>T=!q@-lLeVz6C?SJ>{>!zy-+5p!4U5h|Uue
zk&qA*16TV4=Rw3YBv-EAQn+wc>nSORGwtoyG3ggM@0EY0(;h%`iCVh6AtR?}U}R#x
z!Og?V$1f&+M?z9c`u+n&CFO@Isye!Q`UZwZ##Ya)pWD2!wR3fI_we-c_6d3$91>md9
zw{cB?E)ksrCXa{)1Od(GWxPv1HARzTiZj`%A5gvU;Ob$@I~r(;cBB7Kd^dUq3i0DC
z#1^btzh!x%6E6I0K@9JUc&eq{z}MdE*hQ((^^@^ilwbs_UGL_rf=n*NU4da)PIQ8A
zv+z-yOz}Lb3g*lSZGgV@b;lX_&ugBum8Xk(7A`*og(su5`tZcii>Ll52t8h5P?kT6
z0y2TTv_wS`AFrQOR
z_KpVF_V&WzidtdqQnELdxe8-BB)p(?+3=&=Y118b;c4icNSjV6xmrT3v=bfn45W{_
zFD5u^J+#+eq}~c`f`$>&M!JTdD@+7;`wOh+4Q4Cxr`~wxtg0Q?5H6|+vP|wJtt~Ef
z_{OGWce5Q%i>MzNmMVZiu1kHq4*t~#2CE*Noc#sjF4|}_uFMk7}w=jN|bcYRW+@bMm*m}KVBUl
z`?0WVH{6Ia-^B;)>`JE_$?bDyd{7}J4N^j3Qim$;fE7jES@O>9>quzT1^yTB@OFJ7
z)lwHE;YNY{X?un1)QJ;o>4?ws~;pvaZlqnM6!Wr}kNpi}`h;HY!
zP6a>O6kHnrW6v1tgMWuC^6XD%>exIUBG55JLRJT7&p_wt;aFYsn`a<|8Wexg7$*$d
zI|Efan_GC|uR0YB$-A9_f)y93kN2wq013|g;|H4}GT>W33=*f$K*aKekN#z;km_5%
zWM0M_+^@f#uYT3)RLf1xuF`nvsE9WP<3-P?Z_eIHfXm-@$ZFaD%RX~tUSi7Vam
z5%eqdf3ucs=-jWlM!{p@*4hOwRqy0Ip-nSst$U?rRmkekRnZfpjPq}9AvvS&6~LeH
z8w>QcEecp#`@om`CukQ*wm(85!JX1?8WEvI8I{$t;zLu{;+vi~RL!F9zuRoC4?6>uZN7~INTScdx_+nCSb0`4
zd)0v78R*;Udt2?J?AmEBL1wvL_)BqTym$fyl#RVTb5If)n
z`DuCN#!&0r=dcsizGNu8%k4mF3bnwNnISu-h*l2sQ+Vmpe+?sF
zxM1H5(W(*rMA)WXkR$}*4vNc=80=5i%5+T(bj&|IpIkS-`Z;qeGJ1s)t`VP`wN@8h
z>m?-Cts;B_bq`JJ6YE8^VV3wdfF8v>mgrruQ(i6`%z5RT-0PFfDS{qquy4K(wDt+i
z|9#4T*!)?1a|+m}$lOUI=upzS57J7Y&Y+W(No#x4lYqW1F57u
zYV@ZOXelH@e=>tdkk_P*cWMPfP6U0u`3G(*&1~G&+0uAnDOvhYaGz=jj
z%w#)F2wTey(r@TXrEY;R#>Lq>oq=9cDC7A>7@>XH%>-g0e+EVGGtgvBsYpG6ta`*2
zAq{QGz=Yy)CTF0zdjn@6@`w)b-uiE!vrrsRK3oy#-aZ4p0qf=Xt^eulKavWb{P0WW
zMSuOIz1%;1AsC(4@Har+H=~6*S~fK8V6Pt-+w>y-c(U@3nv4y&X`R->c2An;LjO<8
zzf9i$f_tfs4r=ecIqK0#n4wR13NpV*U@>t$s5QoEz?RyqR>QChr85%*>yEcp74o)#
zrIJ{L9Pd>DCNn;hKsy_G23qX}yy5fuO$iw_$3Fvlz+z{hmJ|#Ee>@4DQ~kAmojRa<
zs}8`r7qC#%QkJ294o`;8t7H2Jep}7x8^5gI-(;3%=VhjbRw(yC&^dS;!0da>Vy^(}
z@E-(@4GwGdw*JgqL11+PL;>S2eJ1EJxx4cGt(KUFSVLLTnzq$
zIOIN(*0t>L#c}SR1_1w~K%6m5YTQ@MZ|G0@EB{fXzmOv9rd;yBc2lA~*#B4G6FqT}
zV8V@o&9pr&69IcS)UFvXo;3fEb*vm)u@Ie_q3U+Z2SJK!dkAW|gv~@M>kz)Jt{WOV
zwVrx*Fmx4pt?cT-i}Yzyv#9S!U-qT`aCB{i8yCA!r5v-yPCF{A?9#uk@8qJhQzN4l
zQ1ps)sMswE?b{KWomvJy*hTK^g&=IQe|FkDbzp03u%ckK4}{V9Y#_nIW5U14qw1A1F-;_wbDkd28o%YeH!2`16|vg!
z5%J;U!auFZiUUL_`!)1wO#O&_y94!I<8>9QwNEaZgN@KOzzY`IDrx@gcZ>9}DsR=Y
z@=2xL%=mKO5Zos-11~x(-n~59;O={4q}8QL70JWele)|CQt@UpiBBZjQ7*Oyz#RX7WxX<3OnXR
zGjcL{b3G}@oElQ8envdfgU#%i&oibYyqZwXe-peQ;MLXVum
zhhc>Kh-{C(+s!A!Mij^U7Gs{+qyq5nZoPO>S63@?K0T9
zA@t0LG@?ns$Tq+_E&TH#`qlFdTARs}&h^e_JIv(d)~}gMbr&2Ss^4vzc5sA$4-0i^+p}
zWl|5k`AX~%@Sgqs8%ibDXN)=D0QX2auDcP##M`Lf#S%w2p+YDeN5^k?jFuUh+-NMi
z)q2T_jMa`#*Y}+EfQ4ver$Kvna^bO?vb(faPVr@6F-P|EX23xIEn7}~(TmU>3M)U1N{&d)u+|>PKLb63(#!ey+h#l8Nmz9#^@>wPkA&Ba2oKUh
zoTVE=2MmBqq)Klm8eyPq$7y4V;rpqiGS6yV6r88%*`U+ra_Y7(AQZXHP0V9arQpX+
z(3H)|-H6RL_s%0F($Pj~-pn_&A&i5P?;La$LA?7Qx7-ZwUXzb~Z}CJKSk-Hz1Kvkj
zPq&wuTu}@snd1RmV>*u#4Clg^lOIHju_R5+@g8vPm);&D@b>xpo<__SDi4bwmQOT(
zWSr#fz;{@wi-iogGookP%=rgpg=`L^wAZe49=$5PtwKPZfy~nmR^a}NV3>>89o`8@
zX!yinl`HR&oWM8gqf6{Sa`
zM5p`Y+!~2Yc3IJKZ7O4;PJZXL!uqz=Xi()hr~m
zYtQR5MbAKzRpf()rsfPjI%44M1}&>f=tcI16)hYSL=kWA+G;^QAF^>`(k?-&H$1h&
z$4MWh*%WY$!)%$qE1YUio;)Tr?X_>z}|5<0asG1r7}9Kux_w-Q6hEjwhS
zYgVMYk^LkC#tUFAFR1jR{GBsU*FEsTDtONhC*S0R1)Rbg6ojb-f`cC0x3o^g4OM=w
zd_}t_J)!`Z53hWROD`Qnz4VTeDGh?`LZFjbh#vbjyW`clK0JK(aGht*7{}a1Eqw;^
zIp!Q`^*g6XYY^*3LaTN6?{0MFINX?eW$gChg5#UeE5--56GzpoSdncHKr=3WZ
z>Ys*#PY3Xvv!#ReuZ;Zto)xRxj_fPx2`D8WMW3_>$PM}lM$y6#Q}DlY&5OTr%~Tkk
z2mlC7j}!M6=%k`≪9a*VH>lTml=Ulw@)MoMl9`lRxRd;8q?gv2h_cnkDWqu7HAv_J&
zI{^}+;T`ldwUaAutW=pgRF)_sRyIYz`6#tER-f?Fzd8O1w)&T6<(-gh4#@v{jeru}
zR@T{s+%gv_u)`t3kl;!SO`(z2)V!u70-e^-ngH)SCMk2ZO*3H8U?San-C$b81zSe`f+JIPQ(S)4Jw5RvUn~D=#lg%wg
zn`8JR3w#e!W0YH(sVm{+9l|RD?tYoBrr$w^rylTm+(D5DgP&2==RLs&$mf;COtYIm
zY=MZHb7yG7lp2QjFZ4>E;DLS~w^^XLyO3-jfSkC(
z6=1d#_RRxT^>B*%Xxtq~{bRpmqc!02pr44t=Z;IgtyfG{e!#JZ?5(hwl1l;pQl-|x
zM^>|?-Q~rC9OnB6R@XxWytElPA1){$wP-<^{pU0@V8^4|9)EYe{!455KjoY*gJ;4h
z{bw_CP1egww|5_Y)l(ksCqC%)%IyH)r+Cj_{v?MfdQ|wCQ)}4~q}bWGxh=}1t-2@E
zi~hTcDf+1Q!-U*jS>7Y3FZ!jyP4q^#)!+#q{OrU;Vd|#!jZCG3)dcy~T{``$C0C1i
zsig`4YgFh>#vH({S<6;TKQhN(S)g7uy=fF^ZKC}RSp9DgbJCo2r6h5kfxMe@HSH$W
z^-2#Y)P;N1>Nh@YZ7nj{a=kov1{yoen%-xI87J5$Vr<}U7Gw|fQ5N$1k=a=f806mq
z6YPs<74NWvUUu(en3adm*-Jg}=W335rRx)k4w~Ba1=)MHKq)G>=(X}tgHSuWBG1at
z_-|3Db13KpBwNPB*;S&vM==xIyzn;PwdnhV(i@Gnt^Ui;*ur$iPFEKD4RY4lnFswu
zp9#6uD+X)MzT4de>pSKi`{*NehmmS*>|HzJEO}eSLxta*vD7-7Qd!d-841~SOF)i;
zS_z#jrrmW|d><^7#j^t50ndX=%S+_XK%M5CnD(V)gEz%Ds*$8G>)s;xh{ohD2fXyP
ztkb-Rzctt(z2WZzt#sG+WLja;bvjho+4jciAgk#n$Cr!oQVX-;24ubqDzj<>6?Fm+
zJ4T<52SadI1BmEQ%-pZDu^w$k>mK367H!Ict0Qo2dsUeFl-(V2-i#OpMcT&zfd0N&
zNZ^5BJT~}uce^x8VC4)H)x7(-d8=1sWkhQ#eA9bZHzPy!y?)N6A#vhVN)043%CTZou`1LSg3eudg0t&=}~E{oTsnQLn}Zqz}7Dyw{usF)_ga&s41luj7Y1Gl!p+`gEfs;TXDJJB29g{CJlWb8$yE>P(2e(
zi6k++os(8<8FkQ6RLvibdmop2vSm|wTrUnt8RF7VdPtxZYXn1HW;
zJ~l{auOvHpch?H3Byn|(7H}TTIOZdg(9C3xI@IB23n)&?`r~`nLC-C(A;XkZVVf8TnOoyN#dY%gx5`
z7KmelqHzMaSJ`)AC&IaX#mMUw8938i8lQRk<@bbNIp;5un_O6=H1}m%$)-O*($Ecf
zY0j?{E?6c~QKo(LnR?hza|U|iiQ6xIg$4+s{nASl-puYb`kY5`klQK;kbY+a9fmjG
zGu9uQ4sD!)NC*+=8#;wyenSJS5CF_4C%iqIu1`8c^eS~ejGZpxl_zsY8>MJCqqN0!
