From b9551db5164d9a4d2fbc267655ab5001bdd1278e Mon Sep 17 00:00:00 2001 From: Kaitlyn Barnard Date: Mon, 28 Oct 2024 09:21:16 -0700 Subject: [PATCH] Updating docs nav structure to make Platform first class citizen v2 (#8448) * docs updates * text changes * resolving merge conflicts --------- Co-authored-by: Kaitlyn Barnard Co-authored-by: Toran Bruce Richards --- docs/content/AutoGPT/index.md | 29 - .../configuration/options.md | 0 .../configuration/search.md | 0 .../configuration/voice.md | 0 docs/content/classic/index.md | 142 +++ .../{AutoGPT => classic}/setup/docker.md | 0 .../setup/for-developers.md | 0 .../setup/get-repo-dialog.png | Bin .../{AutoGPT => classic}/setup/index.md | 0 .../{AutoGPT => classic}/share-your-logs.md | 0 docs/content/{AutoGPT => classic}/testing.md | 0 docs/content/{AutoGPT => classic}/usage.md | 0 docs/content/{docs => contribute}/index.md | 0 docs/content/index.md | 148 ++- .../{server => platform}/advanced_setup.md | 4 +- .../blocks/ai_shortform_video_block.md | 0 .../{server => platform}/blocks/basic.md | 0 .../{server => platform}/blocks/blocks.md | 0 .../{server => platform}/blocks/branching.md | 0 .../{server => platform}/blocks/csv.md | 0 .../blocks/decoder_block.md | 0 .../{server => platform}/blocks/discord.md | 0 .../blocks/email_block.md | 0 .../blocks/github/issues.md | 0 .../blocks/github/pull_requests.md | 0 .../blocks/github/repo.md | 0 .../blocks/google/gmail.md | 0 .../blocks/google/sheet.md | 0 .../blocks/google_maps.md | 0 .../{server => platform}/blocks/http.md | 0 .../{server => platform}/blocks/ideogram.md | 0 .../{server => platform}/blocks/iteration.md | 0 .../{server => platform}/blocks/llm.md | 0 .../{server => platform}/blocks/maths.md | 0 .../{server => platform}/blocks/medium.md | 0 .../{server => platform}/blocks/reddit.md | 0 .../blocks/replicate_flux_advanced.md | 0 .../{server => platform}/blocks/rss.md | 0 .../{server => platform}/blocks/sampling.md | 0 .../{server => platform}/blocks/search.md | 0 .../blocks/talking_head.md | 0 .../{server => platform}/blocks/text.md | 0 .../blocks/text_to_speech_block.md | 0 .../blocks/time_blocks.md | 0 .../{server => platform}/blocks/youtube.md | 0 docs/content/{server => platform}/d_id.md | 0 .../setup.md => platform/getting-started.md} | 22 +- .../{server => platform}/new_blocks.md | 864 +++++++++--------- docs/content/{server => platform}/ollama.md | 0 docs/mkdocs.yml | 69 +- docs/overrides/assets/favicon.png | Bin 26841 -> 47680 bytes 51 files changed, 674 insertions(+), 604 deletions(-) delete mode 100644 docs/content/AutoGPT/index.md rename docs/content/{AutoGPT => classic}/configuration/options.md (100%) rename docs/content/{AutoGPT => classic}/configuration/search.md (100%) rename docs/content/{AutoGPT => classic}/configuration/voice.md (100%) create mode 100644 docs/content/classic/index.md rename docs/content/{AutoGPT => classic}/setup/docker.md (100%) rename docs/content/{AutoGPT => classic}/setup/for-developers.md (100%) rename docs/content/{AutoGPT => classic}/setup/get-repo-dialog.png (100%) rename docs/content/{AutoGPT => classic}/setup/index.md (100%) rename docs/content/{AutoGPT => classic}/share-your-logs.md (100%) rename docs/content/{AutoGPT => classic}/testing.md (100%) rename docs/content/{AutoGPT => classic}/usage.md (100%) rename docs/content/{docs => contribute}/index.md (100%) rename docs/content/{server => platform}/advanced_setup.md (86%) rename docs/content/{server => platform}/blocks/ai_shortform_video_block.md (100%) rename docs/content/{server => platform}/blocks/basic.md (100%) rename docs/content/{server => platform}/blocks/blocks.md (100%) rename docs/content/{server => platform}/blocks/branching.md (100%) rename docs/content/{server => platform}/blocks/csv.md (100%) rename docs/content/{server => platform}/blocks/decoder_block.md (100%) rename docs/content/{server => platform}/blocks/discord.md (100%) rename docs/content/{server => platform}/blocks/email_block.md (100%) rename docs/content/{server => platform}/blocks/github/issues.md (100%) rename docs/content/{server => platform}/blocks/github/pull_requests.md (100%) rename docs/content/{server => platform}/blocks/github/repo.md (100%) rename docs/content/{server => platform}/blocks/google/gmail.md (100%) rename docs/content/{server => platform}/blocks/google/sheet.md (100%) rename docs/content/{server => platform}/blocks/google_maps.md (100%) rename docs/content/{server => platform}/blocks/http.md (100%) rename docs/content/{server => platform}/blocks/ideogram.md (100%) rename docs/content/{server => platform}/blocks/iteration.md (100%) rename docs/content/{server => platform}/blocks/llm.md (100%) rename docs/content/{server => platform}/blocks/maths.md (100%) rename docs/content/{server => platform}/blocks/medium.md (100%) rename docs/content/{server => platform}/blocks/reddit.md (100%) rename docs/content/{server => platform}/blocks/replicate_flux_advanced.md (100%) rename docs/content/{server => platform}/blocks/rss.md (100%) rename docs/content/{server => platform}/blocks/sampling.md (100%) rename docs/content/{server => platform}/blocks/search.md (100%) rename docs/content/{server => platform}/blocks/talking_head.md (100%) rename docs/content/{server => platform}/blocks/text.md (100%) rename docs/content/{server => platform}/blocks/text_to_speech_block.md (100%) rename docs/content/{server => platform}/blocks/time_blocks.md (100%) rename docs/content/{server => platform}/blocks/youtube.md (100%) rename docs/content/{server => platform}/d_id.md (100%) rename docs/content/{server/setup.md => platform/getting-started.md} (83%) rename docs/content/{server => platform}/new_blocks.md (98%) rename docs/content/{server => platform}/ollama.md (100%) diff --git a/docs/content/AutoGPT/index.md b/docs/content/AutoGPT/index.md deleted file mode 100644 index 8d541d488..000000000 --- a/docs/content/AutoGPT/index.md +++ /dev/null @@ -1,29 +0,0 @@ -# AutoGPT Agent - -[🔧 **Setup**](setup/index.md) - |  -[💻 **User guide**](./usage.md) - |  -[🐙 **GitHub**](https://github.com/Significant-Gravitas/AutoGPT/tree/master/autogpt) - -**Location:** `classic/original_autogpt/` in the GitHub repo - -AutoGPT was conceived when OpenAI published their GPT-4 model accompanied by a paper -outlining the advanced reasoning and task-solving abilities of the model. The concept -was (and still is) fairly simple: let an LLM decide what to do over and over, while -feeding the results of its actions back into the prompt. This allows the program to -iteratively and incrementally work towards its objective. - -The fact that this program is able to execute actions on behalf of its user makes -it an **agent**. In the case of AutoGPT, the user still has to authorize every action, -but as the project progresses we'll be able to give the agent more autonomy and only -require consent for select actions. - -AutoGPT is a **generalist agent**, meaning it is not designed with a specific task in -mind. Instead, it is designed to be able to execute a wide range of tasks across many -disciplines, as long as it can be done on a computer. - -## Coming soon -* How does AutoGPT work? -* What can I use AutoGPT for? -* What does the future of AutoGPT look like? diff --git a/docs/content/AutoGPT/configuration/options.md b/docs/content/classic/configuration/options.md similarity index 100% rename from docs/content/AutoGPT/configuration/options.md rename to docs/content/classic/configuration/options.md diff --git a/docs/content/AutoGPT/configuration/search.md b/docs/content/classic/configuration/search.md similarity index 100% rename from docs/content/AutoGPT/configuration/search.md rename to docs/content/classic/configuration/search.md diff --git a/docs/content/AutoGPT/configuration/voice.md b/docs/content/classic/configuration/voice.md similarity index 100% rename from docs/content/AutoGPT/configuration/voice.md rename to docs/content/classic/configuration/voice.md diff --git a/docs/content/classic/index.md b/docs/content/classic/index.md new file mode 100644 index 000000000..deb794d2c --- /dev/null +++ b/docs/content/classic/index.md @@ -0,0 +1,142 @@ +# AutoGPT Agent + +[🔧 **Setup**](setup/index.md) + |  +[💻 **User guide**](./usage.md) + |  +[🐙 **GitHub**](https://github.com/Significant-Gravitas/AutoGPT/tree/master/autogpt) + +**Location:** `classic/original_autogpt/` in the GitHub repo + +AutoGPT was conceived when OpenAI published their GPT-4 model accompanied by a paper +outlining the advanced reasoning and task-solving abilities of the model. The concept +was (and still is) fairly simple: let an LLM decide what to do over and over, while +feeding the results of its actions back into the prompt. This allows the program to +iteratively and incrementally work towards its objective. + +The fact that this program is able to execute actions on behalf of its user makes +it an **agent**. In the case of AutoGPT, the user still has to authorize every action, +but as the project progresses we'll be able to give the agent more autonomy and only +require consent for select actions. + +AutoGPT is a **generalist agent**, meaning it is not designed with a specific task in +mind. Instead, it is designed to be able to execute a wide range of tasks across many +disciplines, as long as it can be done on a computer. + +## Coming soon +* How does AutoGPT work? +* What can I use AutoGPT for? +* What does the future of AutoGPT look like? + +# AutoGPT Classic Documentation + +Welcome to the AutoGPT Classic Documentation. + +The AutoGPT project consists of four main components: + +- The [Agent](#agent) – also known as just "AutoGPT" +- The [Benchmark](#benchmark) – AKA `agbenchmark` +- The [Forge](#forge) +- The [Frontend](#frontend) + +To tie these together, we also have a [CLI] at the root of the project. + +## 🤖 Agent + +**[📖 About AutoGPT](#autogpt-agent)** + |  +**[🔧 Setup](setup/index.md)** + |  +**[💻 Usage](./usage.md)** + +The heart of AutoGPT, and the project that kicked it all off: a semi-autonomous agent powered by LLMs to execute any task for you*. + +We continue to develop this project with the goal of providing access to AI assistance to the masses, and building the future transparently and together. + +- 💡 **Explore** - See what AI can do and be inspired by a glimpse of the future. + +- 🚀 **Build with us** - We welcome any input, whether it's code or ideas for new features or improvements! Join us on [Discord](https://discord.gg/autogpt) and find out how you can join in on the action. + +* it isn't quite there yet, but that is the ultimate goal that we are still pursuing + +--- + +## 🎯 Benchmark + +**[🗒️ Readme](https://github.com/Significant-Gravitas/AutoGPT/blob/master/classic/benchmark/README.md)** + +Measure your agent's performance! The `agbenchmark` can be used with any agent that supports the agent protocol, and the integration with the project's [CLI] makes it even easier to use with AutoGPT and forge-based agents. The benchmark offers a stringent testing environment. Our framework allows for autonomous, objective performance evaluations, ensuring your agents are primed for real-world action. + + + +- 📦 [**`agbenchmark`**](https://pypi.org/project/agbenchmark/) on Pypi + +- 🔌 **Agent Protocol Standardization** - AutoGPT uses the agent protocol from the AI Engineer Foundation to ensure compatibility with many agents, both from within and outside the project. + +--- + +## 🏗️ Forge + +**[📖 Introduction](../forge/get-started.md)** + |  +**[🚀 Quickstart](https://github.com/Significant-Gravitas/AutoGPT/blob/master/QUICKSTART.md)** + + + +Forge your own agent! The Forge is a ready-to-go template for your agent application. All the boilerplate code is already handled, letting you channel all your creativity into the things that set *your* agent apart. + +- 🛠️ **Building with Ease** - We've set the groundwork so you can focus on your agent's personality and capabilities. Comprehensive tutorials are available [here](https://aiedge.medium.com/autogpt-forge-e3de53cc58ec). + +--- + +## 💻 Frontend + +**[🗒️ Readme](https://github.com/Significant-Gravitas/AutoGPT/blob/master/classic/frontend/README.md)** + +An easy-to-use and open source frontend for any Agent Protocol-compliant agent. + +- 🎮 **User-Friendly Interface** - Manage your agents effortlessly. + +- 🔄 **Seamless Integration** - Smooth connectivity between your agent and our benchmarking system. + +--- + +## 🔧 CLI +[CLI]: #cli + +The project CLI makes it easy to use all of the components in the repo, separately or +together. To install its dependencies, simply run `./run setup`, and you're ready to go! + +```shell +$ ./run +Usage: cli.py [OPTIONS] COMMAND [ARGS]... + +Options: + --help Show this message and exit. + +Commands: + agent Commands to create, start and stop agents + benchmark Commands to start the benchmark and list tests and categories + setup Installs dependencies needed for your system. +``` + +Common commands: + +* `./run agent start autogpt` – [runs](./usage.md#serve-agent-protocol-mode-with-ui) the AutoGPT agent +* `./run agent create ` – creates a new Forge-based agent project at `agents/` +* `./run benchmark start ` – benchmarks the specified agent + +--- + +🤔 Join the AutoGPT Discord server for any queries: +[discord.gg/autogpt](https://discord.gg/autogpt) + +### Glossary of Terms + +- **Repository**: Space where your project resides. +- **Forking**: Copying a repository under your account. +- **Cloning**: Making a local copy of a repository. +- **Agent**: The AutoGPT you'll create and develop. +- **Benchmarking**: Testing your agent's skills in the Forge. +- **Forge**: The template for building your AutoGPT agent. +- **Frontend**: The UI for tasks, logs, and task history. diff --git a/docs/content/AutoGPT/setup/docker.md b/docs/content/classic/setup/docker.md similarity index 100% rename from docs/content/AutoGPT/setup/docker.md rename to docs/content/classic/setup/docker.md diff --git a/docs/content/AutoGPT/setup/for-developers.md b/docs/content/classic/setup/for-developers.md similarity index 100% rename from docs/content/AutoGPT/setup/for-developers.md rename to docs/content/classic/setup/for-developers.md diff --git a/docs/content/AutoGPT/setup/get-repo-dialog.png b/docs/content/classic/setup/get-repo-dialog.png similarity index 100% rename from docs/content/AutoGPT/setup/get-repo-dialog.png rename to docs/content/classic/setup/get-repo-dialog.png diff --git a/docs/content/AutoGPT/setup/index.md b/docs/content/classic/setup/index.md similarity index 100% rename from docs/content/AutoGPT/setup/index.md rename to docs/content/classic/setup/index.md diff --git a/docs/content/AutoGPT/share-your-logs.md b/docs/content/classic/share-your-logs.md similarity index 100% rename from docs/content/AutoGPT/share-your-logs.md rename to docs/content/classic/share-your-logs.md diff --git a/docs/content/AutoGPT/testing.md b/docs/content/classic/testing.md similarity index 100% rename from docs/content/AutoGPT/testing.md rename to docs/content/classic/testing.md diff --git a/docs/content/AutoGPT/usage.md b/docs/content/classic/usage.md similarity index 100% rename from docs/content/AutoGPT/usage.md rename to docs/content/classic/usage.md diff --git a/docs/content/docs/index.md b/docs/content/contribute/index.md similarity index 100% rename from docs/content/docs/index.md rename to docs/content/contribute/index.md diff --git a/docs/content/index.md b/docs/content/index.md index f04a1a158..243b56ea4 100644 --- a/docs/content/index.md +++ b/docs/content/index.md @@ -2,126 +2,84 @@ Welcome to the AutoGPT Documentation. -The AutoGPT project consists of four main components: +## What is the AutoGPT Platform? -- The [Server](#server) – known as the "AutoGPT Platform" -- The [Agent](#agent) – also known as just "AutoGPT" -- The [Benchmark](#benchmark) – AKA `agbenchmark` -- The [Forge](#forge) -- The [Frontend](#frontend) +The AutoGPT Platform is a groundbreaking system that revolutionizes AI utilization for businesses and individuals. It enables the creation, deployment, and management of continuous agents that work tirelessly on your behalf, bringing unprecedented efficiency and innovation to your workflows. -To tie these together, we also have a [CLI] at the root of the project. +### Key Features -## 🌐 Server +- **Seamless Integration and Low-Code Workflows**: Rapidly create complex workflows without extensive coding knowledge. +- **Autonomous Operation and Continuous Agents**: Deploy cloud-based assistants that run indefinitely, activating on relevant triggers. +- **Intelligent Automation and Maximum Efficiency**: Streamline workflows by automating repetitive processes. +- **Reliable Performance and Predictable Execution**: Enjoy consistent and dependable long-running processes. - +## Platform Architecture -**[📖 Setup](server/setup.md)** - |  -**[📖 Advanced Setup](server/advanced_setup.md)** - |  -**[📖 Making New Blocks](server/new_blocks.md)** +The AutoGPT Platform consists of two main components: -The server is the backbone of the New AutoGPT project. It provides the infrastructure for the agents to run, and the UI for you to interact with them. It integrates with the Forge, Agent, and a bespoke UI to provide a seamless experience. +### 1. AutoGPT Server ---- +The powerhouse of our platform, containing: -## 🤖 Agent +- **Source Code**: Core logic driving agents and automation processes. +- **Infrastructure**: Robust systems ensuring reliable and scalable performance. +- **Marketplace**: A comprehensive marketplace for pre-built agents. -**[📖 About AutoGPT](AutoGPT/index.md)** - |  -**[🔧 Setup](AutoGPT/setup/index.md)** - |  -**[💻 Usage](AutoGPT/usage.md)** +### 2. AutoGPT Frontend -The heart of AutoGPT, and the project that kicked it all off: a semi-autonomous agent powered by LLMs to execute any task for you*. +The user interface where you interact with the platform: -We continue to develop this project with the goal of providing access to AI assistance to the masses, and building the future transparently and together. +- **Agent Builder**: Design and configure your own AI agents. +- **Workflow Management**: Build, modify, and optimize automation workflows. +- **Deployment Controls**: Manage the lifecycle of your agents. +- **Ready-to-Use Agents**: Select from pre-configured agents. +- **Agent Interaction**: Run and interact with agents through a user-friendly interface. +- **Monitoring and Analytics**: Track agent performance and gain insights. -- 💡 **Explore** - See what AI can do and be inspired by a glimpse of the future. +## Platform Components -- 🚀 **Build with us** - We welcome any input, whether it's code or ideas for new features or improvements! Join us on [Discord](https://discord.gg/autogpt) and find out how you can join in on the action. +### Agents and Workflows -* it isn't quite there yet, but that is the ultimate goal that we are still pursuing +In the platform, you can create highly customized workflows to build agents. An agent is essentially an automated workflow that you design to perform specific tasks or processes. Create customized workflows to build agents for various tasks, including: ---- +- Data processing and analysis +- Task scheduling and management +- Communication and notification systems +- Integration between different software tools +- AI-powered decision making and content generation -## 🎯 Benchmark +### Blocks as Integrations -**[🗒️ Readme](https://github.com/Significant-Gravitas/AutoGPT/blob/master/classic/benchmark/README.md)** +Blocks represent actions and are the building blocks of your workflows, including: -Measure your agent's performance! The `agbenchmark` can be used with any agent that supports the agent protocol, and the integration with the project's [CLI] makes it even easier to use with AutoGPT and forge-based agents. The benchmark offers a stringent testing environment. Our framework allows for autonomous, objective performance evaluations, ensuring your agents are primed for real-world action. +- Connections to external services +- Data processing tools +- AI models for various tasks +- Custom scripts or functions +- Conditional logic and decision-making components - +You can learn more under: [Build your own Blocks](platform/new_blocks.md) -- 📦 [**`agbenchmark`**](https://pypi.org/project/agbenchmark/) on Pypi +## Available Language Models -- 🔌 **Agent Protocol Standardization** - AutoGPT uses the agent protocol from the AI Engineer Foundation to ensure compatibility with many agents, both from within and outside the project. +The platform comes pre-integrated with cutting-edge LLM providers: ---- +- OpenAI +- Anthropic +- Groq +- Llama -## 🏗️ Forge +## License Overview -**[📖 Introduction](forge/get-started.md)** - |  -**[🚀 Quickstart](https://github.com/Significant-Gravitas/AutoGPT/blob/master/QUICKSTART.md)** +We've adopted a dual-license approach to balance open collaboration with sustainable development: - +- **MIT License**: The majority of the AutoGPT repository remains under this license. +- **Polyform Shield License**: Applies to the new `autogpt_platform` folder. -Forge your own agent! The Forge is a ready-to-go template for your agent application. All the boilerplate code is already handled, letting you channel all your creativity into the things that set *your* agent apart. +This strategy allows us to share previously closed-source components, fostering a vibrant ecosystem of developers and users. -- 🛠️ **Building with Ease** - We've set the groundwork so you can focus on your agent's personality and capabilities. Comprehensive tutorials are available [here](https://aiedge.medium.com/autogpt-forge-e3de53cc58ec). +## Ready to Get Started? ---- - -## 💻 Frontend - -**[🗒️ Readme](https://github.com/Significant-Gravitas/AutoGPT/blob/master/classic/frontend/README.md)** - -An easy-to-use and open source frontend for any Agent Protocol-compliant agent. - -- 🎮 **User-Friendly Interface** - Manage your agents effortlessly. - -- 🔄 **Seamless Integration** - Smooth connectivity between your agent and our benchmarking system. - ---- - -## 🔧 CLI -[CLI]: #cli - -The project CLI makes it easy to use all of the components in the repo, separately or -together. To install its dependencies, simply run `./run setup`, and you're ready to go! - -```shell -$ ./run -Usage: cli.py [OPTIONS] COMMAND [ARGS]... - -Options: - --help Show this message and exit. - -Commands: - agent Commands to create, start and stop agents - benchmark Commands to start the benchmark and list tests and categories - setup Installs dependencies needed for your system. -``` - -Common commands: - -* `./run agent start autogpt` – [runs](./AutoGPT/usage.md#serve-agent-protocol-mode-with-ui) the AutoGPT agent -* `./run agent create ` – creates a new Forge-based agent project at `agents/` -* `./run benchmark start ` – benchmarks the specified agent - ---- - -🤔 Join the AutoGPT Discord server for any queries: -[discord.gg/autogpt](https://discord.gg/autogpt) - -### Glossary of Terms - -- **Repository**: Space where your project resides. -- **Forking**: Copying a repository under your account. -- **Cloning**: Making a local copy of a repository. -- **Agent**: The AutoGPT you'll create and develop. -- **Benchmarking**: Testing your agent's skills in the Forge. -- **Forge**: The template for building your AutoGPT agent. -- **Frontend**: The UI for tasks, logs, and task history. +- Read the [Getting Started docs](https://docs.agpt.co/platform/getting-started/) to self-host +- [Join the waitlist](https://agpt.co/waitlist) for the cloud-hosted beta +- [Contribute](contribute/index.md) \ No newline at end of file diff --git a/docs/content/server/advanced_setup.md b/docs/content/platform/advanced_setup.md similarity index 86% rename from docs/content/server/advanced_setup.md rename to docs/content/platform/advanced_setup.md index 74b63892f..9bb413b70 100644 --- a/docs/content/server/advanced_setup.md +++ b/docs/content/platform/advanced_setup.md @@ -1,10 +1,10 @@ # Advanced Setup -The advanced steps below are intended for people with sysadmin experience. If you are not comfortable with these steps, please refer to the [basic setup guide](setup.md). +The advanced steps below are intended for people with sysadmin experience. If you are not comfortable with these steps, please refer to the [basic setup guide](../platform/getting-started.md). ## Introduction -For the advanced setup, first follow the [basic setup guide](setup.md) to get the server up and running. Once you have the server running, you can follow the steps below to configure the server for your specific needs. +For the advanced setup, first follow the [basic setup guide](../platform/getting-started.md) to get the server up and running. Once you have the server running, you can follow the steps below to configure the server for your specific needs. ## Configuration diff --git a/docs/content/server/blocks/ai_shortform_video_block.md b/docs/content/platform/blocks/ai_shortform_video_block.md similarity index 100% rename from docs/content/server/blocks/ai_shortform_video_block.md rename to docs/content/platform/blocks/ai_shortform_video_block.md diff --git a/docs/content/server/blocks/basic.md b/docs/content/platform/blocks/basic.md similarity index 100% rename from docs/content/server/blocks/basic.md rename to docs/content/platform/blocks/basic.md diff --git a/docs/content/server/blocks/blocks.md b/docs/content/platform/blocks/blocks.md similarity index 100% rename from docs/content/server/blocks/blocks.md rename to docs/content/platform/blocks/blocks.md diff --git a/docs/content/server/blocks/branching.md b/docs/content/platform/blocks/branching.md similarity index 100% rename from docs/content/server/blocks/branching.md rename to docs/content/platform/blocks/branching.md diff --git a/docs/content/server/blocks/csv.md b/docs/content/platform/blocks/csv.md similarity index 100% rename from docs/content/server/blocks/csv.md rename to docs/content/platform/blocks/csv.md diff --git a/docs/content/server/blocks/decoder_block.md b/docs/content/platform/blocks/decoder_block.md similarity index 100% rename from docs/content/server/blocks/decoder_block.md rename to docs/content/platform/blocks/decoder_block.md diff --git a/docs/content/server/blocks/discord.md b/docs/content/platform/blocks/discord.md similarity index 100% rename from docs/content/server/blocks/discord.md rename to docs/content/platform/blocks/discord.md diff --git a/docs/content/server/blocks/email_block.md b/docs/content/platform/blocks/email_block.md similarity index 100% rename from docs/content/server/blocks/email_block.md rename to docs/content/platform/blocks/email_block.md diff --git a/docs/content/server/blocks/github/issues.md b/docs/content/platform/blocks/github/issues.md similarity index 100% rename from docs/content/server/blocks/github/issues.md rename to docs/content/platform/blocks/github/issues.md diff --git a/docs/content/server/blocks/github/pull_requests.md b/docs/content/platform/blocks/github/pull_requests.md similarity index 100% rename from docs/content/server/blocks/github/pull_requests.md rename to docs/content/platform/blocks/github/pull_requests.md diff --git a/docs/content/server/blocks/github/repo.md b/docs/content/platform/blocks/github/repo.md similarity index 100% rename from docs/content/server/blocks/github/repo.md rename to docs/content/platform/blocks/github/repo.md diff --git a/docs/content/server/blocks/google/gmail.md b/docs/content/platform/blocks/google/gmail.md similarity index 100% rename from docs/content/server/blocks/google/gmail.md rename to docs/content/platform/blocks/google/gmail.md diff --git a/docs/content/server/blocks/google/sheet.md b/docs/content/platform/blocks/google/sheet.md similarity index 100% rename from docs/content/server/blocks/google/sheet.md rename to docs/content/platform/blocks/google/sheet.md diff --git a/docs/content/server/blocks/google_maps.md b/docs/content/platform/blocks/google_maps.md similarity index 100% rename from docs/content/server/blocks/google_maps.md rename to docs/content/platform/blocks/google_maps.md diff --git a/docs/content/server/blocks/http.md b/docs/content/platform/blocks/http.md similarity index 100% rename from docs/content/server/blocks/http.md rename to docs/content/platform/blocks/http.md diff --git a/docs/content/server/blocks/ideogram.md b/docs/content/platform/blocks/ideogram.md similarity index 100% rename from docs/content/server/blocks/ideogram.md rename to docs/content/platform/blocks/ideogram.md diff --git a/docs/content/server/blocks/iteration.md b/docs/content/platform/blocks/iteration.md similarity index 100% rename from docs/content/server/blocks/iteration.md rename to docs/content/platform/blocks/iteration.md diff --git a/docs/content/server/blocks/llm.md b/docs/content/platform/blocks/llm.md similarity index 100% rename from docs/content/server/blocks/llm.md rename to docs/content/platform/blocks/llm.md diff --git a/docs/content/server/blocks/maths.md b/docs/content/platform/blocks/maths.md similarity index 100% rename from docs/content/server/blocks/maths.md rename to docs/content/platform/blocks/maths.md diff --git a/docs/content/server/blocks/medium.md b/docs/content/platform/blocks/medium.md similarity index 100% rename from docs/content/server/blocks/medium.md rename to docs/content/platform/blocks/medium.md diff --git a/docs/content/server/blocks/reddit.md b/docs/content/platform/blocks/reddit.md similarity index 100% rename from docs/content/server/blocks/reddit.md rename to docs/content/platform/blocks/reddit.md diff --git a/docs/content/server/blocks/replicate_flux_advanced.md b/docs/content/platform/blocks/replicate_flux_advanced.md similarity index 100% rename from docs/content/server/blocks/replicate_flux_advanced.md rename to docs/content/platform/blocks/replicate_flux_advanced.md diff --git a/docs/content/server/blocks/rss.md b/docs/content/platform/blocks/rss.md similarity index 100% rename from docs/content/server/blocks/rss.md rename to docs/content/platform/blocks/rss.md diff --git a/docs/content/server/blocks/sampling.md b/docs/content/platform/blocks/sampling.md similarity index 100% rename from docs/content/server/blocks/sampling.md rename to docs/content/platform/blocks/sampling.md diff --git a/docs/content/server/blocks/search.md b/docs/content/platform/blocks/search.md similarity index 100% rename from docs/content/server/blocks/search.md rename to docs/content/platform/blocks/search.md diff --git a/docs/content/server/blocks/talking_head.md b/docs/content/platform/blocks/talking_head.md similarity index 100% rename from docs/content/server/blocks/talking_head.md rename to docs/content/platform/blocks/talking_head.md diff --git a/docs/content/server/blocks/text.md b/docs/content/platform/blocks/text.md similarity index 100% rename from docs/content/server/blocks/text.md rename to docs/content/platform/blocks/text.md diff --git a/docs/content/server/blocks/text_to_speech_block.md b/docs/content/platform/blocks/text_to_speech_block.md similarity index 100% rename from docs/content/server/blocks/text_to_speech_block.md rename to docs/content/platform/blocks/text_to_speech_block.md diff --git a/docs/content/server/blocks/time_blocks.md b/docs/content/platform/blocks/time_blocks.md similarity index 100% rename from docs/content/server/blocks/time_blocks.md rename to docs/content/platform/blocks/time_blocks.md diff --git a/docs/content/server/blocks/youtube.md b/docs/content/platform/blocks/youtube.md similarity index 100% rename from docs/content/server/blocks/youtube.md rename to docs/content/platform/blocks/youtube.md diff --git a/docs/content/server/d_id.md b/docs/content/platform/d_id.md similarity index 100% rename from docs/content/server/d_id.md rename to docs/content/platform/d_id.md diff --git a/docs/content/server/setup.md b/docs/content/platform/getting-started.md similarity index 83% rename from docs/content/server/setup.md rename to docs/content/platform/getting-started.md index d6ffa7842..bbb9278bc 100644 --- a/docs/content/server/setup.md +++ b/docs/content/platform/getting-started.md @@ -1,7 +1,8 @@ -# Setting up the server +# Getting Started with AutoGPT: Self-Hosting Guide -- [Introduction](#introduction) -- [Prerequisites](#prerequisites) +This tutorial will walk you through the process of setting up AutoGPT locally on your machine. + +