z6&dxk0!csRYFXTZb?CQirWH0XHB#Z+7S)Y;0T+8&@V~{+{^j2nOoXI6jPDB5~l4~_MgNCM8xf=3xwCRDMvJMt+O*s
z+cTYbaiUsX^+de=DjOZVmB7{j}Tg<3!IY
zqpKzp!H&bYl>EeP#33pA)aPW(Mq~Y@ddgI0*4l(euw|98b(TlJ2>R19JYP=bNZcnM
z<&#|!V|nBm-OYUXK=_CM`Dm3=p{6lEYs`gRbs2@v2<5@olS0FDnSfzA;on0!s|Z9@
zMklt`3E{1_O}PJ
zP;yKkcdP-6NB>%b5k^^K;D&U-d&ER?AzK8kX@*25&F~rMvlLFBJ_T5nW|-hzdv>XY
zc&iP%w>Gv!{flqeTUX@1iU0J{{30##HiOch0pX*>zWta2@wj=Ky%ob9vh3u7
z{0nPk^=H-!#;*pzb>KhQj++;q@b5)VTqBN={@4Eu;04i3npSqQQGmtl*C*^Ui
z+5OI!xc`6VOVSJNr2~SWjrqh6DqLKz+3oD7$~qNj$D3i)9*;J3s|)x?h0m2qCMawx
z1H;=Zy-$=d)}c0~5BIFbUu8xvs@vDj0=}ww(){#cm9G<$=4%wo5
z>#jxX9bG(7D6pXeKdK4+$;$9HfOf_%z-*-p0}2{z{S5G&0iWehj~VZ5vI{1;-mMg2<4y#~^9pTL6^(bNndn(ZU~
z%GR1odr%VEsbG(^(%M0<>ILD#1?|Iut2VjpvSGlZwN(wx)hvGe_t=E`*^78CL%^q|
zdABw4GNuN+Q?{`WPFgo}3YWFd@RA-j4vaO;SP&5Kqs**wZh`NT=3V^%Gt+D{5iU6bq7lTcvDJ6bvgTpT
zK?ajMKX}V1@;H078V^=~J^CJSs1=P^8?)1L7brEqS4aC!{)}F5F@on?LX%3L0JKMh
zoG3sI-7zi9NVf;=m_6pT
zI9z4ev&-9la437sucdWERY2#0ZSot9klbJHRzRYT)sOILKOlDzb3&NrBG_BpH|T?0
zfG!09a#SG6WBns&@3E7yF7^_ZIRGmf;y(S&V#S1$Kh$oc3d;XPQbN=|OY63E^+)C4
zp@l^>8R?N0!bKYK^ca@Au;CQOetl0!ut>X8LzdPgYwwrKA@^&w=ry-E(WyyuRiW=i
zz8Vxm-GiH+SNFd_bS}uEpT&J62Nrp?oysliVa1XNtt5hSMMqml~>8TxgCW_$mnKH&MQR+p0lasS{E
z%b@nqQ2BfTVWk%&0Wi$!Ue$jM6rKJYC|cBE?H5fpl7tFJ1;;e5zs3
zj5swb3|Cj`;=Lf-a3BtC^wo#Y&r)Mq9(fOUQjMG*EPrK5!EDNf?iKsrGbv<~JOe$q
zT|im0SiLs0(pr!8q0aVZ2=n8?P_!Kdcw`b=)2>KfnJ0EWy#Uqeap`8z5sfbx)?GiW
z1LAeaH>CPw^G(MtAhm7hTY9^@b@Eearde5Q^+_gHxpsBk+!rX`LWH@4D}E{njr<68
zWv}Vee=aAVQ=dvPf1%jbJJoiucf^}Rf@){;))yS&dm~hNKy2$mmxI`(?wjjlY=*dU
zl;&t1>bazB9irlEHlODiNMdXt;DU3ecEw75hU5<591rbV2=D3t^!NKiy;ouXi;O0#
ztFXdwgq(qg=e>_lifIAY-!%dF1^S=1{psn%-TQChu>Mox{(pfZ`>&SoH;btzs4|?^
zw&N$7-u)T>^pstA%2&IzVd4B*#veo<4PS8+mjc`s`~A8U_GcO!e`5Z;{~T!s^)}Bm
zJv?|d3&fLjOR4`5%>Ff~{onNW&;H$iLxBFz1?~Uyjv4fvfmonR>CnlK7CnXQ++q0eoFNaWg
zHd39Jl!(*GGxgfpo+-wJwA=z%5Bug%x9pj
zH&9X2@#gvD0c^DX&fT&R9ntnG^TEzm=)oC?c^L^IwW9}hIm(Y*+zwg+>z#ovjd?ru
z>+P?b`wN<>YIIF|Nb`gSs~^d}{|5_ySeF78_|IXF|5-zHu4@X*ERd3KBb_~dTk0jd
zbDZfaOjcfu!q?lfDNRQp{(4OZiI2at1H4FD!7Uf76jo(2#8-T)wWz*S+gYHIVb(zgUey)-Z
z(>y66Cvg6{{=7!=YB{+(zr0$iB3&m%AMHly7V|4uOBZ}Q`LTpF?qd+XS$c9*L|jdP
z_X6U%@V4pJ^R)u!VK|}DI4{}pN1}R8e2<+#%n3Lu+d>R=$C*7tC=>2f2Te6X&xQ$?
zmTbHAsNRUi&2q)h-6SQ&EV*qYs}?jSzhL9B&q*2_Fzt*)7WOc+a970W{CKzXJ?P!7
zPKa^DmO1y>xtgI%^z_Q7I*)GeQk#b|;UUayx><`33fc~Q7Wa!XYPu<*21D*^m8qWH
zalO}+ZwE@r5_OzoTFkM}8A0;yr{1>zsaqELDia$nNwk;C9An$m5`QPzth>^H>>o_y=6IXQSJ80?)TLXEHP;C=A{Rye`ClBi*0&mIyRZwqY3j-n@_`awV@gQkeIuMU)K!LkWO@o(WY5K)jA)j7b=#z#=Gr!w#*|sD
zr#7#SR=HpCP+j3qvD+F)trbD2?fVMCw@T(OTTHTj?vMsM8>MkzgnFYEObmfgDg9Vx
zDU$a_m!R~)Ca=3?W=}sIpOnjiG)^E6DK9+im|pw*O-Fx?6$Qw!#K(
zU*w(6ecEv)vM9NBXuv3dqY
z^^))z=3UD`Wq++XW(!N
z)91j}a>vSzs>uIhuWd_4oc(jYA!n6+u8f*o3{^#`-R%jizZGnTGon+;0fqpv5;MWv#|O5t@TFt2Cc&O_f^WkO7@0iY^4I(H-%a0LLW*n!01J
z$2Y0G7T3)bO}TQ}pFLQZJ2upMJgvq>AE!pkZA(KT_++k@DBGp`3`CZeCEM<6CM`O^
zc;5N=7C|>#-b~23gLPu2(?ZcX#@=T%v~cpLeGcDTjf}R}7sgelIjSV4g*r3K6*VE|
z-IXiENC8N9f(kn=@Hkvl#PJYqDQ2xy#`QYU8_s0dw=>?%eQtmPZ
zZw`_Kh1patmER)vkQA3Q7=|pdxSxScEsmMjd>MuMAn49etFis#k`fWem{H3QN*iCK
zRyR!4=OlHXv+3_jY^snRTjqO#4Tknl*}Y1(7lJ3|$_!p!xl;bL*jms1&pTEFC*$OdmAf%~;VP5meT3L}&-p;m)+jQ~%+Huo&OmTrydP8?G
zd7^i9b2ow1;~_CsTQu}kuV;cxR(+8XY53G_XbYql(;*QPdkc&BUPpZu#lZahwcJP{
zNs?-?rD$Xs8>K0@R$vQ9BfiryNyu*G81!Z1pP5*3l^rNyeiWCcS(>ZaQZhn$W!q$n
zz87VA>9Z2WmaQpA*=OCk_{-teUHRo^sau&G6D$Ca!1P-bCy*t7)LTSZd};kqPGQ-lb^uTF}+4cbUh7
z?s}VXR=TMr)36m~Al;Qy*jRP!>zsCLc@>NhuUIhSI59wm$guQPe3oA!frWsz)=is|
z@+dVIxFpPV9{dP4zk@2`yaxYapt#6pzN0)ipWHLT<^6`9pPG@UdGK>Y#M_jKmCxEb
zFq84AO4viexe%jApSkRcucO*7UBwQMXqA}Sa<83h{x#$L%@*JID(Z0%d0Z$(L|Ztn1wuoIvRYp+i07Pco)B%D44m#9m>Q_07V
z+Db?Ilq8XCTD^#OF@U&5U&Oc5gd3E0w;&l@xL>
zW{hd>9Lm-M+o_>hsU}yF&b*7}ZKauI)FSw59~R)aJ2z92wrmb`-@H8As*U@?9n)tC
z!Aps-+kte~@;1h5OPUI<>KHuGprZN#u%Vyt{j0^Fe<%)Z$Xp=BM?nw2LXt`iN_nxb
zfh4!ZX_@AVeF_6!)EGh&^P7L8D}*C&N)_mcnS$T=TVi}Lr?a!|zb<5?Yxe_|1e^kl
zL(P7#H(qGsq~8C1r@ReHnev#xA;Sn+oPiuE@0Rc^@Q3*GY2F*8K0ROW<@M&bg8~be
zxy?9{yVOt_v!-3EC78Fu{=Igwj5LwX4R1t9j=#J11aVQ_5^&
zYJ&3dK(Z>maDW}dWS&-EI$=*LEaYLFzN*TvT_I9ZyTB62OidZDGt^T<4&{V(4p`SE
zF5GCHF^TXE3EK%bylD~d3;!}YfawTxXwCZEkmYZ;$B%pgNC*1NY6=*2xrD2sFHiz4YAART<(c>Q2n+BZohKU%B)n^H_A
zHXU<7p~8!g?$FaD`t+%!mQee=p=Mok>0nk1
zzwWdx`$%)`Vv)THWS#@lCLRBlDj@+y97or&d+t=@E{;52UjMx{^m>ge_cDujavD$?