## Introduction @@ -22,7 +23,7 @@ To setup the server, you need to have the following installed: - [Docker](https://docs.docker.com/get-docker/) - [Git](https://git-scm.com/downloads) -### Checking if you have Node.js & NPM installed +#### Checking if you have Node.js & NPM installed We use Node.js to run our frontend application. @@ -41,7 +42,7 @@ npm -v Once you have Node.js installed, you can proceed to the next step. -### Checking if you have Docker & Docker Compose installed +#### Checking if you have Docker & Docker Compose installed Docker containerizes applications, while Docker Compose orchestrates multi-container Docker applications. @@ -60,7 +61,7 @@ docker-compose -v Once you have Docker and Docker Compose installed, you can proceed to the next step. -## Cloning the Repository +### Cloning the Repository The first step is cloning the AutoGPT repository to your computer. To do this, open a terminal window in a folder on your computer and run: ``` @@ -70,7 +71,7 @@ If you get stuck, follow [this guide](https://docs.github.com/en/repositories/cr Once that's complete you can close this terminal window. -## Running the backend services +### Running the backend services To run the backend services, follow these steps: @@ -94,7 +95,7 @@ To run the backend services, follow these steps: This command will start all the necessary backend services defined in the `docker-compose.combined.yml` file in detached mode. -## Running the frontend application +### Running the frontend application To run the frontend application, follow these steps: @@ -116,11 +117,12 @@ To run the frontend application, follow these steps: ``` This command will install the necessary dependencies and start the frontend application in development mode. -## Checking if the application is running +### Checking if the application is running You can check if the server is running by visiting [http://localhost:3000](http://localhost:3000) in your browser. -### Notes: +**Notes:** + By default the application for different services run on the following ports: Frontend UI Server: 3000 diff --git a/docs/content/server/new_blocks.md b/docs/content/platform/new_blocks.md similarity index 98% rename from docs/content/server/new_blocks.md rename to docs/content/platform/new_blocks.md index 1c7c3684c..32ccd5b74 100644 --- a/docs/content/server/new_blocks.md +++ b/docs/content/platform/new_blocks.md @@ -1,432 +1,432 @@ -# Contributing to AutoGPT Agent Server: Creating and Testing Blocks - -This guide will walk you through the process of creating and testing a new block for the AutoGPT Agent Server, using the WikipediaSummaryBlock as an example. - -## Understanding Blocks and Testing - -Blocks are reusable components that can be connected to form a graph representing an agent's behavior. Each block has inputs, outputs, and a specific function. Proper testing is crucial to ensure blocks work correctly and consistently. - -## Creating and Testing a New Block - -Follow these steps to create and test a new block: - -1. **Create a new Python file** in the `backend/blocks` directory. Name it descriptively and use snake_case. For example: `get_wikipedia_summary.py`. - -2. **Import necessary modules and create a class that inherits from `Block`**. Make sure to include all necessary imports for your block. - - Every block should contain the following: - - ```python - from backend.data.block import Block, BlockSchema, BlockOutput - ``` - - Example for the Wikipedia summary block: - - ```python - from backend.data.block import Block, BlockSchema, BlockOutput - from backend.utils.get_request import GetRequest - import requests - - class WikipediaSummaryBlock(Block, GetRequest): - # Block implementation will go here - ``` - -3. **Define the input and output schemas** using `BlockSchema`. These schemas specify the data structure that the block expects to receive (input) and produce (output). - - - The input schema defines the structure of the data the block will process. Each field in the schema represents a required piece of input data. - - The output schema defines the structure of the data the block will return after processing. Each field in the schema represents a piece of output data. - - Example: - - ```python - class Input(BlockSchema): - topic: str # The topic to get the Wikipedia summary for - - class Output(BlockSchema): - summary: str # The summary of the topic from Wikipedia - error: str # Any error message if the request fails, error field needs to be named `error`. - ``` - -4. **Implement the `__init__` method, including test data and mocks:** - - !!! important - Use UUID generator (e.g. https://www.uuidgenerator.net/) for every new block `id` and *do not* make up your own. Alternatively, you can run this python code to generate an uuid: `print(__import__('uuid').uuid4())` - - ```python - def __init__(self): - super().__init__( - # Unique ID for the block, used across users for templates - # If you are an AI leave it as is or change to "generate-proper-uuid" - id="xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx", - input_schema=WikipediaSummaryBlock.Input, # Assign input schema - output_schema=WikipediaSummaryBlock.Output, # Assign output schema - - # Provide sample input, output and test mock for testing the block - - test_input={"topic": "Artificial Intelligence"}, - test_output=("summary", "summary content"), - test_mock={"get_request": lambda url, json: {"extract": "summary content"}}, - ) - ``` - - - `id`: A unique identifier for the block. - - - `input_schema` and `output_schema`: Define the structure of the input and output data. - - Let's break down the testing components: - - - `test_input`: This is a sample input that will be used to test the block. It should be a valid input according to your Input schema. - - - `test_output`: This is the expected output when running the block with the `test_input`. It should match your Output schema. For non-deterministic outputs or when you only want to assert the type, you can use Python types instead of specific values. In this example, `("summary", str)` asserts that the output key is "summary" and its value is a string. - - - `test_mock`: This is crucial for blocks that make network calls. It provides a mock function that replaces the actual network call during testing. - - In this case, we're mocking the `get_request` method to always return a dictionary with an 'extract' key, simulating a successful API response. This allows us to test the block's logic without making actual network requests, which could be slow, unreliable, or rate-limited. - -5. **Implement the `run` method with error handling:**, this should contain the main logic of the block: - - ```python - def run(self, input_data: Input, **kwargs) -> BlockOutput: - try: - topic = input_data.topic - url = f"https://en.wikipedia.org/api/rest_v1/page/summary/{topic}" - - response = self.get_request(url, json=True) - yield "summary", response['extract'] - - except requests.exceptions.HTTPError as http_err: - raise RuntimeError(f"HTTP error occurred: {http_err}") - ``` - - - **Try block**: Contains the main logic to fetch and process the Wikipedia summary. - - **API request**: Send a GET request to the Wikipedia API. - - **Error handling**: Handle various exceptions that might occur during the API request and data processing. We don't need to catch all exceptions, only the ones we expect and can handle. The uncaught exceptions will be automatically yielded as `error` in the output. Any block that raises an exception (or yields an `error` output) will be marked as failed. Prefer raising exceptions over yielding `error`, as it will stop the execution immediately. - - **Yield**: Use `yield` to output the results. Prefer to output one result object at a time. If you are calling a function that returns a list, you can yield each item in the list separately. You can also yield the whole list as well, but do both rather than yielding the list. For example: If you were writing a block that outputs emails, you'd yield each email as a separate result object, but you could also yield the whole list as an additional single result object. Yielding output named `error` will break the execution right away and mark the block execution as failed. - -### Blocks with authentication - -Our system supports auth offloading for API keys and OAuth2 authorization flows. -Adding a block with API key authentication is straight-forward, as is adding a block -for a service that we already have OAuth2 support for. - -Implementing the block itself is relatively simple. On top of the instructions above, -you're going to add a `credentials` parameter to the `Input` model and the `run` method: - -```python -from autogpt_libs.supabase_integration_credentials_store.types import ( - APIKeyCredentials, - OAuth2Credentials, - Credentials, -) - -from backend.data.block import Block, BlockOutput, BlockSchema -from backend.data.model import CredentialsField - - -# API Key auth: -class BlockWithAPIKeyAuth(Block): - class Input(BlockSchema): - # Note that the type hint below is require or you will get a type error. - # The first argument is the provider name, the second is the credential type. - credentials: CredentialsMetaInput[Literal['github'], Literal['api_key']] = CredentialsField( - provider="github", - supported_credential_types={"api_key"}, - description="The GitHub integration can be used with " - "any API key with sufficient permissions for the blocks it is used on.", - ) - - # ... - - def run( - self, - input_data: Input, - *, - credentials: APIKeyCredentials, - **kwargs, - ) -> BlockOutput: - ... - -# OAuth: -class BlockWithOAuth(Block): - class Input(BlockSchema): - # Note that the type hint below is require or you will get a type error. - # The first argument is the provider name, the second is the credential type. - credentials: CredentialsMetaInput[Literal['github'], Literal['oauth2']] = CredentialsField( - provider="github", - supported_credential_types={"oauth2"}, - required_scopes={"repo"}, - description="The GitHub integration can be used with OAuth.", - ) - - # ... - - def run( - self, - input_data: Input, - *, - credentials: OAuth2Credentials, - **kwargs, - ) -> BlockOutput: - ... - -# API Key auth + OAuth: -class BlockWithAPIKeyAndOAuth(Block): - class Input(BlockSchema): - # Note that the type hint below is require or you will get a type error. - # The first argument is the provider name, the second is the credential type. - credentials: CredentialsMetaInput[Literal['github'], Literal['api_key', 'oauth2']] = CredentialsField( - provider="github", - supported_credential_types={"api_key", "oauth2"}, - required_scopes={"repo"}, - description="The GitHub integration can be used with OAuth, " - "or any API key with sufficient permissions for the blocks it is used on.", - ) - - # ... - - def run( - self, - input_data: Input, - *, - credentials: Credentials, - **kwargs, - ) -> BlockOutput: - ... -``` - -The credentials will be automagically injected by the executor in the back end. - -The `APIKeyCredentials` and `OAuth2Credentials` models are defined [here](https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/autogpt_libs/autogpt_libs/supabase_integration_credentials_store/types.py). -To use them in e.g. an API request, you can either access the token directly: - -```python -# credentials: APIKeyCredentials -response = requests.post( - url, - headers={ - "Authorization": f"Bearer {credentials.api_key.get_secret_value()})", - }, -) - -# credentials: OAuth2Credentials -response = requests.post( - url, - headers={ - "Authorization": f"Bearer {credentials.access_token.get_secret_value()})", - }, -) -``` - -or use the shortcut `credentials.bearer()`: - -```python -# credentials: APIKeyCredentials | OAuth2Credentials -response = requests.post( - url, - headers={"Authorization": credentials.bearer()}, -) -``` - -#### Adding an OAuth2 service integration - -To add support for a new OAuth2-authenticated service, you'll need to add an `OAuthHandler`. -All our existing handlers and the base class can be found [here][OAuth2 handlers]. - -Every handler must implement the following parts of the [`BaseOAuthHandler`] interface: - -```python title="autogpt_platform/backend/backend/integrations/oauth/base.py" ---8<-- "autogpt_platform/backend/backend/integrations/oauth/base.py:BaseOAuthHandler1" ---8<-- "autogpt_platform/backend/backend/integrations/oauth/base.py:BaseOAuthHandler2" ---8<-- "autogpt_platform/backend/backend/integrations/oauth/base.py:BaseOAuthHandler3" ---8<-- "autogpt_platform/backend/backend/integrations/oauth/base.py:BaseOAuthHandler4" ---8<-- "autogpt_platform/backend/backend/integrations/oauth/base.py:BaseOAuthHandler5" ---8<-- "autogpt_platform/backend/backend/integrations/oauth/base.py:BaseOAuthHandler6" -``` - -As you can see, this is modeled after the standard OAuth2 flow. - -Aside from implementing the `OAuthHandler` itself, adding a handler into the system requires two more things: - -- Adding the handler class to `HANDLERS_BY_NAME` under [`integrations/oauth/__init__.py`](https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/backend/backend/integrations/oauth/__init__.py) - -```python title="autogpt_platform/backend/backend/integrations/oauth/__init__.py" ---8<-- "autogpt_platform/backend/backend/integrations/oauth/__init__.py:HANDLERS_BY_NAMEExample" -``` - -- Adding `{provider}_client_id` and `{provider}_client_secret` to the application's `Secrets` under [`util/settings.py`](https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/backend/backend/util/settings.py) - -```python title="autogpt_platform/backend/backend/util/settings.py" ---8<-- "autogpt_platform/backend/backend/util/settings.py:OAuthServerCredentialsExample" -``` - -[OAuth2 handlers]: https://github.com/Significant-Gravitas/AutoGPT/tree/master/autogpt_platform/backend/backend/integrations/oauth -[`BaseOAuthHandler`]: https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/backend/backend/integrations/oauth/base.py - -#### Adding to the frontend - -You will need to add the provider (api or oauth) to the `CredentialsInput` component in [`frontend/src/components/integrations/credentials-input.tsx`](https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/frontend/src/components/integrations/credentials-input.tsx). - -```ts