z67v!|=0;Gea)TT*?9bv3B7mk@x~RpSV|u_HuY(%@?8o+t{q=YTL16znVBdT`>Uv7F
zcH$xrt2hq5TKUu8agVSVNXbtHi1$P7v2lc8Adm7TyXe$oaPL0m3O1zosu45+!o4`B6kZY;@L7x
z#zN5LV$qN{kWaPqty0SgxkOkihFPd@`J8f*25NLpOIm>Lj%3v0lTm>5a|@Hyr@$L
zc$eP-TGxzY9`a*^rO{RJ-!$UH8MJ~}A`{4|&>7}&{%pA-W-$&lrv?gQ#$
z*S%Bn85t1P@}9EFV|ee!=yf#u_SG6&JXkX}WwG>_{>B-oBpi<-gpzB{+w(p%%}OXJ
zl2b5NH!n25uQ5EdCn`(CYCqrd{laQe1R?;ihr1kIES!N_uP;$Mo05F2gTE2sd1G%R
z?DmtzKEm*{~8=I=5wu^~;
zwngYQlHrz*ff!)z&gxB#QKQ#AD79QaN&^fPu&)OUE+o7SO7qTqMj`iSx4X*;N2GO>
zE6~uOHLTl}y(X$o_3CriwY$*V4P$oxp_SnVh%-rz5N~S%eBo=a07EN5Xh?8B?B!1C
zR&KS0V187z^FFxu{mSsM$!FGUABE$8#w~0#8qXP~DyysAESWJ8Og~~5?Z-^i{|cqP
zRp$Zoe1V|X>Q$ncq5nIwu^
zQwl!!q7v7Z4w7wzc}Lok_>=8C81HXKeI5{8N`2}EUDcPh>I)WW3I}GZ
z*qA{2Effz;y6~E_b51hro@aJLaKNW3et2u%yB9~=fwCPn6Pjg@mG{I8=f$Y!1-P8!
zpFL67reI!i+o<)UDKs^jRpcDkn9Ph<@^RxtqxeO-yLgZh)e^G$M{>l03dGOYKaEk&
zNsbB`7)HsXZffmc8p(ED6fg-R^krsMdwz~8&x;Jq9bPRfY7WK7N{1A)h6)Gk!E1yx
zbRr9sW}bCWi7i7p8wzlt_#nsT4PX^>>hm|Ec{FQ5&pxCOp~d?TZ<>76eB)*&qdmxz
zusgsVW5dr>@ns?x4D__j3z2nDr843~sh8|y(eUufQwFu~FP2?s`$?p53m(&*HukZdAYqZCNzxuZao&r0Atye&>qSAB%=8
z>D2hLoiQPF+9yAdrH-Wamz4eox-o`mulb%QMh*GoBCycIKhIGmxyJ>kNq4++G&S
z43)?|1WKRH$WBBFb*ZUeC>?k?ucedjmO$pa&OqVTb6cy_%Vux!UwKY)j}5@+Rf0k_
z;$R%g4=6|vx&0K+=NIO8o0uRy%AJ^W`5S{=lOOatY`gI7NMU0z7WL4+6dY?&Z6mu!g8jC94A%h0w2;MCBUGf=;A
zrjJg*^8}=!2q#dav}Edp+XsxO00QgI9L0{hEC2l0lGB2(*r&gO&a*v@4*2GE1`5=t
zZf!n3X9^tyhTX4^gto!cV+pK{>diDDKdHPokmc#Vb)z(59s5`o2ME6j!qakLqAacAet|^!fY)E_Ot`B43_ztVssyYp*Hh
z=lcd3R%gF^Bic956UTu(jz{b3_IAijdVmBe){$$E+tA{N^K=9sB)?RlpcOAN5@)e+
z43OO>AO2mntbgQ_%uCI+Fhm5U=94FgA7gh!u+J+4I5}oN6|N8Fne5Z~OB+OS*0QVq
zsx@0^-qQQ3!)i$h{KJ|n8lXYbl?
zY6-rW*O;$W_1=6c1TwFEzOMZEdn<|IGn#uG9j|Hb6-1I~1(Ha^KrdrD
z#hy!PnXwd-XEZ!Fi|i!*K7Egt4&wVBPw+vjA!Y7~|DdBeNDYdqp_7^U~TLdISB
zMj70O0a!|X-U7R&TuX^Jd~|vXO1jtbtclM{(`Z^>$~HWI@hfSg5G`nYF`BW>nIl6)
zQ7aHa&PM-$7sX`wr0`OZMn)hFjn&<-bUmf>uLDW$7-lAst1&+eEvWgHJ8<{%S6c_W
zi`CUI)`Q)CVJN@PecDG`tS=&01@-O<`fC=;!vq{5
zd(?T6y8GIn*~WFut~}u5DC9#G-KLa=S<>k82R{Z8fvASZ=V2Tjk^>Z}J;=LL{_65D
z5MRuFT7G@rL2(`~jzBVC_1wL9*P8#!Wv^?)G=@fBd6FozXqt~=Z&I)Dx6>L5i*`ly
z?ln~IIadb^)UzjLJj_h;_zv6nI3J?;7hzT46_p)wz{1
znoCzfyf7UY41d83+%rcuu?qR3;SxV4?P1M
z;&NZ5c-Ft1uV>~hcq_$eX(R9XDg(yzfac4Ct32AVOg=*5pg~?s5I$g8)#E%Ii
z_&GkDAK9B}Z6oMxI(?ru*1z2ol$>mvS^Y*O`$>66b`t!FK!%pyzztz8?i5F0>t&?V9RVvPa0;o
zmNfLr!hk7s#m7adgTs=h
zFsb@I0pel0TC)}yZ{fA#(@g8b+49o@a3xUXE41jjtZ&~b{xd76U7|v1`C0UGlA${>
zyTI_inTDK|Y6#{gQwWiu2;LXfMXfS?4r9d+@B}B_QpsArp12TsD&jwdlfvIzLTuS(u#T#eiA=shR@QLkVZ!E_af
zv;6S;yBHw)fF0O^dSaD7xmdj&Vc!={(T
zgs&SDUnsWQHB%BIRkf_X@-*bFWp4EfN!yQ
zdYgo7`Yhy@umW9Qdf;v77JqnO!pzH>`^kid+rRjiiz_uDKZ*LpocSb4#(Db|H3ic>
zubf|$>SHX5g0`YJ&@Hh&TuR(rbWFWI7=6blCT8dD-MF&!i(rO(fi){RyY)l&ii(iK
z6%L+Gz6>c2vaZS>9iv_|<*GYVekM(Ov7qf-`%A8ADd22@rl4a(&@};L+!-izerzs8
z>bY0sBg5Zh>oC>cPZCmnPiymiCZy(CWYD`8dBqNcgM%zDy7;aL+L`2otddt{`HIa^
zi4$3#E~F2iu^W_?G(>B^h030bC45Oc|syuo;r6rEN3
z+R?$f+U1^VFk?{An_EW7q@n5k9H|^>Dvy4U3EqlyTg2o{)_v~3qVI!x2~@5;2>!Qb
zjsNj=|3PP}-|JoU2b`YDQ%*@pYCz+qRZ2Z$^Nl@u?o=-bC-ToVHfc?i`4(Uo$ntN^
z)k1zv&XRpsz3K}&CtE7l*8p(ww)s099se{Zrvn7d+Gn8cWdGY12O@;^Pt@%F`*B6;
zo{or(*q>Pu(?KHOFs=acQ7=+MA=cyvOQCQPDt|~gl&EjPUSNI)9|;8gV-?UfH63ip
zcQc~4oA&fpcB&`XH8A$|At!DuRoz8V*Fv;E?|5mwIVN*)>q#fE&m-a_?pC$vde<*_
z&E2Dt`(}e0aPe%Qgs#NiI@aR3;8XzU99AD7Sc{f1q+zS0P8Y1gj{#;F~zupzwtGx@Y}C=GXEGf!9GX5FwKf^R4qeM$ry;E4-Z$oMKY4T
zxfCyck+~Yw>zurXJOdrm?m#wXaRx%@MS?t%`snI$@{GgDiw_DSR-B(O**`{`SmNj6
zGoQO_e}DB#vse#JIea87*y3%jDeO{Uy7an|Rkum@nwRx!x#xEw#`)+C!rd{80|0(L
z!&m~QMXss%UcmRhfOY|@*1!pUW}V@cjV9V^&e^m6h^J0By(3d9B3p>_WLlGgBM3_V$oO53|5HZ_~?E-mHh>YvwEi$R25&io1|<
z-`mbg<&_HjPu5+@9n;WI$dY#`pdgp{*6YAaB7~Nu}f>NnfDdaOS8AkTG?*%uXRmm`l}Sv
zc@J!hL7g%=_xvZW!!vV0nAt}nJcGAMxrHxqR(o5n;KgwB)B@+Bn<=QzWE<}Eb%wr5y
z@q(PmrnFiYtzf&Bl|yzQfm=RpWN~`^wA#PTWVl!a_M<(A&!OSoF@_?N`n*W(;p-)wnBY6RN#VOx5aNG43(s
zIHMrNZWS+gx%O@?Z^=}wN85KdkTnyocpdKV&Uy}k7hw**{4rWSH#H}Q;u6*-Fo^JN
zhmzi;r`N;Z6NyC}efB1&cI3~_4aRG5n6x$D>{Xg)c{JG+T0Lif`vir~gj@IX$CljM<0S&F*stB#1Bh8x3)_!b`I=g7LxdL0s
zN=w$pCuGW+>RJsVZ6JKIi>
zV@bMHw2A));y!Fuo(1^t7{kzM_%ZUO>9~51pL}xfe0(K3YrDtTs0oUc?c%OB2~b?V$D_j<`kZ$rqW
zt!`gwQ8T1>aJc;9-j%k+V}f4%0#^KbRSRx?Ostjblh6<4XP?~U-}*z-i@(n|P{Oy0ZMx4uIyF{o@_SbMJ@kjJsxr4M
z(kQ##UfSUTn)B=~2$(4dZr*_>i8iDW&OmuZlcnT#`W}O3MjK;J&diAuZGqUW(V`h_
z+~eBOk0sV6p$Vi6lqXeLM28(
zIweIw8bmrrMY>x+2FW2D5E#0oq&r5syE|vR3w`(A`ks65+3z`bfA&86bN>;C-&*=x
zYo7J2C%@l1-l*>DtW7G9BDROS#*MbxOpZbSdWgyx1{4Vlh(A0MnJ~>hb{oCBG^5;9
zqU7P$r69$9t31Z~m?2Y}J>G7TG64mLIfVGR;WT0+Nj*QB3b>crvObgJ4nBuv^B#YSD%(wmR3*nERAaGFT
z%Au$|lA;$Y4Bj*L{Lc&M6Wk)(kHx`@3FTm1VasWy7ptbV!m$;E*w->5qtcx+E}=Ud
zl_1mEcy0;Ad@URe9;QcwvUix+yiVv`-||?+EqiQ(bp#8Y1ZV+d+9t>
zi}Q`u&pGDD@LLw^EgYhgzrq@Yhp;8=nQD4})eKPzc0{sjDGtUo
zT}cgVWLm=li(ok^BTLy&$AH24rB!2sMnC(SU1}%Fh^YCTQNoeOP*2gM^&{8oj=rjBFucYmvz-7nZ&VzlE*tSq)6FPB(e3Tkd#1
z2-6@{lr*mqnO0?!xpb%;8+a7<#E36(d^*wkrW*u8}i
z%H9FhJ__PLJk0-7HQ673|DRWv1^Bza$w5dZ{wN;f3YvEET~L``P$iVgVURlP*($cQ
zuR32f?edql7ZMO~u3kH!0XebnChA$6-r+;Rs(ld-5Lmc8{`YnTr8VM9Gz)}?gDBTl
z)v1NOsKK4u4ht5kkUN}qw3fr=?p|d`sBB@M`C$JFZ{1fxs^6e5O#4XU1x$%%hPzXt
z{5IywJ*XuMrZx+UtLCPX)6-!t;aa@GTt{Vr_5Z}D#YM{|)4I@4Gx$*2-5V9p{8SiiP(Asg
z2(?hIz8=92WnH3ue>scJ$Wnvs0feMTr;SjT*eUrDfefe~2~|zuh$xIk`=hWh2q80XH4g-aBRi%EU0*
zXz0w)ZhRT95&uGW1(P4sLXUh^3<(qeP$Re*J6pHjp_ZyQh)W8iTaKiqt4xN9!ryP_
zJcN_2J`KAVBKVR~RcV+a%&(oemwnbmAjp36E-L?V9uA$-
zQca#%QLBwD8bYOnsjK+YNtF3{9J2@sEp^YQGBGJgf0|PH0xBB&x9Z?G`YmF=+&c|_
zI*5JDS{yX>eq2oyWakJZlf`fLeLcOZGK8X9)+z4DnA*%v^HWGRv5_1JQpurF&!Y3(
zNbq8_v}{M{kw?Be#bvUW$9L(xJEscEvC^dem4Q_6cs@$8Kzz$X_*kGbD{(x;shJH;
zfn42e(rE}a6}X&3nnR%_%CdLu48NV?1N~kj1j)i+do*conzRw)P+5elTR$~{ZqobP
zU#@A_3$&}c1#=b%)2|0~i$$er+M6eYl$*OfpeE3?y5fjXp^=I7^%k`910jw?-qRJA
z{URSAiR1n3div_MS~V%-sxDJfKa@w``K$6W9ccKH{aZ-LHt5ec)
z8=aEu6Xs-}aIpP-uRR`j(+oMBQ)_CNypOwstm|5xT`!NdzzoFBiqSSp^6@v$W#CK}
z&L?pJ%BNk`_IuwcNMy`we|I9n((H7lXdqFUs5TTvJ14Ywjy}jzZ=S@
zKF;(wNcQElJ@xr19ba+jq+us56sxL~R+Fhw9+x}y?Qt*7$G^EyM^-z^?8peWTS$Gu
zbvPB?3?Fdc#W{25XcN{UCsCp5f*&a_&Gi_Ex}CsAmvADA`U;GPR|b71u$KHlHDby+
zRP-(K;+={)e{SY|ik`<^tKE|iiU8`C@OyWFD8!sdP579R^W3a{rIsy?|2|Af6Qvr8
zSOant(~Kx|_Nu~HUnsSxFKy;&
zvElnHE@ZuEt!+Uf+?18F4JnXqL&{2j<|!gAHT()Ezj4bQH{;AWp+;Lj7gA!aK~eH~
z4rL}w?qlWU=JYReC1lxHKOw&t3YE3HZ8FV6@53EycKckjD$dPoU*<@#3(jqjYftRl
zj)y9tgDSG>^oFo$q-1mfm+zfg8!Z!H1b5!7!>5WQ$;~XS89_HQYr7Ope@BCekK-YJ
z>cJ6*YEUZ5?cSmYqIVBDjyX*0Nu(}Hwv@10)I)RUVpV2S#}BdritRu)t;29I86QL1
zgX(q(;Yh?3->Mxu!GJVL%Fv*)y$>Txj8Zl3@MIYt6A8qw=8$yBrP;pMtBvB`Z6Ysn
zeR*hN{fjo5)CuRx!~tdq}E
zVr73hUMMe9k3w#Wp5{F6acuiKL27+=v@IKL*o~fbdc_p5WlP41qRK;GskTj6+o76W
zJNVS*W~Fy7($ih7#;j>wKA0L4>=ZN;9tr#~>U(I8vvZt(sHQE>{=BDL+k54zYFLSH
z;)t1AUYl^L88xy5x5*#E2TDOOt}49ghv
zVo!TZQdS!n5kL#k0z5HMaAW>JZu)zE{!$Uux6~|53u{Dw!!ZZ3jag7)a;ZPK&ft3z
zPX+(`g}#yEK-VgFa*2Nep#TB3{)G5in}U5IkP#pI0Uh!xCjmJ`i8@}TjL-SS3Id4J>n>Pm)ejiN
zQ(8N`!oloV!js^wM=|*0uMymTx4$$0EQveLKPQVya$&XH$s6RPn6tXFkq^Hon#)b7
z8@H~W?BYk7NFCY89mL#=;T88^ZYnhAr+ke=Io?;?wv&$(H$JN2eaGLf^qpcJb+U=B8~9Qskc3
zoY~J}s4H>|4E4~pA06!X)r4d^l7ZLmu>}%U_hqdoz?~%
zyf*(j4Y93{oMRH&LD%?C1%IG_+Glm;)ZGR9mym52|AHsb=Qvl0kFF{%I{SVaqAGn7
z*5OPWa>wFK$oUqu$CHw>Dmw8q-CI1jKVBSDEZ-n$3%E8rILfCV5*hd;&*pREybI>G
zXU>j))~vt03!eqT%(0)ZH_8x~W#GIZ{JqlL<~4?xDsdZ*CEoC_CU@?>jg?-@*RsI9
zd~Vzt9`l(y#S%!{+YED{1(W0G*jfoI$}i33NRfzDS0^j_P1Bb28?ATsUq<-efZF+p
zm1!~Zk9>Lg!ueULkEQFFu`zdf&)fGi#Z}8X~LTn4zAit=&Geau~vQ
zZ1%8t1%w0Sd&;?@C%;@|J&1k26&!atgP9I1+li0vw(NQJ4@~6ga;-bpkqAZCevWpzLI;
z6wuvXk?2shRJD+~{tA@N&|khIj0`_eX3B{aPA!#`%=R`Bf0{gIdKkDV>bh&cn|DE3
z%3gAk%Ag_TD|oY{QP|
zE7!~~y~k@R4BKtc4Qig;$)13p3%Mq)7?5H_M$NzQgqJYG)SaZPPX^mK_);COb;>(b
zfv-X|?rda9yPKDYKyQ}pH22daKSrN_l)n9U(R*iQpt2sy`8->zxYr{)Cqlg+5+a{B
zC5~xdn5^ihFtfFDdo!Z)E~Bvpwv%P=7ky@rQ3mN}Z89mOGx-gtwJDZ9__=^Z2mi5Y
z+&o%Ad$6-ga%)_4J$xDfg$7av`sD5L-OQTCzj|sOhw7OTb+kG+VU0{PF
z{dCYl)A2&TvtXwwLZPJ6EGFRGl}mcc1fSPWO(Cju3XtMdp&&2f<)1V%>S#{}QHY|H
z^mi(py=M*s_
zBB+*Z(mq(&S3OJ@$4kaLheMrVj37&2W|VBPAvopTp8TE-@00oTrK8Y(
zq;zs!|DidNNHldo#*M&|5fo7{8fl^+9VX-f*KYN-b&05iFUj>1Yi<5gQ6dcIQwL1jL3BA$l4
zUTH_fgRUAWp$lte<~OdF+zk&Ln0e{!PM+=MXG!Gp=6%!`u|cky#nHv?y+$~*wPk(M
zW%9f57&87em|;D_SR;b>s-!j5sChT1Lh&lq_FBD#f^;Zh{5vdpo(n79DIS@+vC4=d