title="frontend/src/components/integrations/credentials-input.tsx" ---8<-- "autogpt_platform/frontend/src/components/integrations/credentials-input.tsx:ProviderIconsEmbed" -``` - -You will also need to add the provider to the `CredentialsProvider` component in [`frontend/src/components/integrations/credentials-provider.tsx`](https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/frontend/src/components/integrations/credentials-provider.tsx). - -```ts title="frontend/src/components/integrations/credentials-provider.tsx" ---8<-- "autogpt_platform/frontend/src/components/integrations/credentials-provider.tsx:CredentialsProviderNames" -``` - -Finally you will need to add the provider to the `CredentialsType` enum in [`frontend/src/lib/autogpt-server-api/types.ts`](https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/frontend/src/lib/autogpt-server-api/types.ts). - -```ts title="frontend/src/lib/autogpt-server-api/types.ts" ---8<-- "autogpt_platform/frontend/src/lib/autogpt-server-api/types.ts:BlockIOCredentialsSubSchema" -``` - -#### Example: GitHub integration - -- GitHub blocks with API key + OAuth2 support: [`blocks/github`](https://github.com/Significant-Gravitas/AutoGPT/tree/master/autogpt_platform/backend/backend/blocks/github/) - -```python title="blocks/github/issues.py" ---8<-- "autogpt_platform/backend/backend/blocks/github/issues.py:GithubCommentBlockExample" -``` - -- GitHub OAuth2 handler: [`integrations/oauth/github.py`](https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/backend/backend/integrations/oauth/github.py) - -```python title="blocks/github/github.py" ---8<-- "autogpt_platform/backend/backend/integrations/oauth/github.py:GithubOAuthHandlerExample" -``` - -#### Example: Google integration - -- Google OAuth2 handler: [`integrations/oauth/google.py`](https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/backend/backend/integrations/oauth/google.py) - -```python title="integrations/oauth/google.py" ---8<-- "autogpt_platform/backend/backend/integrations/oauth/google.py:GoogleOAuthHandlerExample" -``` - -You can see that google has defined a `DEFAULT_SCOPES` variable, this is used to set the scopes that are requested no matter what the user asks for. - -```python title="blocks/google/_auth.py" ---8<-- "autogpt_platform/backend/backend/blocks/google/_auth.py:GoogleOAuthIsConfigured" -``` - -You can also see that `GOOGLE_OAUTH_IS_CONFIGURED` is used to disable the blocks that require OAuth if the oauth is not configured. This is in the `__init__` method of each block. This is because there is no api key fallback for google blocks so we need to make sure that the oauth is configured before we allow the user to use the blocks. - -## Key Points to Remember - -- **Unique ID**: Give your block a unique ID in the **init** method. -- **Input and Output Schemas**: Define clear input and output schemas. -- **Error Handling**: Implement error handling in the `run` method. -- **Output Results**: Use `yield` to output results in the `run` method. -- **Testing**: Provide test input and output in the **init** method for automatic testing. - -## Understanding the Testing Process - -The testing of blocks is handled by `test_block.py`, which does the following: - -1. It calls the block with the provided `test_input`. - If the block has a `credentials` field, `test_credentials` is passed in as well. -2. If a `test_mock` is provided, it temporarily replaces the specified methods with the mock functions. -3. It then asserts that the output matches the `test_output`. - -For the WikipediaSummaryBlock: - -- The test will call the block with the topic "Artificial Intelligence". -- Instead of making a real API call, it will use the mock function, which returns `{"extract": "summary content"}`. -- It will then check if the output key is "summary" and its value is a string. - -This approach allows us to test the block's logic comprehensively without relying on external services, while also accommodating non-deterministic outputs. - -## Tips for Effective Block Testing - -1. **Provide realistic test_input**: Ensure your test input covers typical use cases. - -2. **Define appropriate test_output**: - - - For deterministic outputs, use specific expected values. - - For non-deterministic outputs or when only the type matters, use Python types (e.g., `str`, `int`, `dict`). - - You can mix specific values and types, e.g., `("key1", str), ("key2", 42)`. - -3. **Use test_mock for network calls**: This prevents tests from failing due to network issues or API changes. - -4. **Consider omitting test_mock for blocks without external dependencies**: If your block doesn't make network calls or use external resources, you might not need a mock. - -5. **Consider edge cases**: Include tests for potential error conditions in your `run` method. - -6. **Update tests when changing block behavior**: If you modify your block, ensure the tests are updated accordingly. - -By following these steps, you can create new blocks that extend the functionality of the AutoGPT Agent Server. - -## Blocks we want to see - -Below is a list of blocks that we would like to see implemented in the AutoGPT Agent Server. If you're interested in contributing, feel free to pick one of these blocks or chose your own. - -If you would like to implement one of these blocks, open a pull request and we will start the review process. - -### Consumer Services/Platforms - -- Google sheets - [~~Read/Append~~](https://github.com/Significant-Gravitas/AutoGPT/pull/8236) -- Email - Read/Send with [~~Gmail~~](https://github.com/Significant-Gravitas/AutoGPT/pull/8236), Outlook, Yahoo, Proton, etc -- Calendar - Read/Write with Google Calendar, Outlook Calendar, etc -- Home Assistant - Call Service, Get Status -- Dominos - Order Pizza, Track Order -- Uber - Book Ride, Track Ride -- Notion - Create/Read Page, Create/Append/Read DB -- Google drive - read/write/overwrite file/folder - -### Social Media - -- Twitter - Post, Reply, Get Replies, Get Comments, Get Followers, Get Following, Get Tweets, Get Mentions -- Instagram - Post, Reply, Get Comments, Get Followers, Get Following, Get Posts, Get Mentions, Get Trending Posts -- TikTok - Post, Reply, Get Comments, Get Followers, Get Following, Get Videos, Get Mentions, Get Trending Videos -- LinkedIn - Post, Reply, Get Comments, Get Followers, Get Following, Get Posts, Get Mentions, Get Trending Posts -- YouTube - Transcribe Videos/Shorts, Post Videos/Shorts, Read/Reply/React to Comments, Update Thumbnails, Update Description, Update Tags, Update Titles, Get Views, Get Likes, Get Dislikes, Get Subscribers, Get Comments, Get Shares, Get Watch Time, Get Revenue, Get Trending Videos, Get Top Videos, Get Top Channels -- Reddit - Post, Reply, Get Comments, Get Followers, Get Following, Get Posts, Get Mentions, Get Trending Posts -- Treatwell (and related Platforms) - Book, Cancel, Review, Get Recommendations -- Substack - Read/Subscribe/Unsubscribe, Post/Reply, Get Recommendations -- Discord - Read/Post/Reply, Moderation actions -- GoodReads - Read/Post/Reply, Get Recommendations - -### E-commerce - -- Airbnb - Book, Cancel, Review, Get Recommendations -- Amazon - Order, Track Order, Return, Review, Get Recommendations -- eBay - Order, Track Order, Return, Review, Get Recommendations -- Upwork - Post Jobs, Hire Freelancer, Review Freelancer, Fire Freelancer - -### Business Tools - -- External Agents - Call other agents similar to AutoGPT -- Trello - Create/Read/Update/Delete Cards, Lists, Boards -- Jira - Create/Read/Update/Delete Issues, Projects, Boards -- Linear - Create/Read/Update/Delete Issues, Projects, Boards -- Excel - Read/Write/Update/Delete Rows, Columns, Sheets -- Slack - Read/Post/Reply to Messages, Create Channels, Invite Users -- ERPNext - Create/Read/Update/Delete Invoices, Orders, Customers, Products -- Salesforce - Create/Read/Update/Delete Leads, Opportunities, Accounts -- HubSpot - Create/Read/Update/Delete Contacts, Deals, Companies -- Zendesk - Create/Read/Update/Delete Tickets, Users, Organizations -- Odoo - Create/Read/Update/Delete Sales Orders, Invoices, Customers -- Shopify - Create/Read/Update/Delete Products, Orders, Customers -- WooCommerce - Create/Read/Update/Delete Products, Orders, Customers -- Squarespace - Create/Read/Update/Delete Pages, Products, Orders - -## Agent Templates we want to see - -### Data/Information - -- Summarize top news of today, of this week, this month via Apple News or other large media outlets BBC, TechCrunch, hackernews, etc -- Create, read, and summarize substack newsletters or any newsletters (blog writer vs blog reader) -- Get/read/summarize the most viral Twitter, Instagram, TikTok (general social media accounts) of the day, week, month -- Get/Read any LinkedIn posts or profile that mention AI Agents -- Read/Summarize discord (might not be able to do this because you need access) -- Read / Get most read books in a given month, year, etc from GoodReads or Amazon Books, etc -- Get dates for specific shows across all streaming services - - Suggest/Recommend/Get most watched shows in a given month, year, etc across all streaming platforms -- Data analysis from xlsx data set - - Gather via Excel or Google Sheets data > Sample the data randomly (sample block takes top X, bottom X, randomly, etc) > pass that to LLM Block to generate a script for analysis of the full data > Python block to run the script> making a loop back through LLM Fix Block on error > create chart/visualization (potentially in the code block?) > show the image as output (this may require frontend changes to show) -- Tiktok video search and download - -### Marketing - -- Portfolio site design and enhancements +# Contributing to AutoGPT Agent Server: Creating and Testing Blocks + +This guide will walk you through the process of creating and testing a new block for the AutoGPT Agent Server, using the WikipediaSummaryBlock as an example. + +## Understanding Blocks and Testing + +Blocks are reusable components that can be connected to form a graph representing an agent's behavior. Each block has inputs, outputs, and a specific function. Proper testing is crucial to ensure blocks work correctly and consistently. + +## Creating and Testing a New Block + +Follow these steps to create and test a new block: + +1. **Create a new Python file** in the `backend/blocks` directory. Name it descriptively and use snake_case. For example: `get_wikipedia_summary.py`. + +2. **Import necessary modules and create a class that inherits from `Block`**. Make sure to include all necessary imports for your block. + + Every block should contain the following: + + ```python + from backend.data.block import Block, BlockSchema, BlockOutput + ``` + + Example for the Wikipedia summary block: + + ```python + from backend.data.block import Block, BlockSchema, BlockOutput + from backend.utils.get_request import GetRequest + import requests + + class WikipediaSummaryBlock(Block, GetRequest): + # Block implementation will go here + ``` + +3. **Define the input and output schemas** using `BlockSchema`. These schemas specify the data structure that the block expects to receive (input) and produce (output). + + - The input schema defines the structure of the data the block will process. Each field in the schema represents a required piece of input data. + - The output schema defines the structure of the data the block will return after processing. Each field in the schema represents a piece of output data. + + Example: + + ```python + class Input(BlockSchema): + topic: str # The topic to get the Wikipedia summary for + + class Output(BlockSchema): + summary: str # The summary of the topic from Wikipedia + error: str # Any error message if the request fails, error field needs to be named `error`. + ``` + +4. **Implement the `__init__` method, including test data and mocks:** + + !!! important + Use UUID generator (e.g. https://www.uuidgenerator.net/) for every new block `id` and *do not* make up your own. Alternatively, you can run this python code to generate an uuid: `print(__import__('uuid').uuid4())` + + ```python + def __init__(self): + super().__init__( + # Unique ID for the block, used across users for templates + # If you are an AI leave it as is or change to "generate-proper-uuid" + id="xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx", + input_schema=WikipediaSummaryBlock.Input, # Assign input schema + output_schema=WikipediaSummaryBlock.Output, # Assign output schema + + # Provide sample input, output and test mock for testing the block + + test_input={"topic": "Artificial Intelligence"}, + test_output=("summary", "summary content"), + test_mock={"get_request": lambda url, json: {"extract": "summary content"}}, + ) + ``` + + - `id`: A unique identifier for the block. + + - `input_schema` and `output_schema`: Define the structure of the input and output data. + + Let's break down the testing components: + + - `test_input`: This is a sample input that will be used to test the block. It should be a valid input according to your Input schema. + + - `test_output`: This is the expected output when running the block with the `test_input`. It should match your Output schema. For non-deterministic outputs or when you only want to assert the type, you can use Python types instead of specific values. In this example, `("summary", str)` asserts that the output key is "summary" and its value is a string. + + - `test_mock`: This is crucial for blocks that make network calls. It provides a mock function that replaces the actual network call during testing. + + In this case, we're mocking the `get_request` method to always return a dictionary with an 'extract' key, simulating a successful API response. This allows us to test the block's logic without making actual network requests, which could be slow, unreliable, or rate-limited. + +5. **Implement the `run` method with error handling:**, this should contain the main logic of the block: + + ```python + def run(self, input_data: Input, **kwargs) -> BlockOutput: + try: + topic = input_data.topic + url = f"https://en.wikipedia.org/api/rest_v1/page/summary/{topic}" + + response = self.get_request(url, json=True) + yield "summary", response['extract'] + + except requests.exceptions.HTTPError as http_err: + raise RuntimeError(f"HTTP error occurred: {http_err}") + ``` + + - **Try block**: Contains the main logic to fetch and process the Wikipedia summary. + - **API request**: Send a GET request to the Wikipedia API. + - **Error handling**: Handle various exceptions that might occur during the API request and data processing. We don't need to catch all exceptions, only the ones we expect and can handle. The uncaught exceptions will be automatically yielded as `error` in the output. Any block that raises an exception (or yields an `error` output) will be marked as failed. Prefer raising exceptions over yielding `error`, as it will stop the execution immediately. + - **Yield**: Use `yield` to output the results. Prefer to output one result object at a time. If you are calling a function that returns a list, you can yield each item in the list separately. You can also yield the whole list as well, but do both rather than yielding the list. For example: If you were writing a block that outputs emails, you'd yield each email as a separate result object, but you could also yield the whole list as an additional single result object. Yielding output named `error` will break the execution right away and mark the block execution as failed. + +### Blocks with authentication + +Our system supports auth offloading for API keys and OAuth2 authorization flows. +Adding a block with API key authentication is straight-forward, as is adding a block +for a service that we already have OAuth2 support for. + +Implementing the block itself is relatively simple. On top of the instructions above, +you're going to add a `credentials` parameter to the `Input` model and the `run` method: + +```python +from autogpt_libs.supabase_integration_credentials_store.types import ( + APIKeyCredentials, + OAuth2Credentials, + Credentials, +) + +from backend.data.block import Block, BlockOutput, BlockSchema +from backend.data.model import CredentialsField + + +# API Key auth: +class BlockWithAPIKeyAuth(Block): + class Input(BlockSchema): + # Note that the type hint below is require or you will get a type error. + # The first argument is the provider name, the second is the credential type. + credentials: CredentialsMetaInput[Literal['github'], Literal['api_key']] = CredentialsField( + provider="github", + supported_credential_types={"api_key"}, + description="The GitHub integration can be used with " + "any API key with sufficient permissions for the blocks it is used on.", + ) + + # ... + + def run( + self, + input_data: Input, + *, + credentials: APIKeyCredentials, + **kwargs, + ) -> BlockOutput: + ... + +# OAuth: +class BlockWithOAuth(Block): + class Input(BlockSchema): + # Note that the type hint below is require or you will get a type error. + # The first argument is the provider name, the second is the credential type. + credentials: CredentialsMetaInput[Literal['github'], Literal['oauth2']] = CredentialsField( + provider="github", + supported_credential_types={"oauth2"}, + required_scopes={"repo"}, + description="The GitHub integration can be used with OAuth.", + ) + + # ... + + def run( + self, + input_data: Input, + *, + credentials: OAuth2Credentials, + **kwargs, + ) -> BlockOutput: + ... + +# API Key auth + OAuth: +class BlockWithAPIKeyAndOAuth(Block): + class Input(BlockSchema): + # Note that the type hint below is require or you will get a type error. + # The first argument is the provider name, the second is the credential type. + credentials: CredentialsMetaInput[Literal['github'], Literal['api_key', 'oauth2']] = CredentialsField( + provider="github", + supported_credential_types={"api_key", "oauth2"}, + required_scopes={"repo"}, + description="The GitHub integration can be used with OAuth, " + "or any API key with sufficient permissions for the blocks it is used on.", + ) + + # ... + + def run( + self, + input_data: Input, + *, + credentials: Credentials, + **kwargs, + ) -> BlockOutput: + ... +``` + +The credentials will be automagically injected by the executor in the back end. + +The `APIKeyCredentials` and `OAuth2Credentials` models are defined [here](https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/autogpt_libs/autogpt_libs/supabase_integration_credentials_store/types.py). +To use them in e.g. an API request, you can either access the token directly: + +```python +# credentials: APIKeyCredentials +response = requests.post( + url, + headers={ + "Authorization": f"Bearer {credentials.api_key.get_secret_value()})", + }, +) + +# credentials: OAuth2Credentials +response = requests.post( + url, + headers={ + "Authorization": f"Bearer {credentials.access_token.get_secret_value()})", + }, +) +``` + +or use the shortcut `credentials.bearer()`: + +```python +# credentials: APIKeyCredentials | OAuth2Credentials +response = requests.post( + url, + headers={"Authorization": credentials.bearer()}, +) +``` + +#### Adding an OAuth2 service integration + +To add support for a new OAuth2-authenticated service, you'll need to add an `OAuthHandler`. +All our existing handlers and the base class can be found [here][OAuth2 handlers]. + +Every handler must implement the following parts of the [`BaseOAuthHandler`] interface: + +```python title="autogpt_platform/backend/backend/integrations/oauth/base.py" +--8<-- "autogpt_platform/backend/backend/integrations/oauth/base.py:BaseOAuthHandler1" +--8<-- "autogpt_platform/backend/backend/integrations/oauth/base.py:BaseOAuthHandler2" +--8<-- "autogpt_platform/backend/backend/integrations/oauth/base.py:BaseOAuthHandler3" +--8<-- "autogpt_platform/backend/backend/integrations/oauth/base.py:BaseOAuthHandler4" +--8<-- "autogpt_platform/backend/backend/integrations/oauth/base.py:BaseOAuthHandler5" +--8<-- "autogpt_platform/backend/backend/integrations/oauth/base.py:BaseOAuthHandler6" +``` + +As you can see, this is modeled after the standard OAuth2 flow. + +Aside from implementing the `OAuthHandler` itself, adding a handler into the system requires two more things: + +- Adding the handler class to `HANDLERS_BY_NAME` under [`integrations/oauth/__init__.py`](https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/backend/backend/integrations/oauth/__init__.py) + +```python title="autogpt_platform/backend/backend/integrations/oauth/__init__.py" +--8<-- "autogpt_platform/backend/backend/integrations/oauth/__init__.py:HANDLERS_BY_NAMEExample" +``` + +- Adding `{provider}_client_id` and `{provider}_client_secret` to the application's `Secrets` under [`util/settings.py`](https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/backend/backend/util/settings.py) + +```python title="autogpt_platform/backend/backend/util/settings.py" +--8<-- "autogpt_platform/backend/backend/util/settings.py:OAuthServerCredentialsExample" +``` + +[OAuth2 handlers]: https://github.com/Significant-Gravitas/AutoGPT/tree/master/autogpt_platform/backend/backend/integrations/oauth +[`BaseOAuthHandler`]: https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/backend/backend/integrations/oauth/base.py + +#### Adding to the frontend + +You will need to add the provider (api or oauth) to the `CredentialsInput` component in [`frontend/src/components/integrations/credentials-input.tsx`](https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/frontend/src/components/integrations/credentials-input.tsx). + +```ts title="frontend/src/components/integrations/credentials-input.tsx" +--8<-- "autogpt_platform/frontend/src/components/integrations/credentials-input.tsx:ProviderIconsEmbed" +``` + +You will also need to add the provider to the `CredentialsProvider` component in [`frontend/src/components/integrations/credentials-provider.tsx`](https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/frontend/src/components/integrations/credentials-provider.tsx). + +```ts title="frontend/src/components/integrations/credentials-provider.tsx" +--8<-- "autogpt_platform/frontend/src/components/integrations/credentials-provider.tsx:CredentialsProviderNames" +``` + +Finally you will need to add the provider to the `CredentialsType` enum in [`frontend/src/lib/autogpt-server-api/types.ts`](https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/frontend/src/lib/autogpt-server-api/types.ts). + +```ts title="frontend/src/lib/autogpt-server-api/types.ts" +--8<-- "autogpt_platform/frontend/src/lib/autogpt-server-api/types.ts:BlockIOCredentialsSubSchema" +``` + +#### Example: GitHub integration + +- GitHub blocks with API key + OAuth2 support: [`blocks/github`](https://github.com/Significant-Gravitas/AutoGPT/tree/master/autogpt_platform/backend/backend/blocks/github/) + +```python title="blocks/github/issues.py" +--8<-- "autogpt_platform/backend/backend/blocks/github/issues.py:GithubCommentBlockExample" +``` + +- GitHub OAuth2 handler: [`integrations/oauth/github.py`](https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/backend/backend/integrations/oauth/github.py) + +```python title="blocks/github/github.py" +--8<-- "autogpt_platform/backend/backend/integrations/oauth/github.py:GithubOAuthHandlerExample" +``` + +#### Example: Google integration + +- Google OAuth2 handler: [`integrations/oauth/google.py`](https://github.com/Significant-Gravitas/AutoGPT/blob/master/autogpt_platform/backend/backend/integrations/oauth/google.py) + +```python title="integrations/oauth/google.py" +--8<-- "autogpt_platform/backend/backend/integrations/oauth/google.py:GoogleOAuthHandlerExample" +``` + +You can see that google has defined a `DEFAULT_SCOPES` variable, this is used to set the scopes that are requested no matter what the user asks for. + +```python title="blocks/google/_auth.py" +--8<-- "autogpt_platform/backend/backend/blocks/google/_auth.py:GoogleOAuthIsConfigured" +``` + +You can also see that `GOOGLE_OAUTH_IS_CONFIGURED` is used to disable the blocks that require OAuth if the oauth is not configured. This is in the `__init__` method of each block. This is because there is no api key fallback for google blocks so we need to make sure that the oauth is configured before we allow the user to use the blocks. + +## Key Points to Remember + +- **Unique ID**: Give your block a unique ID in the **init** method. +- **Input and Output Schemas**: Define clear input and output schemas. +- **Error Handling**: Implement error handling in the `run` method. +- **Output Results**: Use `yield` to output results in the `run` method. +- **Testing**: Provide test input and output in the **init** method for automatic testing. + +## Understanding the Testing Process + +The testing of blocks is handled by `test_block.py`, which does the following: + +1. It calls the block with the provided `test_input`. + If the block has a `credentials` field, `test_credentials` is passed in as well. +2. If a `test_mock` is provided, it temporarily replaces the specified methods with the mock functions. +3. It then asserts that the output matches the `test_output`. + +For the WikipediaSummaryBlock: + +- The test will call the block with the topic "Artificial Intelligence". +- Instead of making a real API call, it will use the mock function, which returns `{"extract": "summary content"}`. +- It will then check if the output key is "summary" and its value is a string. + +This approach allows us to test the block's logic comprehensively without relying on external services, while also accommodating non-deterministic outputs. + +## Tips for Effective Block Testing + +1. **Provide realistic test_input**: Ensure your test input covers typical use cases. + +2. **Define appropriate test_output**: + + - For deterministic outputs, use specific expected values. + - For non-deterministic outputs or when only the type matters, use Python types (e.g., `str`, `int`, `dict`). + - You can mix specific values and types, e.g., `("key1", str), ("key2", 42)`. + +3. **Use test_mock for network calls**: This prevents tests from failing due to network issues or API changes. + +4. **Consider omitting test_mock for blocks without external dependencies**: If your block doesn't make network calls or use external resources, you might not need a mock. + +5. **Consider edge cases**: Include tests for potential error conditions in your `run` method. + +6. **Update tests when changing block behavior**: If you modify your block, ensure the tests are updated accordingly. + +By following these steps, you can create new blocks that extend the functionality of the AutoGPT Agent Server. + +## Blocks we want to see + +Below is a list of blocks that we would like to see implemented in the AutoGPT Agent Server. If you're interested in contributing, feel free to pick one of these blocks or chose your own. + +If you would like to implement one of these blocks, open a pull request and we will start the review process. + +### Consumer Services/Platforms + +- Google sheets - [~~Read/Append~~](https://github.com/Significant-Gravitas/AutoGPT/pull/8236) +- Email - Read/Send with [~~Gmail~~](https://github.com/Significant-Gravitas/AutoGPT/pull/8236), Outlook, Yahoo, Proton, etc +- Calendar - Read/Write with Google Calendar, Outlook Calendar, etc +- Home Assistant - Call Service, Get Status +- Dominos - Order Pizza, Track Order +- Uber - Book Ride, Track Ride +- Notion - Create/Read Page, Create/Append/Read DB +- Google drive - read/write/overwrite file/folder + +### Social Media + +- Twitter - Post, Reply, Get Replies, Get Comments, Get Followers, Get Following, Get Tweets, Get Mentions +- Instagram - Post, Reply, Get Comments, Get Followers, Get Following, Get Posts, Get Mentions, Get Trending Posts +- TikTok - Post, Reply, Get Comments, Get Followers, Get Following, Get Videos, Get Mentions, Get Trending Videos +- LinkedIn - Post, Reply, Get Comments, Get Followers, Get Following, Get Posts, Get Mentions, Get Trending Posts +- YouTube - Transcribe Videos/Shorts, Post Videos/Shorts, Read/Reply/React to Comments, Update Thumbnails, Update Description, Update Tags, Update Titles, Get Views, Get Likes, Get Dislikes, Get Subscribers, Get Comments, Get Shares, Get Watch Time, Get Revenue, Get Trending Videos, Get Top Videos, Get Top Channels +- Reddit - Post, Reply, Get Comments, Get Followers, Get Following, Get Posts, Get Mentions, Get Trending Posts +- Treatwell (and related Platforms) - Book, Cancel, Review, Get Recommendations +- Substack - Read/Subscribe/Unsubscribe, Post/Reply, Get Recommendations +- Discord - Read/Post/Reply, Moderation actions +- GoodReads - Read/Post/Reply, Get Recommendations + +### E-commerce + +- Airbnb - Book, Cancel, Review, Get Recommendations +- Amazon - Order, Track Order, Return, Review, Get Recommendations +- eBay - Order, Track Order, Return, Review, Get Recommendations +- Upwork - Post Jobs, Hire Freelancer, Review Freelancer, Fire Freelancer + +### Business Tools + +- External Agents - Call other agents similar to AutoGPT +- Trello - Create/Read/Update/Delete Cards, Lists, Boards +- Jira - Create/Read/Update/Delete Issues, Projects, Boards +- Linear - Create/Read/Update/Delete Issues, Projects, Boards +- Excel - Read/Write/Update/Delete Rows, Columns, Sheets +- Slack - Read/Post/Reply to Messages, Create Channels, Invite Users +- ERPNext - Create/Read/Update/Delete Invoices, Orders, Customers, Products +- Salesforce - Create/Read/Update/Delete Leads, Opportunities, Accounts +- HubSpot - Create/Read/Update/Delete Contacts, Deals, Companies +- Zendesk - Create/Read/Update/Delete Tickets, Users, Organizations +- Odoo - Create/Read/Update/Delete Sales Orders, Invoices, Customers +- Shopify - Create/Read/Update/Delete Products, Orders, Customers +- WooCommerce - Create/Read/Update/Delete Products, Orders, Customers +- Squarespace - Create/Read/Update/Delete Pages, Products, Orders + +## Agent Templates we want to see + +### Data/Information + +- Summarize top news of today, of this week, this month via Apple News or other large media outlets BBC, TechCrunch, hackernews, etc +- Create, read, and summarize substack newsletters or any newsletters (blog writer vs blog reader) +- Get/read/summarize the most viral Twitter, Instagram, TikTok (general social media accounts) of the day, week, month +- Get/Read any LinkedIn posts or profile that mention AI Agents +- Read/Summarize discord (might not be able to do this because you need access) +- Read / Get most read books in a given month, year, etc from GoodReads or Amazon Books, etc +- Get dates for specific shows across all streaming services + - Suggest/Recommend/Get most watched shows in a given month, year, etc across all streaming platforms +- Data analysis from xlsx data set + - Gather via Excel or Google Sheets data > Sample the data randomly (sample block takes top X, bottom X, randomly, etc) > pass that to LLM Block to generate a script for analysis of the full data > Python block to run the script> making a loop back through LLM Fix Block on error > create chart/visualization (potentially in the