z?Y5_c(XK5w5-$Xxt>OQ`>
zQ-_~BF(s6gc?wI}_xG0j}aNK)9x*|(?Nbs;ElH%pe
zSmz;y-zMae%+xoWD|yYO9}D~WXJ^&J&ttlR;|^TVSx+=8keQ-4TZ`Dvx3LrtH>~5}
zDFz1lkzn#>6-=C>ZYRaAhDvllYW7KHX?=mS!_9Dik-5!PX+w3i$1Uv*>7={Yg7+#K
zpv*$Y5dFm>`|82~()&YQ)!vH8er%VL`PSmBvPnxcOmyCqxk-=>E3$E2vK`5q`g+Nl
zUUG|Y_Z_Fj$(>NOG>#!wtJ(Q9+^7QN%A4)9_m8f5UW12(r&YSEZRh
zT7PJ{M};FKnZXMYcMA4%XBM@p?74s1#F?@P0l&kx3
zGlxM)VB6{ooMX?&kwLM)??_meILWPve?=4bU&y3oVvc*qvNAe3Yv4spZWQ_o$IYxo
zP-#QG@JNo3?YyYd4G`MKQ9ACJQ6VeJ^>mdHz}qr%m}Dni&>{D*TOLBtF`#=P}b
z&AtfAO(E6lK6c&1#5M?OiaCk^z#^Ho!eSxrNKNuZsy1X5KO|EmfK0Dq#1Ud$B3eKWQ={sy!q5QhdF<{Pk)W;2b_3A9B
z+@%7muUOg1jyK&?izcc*eJQ{3>*@03dJz(%k};DH=iab+Ei|-`tpey{xO*h0e^i6Y
z_ir9_5p1vrawS~)qBD*dlGwd1CE?Dc~lbJj5^O~K~zJSbHW_O
zsm``3$1E7o{b=syyXjn!_Oda2SN6(2du`L1WhCJZLHFGU+af0jc*qESf+v%`bfOw~3bL-()JhK6WiD5gQ@y1F*kMoI0Y5>c
zy?{=rUeW+EEx|nGp@1hsE#ayB>@-EbD=K7Ei$$!>{kc1&&<
zRP(BRP`08rm0E7GCwFAD4@)R`Ahc&EgkNgFm<3}Z-1zr$gjK>!9;O?G
zw&UR5^u^kJG}$o<^=<9mGses|_z0?G?nh0vRwe97y$oR>LUcL^&&7I+t?wcxDXa1q
z`rEEU+(m0>cl3u2&=wO`o(|Vs*mfrSJMBoDVUA+}!I);mJ#vXhz%4DtII2)}qx!V1
z7#T#@W=(}LHHjlJx-T1(YGlO)tcP_oyc;|j!3F~gbKBq`ll|I
z!6o^UM>nW>Y19u-^(Le~8#3Qbl#M6T5V*1P5-iLkb>CK~+(1V+NbudzTalF_D}=^4
z(2^i3syb$tvX+-GNPhQd;C!!J*RqJf_*9VOjQW&B7?e*%A*|u61CmZIOR!wlXxA76
znl((9u($dIC2i(RubmB{N@1)s*@pCP$#jj
z1}A+c(LWs}x>a=4u2558sW
z@=Q*W@U*UE-1kj$gt--zTq~LiW~7`E#rD7+baWRJ5+6VA<&NJiwLGGv1Mu#<0AKUL
zvdCR6L5>!iJEE7YSh8QF`vEQ(9HjowOExis(*ZzwySYW)o1aDKrE3^EVLB)bZY>4E
zm(4W`|D`;YU`L$^K+m<&MtfEgm&dUk-SDX_
z@V(R1kKov+sQ*Xc=l@tH_|F=n_H%q|apz|o6q{$DrOpgGdFOkxFO%o@0&~aid!e%o
zscPY=>L<_<6#Sg5?ohIs-vk_P$2ESp)9?L9P1uZfg#yrSbVsk{
z=_|sIqN8CSuocW-fF*fV8SvMcy)^$wNTbNxI)zEYdITetpH9DgW*srlETpSnwINqG
zA_5&VOEa&mpt?f1nk+nRcS2S)HCS1YZdT;Aq)_6Ng8j`=nZ{4xlnRoSjXhg&u>0bz
z_;wOHt#F+zaK<_7^VaIlJi>@4@ivQ$>7nQO_PC46a2sCM>8BbNjlXcGTz3Z-jrYLy
z&6U@+!%a2pJ-tVS-4)5T!WX(-@4TJ}?kN94F{`iz`wU$VM$DQed66L}9_ab~VrTQ3{V%R3>6J0hJhrFutT$J3g^
z>DaBvdF;>sjmX8Py78kdPdoulPVN}68Qi=gE76h{T^#xD7=i03{e-=x!V#O1HMXP{
zXqT>4z6tjX#uLJsKi{2{ms{Fv8P?x^k>U3x=YS>4Ynqp#n>*MQOueUd_kuUv&Oc}(s3|>d;w+Z#D7JH;E(V0&vFp59qKCY0)7D^%^26AygN}{
zR3)Ht{e;zktGlvJ*7~he%5u;fNjfi{Fg4{si6`#>G=hv1!q1(l)&jV8_*@kPJ;9-{
z5lH(;@ASm!csS2*)0_-Crq?dFJtH~mw!1SCNEL+DRqr$tXl*CLB#V_m(K!Fe59iI2Hts)q5(r_zQ
zqa(b#MTe2wrc$jZX>*MG0%S=(rQ2z*_`o7bg!dR!lUz8&O62wB_glu3nI%ERxw2I~`#;fe_G7&e3Vh<6U*LV+*b={LN9k_sN21xl(}6I#QL
z?mbHAPB)>rm7G0_uPrbz8@}%|A@7_!f7Xn`#IyrK6eF8F{h_Kg%fZ3g?d2Y?O%qr9
z%MLn&*d<`RtSN1rYl(XW6tA#Z4p6+2SVS<=0ZY1XgqR;B&9lyi?8BsdBgaxILX7Yn`^|!YPfC)?`Vd^@=)0JS&>$(
zN{yp`eQAy_V#G!G>#1hVh8p%0wC1d2V}YnDM^@C*^T;aSQYJ!Gp|chzE2SiQ*PeB<
z(o`E8p_$rYt=6r_b1a*j^q|Pc+q05#QDe6yuC1Tve8_5*pvTElcTV`l#t2X&gob_C
zp{0pN77+(>!0mLh*vMf;(5u)(5T!eovj>ZbN!o|WtYZ)Z+e?N%kjlAe8-Uw-BcQhWSMky<)CyM^tN
zT4_0)kM@U4%siRk6RYc8G<08lWq+z!;sBKHqsBUIwsR=Zu97DPE3gvqU%5YuehWD2
zD@>XB2FTM(>g*I!T76>kB6AT5%14prTM-ao@W)LPhV?u8q2IYUQA69n=
z-^u#msIRUD6Nj>X`D4dV4rPvRH_D%Yw9Ufy$l%TO-rL_f0uK8YnaL#^
zcR%e**s5E6w5|AZG%kAGiyvZdu{Lj&ZyDV9dg#aQFCy!Vhftu}=5M
zDd#aCZjs+cFv$ufCKrx;_{3=Z?!g4od#r3O$?C}lmr*|%0y-hJZT3j!3^$6JpoMUh
z_XG|3gj>3dD_rV{zhKZyNq!c&E$-ZjHN3^jo5h>UPchx^1vjlJ+~ff5Z;sJ=)hlo*
zd&XG}<}_&fUIW|*&~<>e?A%pcE9MlE?DM51)=F+k|21$9mPVuvNAq;rdpEB`mb5@Y
z;)y4nqupa$MorUZ!F6Oj?g{feue2zq=+>r{lpFEBNa&+5Y5dt$q>HCYuN^b%t+v&*
zRsA<-7u%hWP7NR7O#`=vS+-giH55CN*%da6fzqAOVeRZ8sgenZ
zW^26YDA}AUM36z}YsT=|)(W+kJXb6_ZVbbau$(>U(sqt!&-RcIQyVJ_Yx0}3^FAU<
zi*9Wie^@ZFf}-tzxzFjJRlV_tstlE>jRcbKeWwM#r@z1a2ng^@y;h4({oNlBn@mgN
z0J(4e#|xDdrcnay{*~zT;nZ@7k?rW`Df)oR%m7AFZ(kbgVgC&$9Omgd+w_>TSL$LJ
zJP1O~#4#fY0q__#wd>KCbD+cW79h+N`DOUnj8gxYV9{yB03B@QNw&`+C-%m^44n~0
zogB)@lr9V=wFM(Smg(gV+%0{6;wwIdxzO&>tp@8IpwDCy>q(rYI>PTtLF*Ci=FLmn=6Gk724AOnrg~Vi&E8qT1MjTM>)~0NdVx!i|
ztfG}AKFt@x&j8U_DrZmKS7NlAq_!VE)tDbpgcbhm?y;49y;ejA9x=w0v1rq6?J*lx
zvaMYdRDR9KGHQ1anR4Q}8h<{uT`&C&2cNJ;cbl^4IPUm_=ZFbjIF;qD^5adpz?&+$
zLf2E|%l$J`QYGxg0@RN#Gq(Z{45EZC8q%=L`y^;CaWD;t_O>BN>-%TPa8HN6V=u`J
ze;CT2IE3Lf%%*^I9G7?yI$7RcKUb+RpE$9MLTh?yI|B0f~l@