code block?) > show the image as output (this may require frontend changes to show) +- Tiktok video search and download + +### Marketing + +- Portfolio site design and enhancements diff --git a/docs/content/server/ollama.md b/docs/content/platform/ollama.md similarity index 100% rename from docs/content/server/ollama.md rename to docs/content/platform/ollama.md diff --git a/docs/mkdocs.yml b/docs/mkdocs.yml index 9b1d7b6f5..2a8d91d22 100644 --- a/docs/mkdocs.yml +++ b/docs/mkdocs.yml @@ -7,49 +7,46 @@ docs_dir: content nav: - Home: index.md - - The AutoGPT Server 🆕: - - Build your own Blocks: server/new_blocks.md - - Setup: server/setup.md - - Advanced Setup: server/advanced_setup.md - - Using Ollama: server/ollama.md - - Using D-ID: server/d_id.md - - Blocks: server/blocks/blocks.md + - The AutoGPT Platform 🆕: + - Getting Started: platform/getting-started.md + - Advanced Setup: platform/advanced_setup.md + - Build your own Blocks: platform/new_blocks.md + - Using Ollama: platform/ollama.md + - Using D-ID: platform/d_id.md + - Blocks: platform/blocks/blocks.md - - AutoGPT Agent: - - Introduction: AutoGPT/index.md + - AutoGPT Classic: + - Introduction: classic/index.md - Setup: - - Setting up AutoGPT: AutoGPT/setup/index.md - - Set up with Docker: AutoGPT/setup/docker.md - - For Developers: AutoGPT/setup/for-developers.md + - Setting up AutoGPT: classic/setup/index.md + - Set up with Docker: classic/setup/docker.md + - For Developers: classic/setup/for-developers.md - Configuration: - - Options: AutoGPT/configuration/options.md - - Search: AutoGPT/configuration/search.md - - Voice: AutoGPT/configuration/voice.md - - Usage: AutoGPT/usage.md + - Options: classic/configuration/options.md + - Search: classic/configuration/search.md + - Voice: classic/configuration/voice.md + - Usage: classic/usage.md - Help us improve AutoGPT: - - Share your debug logs with us: AutoGPT/share-your-logs.md + - Share your debug logs with us: classic/share-your-logs.md - Contribution guide: contributing.md - - Running tests: AutoGPT/testing.md + - Running tests: classic/testing.md - Code of Conduct: code-of-conduct.md + - Benchmark: + - Readme: https://github.com/Significant-Gravitas/AutoGPT/blob/master/classic/benchmark/README.md + - Forge: + - Introduction: forge/get-started.md + - Components: + - Introduction: forge/components/introduction.md + - Agents: forge/components/agents.md + - Components: forge/components/components.md + - Protocols: forge/components/protocols.md + - Commands: forge/components/commands.md + - Built in Components: forge/components/built-in-components.md + - Creating Components: forge/components/creating-components.md + - Frontend: + - Readme: https://github.com/Significant-Gravitas/AutoGPT/blob/master/classic/frontend/README.md - - Benchmark: - - Readme: https://github.com/Significant-Gravitas/AutoGPT/blob/master/classic/benchmark/README.md - - - Forge: - - Introduction: forge/get-started.md - - Components: - - Introduction: forge/components/introduction.md - - Agents: forge/components/agents.md - - Components: forge/components/components.md - - Protocols: forge/components/protocols.md - - Commands: forge/components/commands.md - - Built in Components: forge/components/built-in-components.md - - Creating Components: forge/components/creating-components.md - - - Frontend: - - Readme: https://github.com/Significant-Gravitas/AutoGPT/blob/master/classic/frontend/README.md - - - Docs: docs/index.md + - Contribute: contribute/index.md # - Challenges: # - Introduction: challenges/introduction.md diff --git a/docs/overrides/assets/favicon.png b/docs/overrides/assets/favicon.png index 9c60eea982b46802df3094fc11ca76824631f089..69bded6abf7e60cf2fc41fc9c8c33cfa2f1e918e 100644 GIT binary patch literal 47680 zcmYIvby$>d(C+TC!XmYVf`W8Pi*z?gqo8z`bV;*INQjguDS}8!Dk-&#BA|4qba(ge zo|oVGzH=`AfNQU1-shRPXXc)HV|BGvNp9c24FZ8k)YX*qK_Ce5CjYe#wGj~ zE?S)wDM&-CqCof}<%=%!eM;gE{0x`8*YG+pW#lfji~S8FRaY)iW7EG9+y;qXQ&4ra z(9KVia3Y4$N%C3Cs1U}`&sSKUUS19kB z^QG=E$?oo0*8MTo`vcNzf}4cAUCa%Px}RaT=4ZDJbVewk4++91>P+L5<~Xyz#TP1z zI4th?ScX7E!v)^TE6Nw+`M>Zo?h9?Y1MbS~Awkp|%M2TmimTOq%ELpaG44VgNx&AH zk))2-;u}-WWgC0%qMm1+RfO=~Wohv$W66X;78K=Ra_27W6wd`9FW*&;4(3s2E(PxC zW^WMOuQ3EIr3*aTXmMbl;&E?5To#Pv5Od%@5=s)aFK^1t&Iuv3a6t!i)i9(1GnmJk zgg@-ST461%LFB*=O7h?rGGRjD&GV`ZWnpXShyuiNnB+9KnV-!BHL_(M&Ealel78?)zMVmE0AI_5 zYGpmIlzOOWR7yt4Pae$;jRg6>F00{&(J;cN-)?RBJU&+5=S>O~gQ{*QRF0^>ej^m{(BGW_Spg=P4?5~AQCJ0tP~?Kxc}Q+l-}$kKp-$8TQ7|?W>{EoFK7kypy=>gdL9GUC^4a5En8D#td=^HioXTs) z`9~1Jsyb^P=B!Y9hE9oWQW64^?Lq4O)n4{>g;_*jq)Rh_g}`6gr}a`nxDwZRin<4v z&mX`f1a~*$1`KGA$7*btj%ipM23DAOdkr`t3ZRE0Re7l>SUY+!v-cY>#U(xCEG7CZ zy-HpGO*8Czd*wQh(nlpeug$+TU?ITdPRBIqWF&{fhc8r%R#lrxlOk5XOF09S4#e3r zb3pu;E{;2hwcan0C>HrZO6boBBpkUv@r0Fx5mK${^)ItFmL>OKPj0FXG!7ZKon0jQ zkw&rcEGKiq)lG7l@?NRn?Fd8pA+`GiYG2h+WO9;D4S`A;I}zUkMrhFsEX;%uJ7vHL zjTX}Se#M{*vV7g6lrVe}6j$Vft;DkeOB=L-O-3ncRsuPSkoiA@vL+_vW8tI(yeCZ! zJ4vXM&d7pO&v!>H^K_74ssqj1I4`@N~1RH1Bhn%7~ zr^P$-$UFU=^@))QZ}ckjdHJdW=VUc)^|np;tS1fb_<3 z(LEuS@UhxMaZavtsF%o6!5W@@gD^nOHO%r~^_4|C);8Y@m#bxDgytIa@CM#hW~`$Z z9ar8rbG-bfB!(|nZUmv%m_R7d9JP2{2vq+mnNJ#9zjP;kyZyz_^9`_Hw#|W~g zEL!Qp9Y@-=ZYJRujj)|>x4ub^|2X#m4kA*;ZNn4wPq75$#Z+w{(L@wzJJ0gnf@yl) z=k(nyuxF$<(bAck7pn}lclk-u;)!`i;v+_xsKs|;()(mw){1H$d&EwqnvDC+ZY#^Y zPvmCyzTwQ`tv}U@@4AGv4HOnA(4VR^0`s~wsMl~4wd%9 zOn=Zbl&DgmmnSvA?|9X|(Ut7#wvpB#7F9fhvU?X7ED{w5K1~@1ByAKDfAqoixjpx3P3XhT_)=}LOo$$?qjnDMW z3j()RH}_wdM{Tnq#!s={hHP(H_TyAWrrE`voKI8k5Cq});E4(;9` z_jzz+(6IZJt37d!!KWS5qx^}#mnrgcL-3U7y#W;Of|aO2D`t@O1X~Zk;?(fp>0!iE zqHOqQre+aC!#$Y-!9R{WUGk2Gu0)51T@FRLa|HEe4g@5%GyZ)#V9j>Ef~=5m;OGwR zG^ymS2_tsZlsy0K_>H5UCiB1NG)M;2RIFBQSg9`>?%ZllwnlkeqN(<}+enIFaMo7(A{GWPckJ8*}^*SYHk!oTu}+!|w_70L%d z!{nF&sY&+apSKwWyhCOg`+^D5I|VBFa~Od`yWPs(TO^bIu=PF_QDDGiAidK}y3&9?cIZ{`_^#@sb3@8`ZW| zVOGjOl2-z9e-V7IX@5oXcZ&y6GH$Z{*I!#(r%@)7tgKkXn|JZxr{dC=YwQR6LI*lt zbhdTIk50e6e!fdHOxO)d@ZC5$TJ8PfHf{lN(uH$^fe2?1V! zZt@UiM~q*oU+f%{qsoj$)b~1@8?@G|1m1;u0V$f1h4~rGbm;J;@x)h4c07y-JYpAF z|9J3Dwqv#Q+%{ncEr31kBKu8oP;{#3mNx1%;*5?J`>kQoIx2v$4bfriJa88l0$%-- zT-b=3@ggcG9F61nX~=qm5kqAkN=8PvjvQH!WgiLLmm7J-%utP4a^d$Mh@Sq{f*1}Z zQo_=fNi6$cWz1=v&c_8+x6NGf*L~Id>#31;ioLREh2%dgxD%S-^N~{(pZiGi4tsed zDDsc7!NV=Pt^I-f>o_u{vxb5h1 zek-HW8{VumtX1E;&h5Vz5oAp}EeP3WL}&6gAHDeFq3T@IfN5S8TfQb~%J0uVx)4t^ zzW;(qc>aJ!dcgma;Zt1Jv|CCR{))-EETE3-PSd~|B=X7nur%>ir*P|Q^cH(*1N?(4 z*9!rMa8L1_%5uWVZ0U3B{1U_mR#)nxWSscUivjQbCkYWP2q7;kLHOWR3+{1Ml!>l- z_M8pxBdeXM%-=_=dOv?|F)2gA&*5;${Oz|xFfzUGkLTnB_hc`cVyWHk_Ba4REIYZD@-=Vz0sTqVyKW?C;$I_|?ko>{&C0RVS{gh>XYE+%NBAdr5By!aMFNtMgmB8k2XwUKJ#<)tr*93!>C3O+o4{JI>mxDe^~l3yxLA z$m=ALR%lT#=`q`eq1)G&@w=6=BiX@Q<;Uu|=9=V%%s2K)wkOHa7xeJyl89xIHzs~t zj?Zpuvp47elaCH=?;zWd?~ThKYh0ovaGL(mcGb@gsypzC^M5fRQ6)*ZmgX0_kr@96 zMdS2$VjQU{aKs!gv(hJ-wYp{sB5xfxW+fE^tORk(eF@*A_prT?5WCk?Ef=#2RJ$_y z_NXl0=d=Ny_o_5Vu--LvA2+2|$QxVFDowSbR3svV;2I z3CXi)&o<=NuxSL)d>3HYR;DJ!R%@!bysPlq!%!RbZiE-&e3tFw$^Hoc8r{FclbE0Y z`Ww-7oT*}IPw9>pzPv*$;Wkhna_bZua)7Po;HKOga}l37Z2zET{T+tfxm0Q)#)L3m z!rM#RrpAfQ)gy&LA*lmZ?BQJwjr2Nj4nz4hvCXGZ>&q4WYt2==h{`YEf(jg(OK&hE4 z21WzCOgoPgpc&(d1(zU&Aq1~4Ei z@vOL)5JZpVby6}1@e zQILZ9Cq_S5da|un=Bg6uDpwPq(5?n?{kzeVpd3M04SSAqeE>&p50kdi^K1QeR{-5k zem1|N#_IL=LH$T8U3%IJD4{gqX$tN?_qhbVuwjLUY9<+d+xbPS93my=|<=n(56x^+qE`Sq$fl%&yS&%m&wLC9|*sxmtGpDsfs zEjg2Gv-YK)<&>`_FX!2#FX7AR)mg+}L$(SDn(}K~%3tBAA`ZmW$p^HvTrA68W9>!k ze#;57Cme__xKin}V$E+S4FQe}+RUp&_Z)e)11^>@Yn7zOh)iBh&Q}THD=8n_FxG5n zTtt*|Pb@-N?`&^k>p0*wDc;lCP`~|gmcX&G9rgK^jFB<8E16|>xXo{QbOC=G;#C6H z?yvDgp7s7q`%jtcF3o1|@C}w|7c2S{Kr_Fe8gnD%h)pGZIPD?9)~n1pKf7D__An-g zJzn0}HJzc!lm?AYj%47(^~ffcz~$e39S&9|LuY|iIW=Tt@hX91!*Y@Z?-T|-9|?#8Y`}o1)1vi}-eVp%e{w_<3Rcy#K`S zpYV1ZJY7eLnDGW`9N#~oaOy~^jB<9o;6^eSNpicnUd;U?tByT$-?yrp@ylF>pIZ#9 z+j}g_TKLZYa!!5f$2K|PZWbz?kX2IQ_cURMz9;w*rHTl!t?W8oKWOn_%s>tV_ zi+=~Rh$q6rsl=By9p#sqmlnNQsx6_@@|Q$6x#*|+FZ4@~ll=n_E~dN{)xF-jjR}$m;RoT= zOYM$r_ag$CJr|ozOjtua);p_Fw_)IY>b#S$5S+ZV)tdg`Bl95gfp<_LIbv^kgUnHw>!l!NIaDHlHf&=X@da zzMcJ1Xo20f0O{~2e=x1*`q8LTpR_wyi!NzTLBAPokzZ3Gwc;a}JJ6(yX)2&jtZ+N209^^vr-FoYOn? zd4_Xl0B!uaGJYmO(Md1sl;Bhj@2^=zNA=$~*Zd;n`LcDIxuj$JYkSQF*{@hR^~^^K z{J3Es1;RwjHa0Hs9>2B8Oo$aSQMdW{fhymW<4{dS`(Zx@}7Ea!*z}Z zv#Z%<3RMW4P}#5 z79qHpjH)dVi;1F^*>Ici{!B++BQ5fMK0r~5JMbUlDJ)A3h zE8G3OW~;vh@2p3^EKqjYBc#JbO-n1K@sv0~6IC@?Y^Zw#hGOPj*`Jkum*_g$i%O8Lm@ivuXZpZw!%cGyp>t zDB(%M+O5k%s-Rh(`D^$rMJ(k(4lptP-ZI2^D^`O3ul2Ua2PM?D^e{weGb?N6WZK-@OS0(PvL3+wq&?`2PFR z!GWHR^H`;+ZCm|2!(T%1=o7wi*s|=nheGR8Sd1{Gct2*7wS%=FgWT48gI_nRavhxr z!Yt_iwK$n}D!2P+emB?qxi$C$x817MiTEbTd(NG=et>AI+m)9T1FoWUbCs!4-*Otw z>MTKUkKY-8+})GN+u6x{iu-f9D~M1wd9khLvzMPdAMiL;EyBeXx^^k$eMn&@nL!I| za8P9nuFO^N!GYNx>Q&d#9S2UmuDbwWyR;d!ymokeKuW%E18Pk+x_s_95qqtZ+OLG8 zKFPoAhb?E*br~B8OmXyHc9Cdrr@R-bh^Q=ES)cKiX@4@gRXHo@s^eWqnXr5Bk0y24 z1q{xdafw$_5m$uaJ?*1vI_{KWZ!AeExfK@~Aifdjgox=q_&cisqquV?3N>`h9f+ol|dY4rU;(=VP*Z(=hJ3QT!%^ER| zbCo}w2^&0Y{qUhNEU*-H9(ZC;H51lwaDv!BaPZqQsgjXZdYdEqQ!{2?5SFDcbeK zMx-m~>|_Ra(qi(msPnG=3`bW!t$Z#+SPt%^N#K&RJN{DFvaGR@U^UJ4n)WiPZ8=Br zZVfkmx>2RJSl82AH~bZZWm?bkmExiwVm00$0B| zX6Bx91S&=kcjN_~7eH|t6{KQ~?zvesopZ9Kp5~LF_X#1u+z08dOMv<^sntv-DBCfa zg6PPqU;JbHjE<@-?ZXjbU{fwtRlhj&U>>EN6zbm}J$T{ODto?EcAu3_ba!LDzP0v< zoN&N`B}ku;Vn*RZf1E-!IjHC z{N*wSc>CkUzCXpagW!u2T!caMH7%hT?&A24aDOb;I7PQBbwBODP9@Lr^-RXD-hVG^boAH6l~1v37`N`~W(0 zeB7WnqeSRmuRs$`P8jq-m`&)|k1u_hQ_tt}axBvTQX+KE%kYlo-0#NPjKW{ zVQB%p8Mx~_M{HHlp`uk4mD2{zQ~Y20;kd)+WVd%QM{%CJm_NsN;u#FTZ77{2F$*kE zJFAYZ2QyH{d%X*jLt1RvS3>Y>Qn`1}WRVuPZdf4@eZeVuU03+y5n+ia*can=i=m;0hQ zCmqp@xmAt5W75Z8bDUF=R;@2ELB7uNhT&k|K#`tQ{FpiP#;{*~)OuVOe8u}TKrSB< zvvV#ND3>%WoY%Db0i`u}RIqUYJ7q05EbL2ySvAS13InrcerYn!p0e>z@~?1cuB>*L z#p-H@G<|~J^II98Hx0-kx~~=IRXl8Tq-CE(N{26F&JdCK?rhDP$z6$d{ixdn&bLZ~ zj|m@?2eNum!=_m~L&ui0g_zDNYmE9Y++WNsujgEh{@#39m5&sybaIliv(o(TV)(a; z3Zl!ETJyPcMg_|zgkv&dbH)v(c;m)WX;Sj2)Si9U9IiP?NgXy3dG?_4pei{FYB?bC$zTsLP-nhbI)HFy+#={MLFy9oYADV$&m?j0`Hc+Ux##h_#4UY9!UdTNr2l!C2&GuS?u!~plsI^v&4HTPt zAb4-?*sSL((8cov;O@1>&Q&@#)Z~%qv)z=W6pGx!Yhy>Vntyyy#-Vkx-N`^&Y}S~ST_D2Gse7foT={Mq z^#v`o7*%_`2$R)l#dus4+uJtXy)u@DOC!!zBL16p)mgIpU4O}1FX2pIWGdaBEJA#)Z5D@C%|F&az#4wE zs!Q%1*OzpD_tMve7YjfWq+c04bzd4N5(^M=7z?ta>PicZ_GUlo339CuHZ)P+)Z1Q0 zs_fyl*wFuQnV`O7AO2NO6()BlJd5c{8VA|GuREg9c(My@o-@ zT%O@sn7Wd(Jen$uMSFv$U}Gu*Btq*Gpey64Q~%iUC!a`I)+gfixh`7RU3yKCMdFXA z>nmEKZ^VDa0{-GL)%$%!ROC>lrKb||LUakk(Q=Ed83%AYvWYLJ_)u3dvo&SuR8-K1 z_rZlZA7_40>n-nr#({a0;qy&R3fyvi+SRV{K9;~f!HefoG)nd-77=0xzNnJder_(LbIsn z&$0HE4{1VJ4-x&Qm%q?u9qc=FlV|mI#$2XsFwOIm&fbA6TzllNw`#!&961@=+1r?+ zt-CW)`pril=Xi z+1eF=EdWwyo_hxC`^roj#|ZvJrr>IS*nj$ygP?l-nol;4T}7&FS6WR_X+WLemVpG#~&9Jge&a6`VB$S z6=6m*JbsCnGQU&qiP9=4W)!gcxmDYYwLhZq>B=P(KjO=x-|7A-KuJ8a_m4W}ntv7J z5UNwBnk~_JhFBb8!cjQC#sheZg{l-gu2JV;S7}@XVUP*H;LyPx~{JbNxL1}Df#Kpx4gF)1x&q6^;HAQVSZW}X*Z-mVV{3_pL_W@<=j0q@v>(W77 zRwAYM|ChctauOdcA*qQ6%)e5aSAY@ED7bSL!IAuGq}|{+#_n}~E+ZpsnbD)YvPlPe zP#eKV@*N;sG_L>cV~3e38r7Gw8Ct50R_K#C=5L6AhOq4zQCd?~lCp4*)|K`vQ53b* z)w#|a&_w)$0YYj`^E={m3EKLG|80hNOq7`UQ7&{X^Ji3Fi@?z6uQ^0PxQp1|H2pH6 z>chL%_479a_j8f&JPR}UGQ>c07K9(5E(6RXzL2Taoyo5I5q0BKGXBEck6%tD7FOnx z@p7}?x&@X;!^3`s4hv@H_7d1Ay?c8_bM0{YKo2GOIU?pXvdXL9|`J|~@13{6$CvImDKSO8Vou~f-n4y4%7m7|k z;tx5PSp3i^zxzJvw0?dTfqSUbVZ~^AcK%w6Vx4v$%mZbd$N&Ctw#zY2*7hmu$HL}RBn*D^7H+caL^t!6yB&c)hs=$g*F-{5rP`aj)=G>hFXzckex% z&HA5WT)mX8NhTHa>;14_{p(mJAz9K~t0Zdk<9~cihjT7Lg4M=d4cOSY%uusY{M!wO zH?KS=5kUr` zGl8w}_TfvvwcgZK(^(*wP~BEx7!I%Sb;3>PZ=T!*JDI_Y&7thqnORH1W;m;XSL#%U zI=Aw1y=g7D83Te9(Di04QI=bU`4-f~8`zwv)6v@)8I&#mWa02L)vy^ycQ;2mzyZN( zI&no%9G5J%{BcxR=Z=BGZQ$5)>M**?>0L^6u+XDP*|^r}RiJwfd%X(G6gViG0t-$REF4mc(Mbi(^A#&Jo!6sdahoaWUsZI1swXU>%z0*7cns2YYHL zjb&tNfk8&&(2?H$j&oDHY2Wv7oXw9!&$m$pu5D!>TDOrX<_slDiLD^q($7?fuj3MJ zqoy!957(#9zB`XE*_;2W6tsG(I*S%SQrH*bt8X zjM{O%MS%l1^hxm}>w|Fkb9rllC7y`@Io_04{CM#`kb6xdhHz08$sWbIZ8goLw3?H{d7FzEQj#^^;_ma=}%&nyC<7< z+WR%s5V%yNdUYBqsvfV?02rL` zncLzwZCiZH$;=qTfphBnUz_D21wlB8+#E?K-(I#A#hb}I6#N>^I4;lxR>-4)I%pfbC=SMMpO z*1cs4?z}olJ`6`P&htT{uTQuBw#u-#KDTucU?|Wq~cm@6ldLn@4A%JqU!9*#r*mPd4-^);yWqvdgI3xe~FSX z@l_!Xa(4&)zMsuiS{}aC4Ydp{TM_7gpBm+Q;|zCASNNrhV*QYP8eZmK%hYGmG!R~Y z@P&@M`;5$lYOmL0pnx)x;sNdY&60L_uEvIG?j8x40G!EjB`(OeW)tnasAt~?!Bwz@8FCanG{m6XZdKS@2kX##BrLxve#*Q$ZcGDo|0@zZ> zCzkysp(fH{;ZNfTRph^1OCa`)<|2EhtoxL$GWJqb(l3{P^ri(?H2dgM0rHiZZfl24 z_BPYA?LS19e0jwli`eEJwUa{n@R}@#<$o*WY^Wx~qEd3}A8@B$h>2&RhN#W=jO0zXt>U zA=FD`i5p^YjQy5kIM!;dO2)+v92(2yvEN?7g1)YAiw*mp3XEwz>sxR>3?P|XRpZx3 znHDExw>JL3fa5aVr>@?l6pUZL1~jm%Z~CN7zS9TayrXPOa)%xJ8KG%hzC+6lQ+_6& z`;c60eLF!k#&L^T^o#_0tEu)ZM~er)Ra^ot{t6pF;&nf_=ZoI|9P$TAb9!6(r<|*P zW`c`lCPluOB+$}GEQ6|B&6kMxcYuh%k&8L%uidiznB__DHS3%Ruusv@VBFKSTDo0X ziXj^j0IL)p(c6z67-n|3ulnOI_$!ubP^70=LJPRF4M#tUV4JdK}9Nvh|^JSY}yrI|pQ=B;}h%*k;na{1|L|5}9Dsz0!1z{S^c zy=S?{(BM3@*`?!J%%}`_Xs#x#SR=eqvI%?Zr*;GkIXo}>MFVkIM@s@|P&(dYMz&CN zPLEwJe|m;WAY4_KJSJ>w+*K_QH}^R1S~O^q7!Ns=&zM2lwRw>H?15RF^K>@m9)`a- zJ=XNbRpZc1luhjG?S7@AV$;9+FWmf_&T{U!zkk2TprhqZUx{U`dq@WP4$-BbaXpc! zE^&CKR2hIDX*J33oY99VJv+%zFU?zW;bOXprrd29om9z@Yj%|#iqY;6`;?TB7w>>i zN6XpV?3Lx$XCT190@ok$WgzpYa5&J!CDv*nM#9_L)frSPZi9W8^MJfz%j7YT-zTmq z<_ZUQHD8G&V8vTsinSbe$ra&;pe}}^uT9=QqX!6@_XK_>Eq$)@>j%5zV;cI9u-v znMuEwgaZoubMxq*e7v7A``bEB$-Hd&zy=uX{ev((4qtl6eAu$EwKhOQ5la2=O}|K> znlV&!c0F%WL75BGZd`uHds+YiYQwOFHIdCq!K#1g_H5tS#MfOom&GK#`Wz6=h-k|_ zV*g6#+B!Sp{AiE_xgJm-eus;BJgUTN*-Ei}Ie+~&#|;>CZU`NPa3y&r#WDm38h3xT z&OZcDDp#t(WY-tiz<>`1e7JH#4A-D}&iiBN1hl2smb!dX?<9o-hJkZJNyUnmUNTE? zcM5vYU8ni2RE9(^S3A0P7V2ovR0-)a8pI0RR6*NzeS2ky&0B4V!lC@9bpzb@r2=OyfoDk>6|Wx_wY!6K}U zTWP0w{N)m;!vU34VR7Lsd`=dO%OBPIGi%o!4wREHdI*8`MsV0YNw%~7md0a_@>kzV zW5Ai!md3*yZHgs_Y!ncjo(s_;EoQ(W^Pb@EmkL<_^;S z?zb$x&lCD4gsQk_)E3IlHSvAM+W){&no2h*o(M+wsKk58AKE=?CHHIRw-4>>rLb)R z+=?Wt5pgcwM;esnWzrf?44<+oFc(0^SvLL~nPG2LaQKw&A{vy@$G;~VV|2SgE$9@nRN9uiX%PvUoalQ)fjLh13w@v>0{(dPM;K`N>)= zZ?P+;@fKQ@t!CM@e@yDV>4~rPfQjQK^>|^82 z>$clvU*Y-Eya?&r?20m9=W@Lbob>{b_|kzNCVbjT2n<8yNmY&cBP{Lk;&Jts#z!J6 zig^-2cOh`yrheCj@)IbE90~BG>$B>mH91hKhvXVStk5x6&DHcd5Kh#-_r|?`kN{4c zo4L}t%`nca#j>hXSV{Mld6e$vSYXG{XBoq22Z004=Y+^nZp6_M1YY>5`jJ-&L!bu3 zP2s@+xzgEdB`wRbRsoqd9CHBqo4mzmHmL0fFVXd|c_84B1A+ogdN~9?b$D{Fbof!? zGMCT&ota4E6U7{qk3S<7*sK8?AQhpc>0=k%4;o~==>#@4#LC}Gkaglvd z%KIoZS|#Jp@$R63M&IhfzGAzcE-Inpd1eO#G9N&wcJ}}SwE1z`%!ZH0Zy;DkA5ek zNJ8+?Yzrt{<-*THLA5GrzRS(Uw^FC@Q#W1mF%Rgj39My2hQ6+UX^mif0Jsa1ANDq9 z))PO>K9ghDQ}jRw{JVAF@l+7@ec|s`PQGm0=up^wd3j5Ktp3?P&bIQWwY+}BBC<|D zqNsQX!A4+CLK*jm{Z5PAzWb9Ie`Gffru`@B0&w^{LRK5gACh;P2lb7Cq#q{72)A(! z&fD@X=IuQ8WAxJ)P+81@kj5HF$DTje(Q^GU_%eX&`AwVhW}pH({-L?FYo1JRnK&cOCIdTz8fkA;0=lgw!e=sA3p-UqW9Ty z{CmO~SZZ;)DZikBcvGtJb`nUjGGF}Fxfbw&N6g3#u87!+Y`P~$B}Z6vlo3uPl{o0^ z7VKb(W&=$_`@VMBrj4eo0$O0yg$cyV)*opj6X@w&w!3F>p!-T>Jn2(DFLLKgEUj)_ zsi8Xam|piGXuv#Sml*6(cb2n*Dc!KZ%aQW5J0=51s@*sen3|jhIG!CSCuI?nLDh;4 z#Qkx*=PoVW65vd2kW#35L(x?&2-bfR)l#~U<_2oA?#vbm%U@S}h)Dw8BeoZ7zt}4i zWatE9l;{9Y^0*g<>Dv(ti*@sDKFh%-y#Ickqqh+qqLmCVerX2pSBd`6Cddx>VG@Vv3K9kfGdZ-VT;wBegG_vIup@@kbhx0Ckw zfq$enp;B4Ck)R1^&(>mf7G;r;qhSfa)hE*rQ{~`=A?{()t3+<^WI0sk>a>g*O;H1u zK4h-4C<_uY8&4ylBR&h*lQjw9{eBkCMdHPkM5d?!Ly|F0Z&(|8FG|1*+}551ZK~V5 z+Cg*-Ilo`Kr;0L`l4b7?zQ2>KwcW*McN5efVcS8cU)dkm`Ph14*Mhw@z2heu5679s zE@O!bNq$UPTX7jeZu|z+ANRG|_?Q56J#mSd;o|Dy!~8*HXboXc>wnL87L~>9?!8VJ zD!g$PvM`41@e6u#5d-Z=-{@cG%-@D;zHCuzI$!!axvIGWl{C02{9U4*eeJr&y$SY7 zyQqMs6-?3eV^->mZ=knv?3HgnZhhGt`Tz)fPaYOxlT8+OF@FfybZ_V7Zn2*q4|X`=KP77#&AVqHlJTo1G2!0QUTPgGNE!+ z9etQuc7)2PvMMxQAd)z2Mi7KaV0vaPOReJW7$2+Y{){@;e+DEGg}3PQ7o9Ra=g~&S zbJpL$d?lzFdFLriGS~5IjU9ENn=u%0IKifjdM&uDVkEhEw}(?=%l;oE1qF^~PH;dl zKsldHZXky9CXEO}Y;+bg7ij%XPyK9CV{6)h2rWLeuSrDg+m@}^BpOs)t^#S-5!Q46 zuT4gP4ASZrA(9{sOT0<+LL@-hHV!1ul6S*1b;91pdDh@RumFv?uu+!K8sTmi*Lrna-k4~D*ddH2ehlJi`s0`* zR?e*qEGvBRWnJd~3!{$^k{& zAa@?iI>YulEq#wG#(2)%k#zoxm{RT^pVFJ98gm{@7L)=06-gWR;%%+N-~n;WtNN_i z5xZVgg1`cafuQ=C7lZFl`GZg~_TS@yd=m3ec` z^Z%6(G&UEa%xkoZ-yH!;2){2`TPBAhlmoB zeQaAhBE1n{vXD_xG0ZUkM3B(V(N%(^ypK=v9R>atM1jLc^p%Wo`HIuthl*W#07$#A z&&lF-1|C7tkG~~W&&vWB==EIwSI-{?l|EwlW9B!2GW_F_VDLWILBD0M&sz%*x1$vE zQq!FsA~0+KmM+NZd7r#j{f+JDl}?%g?i-b)rBy95jO%HdDj(7R6Cf7_a`f>jEVa&1 zlcrRMAsq(nMR}lT>hyBoN6s`kl|r3fFyyax*)e|1A7)y-Qc?v_S%&5lLdb^*Br}Nm z)eK=i45$=;KK%{QfVbTXJ()nQXqWPoP{@4hSq@6V^5dzWWvx66DXgf@C7=7>>yxpD z!lUYdE;K)}g4D`EGz>f5b!T(^t3<#{K@vyYCV&F(0QZ}xgcw0{U4n4;4D4>9slO0` z_ix!dXWPP%iGcqfVpDA8f4whZP$NbPm)^)$fLj)}@NYN2MpFK*JL89<$Jb^+BVZCK86Ct3WDmZ`;bMT3)@Puyp^pUyGZLjjlezN8swG#EY8qF(O z82oz;xKf4%IZ`Q*C>JME5lR%0aysr&Di0i8!Q71p53_PvOt5-lKXY@8YcO{SkpqAqJ1pysV z-{^=MvSe+6;Gh`9W+{TMbJ*^NjGmp4QU4a%J=EBHxzmXlHsl_WPMi2%*Qtsr%B`P?D6?iu)6Xlha$efq~;?b2kS`qMa1`dDyk ze~6)-9}x4lr{PdEKm$DDLFMq+Rwdm*g5qzJE5rAx!$N>tgn#bWsF1vrRXlzvQQs$9 z&sKYzDX|E4?ryF#O6a#V3ALQ3a~nHEy?)$!4_an=Y2&pJT?UE_;*32GN7_I;X1j!c z$!xedSY|zfBZKa0I?Me1wy!_;BI@g5v)9iO;vthplZuHN;h`^$A{tU~+ znC_O>4K>~*vpax3O}%AeE|-$8lT*f#ELS|~v*&rT7MMY5F7X-JEop4J&AyQ zEg$jeTe^`oV^dqW69oOnM?fay)XEkQd3glA1!D%R2ZvrOAu~=J8?K%u$#iD9mBevg zwiFeP@tzchIE@nDAQ~-$(3qBK&)X6JQC;@TjH(~+F=!q)ZS1MO3yV3t$v=#_UU=7~ zgsZccki9GBiH9X|Uj=PF0I2$9G_S>ld#ih{AvOMK;pXEhfo(Fg-; z6a5zG4IKme^xVMVty+&|hkG@pr=IVZSD^pd1)yrmnI0OH1%`EoMzZ!g7)GzC=Yi@? zH`3jp9lvGV5|n8Btn(P~CI$;d9@g$1NgOeZ#fi?z^0OyIXQLmV3qVYI`^^pgH zFu!)n5X9KAl{@d-WTi_2TOWYniD5eY-SF~XJ1d+I3_obQ+OyKa?Hr%Ibcb=$F{j@6 zFjW}gnaA?5pJY8pN8Kdz%XxF%kDe%qGA1FRgsgZ#0!ZK2Z8eK_Di2U^h z5i$Kl<^Ig%nf$VYpsrM|N6ZW0o{T`g;sd6GXgj%oF1v^1&4PN1=8zEC$%?C3gb%*# zc#&!-or!i%6IW9W8}Qv5)wlpkg-**f*hAf^#-A~h$~@Bp{YnA%^kXtOS0v_jq-60U>(!v^NP{Y^PzuH?Rh)!eFz)YESFHDJA^jO! z>Rq6Ol>_`OJON14xh9;vbMf<;mE)cA#GgnPmk+YSff>9z^>R6>z>zmONtzax@zO)> z8MWa5{F{^#!f&Z32Y;`6M3OCZA8yO#i1~j#A7n?=!v6r#09u@K=&agR!w$vrn z@ZVfWZvQ66f@$wP@$auQy}}mJl&T&NYBo!&O@|vLvJyF$JZwL{PnoKx6lBM{@L?i| zb=)OtWR|(^Ki#?A=+0HCK>EL9-?Tu`oB3^6vXavOeu8$rQ9yig&CQ7NKb1#6lr+|c;`$NOvH**813I+N9Tn!4YzZvd}c$o~pZvcU-T!;=GZM)37{u2GJbMov z2&((5-$I6V2DLz0KM|z*7?SWX={|@1sgBa%pr$}UVz39#j~U#C5*(&h`BR7%@bcyu%EyhD9Ys@X?{EE& z$jlf3m2IJqve3pMzl-G0yg(aT%A};*3l3z)3WJnR1+wO?YxAHzB1UB^xRVMm72x~|~5kE#&x!kE@^l&S_R$v(wv z0X-c|`t$Pt8A;;Jxc@O3LY1We3!Ug|J)9Te6x7Hfr}aiWEOK+zT;R9S{zf4Jj>R~o z;Fi33`Kv#59D<`kZM5g}^o&qMgu$<2qtpMlJbVlTJMDuM7#a1R3W1}UYJ2~@{8~J3 zMo%2fapOWVL1ZBMn9qw)I1lmkfu$=V)8oPpBbGN?OUg||fvtD2rNe*I9mbr}#!}t9 zI;OpeqXAxXulyR3PGz+Mo_$Q7xPR%y^aW7c>>zU?+WtQ3g%0VO>;(=qwhq8mN2^ML~rx9qQn12t3sN*lFt zKfnJm&*UHC*4@WpqDfon-;7B6wLWAVdM(r~FvdWXgk=AUV<>FsE=E{H`ab{-F?0La zZ`+Hs5)MWvh}+B;N%|4Rpzfb;t}VE9umKOe8;lhGkG(9&S2NG`iESc{uZLo9H~uFr zT!=X!X0os#{5uz9i5wF}CKUyTa9oI%cnbbc_40n9k6|rN>mpH`nxk{be#u3Ltc3G+wpNc?3INW~2p@ceu2K^|#nxfhVUyIf2W=lB zx3EZLLE8nTPAnqpnT?mEo~(N~AMWC!&4z|o8(4LAv?qJ4r|9!#>xrj&{A>`8*J ztV`mQ^pBIa`Fcm$`D*AWoM0~7t=nGCIMj=(U#3@daZL@@2VYToyx?|YH^GG3$c4hI z-VT>f!Gjb?@xtHH3t*ryKXUbtE$6G{+UuJkNTnV&i0OJyNzJFU_w)x^c=5cv!%l_g zUqJ)@Y)h?x62rX{KEuDL#* z_^))75;1;JfcS)&wguGx`1lUW35CE~xQ%FlSj64C_CiKYZUp6-EqLw0%J8VN=UH?1 zk10?yMB-7NV<3HVQnFz5d7vW(#=&+Nu!;5yo#ur1u+vA-6|MXh&M(EX@W7tLd{Vr7 zv_c3y{TyZCW*>mh3)Z^qvRb}!;rY7XRcOgsczti zhoKxc|Ch{&QF+=M9L6qzvM>`o6txaTDv>|ncwVJ)PYYtcxTGQZoSbr2O`7Ft;tzWu z@_?~;St7OgA!EUoUEg>(LxKycxmp^a66uG-bjd%t;ABI(Red=;7mjSkwJkuw5+5*1 zmkt=v!&P2`?7t#nM&4!e)FQn=4=WBq&pxAT(ZGd)q<3^MNFtPt@&c7s5+l0!?iMW0 zY*L!1acxH${)nwKOH8^^c)S8Ttpr~?R;kAx)6+ME|4ouD#)C^HT5NjBSe_PZ9b5&M zZLv{GpA@tr2cF6Lyc{ft?9t)OG?SRMBlC7y%VzkEFeZzN`g_TAhC&Byz5kZ{xgtV~ z2lIZUDkq`B45WYZqDdg-KO;D;{t zc*BNPck5W=cm6}iSc)P!Rs`UJt{;nx?YH0w1|L~PBB%GMAW{)%gJmeinwQjR5|?$<;Ph@s#YM!IrhAdE38rwRm65-1~(8s?)wc4t;~W6`MKcxm(PS#ox36&<;#>V+jYqSl$wOe=0| z^79JM@z_ls6s#niExA?|$`RmxoAEQC5 zhEVv00N{1WG0!prozrM?a@nNePHF7VfwfPjMy(hC zf2)8TI)|@XL~U&0KWpp-Iv?V!B6y6U@{ycvp8AAnDKAgL@*+&gyr{2q^&^J}>b3bX z{?PFCl8pAV>evlSFq+kWYhV=`OyCu zcKq?Xs^<%A(!ciQ!T+8HrdG)5+~~1%7ds%>cJ6t4qe;uuad-r0!>20Mg#~IYa z&fbl$b-bE^M+6Hke|=p4lN;*$$rdw#;@E;JT4()A*7!M5B(@6E-z9K;;>Y5w7g?r; z6O$Eg(xC@VOVh=_pknGmL+_I2A9DvCf?*Y2prKi4c-tiwAI>?wiNwFL}WY2-3-G|a6D)z9I$=z7TtEvO0$X~I}@|v3xaCzqW zD1~R>-Vr8ED5Z=NW@`i4E zH+??y=Vlz>dMj-`X+rp;OB=K=vuAKK5(VWWo}=f;KOuN6tj^K5^dWsDG~EPDCZwO^ zJGUnCq3J-nO4~oxR$>zLARypk(n4q<-r*mb`3it_cS*+ama&RQ$27=sZ;i)lkVGKXlZNCeK`K7<^|aQ#%q z%@!zw5cB0+&l~VcerT%^=R@J2d`H~kZW~(^*is_I@s08_QS0%v_yDW+Fv^I3Gy4{RC^GiB#$vIcd+X zOJ*B>7ehS(i5Jin*S?*3aOyMzB+Xpuw9|y~X)5N>*t?U1tmw+KH`nM4w>c?S%T;+F zp5po*g(eR{yGj%}X#Q{8WYx&BMMSS3x2FIcR@g&SZH0WaC6E5JOuussLyM4G=}5fg z8-92-dXt_ag*$5=^?Cs7`~W@7Pn3gilD>4Y_)d$F{#+5RATmJ8>VvW1130B6kW3=Y zem_kQhn6yC+Sr%vVkv3)M{L)48o?$rZ4Yeok{6zQ)ST!$YjtJyX|?Dz@hI7oa!ofS znTTbSGioYW+JRAeiGkJr?S5PCbk2PdxmTd1;u4Q-jLgePm{5jwu z60wL*bA|QgOt2$)t$3L0XbboZYPl#Oq^jTr*=6J;QnjPOB5UE{rb3yQMb&CV5!ALa zF`BN?9kcsu^3eJ)KjCR}^&(`Ulirkd`t;@FtwO^yylqREH*@GEUD5UMOz&u3vNmf+ zHp;~Q<=KOS-Gv9?H+gS_i~Iv-dM(#S)?4!VsA~~y5iEF!FclX$^Rygyy2SQi=9c)& z$(4hZKcl^j-_1Q3L8^z>j~{!f321P7j*Pk2lxr00QYLW)wXB6s?fQgN+2y@?UTC66_uPz zOVV_~(&Hi8e!i%6=XUYJNqVBuPm#m?wCAAI+3N7Fy&H~pj__w_*+n}?!VJkTkAV4K zuQv173@@W^@3|M1W!Ocpq`~RpA-!9m9P(QXX>GB&_0eyxlZ|d#6*Pf_jQ*Ux(0{?? zsA7W{q^`4l#K$oLIgiD$T^3F6@ncUrrsL_v!19y78 zMLM3MyOPk=^pbGdSGxBU7d8)2!yeEP7@;LFKsX|Up;U#w83Rl2l1sLn()O?&Y*;t@ zUkPaSf8MZRLG>dH;isHx@`>+zP}H-SeAzR@zobmhvYDepxGINf^iO&#x#!BG=T0*j z`BS8buL~%D#}&m9b@d{~PCP_H2Oj=9Epk!U-n^jgdKpIrRy9;@?!xsQ=p)XvhaL19 ziwLc-$ao9=X(Ki@M!tP3b$``->RjP^m9>8z{ZK`?uiG;BnSC4g$bd%$EP7ZdlbvB2%5~%lON7Va+kFnCYO8<&@nGv zIZmx_(Qk2){alQqD25fORODFR-Ag3C*8C)|Dug-(9qHL@^YW z_i=?t(w}^1@G2Y3kX*cJT|`G4gp3}i|~7EXNw8}&z~X8nRxJ)dZQ*64@SdG+&2HZQPvqN zuvdE-7k@laWrUu+z)s&MiPV@D_-GETDF@ql3><_Ds_)hvEnbL{RQvY?yE4wVT=o1A zV@aH?tiZQZ6R&?3(&0Af$vf8!>XPJ;0EfLtSyd84sn2|~%}NRt)x(Q~YEHEf-&R4# z{H9Q+&aTmIl+za87t5!LudhlSrrmo7+-FdQZfX}b@yEo&({{}U12nneQjK!>jf^#8KiCQJZ$)LfWWWiF4-!R^+Zg^5zXV-kEDWUn|_;I(6GwSJNFBnC>}GC#U(tfT6gR2 zm5nGk3mqWKZb3B9{W1Hbr~W8A?e=wn#n+?g`HF1MmoeRl^+p|%|A2TYuOd%OePgL! z6z%#(;neafjQ9pwr@Jg1zWoc)B>{)4Tw(+LU-cp z5P8U|dNY@1wj;1I9eu*7g^6lHoOB|+do^dX@}gACJ@MafjfLGX*X9>5VD-=!G4bZo zxqI_7dx5tGmQGVR>aKEAGa>HOZofIKvM@0}R`}{26A5GFmK=7$W#qVr|AC+tdD!fQ zt)$J3yDz6~KZ3fGf)|4`_j2&ao;j7hf1O4~>K<2a+_gvdf{}Ab z*m*hBdGJpuXY={5zV9OispTJ2B9`VVS;?Bb*TvKFXLb)#hCrV-PAPS1RDt-q1zfW! zPG9(J|E|@Ci@KJ@G>_nEP=~S_!kX&3FhT6}eAeywR>D`hLZg7h#(FhkZ}KSE`~=kL zQ!V^fH~q*v{uexr1EZUTmYzLVEdF>24*eMH>o}_a(&;T^aex@RHXb6!y5CdyOp^ID zHh@6KSEKSAM7qM*5NFjoku7Qb9(HNO?}fnSiO0rYHU|}e$ZRu+zXv^62)9UyQ9ok# z7P+B^a#SbKiXN7nmAJqZ!3=kEx&@>-F*KlkPG~vYN@*1)yJJlXCX(R%lCeyqvhc{F zt|8^iqb#S-4}r~>ZC`Hj4`O^|c?~INph`~anpL75k}e@=cnMMrO7Bd6v8PZQ)H-2x zUQ|9W%d;0dg37p|c)!?o>JYt!lZz4xx-7^r9f%3P5i~5R{~0>Y?GULO#KpIB7!G7e z)-dMSl&suHk?fYhkoCZ8I|R3tyj>TFrFF`#wR0J3f+RYg5>H(yyk3BHppZndgzyIy zP~-ISVOMO=tmqVV9JAeNMt59hGL!HAs0#~+{y0pGDU1}QUUN%)rbkaq1UJhX@xX*< z)Ps)D&(k9xlEoENLj4+3A3N?>+7UM3rIU>cF-x3paNG$XpH9=~ebVQR`f)K!hqrUJ zyvRF7J=0g2&^uiMHeHEvmS?sP49{$EEaoA(Hz9UY^cFI-bcz;(#l#~G1UMQQ2r~qa zKih%ZXo6uDz0m4Hfo-6zl(`v0R`OK(DQ%1cN9the=W$TTU6>R6G;L_{EOGIv+F@vcPmB;Eb zDgZjRdtW0lH(@EOUUjKFNC5w(tYaP7?aONnp6`(w!F{il7Etuva2d1!R=N>k1{`16 zOl0$MDUpn~o!sgBHHhx-<5JSy2%ODIUJ#%#xo11Ff(L`>*8}eenS#>vy#~bbddV~5 z7y^CPiY0<#WKjZb-=1$hU*r?Y)rj(8>YCV^eg-zr2a$fgjtL7j?YZIof`@F4bxpqb zw943ba-HZb&(~dByv1b3a#LPbvaMqjpC0XcfdwlGaiGIm53jr!m zs*%hU!6hSbx($9D&01Ts#pn_~AOzU;wPLG#=B}Gb9{Izq?60TXk4gC#4fgUqlU&Zj zpOsWRvXT2({n?X|G{d3V`aH2hrcR=n_XY>Jz*rmoL;w5Qh^gCP@p28_l;P+53$A&p zVHkWgH<@ML!>TXe)(sYl>4X}x&klTq*PhVGWqXw!axF-w(c62ihGH19bPocD>bZVg zu&e!~%yks!V{QL8>)(ML7I>F|ARPlMDq5yEveAY>sr$LqJ?)4gA$Qxp+9$`>3+Pf0 z8vDpzzfQ#KR1u~5^(`!88E+Xfede93db9hb>M0KY%+kG$(GIUO;NYeLaa*Q|*`}Ip zws^OGoshT}cGubL%cYvF6v?K;H8qz$Mtz*)x<7oFk`3haX8jm$nu^ecD5f6KX3m!e z_?ycQ8*CsLXI^H18SXG~Nb=*SeWBb+T$7V@%;@5-`_lA9>9#ov4Z&<<2|NRSHbayT z#NB7X(#XGk1100-89~=Y){ki?kuiFWA)qbm#Q1~F&msy*oOfCAbyP3DNTtfjz_SvT zfRZBo1if6La_a3jE2CY@T*8MF@HSohaN076d#WL&XbZ?~*4D_^Zq=jzrH?he$ACNX zbfo|hrngFv+(FfiX{SRvvH%|#mW%>f!*?`Hxq#1-JQuJ4WF7G$&i?=p7?a^+7WNX+R~XAxqA<)&8hd16)%r1jL4d!LHA8fJ%#z17tvqJ(UZH z7P5?&t5JXv#GTjmYoWp@z%mkpLOet*9y!`!5mjK0UtM0k0BKEx*u|(=*n^ngGIt!e zk5*2;QKsi>s2wtfukslxGh0O}RGzevVh%0jgDup4Q9}_r!h;yWK$)1MTYIbD6)~2j zVO&ao4g({&SMjTexNH5hh9OQjSOvaEJeQ;W7FH#eU1sHb_943tVl91+lGD3ku4X#0 zPmt_Tws&h&n{QGe%3a<-C2Jx8+O^6(Nkm54l20gpC?2!D{=tirel;+WxUx$;=X6*q z5|DkPg719^0~FIDN-!3fstaNP%Bs?IuocBIJZF?o^fe(3LNy1}O!qb(!-~ex3;*yU zwqrTqf4u10cPkgVpld zp8+DD_W`Qa{v3Dxa5lQ9Ik+JXC666z_ze2E@n0z!Kq^ljg8rpvM49nqDtqTNjgtxu zUZ7gP9_YEne~BB=>s~Q)DF#I1^MVn`-tul?=VPce{3v1YzE+X_aiY?_j|YQri~v(! zVsXxm>NJG+hlBoZ6P?yZo@1zGn7bjFa(WCc+2e^>68j?6|OOy6+uHRWkC{`?~Y0;n~84 ziLltGJpwVq-{HkIY)x?i9c60f!D|ixl2&X|8d3D$rP0p&1Yb4&h#^@-t;2lX8or_y z$59Tpey>s=orU8RKwYJKRxF1!AWe8^u>ElN68=Y$mb#20EbN`Y)O0t$Y!(b+16jN` z+hScTck&V-DzQko_p&U#@}1Yro<%4S~Y~ zgkRo02mQPI+pSh=^RZR*g0pxE}2 zBTwB;$TmN6)@#ji&c~3(mqV%RH_NPS&Q8d2Qzy_4Su3xMea>40fU*&rmXnJDkdXNcAHt?C&_(A6JKp6ZH^nq>P5nc0GA^iz zkl8tkogi?QZHkewm#OSJiZdQ~et`B5L%z!_;Rj@fswm5Y9vou!b27(L$s7cz=yAFW zH-_(LjC!pL1m9E*Sz2sy2S+t&ztdFumi^&y=LkBZUWIVj0D_l#+by}c0R2V-zw1Fbq46;**BGu;hQZ-< zfjL)EbfO5F&+~W*l@GCGITL=RlP6+YQlMG=8ll8f6`g=3%a9=29YNeRM0@5I4qN9J zqQU1(Itq*qNQgrWhwn7)>sAP87C=2-m`gkXdVmXC+pbb!-R|K&=TH~L4UzrK z;_($Av~K@>l0Ez09kVT&l6hF@o)X$FVR|Ph9tFPnQ9Wi6C7MC1+FV)`a@NtnL$@P3 zjFx{m5805ayU^PDXX=n)Horxpr|-sOmi&16w_@y$5`KddNX4~kTH7sPBiHwwS;%IV z7u(SU6k~w!@%*dCSw(!{Zi)^Bler$sFNo{gQVa!z!2_t7^~cwPzkc?+P()sUaH82i zR^xm?T%xPRk?mu4*xrYZ+k}e0JAzKl+tXLW?ty6;O80fzpxb$@ekV(6E-RQ7qql$n z`eUsd2O|3U-@copR`S7_>rMw(J$HGi)sP$Wk3E^M+4lL%n}X=@8w;a`WoEYY()YeQ z;s{LS3RoktVkq65Zb=c2BdL4$jol~ia>YJ&zI99B9=iI~0sQUb0E;M1fbM_8x5h(W zxf`b0T}H{z4D6woP{at^1(6-&!?WZ3v=oVu9eYOUeu*SmYcop* z^d$=7g?H0cjJ2@fdVgLoi^8pMT?0|&GpuoPtDb~LjI>9!<&AM zQ)0`T#=smMbpP!ei|!;uxP@*hZB42K58k2w=kH3amM)91?&9ikKPu9tJw#OJdZcW4 ziw~gkv$!J5uLBm1u|`7~fUmu-FR8rf4sviU1J&3`eV4GO&l z@F%bg>6gWEAeWt`7ri1E^CXO}=*YW*_11C?A%{uMge#BoZv^3v5ebM$fL7 zEcj^Aa%iQqu$CokP`eR@C-aC3rL^4U0zfbFogiV;j<)q^P1rzr?8f|+3S1B`{IjjJ z)?IK0l?Z2L`dQWm6Sq+l}7J0bI|3TtkR4U*UReB2$YyR6ekPzqYOtXsL-9Le8D(1Cpy zEi#-#mysxJ^Y>Kd=!CK5Yn|*R62noqfqM;1gtYPRaimWgat0n(T9?Y9-kaqTI=haglkszDfQ%*o|q31#Ch(Xet<)v~I*sXg(Pu_~z`T9kK5FTkx3#c{6E)?l??uslEp~S2%pd(VCdzj>^Oa+7 zYqX$+YTabKd*rG?Kf;mSPh)7B9Pt&a!^@AGvF3@`wFL0V1n_v}KU!mb$oS*-fWzTi zGMR`fRfBt>Q&1)CTTsp)XL35pA)(M?#}-y%JeR3=={=I%QBBFpy?GV4Zp{}o#OyqR zdM)7c3TXkkG+Aa&gpshVQ7>wRpZJRZ;YT}N)Sri0q1f8PP{K{CG8B(xd15qL#TAKL z?o04=s-KXflZc^%_3^Z~vh>uv*m8*yKpbXVxWq-I^%k!35c(-Vp=D(dw!`lpl$kn4ZgiucNO_n8ZxS z1`}ytSeF9HrTA&b=~wYY5q9PSY~h@cA8z_>_R!6E`s!d~avu9Hh&)n=%*J=&yp(|p} zZFidXj309i1>fTu2chhkLJjan(`dJ}JA^gqX3Vo0lxmK3mB%wiFic*d_kPLIwdmC* zEK@$UBv`ql0vdgAwX&%92@Mx$g`f*`2;=iRyqAvDyJG5z?|l%-R@sTYhWBi z%R!H@79hg&zHo=3!oG(6wNn9a{2JB)tR3iidE!km_1^#&2--c+T^oChB6?S$m}bPF z5N~2hZwP_&rWRKK^I%Qa1*ON|X|+V%x<12189+5;o7f2Wbc%9<+J2QGgX*Wp`g+%j zk~iCpCN=8}<`x3@oAE*xz21fM9Xdfb&kuNDxIxfxjFD_PeZ3Il4}}^FEn}b1KvokA z2E+-UH*1t~<{s1@DKUHiJl0FVquE8R8#H=}Qt3yCL_{9igjk z>~A=c(-AT|&*NEaC-$%b1=z1qeg>u-hPhJ!^~034RH(7frB*=z?`tasAc3dryOsj5 za*vTw)wQo$*#Z!=?gnKEPSWj#(kObj7k5hhjKlyrpT|Giq!Zlpe~rWmr6iz8UY`eF zZFd*_^crwnAYlou?i=9_ync-aExNM+iTamB#0h?ut|Q$>frDR0bgGfxm3fFn^_6xU z79ihN@OXZ2xCmjx+(7H$D;4Mto4`BlC@_hGqx zzGaStsaHgS)6nuns;2-WFbu}qB|ZhPZH`vg^UDUXHikML{Q1^XZhmA|Muihk;tkhL z=q3C(&i*6A<;d;F24KVlSow46sHp2Ax&6JTg9%S% zino1-On$6Y6gwLGU#^k>@V<}yi9Cq!b{$vz>3sg`$C1m-H*~f zWUoI@JBBG@fjb+Q=cvP|wu6vmP4d}bit6#<8#jBs?Eobx|MNjagR#b}k8f1i!9E04 z6w!2Pc92nEf7g?`vtc!K1|3b9u!w%pxOt-(=lA)~AsJ&RRv%ABQr=Yp+Vwdgl5?sf z-S&=>eV73|fzSI`M5nS1vS@tqPq8yXy}0HVXzjgu$~+P>4!hIsn_#=)6y)J^OkP;C z303|a20{(t~Tk)w^z9c zpfagqi)bX>?o?ldG=&X9-bj|4*z6_h%1Cnl_xJPm6_Rh2{PhFUbZUzs`m7QFV2sD> z79Fp?j!E#(7_o?pI9@rw|#J>76LPY zO&(fA47dCJLg)L2)_=vN4AXFiJ&eGCbO3pfJ!c(P0-2Tw;I$y1vb0N?SmDfJSc|S! z1fCKbu82Ws(B)_Rz6b%>c&1g%E*9{+O#EXldrdpP+J8NG^t(B5qu1Nvxi{~qP~f{= zc;y!Z5~qkw4!E8l;D&vrTy-2EP@^V~-SK2^)2z##uf!1xF7di=y8P)}MbWNuI4D|e zUKd0Zv8a+%KnH>1^i@~$M|8AvuwK^9yn3XC%?&i0L)!DKr&QS0vz@{o-+S#JjVBn| z19I_273-gxWcNnf=*ySgv8`Ky%Ir5XuE~C4(}~kGITZM7<0GuN;2(jOK+!cphSfU# zCt*8uGe3C|=_n`DqQ4EAB3`cJ5;Hrn$t<+z1v_8`OBEt`M1&0l;I)_97xWn2wJ7gn z`nf5x->&%0#Uj%`d`;SlfD#B4HOHjm^g$yo$tQxyg-1eG1|xkx4kX4TtKA7vP^teN z7b&!H&izxv-Uk7FXcU1&lm4bjk)63{$*3R<@i|(xc5=!;>%HB8!5$C!MNGd2xvt^s zIxx}qd<7wf8l5qn5(fbq!Ix$Fr{as2!Zp}_rHKz9tpJLgYlJyDUa}WJ-Z7(%%J-u1 zUo6XeKPX}d82CzPTxDG zwN^u9wuE3}?LgsA!6D(G7j!FGsEa5coP{&23bU>5r5CgS5n!lydJjvC^bG(m=VMwP zv~M-C{XpC{y`G8aM->h*Ab!%^!`%Q?9tI;V-UVR9Mt|!+T@rUW2q>As;Xn^h5=R4>`*RBqLrCzcg$J^e-tCn-yK!XAwaJc#&Q*5A zDl>B*Y0ZCh7Nm$g`0Ys>ij>ZozpK+xZ;$fK^`CzK_C|H=$oZ?1>@E-PJmu!ebUAfV z)}oG;=&*9qVNB-pUheATdxNMC@#Mi7lY35KEr;m*_qf4q5);V#;AQ`}<;IH`)j!`a z*qqMZ!So#6E^Ih)h|coSV>?9OuK935GOFa&cZ6X!kRa{6&QAwa8IV`MzH}RfU@kkJ zZzpX~U@I0l4q06;yWb~$2E5T~s_^o^Gb@bG7kI4oeu?<{wlJ4xv+C2;-5eYO!^dg& zCwWEWG+n?sB)fse_2`>dfJ;JpJDR5d^n)<`=FGdW1#1A8A4IM+se?XbUa+%&II7}1$~`N8@_RxBF;#(aPiQs)&8GlCWK;Md|< zutY|Dw28t9>p9c1MmKSNW_TLM_ZtwmH+1D1IpUtn9}N)FyWjMnL+j{;jD%0JBKx-B zN@sf{i_Ze*VIuN^0Q!u1g{{v7HsW5M^fUs}epJwnl&!2aRsw%}Z7xw{Zi?mkF!08UO{R38P3R1YjQ|-cWH(EMx>h0jvrBN|-zZQ!06qI82}< z)-WCPF?Hk~R8gpbP*)tD2~@v%t2Is!!P2L%7D^j9y)~4<97r(qBb3k5) zpO8@9(&;V`j3oe|TO$NOhBGsgO25e|jT`s@0eR&_7)d0SHVZagye~@Zc_GA!F}wdvmf|!Dk87JnV&?`XEB%fzp^>TMX% zsdN6V_xZ4D;PEY6bf6%6>0dG_ZLI3KCs=cb@JE|^hyXD31AWOBF)R?<_+2!{Z(S0& z_fC{S7E1rf&IAg26z}C_NAHZP-`nWE&&lByqznu%K@V~v&2DZ^HH%Nm8DH$rZ%c62 zIx**Vl6F6{%-2m$4$o5x4v{jYrTDElah#t10}@jZ$Z)mW&i4=o*PT&pAefV8JpS}L z5Z^w4?BpE!V~=NmNmH&r$jQr@jk_;%s@_2G5=9~Z4LcZS@dgS!ri@p z;*TguN)PEc+gQ=Mu{?Gq%@TXU4`n<;kh!u5IAR{e&j4IF1|Wh(pgiLBytVjskkjQv zd^e--%#Uni^SFh7>Uo`~l*ATE{{ZoBhMh|3cpGI3=viY9JH_#rb?2h zhdR&%0NbalQlvAGXe;}o^PB9V)4dP!mO#Qe;;6j_|3wufB)8$abh5qjb@!N8Hs>cB zn60MgjTcGJL60gdGf@)75a}WoYiO1&+W6(bfewyZG|r26(bclqqO)qESlgP+{@+sG=^I|0|u;Utj7 z^1+juPxyX7e2+Vj7mWXCv_RZLhc*(Tk3G~yEG^5!MoA8ca$AtiK5c_`yXwosu-~Q5 z**L!FF2`TLhNG zHR8GsYosy%435fUsXAjScl2wBT82Bc^i(rE*P zb`O!?hnbX+>zda_7?a-T@MQ<^14Dn6+@|a$_LvjHO_ox@&kw=Eig#{bvFOBN!mc4G zjkhwP*|eS&<;6SaLgE8KMwWE&B}WK2&v*0xG>u<>sX*tCj4x(3UahgL0VW0Ce)tk) zSSxkzcG6(LkoW03+9$itFoJ{&$0l9ynt=6Z^;HB>tISP&8{GLv|2D%>NgJUX z^~u*{K54}MymO)YH3_8tH1=^l&ZCzC_sOb$8z?@mM%=hlTi8ce|KYDt4u72*;sYfb zQ20>5STBGD{prnCd5~2(9g-P~?Y1G#ps1MV~g(?^{1yAWR#@byGKf+UIOY67kC!`7P*Coh)%HQ~9JmHpK9$fYxLcZ!>t zA;3^(@YqOkKJ;8*VB0Bv6{by!tBXe1;so2}BWENBUalSt_}EL$Uji0kZS^l$Fu*_J z2=OA;f)S}OOo|+bz(P6(AYpx13#w2xKf^RkWW0PS5H;>maD~0kbgjeegYk1ekyH;f zyz#nM+wJTop7_1^*M*Z)tYH0@|_@Cn9In>jLzEpwCbEm4(!+ zOMYNAjKKeT0bCtlHzNu8t}1a_j9z&ef{zZ>z7&k=YMp$*AkR2;K(m?D_-sJv#LkqQ zVQpcBdOHeS3{VPOJd;8{aP*B&5x3?ZKOxs(U1VZAOupAD6eT+_@MDQ-$`CE#x%hQC z3vuE5S>8)YY9<|F(%l#ZJdN2g^hs7PA}^bF;GoZ{l9mLYOHJuAn4K^Xz@nnoe)XK` z*H5Yo|E@eMtX0erMZ*R2{Z5q&Tg%$-@E+W0uo{J^!d@`gyfYj9hOoCc~K*JJL& zwx>_&ZM2UL`}w289IY~x3WznV*?F_a3pnut^qgvM1zD8OhT1#~n=DtWP}}5(KBOzA z<>K*tU$)d=c_4;Dw0n&}gfB564)K?0)-74S10;Atd7Vd`s|xKHKN`n9fxvcoodXoy@)Tqq6z(AeLM4P0v%Zp5U2n^{CH9v{9x$n;OQ8Jw8j`{jQo+q)ke`ieB=^vA8?Do?b zVQ@yB3xq#4g1{LI1TCo%PQ>SV`3>#q90TBYk-#NHp5t_x{fuBhus)IZ-6<_N%-3RI z(6dBM*`xQSeW|pZ?`kXmkD=yIQwMsGEJjZW)|_=$L0pLmA~oemWiz{x@i41uINtAi z>^?*n%cdmJhem<;5Wy21c`XbCSqVkIG&O$Gu4te0QPQV^EBc?6GvaTvA?j)ji{VJU zo7pUAfMf{dLNq?f0SkVpT)r#e;TPQYh!L)hTf@=no@&J~<)?w0r~xxbZ8<+K7^?xS z(+N5!NPWy*H>MNT!OFWaj;}rA&2>yyXAgy+RSV|!%OBBBkQ>?hQl_wWYWKk@0LM>B zV|y%w0TXk%4xnxqG6+DmHAJ4yc4-!T;AqRaHQiHbmr##>K9cG}gWZP#e1IxZgeJEM zr2%5d*-v9Pykk}_f=4()7E$rsXV4F=p{R*fER@I>qYqYnPt~ZIg&(}0omTW+hV8v2 zL#iVZHLxj9bFZS{p9ZI!6Jdkv1;sqWB~UK+X-zMwGWWgUvIX-s#&g2!S&yVVl}APP zMeXwLvG+tI3OGTsWKix06!O+usAQ7j!Ox1`OYET&iJx6w&-e|)^jt6Vr&30@~ofzujkkL#!Q#+?&gcKM=z=rtyeMe6g zAk(8HEcC)bfB$e@jX_(TUKfGPjpqL+_r&)5o_&Plp%8E)QKcy(Y^+B0Zvs6SMfN!7 zwP&;_p^w?;d-VfbJo(*)3$Fm#EBoEAnBT9MJMPKHKK;S4@gRm9w0igR9OQblTj~4h zu8w}^i@8IRM&80i+EcJ9N|rlOihzuEZ@Cwaw^n&#YUk1wk)jOBYjOKKRu>SJOpTTe zt>|FrEgH_~X%ACOI!v!|0N$eAe!eTRdHk!nEislg*L>Vb6} zf@HMt2M+}07ToDJ7*qm2gI?VR%nv1>bY1)ZYC7wvsJ^d_-(lzmMM6qax>I5Zm2N~N zrIGH2p+Qnb0g)D@LAqyMUV&~;0?NJ zkPVIP5SY@&g}g}TFl$#PR#&MWUSAjc%)J~kTuDtX9-DRIu)kSQ7VyM zA3OfD93;U5gA3&vSh?dZiEEfAZ%KHC2&u^Z`-4;;lq_qto053s?AJ3QI*U*O&pQ%| zT*Wm#cLPJbts;K;ejXg;G~~3e!J^qZ-QdfFN2eZA9HqaE-q8OB&V3*q%PB`O{ka6U zD}U<|dqM0WU*U!SkgQ-Kqe9OI*b!}IjF6bC+Ig7jKdU5oH%j;R`!>!+b?4o?i-n?H zA&2bulqm_g;aM;3goCTu9`0^p9S-(A?|*fY9kK;bBiBtk45|;aW-HQ~_V0YU%4K-m zXF|Sp8*+s-+B;!!h5%lm+ru5CUj#b-@a(!n$=Lv>X)*2-K*t9Y^jw{<3B^tYAe2$uWnN z6;nC!r690IpMH;#=HdDom)6=Kl~rH&4#?