zs_C%_=ucB};*K)~GlgOaS0s0#-H@OEqu`(C=<*&YJ1M(xK?cK@gHwJH#9tH90`qt=0Ct>2sa*z{1Y#W%Fhk9MScQOt5vr=5Jp6E-^O;2{C7q
zP7X>FwrvFp#BW#v5fje@%jmIQ#5fkDfoy>8Ch
zA#U0ZLi&Gu2&6{)1L!o7Vh3lWJmF+e++Z+xN5ZV^xsN|T4BnSUK@Km9pB$2K3>cK}
zqOh}I|6s>%c;}~QKx{7Bdw^4SyXr*uZo@s-nuaAG)EFHO%VPKQc
zsRQTx&^G_cMy4Nbm(l+Jef)i!|C#$3HUAfR%a2ZZBCF0BjZf!J$^($
zXWo6v#-4wR>IDK5iZOFm$?P34HsX6IB>RRu%rQ!RQ^D&2liD`d=1DXwF=L6;`SHC~
zg9D!feVr$jZ_2l{Hg00`OKD{)!0C*7|EoaQ-$}HMO9ZH-u%K
z(|Bk#SMlz4n~gK%4GU^AfAnl=3DqlTqxV;b!`QK^UL;qe;84ax!U>(!$~R?tj6<=h
zp>HFsjH8yArs@V+Kek;O3}C*o{`txa7~Z5^2g#V}>kow^7N(BZ<=Ur)!QnVWa{_6f
z$q1AjJeT`ntnUA!4J@*wg_f7jqF7nHlhVK8@MW)wuHP(3e6qt@boH{B=qfwhY*2&QF>_d3N0iJ%Ka>BM0;?df7GJR_mUKc$Y;FqoenbX
zQ7;d<5sOpfO~)2kh^Liexe2?|q9KK)m1u6+#^q2w{x`E&;s`sf>ELVu<)3s*4g=2t
z{rfOupna^%ZAyu?jUJ%jMJL4h`pKXcZ{zzA<7wRf
zaE3o+2q5WD76yL*9S$){;r<01$-;Q?Q`Tk1Cy$C>ostGLkltKzCcX04ejgY(Br#}5
zeD=#qvW0-(=_&iJpC^oGchacLbYg0
zKS~P{Y@WYpKfU$US&fI`w{+2%toULos$4t%s|5iTcWYXY3@e*+6n+y-$XL-9RNeRv
zkt05z$-(5zJh$}BRMjC|xB=fI*()#u>9RvNAkiG=S2L$T)6AnBq446zXSOpgw+9pz
z#6JfcQZh;1E+$$|Hk1^Mq(Ay4Bv?%{Uo+_tC2u%J2@o1r1DAZ|(S$A>-mp
zpkQ?^Si!dooluN#+RJQ-@Op_IE=0B=QgW&IMZ#yd4&_7V!a$D
z(E2fJ3Q{fiaG_JPMd6kc+f9?_p(&w3(wt3P)ikm=O&QGP>;2(I1OdF2ItSBQtM2v~
zA?DaX1`XFD?)SYUtwMA)SG6r-k2pko88(Xc)lAGxED%Ensv9W^>iKhsBc2##nC+@{z@`j&}4sV4CKR+Z~G&8Cxm5Ccj-3~nn
zC#AKTW)nJV$Bp)?VZ6!zuJ2>djIU~gY(dY{P2@9Wow+hM83R)+Q&HXsLE4b-p^?Z9
zEGP=cYub=NI$EXXKH<6l!GPgx@v2!YLYH(L6VGji7f)m5EEZklAnyGIQ3?hE(2#Kyck4zxeuZ0>$%;hgXD~N8}j~JrB1CGR`jy_>b}cfOh~Ozy)9pFa>~r
zjU!tEDd}j`*3QoMB3xW{PMoIZ;I|f>W?)+`cT;;VZq8?105M5-ds8!zg){wI3oD?V
zILkqOGYdV?T%1LRU*(yKy^Mu5P{GsDLeo=K%ghsGCTz|kDS?MA<}TuHYj11eY)bEL
z`_9fu#9jRHuWuJY#(&A?dMqL4Xl^N@@j~ujN+46>kN@Rc+}zwa-FP{{j#gaU!otE_
z&v>|acsP(HIGjA}oK4+1?3@_>h
ze`{{e&+*oj*PMf!+uW4H)SQ=_!+U;E&-~C#*xfz$m
zI}2M2J7*^(H@LYTbN%+w|5bG&S@e6Q{~`O4G$PB0$T(VIvOShI>voW%=_4cI7l92li(BJ5|WcpQjn97lTp#KJffmuq9G%H%=wt<
z$x}9VHcEOfUM^N17FIUaUtWTOdH+5(7B(>s4lye=IW_Bl`nqcY5Z=EBLQ_FSc?`Hm
zh=NLpa@PT%L(Z<~$XWZBiThtJlzXUX=opyyv9NKF1u6*u_fSw#@1dchqoX0GY!n~l
z{{d)(=no%pOJfkJn_@n8Am;Il$-K|-ysVW(WB8Df_pPHp7B(pvIRzyXGs}~wtbF_e
zf*(s~o0(f!T3G{aoSa=;-P}Dq13m->1%C_)jg9*hpOE-D
zDLE@UCpRy@ps?tBc|~PabxmzuTYE=mS9ecu-^l3L_{8MYG;C>kWp!c!==kLH
z?EC_LdG!ljC;-&o$oieKKhZ^qr0X6U8Y&v*FLa^Yb3+DHLNxS8+!zm~)iF&Sh#vF!
z-6wt?lUde^#lWj^Nb=Tk7@L%lZ;9#X7t(&E?C&GY{~x04cf$USE*Jn86$RONsDuDX
z!0{cRY7R2=W8va(__
z+qyeIMpE;^Qubc8$;pY}?G3+<3snIH=wFk3K2gWk>~Gu-yQ1~7H5ALD
z;-Ye2845hr$Jio0^910D(S>
z2UjE&kQ6f)IuqNDceacDv+QC;No2U-9wG%rM$k6t9{;NX5Qf!!tDcu|_3IgsQAUyq$%{
z*Oi|fQ9yx{eEq#0Aw98h)(eR){gB?Y_w5c4?NWW~XtK&ikutiiCQGbs*(yu?3gw?T
zV8$bc>5gBT-2uLRFuB5+{Gn)~>ief00~xdQywx2w;?DfgpFJ|_l?pZjIcWGMu2Ya6
zkr_lS*D=Xq+~iyp-SLB}J3xu@O`rv@r;p-hR_KqZCt0C@f97yp3S&ShhC|K?!tMal
zo|hqPTMK{Av3}U$b@8B{TOO
zVCp&7E&9$nV`)C*PdU)cvLQ`Jm;BJTqn*~~NZ#h~Q)%48EUkYhxk_Vu2QW;$#D**W
zttfv_YLv)1RhPAVN31Rj_-76tI(vk70LCuK>-(~IfG>^S9r33J|1PUSKzQ3uwc_Zd
zzB?5613#5kP_~Vf)Y*k?@J#@;c_;P`@W=$~5Av+7DVRT&9{S|#fPd(dFZccsN_`-p
zJec^vX+Q4M@svw^%{@b(p^j@czPz-&StFF2m(Q(c-w^2yex=`I24ZVwv_&q5u2dX^
zh#387qk>zOIXcuL%rR#p<=2cnMS3Y4-qDa-=&4H6R;s63mXbbw{>n9$Mwa+B!LRMA
zsXp=okBY5uQ)`v7xOiJJ2PjB(oUQPUbGdlzBx*-VLtV#AvD(})tH3(B2;?b7`9Zxx
z`H3V0b+k179e_2v6kgTdUQHn5MZJH5*WUNaHGGAIQRCAJiECFT^QFevTvm*(NRY7%
zZm53N^FNUJXR`lirtw*Cvm{qPci#a_?>C>4m41$GO8aw8RZzFMO7BGa(=auX%xI)+
zM331j+^7A6va`hUhyK`Sd3qH$EK8E~O-szSD8}hDs5K
z2cxS747>(E-&O7~`yVqcM9uxc9H2PEywZ7kc&%PjKE@>`>N^c`>|#&dt%?hYVw
zP(1kNcA+Z?&4@g0AzWAzN@Q4R#xyBMyudFkdpfp%P^nDwO!oZx^Wuqt4Or5Bybk#Z
zq2kU_SDSZ$YYw-S{is|ZR;`k9*{8*9^8UuYYLVQCv{}+>9qvx?)((a41pcJu2rIRU
z;e*}v#!%7x%4kMe#9O6Q2g9LA$g1bhJHTEBBD^H3Q=c@+E!SjWRlm~ZN9OKIJSF7Mfb8YRyyaO0lH|V*bAG|*(
z;b2zGbnlv!>yZ>H@oJHv8F4Ye@0G|jkL%;4kl|(s=c1l6kxPM;ESOGZB#OU}Vo8SPp42(I?>pud)d$>e@0KN=dKcyj)3Nqk
zN?r`O^|Y>NYeq}aN1i1vtK8!v1abh`E>)pTPr1{e2l}|g32MqI9fGcs!7%`O3xClQA(S*d&`@WuW|3G>;n1{&-Q%0%)
zl&iULh--3Hnj6jXtU??Xm^mqO=+q_dtdSGfzKb>CnrQyk+2Z2
z^g|;hmjY0iLf(so<@MYF-a~ERlNlH`l4n04o6l><4$YaWs-LToW@QwomAuU#?zoz4-Cr+_lW8m`_ch#
zyh$tqQYlc*F9%F-RPF##14Or%lF8~8OAs<^4e_e5!kOEc1$$0R+Eo%ENu{YHa1dBb
zn)CDd$@2^gM+=^ZJP+pOFt7P?2Sk7duT8rXNIXijYrVQ8k6E@&&bqHc^A9f&FX~M$
zR*?f;1PBE4PR@#?y!RUfF9m%am>jt>%Lwf+HwL(P}h}Jn(JBMRMfUC
z`61$2f}OXVGw}6wXL%~V<&kIh^fKN7^bI4%;?J$Udo`4yL!yUjL8k+bY&|ZdX`Iki
zy2pyuYg?nVUnVv00N4vJ5Vsq*h&#ZJI@DdjgWs$e?s~CH_QP@Uwp)CNqRT6sz@~ai
z3?lTbw@|OG5jPY=;k<3*Qk&2}bnO%8K&~Lm0L3)TgW#ope0;|K{L73c`r4QaS`i4Z
zer3$k!2L~a9(j({>bRE(^Y0339J`W(AmjSSP-3WAtm?7d1@@Dyg6RJb)+c|hGyc!6
zgZ^)d_>#JflWn=?}E+W7qg2O_}hFTzwENX%G(lW-9hrSilP9Bi`gp@;cvCb`WDEvF?g`pcXm6
z3%inyK@g|DES=y`HEzjWMqhTrC?>e>)4h!;h_e@dH~&PbQa*U3gi9%+Uy&m!M&(y4
zQ=KZ$Uv`kc$G+oaEOrNo@sQkEg2a@F^jEq{{4ou%ojENI8!7zt{$4`%x8ATN^zk2C
z<^Ju8_n#q-GT?rKkVUFhrGW^!njOyZDfl1KV<^?X%IKjN&zzbeev~pyfTbj(Nhvg&U^!PFVbWE{M_55LWJI^|&1U
zNtv17iObF@pZCk3UD?ZEEMq*K*8$fuuR`fqsj`YO`!80=FHLTXfJUVT57)ANb;K=0
zPfY0A3*FUP>TH_=zi#u)rRkFxq|8)#&^dtPr`~)T6gn;B&n)ofKeQ|?H3VHnRu-sG
z(koBPJdfiwbG8P+ga;R-m2(j$iI?$iUbGRi9`OtQV#=VUUj7e9T~)B%S5!gagLv6O
zRG}WaU80^|u0>agkuE+?EePT$g`{`4mMH4T}sqdtx&)Bt&rEYN3<^k;L2)6t?oBoeOLskgRnCHH>|dkv3TR&mu6
zlPTPyluLA-(Ehold0$`Nq+uT}yw}}qx5buo3&Aw4wgRhl8lKWv>3#djmsBZk3@M*(
zjp9)WpK+9;e!@~nQra6?IKO-nXd!AvmLH(;u)v%#F%~^V}!v=dy
zS?kjlm{VkWqZgIK1$5&raR#f`mecrVso>GqtG$JeQhcBYMmgu_#0jS)U5igoTG@&A
z6I`Tujh!)P^gf_V$!@NBL_kag#sIgE|>O^{lgRoL6@C6xn75ks+=`o~Do#-ocD+20
zy>!g4t2Q$>I*e+^w%tGDg@PD%lPfzQ7G%LaDRvlJM0C8=tLbp!_!PH2giM^xgt%Sk
z@xZ+qZFs5+p3P22T}ez{?&!=E49NS@R|_Hdj;7-ZqBMDk0WEG{=J0CsJlnD1*}|im
znHOsBB?^B(pW`rOtio5i0}6<56cD@tE}JglW6kZz{haqn{X!5TU^m<9d;UP=O3N4P
z=$$doPn9cJ<)@>Uhlcv)2hZxx2Ql2~Pf0?K@)3%vg@{whD}IoqEExU*!9gV;$}Xm6
zo=E74^LcWe?qly&%nOpy=KlTW%TU`q=*ji2Gp)qYnD6~j=$ZlPm|cm{2iJbC#CTcu
zeN}vbmm9Phf4klIa4o3l+h_E(I+ZojxpMnK-Y(zDb&+#1%_8?nLrL+3;5?PQ>GUp*
z4}I;kJ`0(-6O~kRyo+=}`J;9l6`yY-aHAE}Lt}b+c72|N)}7e;3L?!ZZ5rjxFEWt<>XS@`t`mH*38rUOhvDitjcyYVkO@OZeoyE&C#ocpmNUd`8}O
zO=Wk055XoM1U_+k!@+%hyj)k`rtSdRq8vb7-FbKLia=tH>@ptsRe8;5ZraKiIN7x3
zAtv>M`A@PeU{M4a5<^l_G3i@~X#R1A&DK|OP2<(+D|Td=j$j#c-eVrzlTa289f(b3
z6&~y)XM8WrTnsl;JyLBh+VQ0rYcPP2re)4vzFnX8H#W3eLsfEm4;U&O6fwhMD~-Pc
zaGg-aXACZ6T+-IddPD_^sTXW|(uLbr7EkkQE+u93)}$hiwO~rcv>Obije;)@6F-W*
zUaI$jXdIPP0t-z#I3ES|InY!*3Q~Xh#(UOGENVx#?sW8MO1SA3f;aT)D?atkqy)vG
zYW7!MdE!LUg0ErEn6TI}=?T29CNU%!KxRw0)un4`zM^-4-3(D{n8l#j4>OAHrSz%g
zBKRGkHF+@jaN|`2l7r1!V~QcNw~aUH?=<}T5ns)?7Z=aO9F&xLcW%K~{_!wvM3p#L
zxz20u@!
zcu&QN7Vk**yZ6hA;TzFqs7!~5%HG$AXJbaI?o3h&tUtai`i<`F=?l%~0Sg#I8{EC7w>o2YGWrzNSDbZX>yPd-2O5
z^kzarJ6nd${>`feNTsid19GYk#ShImR(4sqI==%<2U>8@C+{{h=a(L
z_&oN?DB}+tuV^xyfHEH^0gA0q=>xW=#?pn3!hLavv+7B0Ax;j@g67A~r_99_2Oj_c
zsmnFc`$(^CmX*>6b3Xm}#K33LTwiJ<`C$-3zcXeiNa)BU%B47}*=)&&$J*=^RA?p|
zb%QP^QRzaGo(@4mN09CCt=`I#D_>Ecg>2C%teurk36ipjgjrZv1;=XQLOHG+uU~q4
ze3`LKo_q2^+Z~{MKFlc(?b>IUw5szG`+(tQ>fk+AF>E8?Ms{fd&v0jgR$!vU#eiJp
zN7?KT+6kj<_AK@{xK=d40j8&}FQ4}Iac{$Fw@LcoGo57VR2fpr~ihpB@Lr&)+?LIY?>Ip9r`9_&Q`@cEo<@28g@unGrT}{IddE71tWr-tyh)84
z;vbzvJq$p|4wxX?-AsPy_$cl+lCAPjjw4KEi8Y@5-*m}YZ}33kRmYG?WTP?|x+L1{
zdjBZ{FjKvyLzb)g>mA@7`m0OK_=v+l=(kZKgM?5YR4mZxJ+Ibi%w22p4l5Yre&t!r
zb@TwhL%j+F3f=Gp7Tebt<@0vUSFfD9^JR?FWJy$2_Fulu
zMxv%9Ocb;nWi?|rRK-(M-RrZn)BVJXHF!U?-t)LqUro|L;cEKz@JIP@-}xcmd|d6K
zLyv-I>Yu8AsLlh#cSfG%r8Q$-U=<_S4unlYDESf1hX?ATe4izTJ$Skp!aPfM6>^R{
z9=yLOk`OsWR6_FLz8BKfk?VH=*@noC-R}xS11R=d+>3NjH{w@mRR{KKDh6v8shFV^
z2M^uL+eCEIvY=@Rys2M*o=bKVnXBq)J$5D|)Zk%{XxpQ9yOZc+~6yqEgnRVqNO^tAP~
zkF_@KU;~mQ7CLcU)?-sgn(1()*?!g7*t93nH?l08JD+X#tSzFp_XeHz(a$rggAiKc
zaSq#?eh$e!NnoO96qmQVF;;KyDMsDZQx=ei$x-uVd|xFJK7ZdVRVsRDnX+7py`uvY
zOJ*k%5ueEBfNlF9>GaEQ(3;G%_-(QcKgp?Z!zhX38L1ee@RqC{W9Os6#wXak5ae>Q
zkv%r(1=kl<9rav647E4>Zg>l~U_sDj{D&2Eu&a(x^Pt(-grkDCFXBU`KXuGx`Svbx
z5=}Z*n^>%qgFiOWnQYaGH^VsBt^4E~HHBgn*FMsq9B3gP!A-O#8d&oCzi{%;Fk%s#i4UK0oDfCk
zQw-S@7P>QNj5px|5poMkW`Yq6CjkXE~gYAuY$ff=uOAfpYz_
zA6;PdyU>VOH5%0ZtOR9*i3l{DhgR7G!FU~EtTEI%aTU0$XFH*L);Vd=HAvSTVEut$
zb6PVFB&G+hEHmajMDAX7b+yn}J!ZEYJAP5@5lY~1|3F@%Cc90aHapL~y|a390`a;j
z+l@G(w9(+ls}Dr)9fi+2UB=YCd?s$1dt7f0F7E(I8&{G6Toz+u9vEd<{{2QNq{qd@
zq|d`~-cW~|+dF7g1jA#N8Zo%=rI+ey6kz!=wncc-8qWd(IVdh8G4n)p@`hzkt7A><
zp;yffCwjhQhWLS8U*x^0Ko~1~>zk{OYUeUuuF!Eh97K+5b|l;Z2yaXmcc8Qz?dP6)
z$H8eVGY_>JO#MmuF$q^Azq*rbSG=0r#}=rJ5L*(P@i`3zChOZaswsevg-n-Ze
zz6>6!uL*dkzK>J0k4eFZGTP{@rPcVl;j8{|^@dZxk&hBbt_CxSPv+J>C8kTcvkR^GrLG{GHJ{~2PW>Ig
zWHOUGG%UA72;
zwDuEEBB-by`@2qYnz523ilYdfZ-YdQMsA#z_mrg?n}=;H>QzLw{9V5rR3vMlTI5Bbt=3g+uUb?8kImSr~uVC`ic??_+i3QpdGkvht!EEaE*CpH)kSCC>%L#7^}
zU46`yqE%B9;Z|%H@e$^$;klBpzf+JE&Gd@{Ci&m^C;fL@YyaCVrzG+^6BTo)YM?pG
zb1G2_p%bQe=2s|V8KwgO4Na=Jw*Q-!N(6opz%@Xh*Ua
z(i2$0R90hx-@q1Mwtnnj3flNpPVfG<)UNoD(&A~r5l)wgBPe=jqKT@qmewLSbc(5+
z57K7zdG(>Vs4LAA{g8UvI9UsKPb2Fsp`#btWeYEMBbbUUg|(dR@!2wDwp7cHn1c(9
zePY}7NB2uSZhr_W9CVv}q8_cv^uZHh#X8vCD|u60Tjx))FX+LoNUxn}-EWH-=hxe<
z(QTEiO(so&5-rvZNyOvOL{DkEx+f!BVGQ7X}j;I?yHXRgZ+zp4#E%
zxH#(&ErUsB7Pbzh7w-U+aL>V;*}V1iV-|cm;a3uNYBbG+M$RIEsTW9-muNjk7o&k0~C`_Mk;(7o!gdfXN6X^_yhV^3MiO&Oe5x$lA^5?nu7}piFj_HApzHfYrzZ{Th&d|UQ{?AT&GwEUN3RR&0^_#v-Xz-dFLyD499Wa?(F&3bIvl9}(2)&g
zZ9WhDJSLZJw|EVV8X%pwJ`u2@s@f15@+|gQR>6N+OL~=1odumVnVxQn7SZCED-H(iIG>k!3y5r8p
zH5~iXZwy%a4df()Oys(Ru1=(A&`AjISbzn?(*h!cBe6=7H}pq#SJOoYlm;=Y!Fd~s4G)4`a8B}VbZ8%qpQcn+*MGwLYJNM{0J`BJ3g9ptX=3{M
z#mhNGe{=yN
z-S@#!C1XVkr?e!=J%Qm(bc?MClZP&IBI3qsgwxigA|Y=_CRO#+wT?4a&1xQstO(t*
z&c1`~D#efZ49R9STy!)s9!bRIyDRUd%Ld+_myzV2x-OYVD)bY$CZN4JX9m}q{A;NfAhdK1&YxlNKp;jee*U`?a7qQdE
zdf(@IaW$SPidC~cMvjB;yD@+2WCACDZYLW-Z_9oNzWUk&nSOI0;iAf6e2a$Ep2x~%
ziM2lcJvj6~TIRnSzt$OQt}ru!v`Oc%9b98hf9}3r;B}aTuUPn?>%r#za^X>`0wQal
zbP`tgr|H+!3K*h&TCrPJ7Y9M;O)et-cKiFYZ!~<
zeqdg_2vyrMA`9ZAuRlzW7th4NEdQ7{8jaD~y);t6$+1*lnJ~Z3pFk0v^RfNWr*Hz0
zUj^Lu>!!h81JepAZRU-n{i261bu3npboX0$zPdYq4zg8OMaPA-ACk26utvejq*l;a^Inq-%GCb#M
z;0}4l?jkA}SNrE{IT?E@5cc|fHM6&YC)m84
z^7OR9Q^~cfNbUfzo2cEeY=^Z=gNE5~+d;X7Qhp_+t&-o+
zqI8{pE7*d#s)Qu|=ID(_V^xYmSuASS*DjW-=Xo+*T;XO8XFG~3M%dkk@Y*F6o-3Yr
z5jneunXnQfTmruvVbst~G5JN%fLbR3JXq=ot&TaXYQZ9YV%kd2Y|x?dBVeOhyy&bu
zkaD9-M4*}Z%a1FgifBGtf*+k-f~06>ljON%b|!3ys+(aAl@pJd)8^G5T%@f_cxYSt
z4u=&Y@oOE&{6UGJdmpS=1#2}wpaRUJ9WEz@vpMGSyBevR10Nrj!oFH5(pT4nzNmSP
z=KtaiK)TyRddp}dOpF%ssA05{LsTnKaol;r%+@sqOA7Z{UynQegm20s&JKiJ*F8@6
zh@-362SgoH6FX-nr)>xb&BuyV+)1`~9lu2AV>^Il*DmK$o3e9|ChDi9AXspiHw)dz
zQy;md63L2{2#h2a9#6o
zSN{4!ce)NF5Cp1Fqn?+~nJd3c_^#K@=trH_Q%)7S%4aLGZ^U1b
z#E{UnMyxO_f_qQ8>tHll-Bsz9@;X|HquQ?eM%8M0DEQQtx83vIm0gUfmEd`0-DSdg
zFO?ZfxL@k)0|Kra21yD&6?dlMC
zOb9cZi>>_S6v(Or>14SHh*spNv3cM&1;m_|S4M%_uQAw;Ko7W_pbI)pnOV5k5^>mu
zXSnd(JWeZ7qnb!{XecRxYV4$Ch-7=7;t*OI?EUD0zphb7LBCk!=ac?e!#?5nIU2qgOfL+yRjj>A!fx-cf421)h)fVM|5(-FATa5-TtS0J@cajM>?_;J^1J)=S!7O+LP#{FA8ZjAY6~
z=J1sVV$U-nZ0TS(3Vj9+*)w*Fv9U!$T3mMf_vflvCi6gpVxUZzyHzVK;o)GWdq0|I
zXfcQtS8fumYRTWV3|ERV`k4ywsj=PM3K(f!ZQ)h7!5PuHPP`%P&76kkcfoLBl71{d
zcPlR3ni5})#?CPr)!ZkF-VndmUW>X(mB_sVyx~Ivn_-6gdIqDSeM_wcsID+0nR3Em%*rN#qiLX4ocEmf#Bj+EWdmOQi}w!4G;qXKFb{d
z+L8^DNJf|(%7f-yMt?C?TJ+ulSV+ij-{T`9zUHUMjvw&L9%i(=-^kFx%nkalgM;Xg
z;}WKh*P>XexMwPzJ@MPCh8wKK6dg*8Y;+v;`N%NVFw7lW_R**xzX#mMq^&gO3KmIE
zbMV0MTfT1ORcK7hOX%e=b9;UV@G{SjKVJ!rU{8#@>6_TQEbZ&Y0dxFlE(cFHY&`U7
z-W{GvWXd&>yj5VBywZ*z46#~nHZV?d(kXOu7~dcqlO5H#sNhn
zXas|cF1e%!h#iPmc76M>jhm|yTc2?90IOF$CiNus`Za`bR^Q!8IrO3xX(r=0GST6c
zOPF|Q1N5v$|9mg+8SiNG8Gn|W+lh%!;P&Vw6;lT1
z>dK#cVR4!TJ#kx15@_!0PW5uw(w86^zL7CI!dTG`Xx&=c<5}*^uBiyUB8$}fv_{Tq
zpI_w=ncwK_dEvh{eg?AjQW&@VS=qC6Y`pwrLMsMel!hv(b!+Ifa|nM^WIn_5!t)}O
zlRuIB5=c;`>>h1Sy=%Buv#sXzDlNVdLX|K+cLltH=#x}L5x9Rpc^!nO*~*az6PBhNd7Wv-bTUii)-fezbQr1WlAU5X}>
zZ@ZaIhKAVWZcZ*oFl2|F+)}i;Gqj}Bjrrn+&D=_e7SHqx8_o_2_7WPkJ}2r!@oVX1
z`QS~KX=8#uqdCr=aQ&*TEFQMKaS<#3jpvJ;*q2w?0G&*Cw|2TIT8x19X3({K9BTV{
z6q}tJsqjqP?SZ~sz>>+X8lZD2&ST?4RV;qRrGZnko2a9w-B$Xz3(e)F=N=L`ZsE#z)Ou+3B>fdFldM8M<%8
z<^Sj~Z|}u(mg<@hNy9?g5>Xl<|W}u5n$>W(sd%A|;(-+(MTjVz_Thlf2})BwHf&y!h$3
zIyPKDVNaU4cB)0HgE?QHEO-8B(OuSP_-SZZX_1wU7CMSU|ioTW>GAz0ps
z*1%8f%BQ|73*=sP_&``a`oVkv$;qEegut46N0653a8|o7Yk#jJ|A#N7Bp*rxf!ZitI#5uxzvvDSa0^Lt
zQ*!g6yTRW%m|Ewfp*+86v6qq{lP*k|0drR>@{2-5EFCg~k%I6ns|m6qeszm+W29J~Q%!*&+P8o2Xq
z9?MrG3@#2l6jpy?@|n%7XcafgCK$QBEoBKDLW|dQxRy%kT%D)_bRhZqmOp_9!!bLO
zcXDgGm6NFVig@_+>TJ!nz|
zpfn9Fe6q>(f;{QueyLV@Dpv47mXzgXAgM8FnDMva(+-dk`>Z?~w;cnIEhN4!uzPf*
zk?iGR^SN6De<^`a=?}!Xf@aTg*_sEa`oJpn7K61k#GBY9;_ZUL@R?qPk