{K{UBx(+6s(eJ{-R`&A1re*(7{*GJhWB zap3PTaA2dxn4iiJG{Y^nv)=#Xe0axtJTGYwmhBI~|KoMmhgawN?Giv@4h`=MQsAeB ztC_w2dntJC2l*qq>SoM?^ZrejIG((SW^f9n&-nyKg3$E>>`@(DhZ&&1orml$*g%k< zpc&%trmuW4GuhEbLfPAfXE6RGcbqLW%FPGH`{c^g;30Cq-Pav>?h62Azmanx<}ybi z$I^OiL$kFTsWfd|!d{;J2M^oDd_3Tso`xT?*=o05)N?s63%GMMMCX3zd(;&R0Jjhz zs>0w^tiPQ^iyTyP*O4EBX;u>M|)Ayqm_tiAT87Ov6s?R)*iw++! zdr%y}D4uCI&p3E&Y{1LvTUL)Iuu@_t_DY%aidaRC9Rbc#b#96FLjfUl#2VpK(&(4S zOS%Mo=IdPTml6;l<1+&$T)_z}FJDd`cEy;M2`Dq;KkYkr)xYV=6S+4ekNW_CFR{wc z-^yZs&nsZd_CVlVzAr6(UZ;qU|7<>g(?c$+Nh+z#6YZE_KIl6t>{(h;l0w?Q{Jv3w z#Ke~SU##HG%*{3id@_iY1l#dE`8)dd> zc>^;V05S7R5!l)j1}P6Rexe?W>GNLxy$xuxtx7q%dpA!Rc2U6SbUN+xE1ElHITv7E zYPxPIlQwhXae6d$gj#4OD6g|L0XU&t#=*weF$5c{Uo-7J4?gNqqp<(|m|gMLGAWzk zbS-}K7|(-SQu;!z`IPI*ed${*ZEAl08ez$cAeImaTJ6mgmlzHR*{p zZxNw<&G^}sD~AOE{<69gD$sms{dlJrN+?AdF*{rtH1900$=2L(5p zqrr6wD0ZbC)lU>aYdJM0Fw~F@t8SMnd3Tkm2nm{$fUZgx9l2Cpj`$1 za%hsx<8tRMUhgX75}Mm#r-VJC$aqORfg0f!Tq_`v`99twxwt=`?tw>kJ_QYQi;%iQ znTD3WIZ*>gN~44+ZvRNq8Yg7*fK4@ylGAy~L^xukj9@3e%4+P(asZ3w8H)?jRNQr| zO3MNQKAGX;1|aKxd$>Bz?1liIhp5B157v9y?AWUt;zb4|`rA>x&E$yrO8Pgu^g3bT zV1mCk^0GCXaTj!A4_c0}1?Srv9UsN;_n1ZCgJ@@n{u@0Wyn4^!`oTH{3znP(zD>^6 zoMinu;Jh@^2!TzOy9L!i{4Xed&w_+br-Xh%;G3XuQZTC9+KD1SlW^^EJB0I&#d^ID z6^uSGZrUpHD~7<^6A+f-0;us7vly28=7qr1zSs8UAy0-piIiK{HHjVZBO`+Cmgl}* zR)nH`4w-Kt0u)vtz;i!?3>N^-TnJqe1b9!xrGYFfq)Gd>e+>a2w)4v5`I?uB`)90` z25JRjD1D>)D|SHi zaxTQ50I|)riM#iI%6X??W#~l#Y?DWXIyKjgHnt_Ki0$XVQS;+GI3@LOU?nPhSe1cc z)Tz5w#(#S*To2)SL@Cxh?0^N}*4-pRp00-2tZWhV(><1vh^OUJ8?dJ`dUL}}qHDKe zI-M_=y!7Gpl4!p=29fGdbm*4$cpv_f2hP<`#|x(DOddc^wiqs8=TC@Gh&V;uQ@S@!Q0(*x4AO?p6OOOTkNHO;A%P)`9*N@6}x z9Yq9mAh&}Niz=zta+>%AXZr41%k!U810Z&!_<`>IM3>8)!SoSPBejP#a&6T;#AfeF zM`7*4_s)mi?>~QX_0d{>!Jo3;gEn)AxFt9N7BbgBe8`Jza72PtJ_jWlJ^L zZb`uVf&j2s=cShspqFO?w*05Y`AWL)u(Q$zdB&s_$F=-x=W7p*^Z@T(t6={<${P%W z2d{@0;1Ni|cTlgdqLAl%HH!BYOkV9#f0v>L&=q$4^e~mRuuM-M4utT5`aA{bPebzN zAv9on9W0pFLiqVt%5iN9h=487gPU~u)dq_?nij6c=4NKZmzVR!;a20!Y=$a_032Etj+0kvM1| z1CpZwK4DZ`->S3&up}mZfJy@Fz#&&mzVbmT|b?Tf4Nz5V~q;% zNfu5Z7Y@o~TP4|HbN)~j_N-D}{5S)@;vP&bMPRuA(iB{JaLJis>TOmL0CG9OhhF+b ziYFxVqD>d$G+7nK9SMO~fe*2d02hb^A7}gj-FO6aU`N;PXiA~cHGR1_b;{{lil z5bk3D$KW4tAqCD54|u#p*yrnZ6XIiAZE+1=Q6~rzp}C7;iC5;+pHDWJJPB_Cv~Rf= zK|Q$!fuk-#4_yj9fTW3l01?!-@K`fz2;4=t42Zq?UOp*Y-&JK8X^m7}T1%#0gNXXh z{GdjnC;%&YP9W-|^?W%HRUs2o)PrU-{SeFpfYoOO06xz^2vEDMLLK@H-(tQZe-VU& zJg#F~9C2@GZ{)ztsrrc`vPc9M9DQi%Dh>n*QI(5O`Z>t%bJxcJg+T4|EG2l&!BsI; znu%UeY5hZJ@dsnRdUy2`Whe1X+?B1rZvXk$NYJljEQ}#n05I!^0V4+BedWMh=G_V_ zpau0A{Vw*JQKJ6uE-3y>wfVuh!Z`@c_KS~j+O723jiujh18&u!o{`$LeWsKlKs318 z>F*PN(lgc5LmZ1>QEoEjBs5>z@1=p2GQEMt}AETClfoDc`_n`E}|iriM?o+o3B>O++VH zl3^!pLXx=hU>GopL~y=*Mo;qz56A(lJM`EZZhByr^-JOh1P86akP5TRz;lJP~PPR`J}%@-)j2b+hwLg)xj6Kr%xBVm-i; z5DJ*TCZBG+CK|-kxKy0VS?A#$`KK7TZS)IldTEU934Ahek<88Nuy=^p^=hHpbe@E# ziDj1Y_wyA zi7aZroX1J5D^M))VoW|=U{9)>gEmvO8l>w<4luQZ?4 zz|Ib--F=Y%!O?ds0%u{_6Vv$2kxOcJQ+j>X_!&knYHQp*PMe0kn#$j^$K}@PbhB~8 zKxx6J6an%CvPDLRp~|E1=;hYGF==#yW+s22#LO}@z+cF7nItem)yJTn0Taiev;#B` zZEhD({I?x|C@!!jfjtG@z5wF*Hl%;QT*0Oz)G)m3cYD=q-)N@#kh~h>(vk%z{j2Wa zgh5Q+rImY3xr0f9P{LR_t9;r4hF=5wf?J=0OYD@_6O+}KjyhGBd!8v@X;F)n>nRHavlX0_X2}|^_^NT;XxL}{J>0^^^HRH7kK+*Y`K!nz)=1J_JfLGP zTZl#N)^0n_X|`=1H-jcZVk%DJEQYL^YE>rdco$J0s3v#%-#?^7~A3e z64MbqiK&Giy}QjnQXI-Xm0=mbN*D+Pl#!D z$v61*mGYS@!Y-uB6bz+-Z*I8b*#X@P|o* zyGsHe#-=S3JFzL_B_C?y8_9R$vdDx9y7Ul(bdp)$Tj=<_eW?VDp`z^en8A;pzvqc? zcq_bricA)Dv-%}7J`MOMp^rV(2lh@p+VM70q<`nkK+SjcJAQy$v!}@1nPV_j_3@LnEjS_aVz4cs^$6KQK3(kFu=e{CjdJxK-$elY$lUHEXhbhS|@H)m}ft zHfmlA02Dwj{TRAJK27LyULr;+QDm zhtFJpTUqx^hTZpiL@HDGMg#T3e#IQzowq0-l%MOkOYHZy*VtT|!4rf`G?t2*&ZVA8 zrL!>9GX_%goo(CmcJ)L(I;}iT@T@z+GV&D6wWUCCjyz^^#R-4N%`aymB5#*=}C=A zQ{JUubiTx#^XYqTck;IRq5n)sK~uL%MVELZ~AqJCc8V4Z5)k(=1P zwGRh?@96pfu-S$D_Z+kzHB$rVqy^+#M7~14H3v5j0i?wz6Dtyc^Ii}58Y9xPh-2HN zc1+6A>Y7X=p_VpVQZ-_kuR#s(0&puIl-9F+qaw?U z+~eMxaDa*1PXOqNdzA{tw4|y~(4_J(q9gW78k0`e*0i*Q^E#Dz=mc=56T1vdfBQY( zUVL01(egs+tk`93%jLpuP?Asc#r;%HZJ=Jz!_Cmk*6(a6o-Tk|@~pw!LzX3U2r4S> zFIJeO81(6^;Kg*{w^k8<-=Un5?O1E|i7YO(s&tIxqQIiNde6U;!WIExL5{C~94}z= zF45CJ^ZyW`@!Zf)VK@I(?K^HPz_l70|Edst&adr70PrQ|{>u1gpfcnI5#Zjm61Oxk zmk=RA)aex}?oHN2N=vfR+)|z4J3&)W< zI2si4M)WW*T3}1iQO$lKI4sUI?6g(nWRw%*D#v8GRpH}VhpF(l&y^i5rR&nW*bCe# z9{EGejqB5OWB!MCm=IVV{l1oUP1I_1#j7U||H zsV=6hbu?uNG)4caSJ7G7<&G9k92|R|+k9mLp~- zrJAXsUUP+MJMkkiVvi{S-2=ex@yPyL^1EPuPN457bOp*qpgz> zv`b*B_1P|XSgm+5?{6-`SN};k6Twx|oGUrghx@3ABA(T(#!ZCQuO5H%Is7If9J*al zWtFKikO{F@ZrHgwD{V09MFjD-jqpf90C$`n(2iI97hVM*M?UXZ`rj>%PHRu{c=Ny< z@Qy#Ip#!iw&BNHpq0cS9>hXUr`fkN+0F6K!JUABi4dBDRYQdwx|P zvm~iBeJ%uH+mNci2LQ@eMR^&W`8mZ-`i$BS2jM(#`-PnyyN%B-6n~nxhDoDLyBHCz z9&KD-n?T8SQ`4D7TQ$&~99(s9vx+DYc@fkl5g9B&RTxlLopKg|c2$w!DLN@<& zmS^#0mY=M?T^kk4QFDvs!Eq1|wSg%F;j8aXi#nM^Z^erJ=Gm_$62$HPvPrxo@bV=6 zRX^+h`vuTO!F*JO=50Hap0L}58WxApmF$MIqb92|BozWz{Wg7oj&F?#zL#jgXHT0v zG_(7Rl#LhpPxpQB&czNqkNtjq8p`qcBth~p8A~aBce2sbI{Bbi(4@O7JB1wGYzANV z>W@~T6KpReTn!oFxtA>K74bxO@szaoG?=PU0lvt-sTRVMzPS%0U`(_%yfMj}u}w+) zsrVVQpnabc;03GNVM!O(S-)VH;$YQKIqm7RN{lg9R-90R)`(%p^aFHB>wbIkFay!w zga5;n>G77&b8?b=8I&H5iUfl&jHw4++PW=OhTCo@%Dp) zR*Nr;2J@sVZdy1v6%iyn=~p>9e1Hkoh)a%Lc-xLh3bo0@ZHP;Lk{6f-GJ`eeD9F*D z|9IYgp@%o{!54Bjd0IYfK+Gh1ROto*#x@!o?MBNX^*t+*T{f;Jc8!6z>Ze2GMlxJ9 zM%?-%>`K#gSXX8M=Hq?=)jOk-(?s3VS08sbL*-7Ai*5TV(w&n@$_F64L!L5dByew@ zD933*u*1r0f%cyynhKcLvXM|TFKDatIUCZODT3kKr(u34f%{<*UmqQC23O%nZrXqps&!(XGcW>@ar^P9L+*Qbx17YI6v4KIc+=_Vf5gB}Dfv=`>X$>a>W;`uE zFOA%dJa8}RojzKOmiX-Mt9a&*oBqv`1Hlos__Y$X&hZg!ZBf}oQF=(SHnC<=eiycC ztDQo7`$s;;Oyc$l>g=y>Q~Ua?<(1uimQMspbvA$wBvQWATjn;Oa>C;l*#8I;;pfPU zKuY{IUw=>ze!*}yf4Hdv*opXO=-ihZO+Vmzi(UVfuMe zj^22v+?B;Fy6K@V_YPw>X3tkI=U;_#n{S-Ln3nX^u8y%DvD}|@iB)~rVYC5LXqG0w zv(RD-1AQ-V$zLJ$RxgoGD_qwfE~iRgx*Z<8d8OmTA2+)B@42th(4f9N{OYgU3Oy0Y z#_ov8xjY&7C}GlLk$Ku&Iv)$S0OL#Dn$~QsZ-f)IbxYY~2hT@V-_M3`94 z77l%_$!7PuBu)Cpsmg*bra{qIu@X>f12acfj1|by$WzvCJ{NQ?JLOYTPTTE!k+>D_&?QN6BglZQhp?Iw za3WnSl}7O84CMzxpm3b3k1v|C5&#=)BtvI*HX0TO?CN8Rw58J>tNylyO?R`U zayb;4%oBiK36c!*n1zXQFbjRpgj%?fL5hJYSIdP7@aofv>6gREoo8_Wyr_u6Q+gy^E31+Zb_5u(MB z5GS`&u`X{qgZ~OjCvp?Yio<02se(RJhU6)|n#NsKnJDD=HXl*~S*zx`Z`Ap%d#cs< zQDnGGK&|&M!N-cFNrag!Bb#Iq4+;PFyP7=g*W!1tu2>GzIt(8hk{2zz*X8QWQ!3F6 z32*aeoB7a!$d-MVO{-25FaT5Vu8h(`c0P9luk3a7zezV<=qHSaP^d1sZXWSKOArP@ zoA#;eNdQ7r1T#yEMzYI)1{zC8O_pnZ>L-hPF(43h^4NHLSbLis_-Re1@?O>N@rryd z=txsSTYmDgI~}$2wz8!sR`V7>x?c!1$gO-WVqLy0k_GiUlkC$B|~6((A$bgz3}& z{HPV!c_8|%qc!<2;IofBZiQ{?(zYT#26Hy!eS!cT#ZBFLbkYP}LRNFlISvJwo}V{% z>Q5$h5Qv%@0J%n9*U`Rs$HE)Kz-bjH?Y^OGrKjy3I^ zQZHx(Vs4#vGci;qq%uBC~h8+CROUR^`&}Z0B=sV`nvwK z`tGux?PM~$-`U!2pcje^)8_0Q_ ze)L;+|4r&O*VsRvVWnSxcx$OhloJ%w=1t6}Wq}Bz2Kg|vwb?K%WNGnIW0tYJ!_Rt< zaNA!3*Z^_lp9xCBBB-<9FgqVWW^z!|>KWp3p8SAGE3pde!>KpZ7lFwAi)`o^(Nt8YR&~bvn$O$b5wNN(beH0a$dJn9`|3hH_+5Qb!)I1uK z$entIBUcO1__|1&%tft$WPrbXlR|Z3YbrCn+QJCkPVy!OR5FZ2$?%`%jBZ)Pc7lE4E%~d8iZ2O@6l`Q zeAH7LwsUs0oa+AyHhNrk$(Lj1TInKSHeu3{q5^iZAtn1R?aL;_k3q>J{Z$a9f`N{i zEm-x7e{nWWnHCpa^C?^3V@`z=dZqvRDRN0bsp>raJLj$k!MG=ggxN%k3AF&@5e6;c?%#HD;B=Nx=Q< zmqD0+j@vi|1G(dRzf)W`9=oT%{@uMr^nJGxB~F6jkuo%2xJD{)I3+Z_qFbJ6R2rVI ztEJwy%4GCW9Qg4x+~&0pz~6+v1u57qZaIP5p(?NvG2foPgS}ou106}s2S&Rs>Lla3 zZzLamOfVs^9~Y$kEMmP?Wa=)z&GXgW{DS-U;JwK+E~tQ&T2vEekx!yJqfs|vnW=H+ zyv3%6Y#_L(A;LBb#>#aL4mk*Lpy;H>!2yt36Oa`{0!=XBTv8jgLp4q#rDjeLZ9q6X zEqZx9T9gJ6e8%dbdZGT$6~ibOFPp;YoAJ1DSZbH3$l=##bXdxtSz2^>f4KXag!G(t z9+io6aNMI$3KpzBPH4Ic{i2xb1)E@*0UeC(;{cy^?F{Jyax5=Y{w9-44#YZ|U>60D z42tTO*I-wvZ_$~q;p}*gd-G?oY^c>;-LE$9IKp{+)XKF=!Kh?|PE4 z|1Dj{33WG@nzR3_6#V(mpz%D4(^KmSGT`+TD3s;*v2GX-)AVYrvgi^*^1InH z&d)JX$FFK+acZR`<07ta%eL}u`>SMco%4IEaOK6j*K3PoG|?o>!r5_kZkc@J?rkJ? znTiCuW&_noxB&=VSF)h;t<_R*tsXc_1@;A;6ts@E2k!9)5)VIseNf~GH(1wHF3Sig zmSF6Ng91@ptR}xL7~tCgDZ`!6v$l&U%4Utw-O6S$sAm{^q67fIG~EqzK9K5kYg$!K z?GC>*r?^upz$%@>$bq2ZT5#@zeDA9lg9n8D;{S`S;N7MK-{S?2VBKI?n+&<6Ldg(O zQ%_XD8v!&Sa8SxO;pv+E-uDTBf6MG}M9%PhGgW%C>lD)C@0P3~WB?=a2_Z@D6g!}S zVDU3-+@c!0;^WmBr_YT&FcCL_;EkQ>uZP=si}Z#$NY-@y;p^-B(u0u)_z z8m}DONOJoiopC9E{5;Z?yoI*_=Z1LL8Qk7qoNVO5aUK>Ld*P zYUS|)Z-yw3a!GEpd#gSg1EP!AW`9Ss0*(%(Bsf}1Eo-LQzf8AI_9pYRern5cHERyg zlE>d9dV8t%_WDDSpb6^h_w zf&-)fECt_oa)&?Eo{`;Lj~0(la%}#%FuFSHV;fnU4HpSMj?zSz%Wr>6qy0GfZoor_ zlt1oOWre{U&*&(QK_uT60^Ng#c;~mksl8Y1?$o+|C<^CmMCz{NpUcRv$WvCNC$z<~ zbH?M%lM=hI$7RNJQVR$yvwC97J;&SO4sj`ASzM}6yfJ@ap_g#!af<9zf< zKeK?-k=7X{hgHT^(OeUeGvO zLKfJ<*5iB>Ar$(4QH}-=C-k=CYUN#CT)8W~<{Ex4h+^Ok<1QyJJ^P3`4`0uDKYP|HO%k` z`Mc!uJxO;Vyr=20e564IC4v)s&YVXrwn4uJgYe+7Z>)i9tU)eoTM};3!8v}smU*hG zf&S(LHJWjS&MIAy3zgn`MKk=>ClJxzq48&?bB{z7e|m{oNz8TK(o;~rsyRP*X!iBX z4sATT-V7I@kBmz7#>M6wg@RTo%+f3UEv#@+w0f&%a_CHHXq5O zSrGNi8Ag$uOe>Ecd5zNBPWf~U^y1J7ifXDB183;Vt z5o{{}sH)C?L^PB!x%)@Me_NdvMYa-pf646Vm7aYrH&R09s6RKH*WlU~D3OK=LEF8} zCR&+;agxJf?)v5f(Sw{{#D&oe?(1vR^I?OgCc)c!e*Oy@EL3usT@*rqU9(uFLU9WCY zT+!43m6q!kMA2E!M3f#J0*hhM|&cVB`T2Q4-t zYm;U427XYmj*(1aYeS8|FEh!spVI$W-pX}Sl#poto^AA1lOLz6HREu9BGOzhxiI(6&aF314sF)`_zmj>oxOQRDBt~Wx4Za4s|HFM;7Ojd^w@_ zz1x?+mkh+doId!aODiZ^1t;aQYqO~ zlf0+J7f5cn{89Ct%f=d4pnYL8wO_LL-hJqupQPu-qF;FpN*UQ@?RtI{vQ+u^;6%?= zXr)0g9%HS5s^|vNlOR(x{3#Ab`aYRu!C<_thj2_H`gidMc+W0!!LZk|oPof1`jOy0 zyPmm~`41BvA3K>=>751Xr|_(*1$_U;l#db!SJ*czxlGwLG3)CxOe`PD_g1LlQCG*m zGMWMYmk~Z}9;eEBb0mLD*VAsFPPn1G`$)D=XcLySO)^CvK38TwQGVzW(DgkmZ@wLj zF?{rnogz);BK)uFz}P?P=Q}bwbP`GTlbxvPF;*^mqf%Uy?^N)TPG^~av^cS-2|~m^pSuil-i`x$;m@B!*C~N ziA`bLB2kQs62q?(p^HDfXB@QNTJ4=D zZ>Nx-am(A!6OxE(3;IG~z39E6w+$6WM|6%a4u@$v%=A^i_vVtU$t<;C*&DuSND)goNt=v|#Su^UI9HzsI&=na(=ZQ-*IZ0LJYhPL_tumft319Klha=@ zu@H;LuxBZYmT%?SBA~~UadRyiC^7%3ydsp8-1m!U&mT1XwLuk1*D^<4)k-iOM)%ft z_DF03W4kT}Za5GNQ*W5cwTEY!{)DGu+=>+@7o1(Kmez}q+&BaZam9{qE;LxbdcMc~ zCT)O|pAK8yb&#-dCsTISp1V|fiX`cLr?-Hg&Lo(9C* zlVru>{4P6GQF%B&U3tkqk5+>)Xh)L#x;jJ)dww|5AGw{y*|5=2K63~9RySSaN6U1$DHQp|GA zfA{xPe{J?Pe*+v9zO&x-CG+{h55hW$(Tm{fA0utc!{U`kZ|#cTN_<92ST6xTNXvcN zf9s28zYcyn-pc=Eo4;>3{teBQ$L{ApABQo!(rmHEu5#l~Y!>b@x8e!}XY!5L9&m>+ zuF0>A=$&pN9QJoNJoB=xWD?ogiB1o6r8Y>~2B4}UKin3~2$-Dp$opy@Y~lk6OKCpd zB*3qq6#XI;Y5ykUC-Z{pcpE3QtkevRwz2zb9j=)x#@lF&K?i{^g#E+~xM#SDH^^v? z&#-OzEy13ty6NFrT2abK9Kj7tWMJJ@VNmjv?qt9$d4@%j{CAhfd%6}zGsja8C-Eml zr!=S7l>#9Q^LM(pA09!Ess8m$QruIC#=zXd|D_YVmzOu^A8g6!$`si@%eUClv3||9 zxZA3O7cYx+I7ceXOkV0l?8v5+!x8vb;%g26vUURH1Anz%KIQt7yB|q-HqywX@#|h- z)^~vb$toXXqG7GHd~6qsvly{MhIWoSK9u1}dJZ|33A$s0lafNywpoBN}z$N}YkoWN@`5W|7l&Ge`|P zFAi~&VV^;SU(w}PXAJHe; zr)_7bchT}642R2c9!y*#HvL-I)l6)g&vMP|mZKZZKl{^ZkDtl-FPt9NdWFQt2r$Pf-z5Vp{91qXETd;iTS@U%M{q8$#38HxB>^joq$-ah9nnyv3yPG`drMFO3kZr7DFRB7gd)9@ zAYG(Lix4S-)X;lNa)#%;-*?{c$9L9x*E&DWI_qRDNM>fwzW2WO%sqFxu6b#or*Vw= z3^M@WnC6{-3;{R-{e;glF+g9_!x2W%*HO1Srk((>upj=xKvD`fv=ioOsBse%bn(tX z8%OMK=-vRJD4ONKh8}7XY`G(FXL6{~Q@E3ziPYyIVb+Hn9l&g*Sd^j^U0;kJ;Q5tZLUyss9GF%$Z;(#Myi zFFYaPueG*{@+Pt{g74=5NK7xq;(g|N%jC!L!-Jaw6P;AYuK|dIUQMYUSu`qXT7*|X zV8{|kS-=72s7d^4&H$yl{*>|N=rNeW1fFTC_Qdu(OR@1rCuQ#%E(47F zNdO|77RIfGM@#!0X+D74fR*meaQ5zX23Q0P^Ann%oy*l#T4iyeK${Mh`Na)_@Kflt z!~K`y_KeA7s}DPLRqRJVkv%r---KU}jQ9gVMQAcQV$o}) z-v^a0Bm^DdMe}iM$#9T%1QbI`X~TG+tsUb@@SA?B3@{BCMvMsn2D<@)#Onh9>_Wm= z?aGKW9vqd>pO1x&{^SAHCf*?3*LZ7NH7`0EFSpC66|>~+O$t$cy!bJjHi6nQq!Q&1e7w%h$x9VRl+fOxlR%2q41j>Izn5Lt5MSRf zh0T3TeEJipxk63(k{6e0MJg)Sb{YMCe69zKj3gMw5*pw2Qs*sWWW;KpDHzv8Ud7$J zsK%;A2U5*gfY-$LhK5#mw|(eP7#3SyzKW+BS8k`#b}c}<45SM-F=5)-1$xkx$!ufB z1bW1wMWmO3Bq-km7?+~}NW`?O5_ZMZF|?g%e^wav45&1K&SQCZhB~y4a0}`92%wfM z#!wfit+^c1=H6Zf4DM?2`I34;|a zXBd$Ge1Fd>VnrQVZdA<^?7X-c3|K~q17QDti^2%Pjgc~kdr|=yXDwJ*Mn4%-Ra2Qz z_76O@+zG?Tl+c6mi$+Hn z&^t*#egXyzO1%!OIKLSIYi#-Bz`zoK0YJnYiJ}KWx&i>h4hQK^ypbrz|Fnr@+M?0v zm^l2>I&N7y(Eq)^DKeN4yI0C=xMr-t!xcK)&he+8-g z6|y3~K~U}N?{yf|x?zFFC~}A7LXY`&6kr|>cvF&r+KH4SKo9`tY)CpU(n4{IQ{*Eg zsdr&(vH#cujVDlV|Gm95Wj)OdhWYekYpWct)&TYQb}+eRURBLO!op*j@fEa$ulfrL zUFpHA;_Ph}19@n|ndJ`UaB#~3r$rA8WT73|)z_Ggi0G24x9M&{PuyFE(7yC)YA_H< zgoGMe5b;L=$Hj2G>k-UTXpHBso)w{^7tzUr1Cf-&28Y^9m;Gjy{sbECeTGnw_4CN_ ze-CvnY@wjJpHyJ@@7vvVfeaWBmbAF&&j6aCt8&WW%rjMt$A4@!n{q*qRRcj7^ndEt zuV2g{3~JNtKqOYcp@-USXt&yx!KBBR0f!c}o#V|W3jqCLHz)qz^O~RFw3t{qv=7@V z+z*A(Lo=Gn#pXlXVTY}N8l-h-O+FW(1pxEK!#x-{6TnQAo>7D54aqi0sV+c6XQfL) zLZSI#4x2nI)m^_0MBX3%_P8IXBBfSz*bOiMhl2$iYKP-Nk>u$9W1P^qAba)yNc+#t zpebOndysMu`-0Y3@bLfP5TH*w_#YzwYlqU6kemBZ5C3_N{~xLUlL7tTL;q}G3%HOKfbQy*GMWO2L1&AAr$1+t)A&EM0HYiQy7+7IW%Jz41U!NWUI+iKME@5S5 z`DN9kjyT6t!&@J0Tjd=-3dnVyOZjnmxy66#T;%;Nm|DQyK?#fDH?apelm#v0`iBkt zR(d^(yZ{8q%L2+vZ!f`oHs& zrVFpgDlwDEzO5eJZPFkycyG?ff=Ik+f#<0fXp(#=;>LleT%`JIxNJ$exCU(v$stLW zEJ|i1s(A1n%`F-9o-CyRJH}|8Vm>o-ig8PW_G=_q7<`Vywvam#d?z*!eBy+te8%C~ zVbeP^zft@u2LY=yssdp;kJy3IwKr&uzvEyC^mzJ;ej_NCc z8KgybG;we#98LA4Vy7JE8mj7x+^>$Z+pn5ozg3?RI3s!$iiN&Tc=HRdIc$nr%=jh{ zcUM;@kX+L7q)TWDIln9dQNdr+Rq?I8%(7HSMSp_09t$E{);{{|uKA?AKpUjytsy=s z=gO-(CQWzzzSPBd4De@3BR7{~BzsE4tV+2VWRaU$SkN}+I7UNc=%A13dU6Q(XtW~u zdA;i-#e?&Sz5kqj?)2yHi{=So=AJ5!g=Y75k75DFIE!dqJvxHCpYCfJQbrHCR$++Q zjRTV>E==X&5$G4rQGGN(cL~YRw)c?4E25*;h<<(i)}Q6?FH~>^BAV z;$Q+t%;?9xGoAK3rl}N3kNU2A+&v$~{_5bR)3NnCbNc9NevGpi*LyiRRsP8xSE$zt zAh@_%I5_BHEIjq;bDXvZU5d#T!qkd(O~+(F7M(l4_ZlCa$GYhYxv#D6Pmc@8E;EdW z84|c!?*!V`blIn&L&`#SRvNG5)u%?}gP7>Ik|L_o{V&g_G6GvYd&{O1SEOoRCQ$A1 z%zj3;m%5+TM_nBc9PTglG=o|DiMr)Mmz3iI%Xhp1(n@xFDY%zN$?5~vRrC0I>cIQ$ zsrln9jUU|R&`CSXx~OzfCJb8pGPRXV7LZWc6TvO@%v_x#|26{LKO6^WK8ascxDpgu zfY$XlPX1ojNDJd*fb!uXkv8=_9qW@FB57LZGuEV^6F1&?IBf#v;(in7Oo6I{Ky!7@MGRED(vW;$vX7HQvGR6m$)_+ia|x^-sjj=RUV8t z>{}q8T0Q|^wVu?AFWw5sL~&gy+jVoFY$3vCb!v$VO-I*_ZO#apE}@Oe=?$LB!|MA+ zYqmq$_gf2jc^ATB39cJSv5$KKVY5H_wW(Iw9ngSTK)i)@bo4#prMteiwM=_(HD4hP z*;7^XG$kx7@k^;xOgR%jgI)zKvuMJIyQFk{eY_$04E&~?*02`!oJAP$t^`6f{a04~ zbwlkuT$R=iWq$BjTHz#?W+p_VZP}&>X4RFgmMJq!!5j+FLHhVdYcyAf<0NO7dQMe} zbfhV=bG}n|QfU^`PQz!dy6|p?s6s~TW}r}N4cxjMkIpZ^y2A14CxOJ^g)_9W#m(L= zfX9l)G2hFRL=TUoP8_|+^_Y38?hO1}v-B?d&3*6w@?|lX(b|}Mwp~crSM%LMZ0OsNqIaLHf@u6>BdEK z?pPYv9i#?wbMY1uUh6|oaJ;y*pT0oLx(_GpMg?u6N98a6jnjRMm&!_PJ-Qh9I$EVV zhBL`9l5!o2pl9sY(6@6>rejC1FZ->PT>^T^ZuGG<8B-(p!ZO+I*v!{!y#+KBX;gQg zqD!?J@xZ(Bnn#={JZ=~n8`J%l7MHq3|3sa8im1m4Qo219W7=s$_Xor}8~V(ZQW`HC z11Mm4d}u0~`8T(#CH$5d-v9MY3#lH-D@60$7Sg4X7!jjCMfZa4!($XURk$t`!1-Gc zpUtH1(t|?#fjiWF!e!0L4)pKJRU<#$LUWjbRsBI&JI(3NOI-_eEc`vFediM#9C=yH zDkr(5owgel92TP~vx`?#U0hmfLRdw-J?zQF^lnH{-rEDGFWC9{xNeoeMIme<2N2t~ zzaM9~MSTA>*3q_+htKjxfLhbbF(2-??|&396r0C1Z<1g8c=q`KkB4hxp;y9zdr$>m z8j=K)FYy}*dUSV}cX+FDsD10DvClNHE+bk?x=D0IG&$9q7y(q%iycIHTp^>Xa! zy{$36C|EnD9gQ$-?oPZ3kihNQsW{0V&3go6mv_P0z&ftinO>shP+-#Ad}=bshTUm+ zUVdGI%CVl3sk`Wc^)B%W%JZ0lb${-c)eZ&p<6&Z#tn!Gt66iTmFqBp|QjW&;VoCQ9 zFE~t^;wt;ymJw&~>5?q1AdTMwAvQkyTY2@qk|Tk)f6B|~F^_s2F~Lu)JLH_o?#OtAF~&*!nUoK>p1QLE-F8^Q_)t9M@0tb0s>x{wpZ3uhvhU_GNo5XKYOx^#9`Bcw zk>1nibBI=3EWJY{RGS8rX=%oHr&eQ^RBF+*B~zE;xj#idK|y6G(&EmyIv#2N7h)Cg zg+;yq1A@%&ngk=m7I(+n%driS>U4$IPx0^g*ua$i{Ef|T z%ouYcV?ANY6=YpZ`IQ#R{SqT)zQi#$~K#8 z`rNVO_l8q<9{b6#{&FbCl1VVXOepPq^!!)ezaI_|5+faxGLsa*_a$kXJ(i}Fmjr)| z?v7$H*)_hT7Cp$2Le8Y0k|k;nBw36?x9*5?B)TPepxw`opW&-BlOK7ot0TmM@^sAl zi*lS*HfI@rJ26G#8D?apqE}^9itw!Y=am)L+}$y(vBqFhslcyHK^~laO9D({@K3ay zD?dZ=*V>u#Md7oa9quEeqw)d2)LIRZGis!bUVa+c;YL;W>H}tYQYjZ6g=eFC!O|g_ z+Ndf6xSwpIPvA|KCHp>v8uc66kL~#02|Sq2u#Xcma(l?zNt<4^DduX^DygdCy6t7h zJ*1sQ8`8g~9oypQ{=rHX_a?eZ;_rv&gN1c}EPJ*}? zv3plwKP019{TiQz8#hC_l2w4`dValt1nv?hAB35)*sL&Ik?cz;>bOMFnh7T{6K>tz zODmj=B`cBh!6MV8p>!t)7a77ljkx%X(R2f4Dj`vrzgb^-P+WB&3CG;e{<%Jn_UZKQ z6XYp0_ei{)iKprm6fmQZltrWb04R8 zSAIm+?~5|UeHw0D|JQ^>16bQ@oG8DpmMP% z)-Qjg-Fpo^UvxY@js6vPy^^bqKP7kBHR7YJlhd-e%H9+H+vy{}gL0~UB~vY`IIj=( zdnW-t1`(Zk>ey&a1>f4$G%ZcHAVKFa?S#kf?+*Sx%P8RbGm|!RPkvY`pKWk^DeH6P zedOvTp}rkgg_RmF2QnAY-y3Z7#8-cvl++Ms~;Szkh%Z-;U<<(?tVQb|cK4qaWAIe%@w;-cm`1a!d+#>Go1ypw> zq!v#q|8ZGg`C~N^J;FDMS4P7A{#Ja99P;N*7gXNzdLVZByN0amgWr|z!lUSWaSvzK zmyn#j&;_OOd2!=rvqlpl!xl*jN$9d=JDH-#E=tX2~q}3jA;6W>y=M8qP=tr3f5VUR`o6;iJQiU zzRBt~%~7nC-F}269kuE6Ngk={`C&JaONUfeX8E+ielsDE18n>Zax&dhU2D0%O7>Fr z%5x;)^ytzrxZm_Io0-(Gy0sNZ6G+=>P2RJdb!%3v({p$#6)$K#Ao@H0QrCo(okTu% z(>pbSV=rt)y}IK0k4M1=_Q+04FJ3MkU8^Y?ErOc{9kF1fwt9@zIe*cSq z$!%3L8uR%p(IX>w3ymyGh0hop4J=P0P)G+@q-bK;o5IGF<)K1sG{2`ibrpK_b!w1) zCFB5E?>fF}4#tR9?_dt0b4b|1*7*3BN$krmN&daNsXpb=MAM+2;F#-q5&Z@zxY)daCSxp0~f{#bG%4U3Ly3+-fP=TDAL;rk*WSw1>t< z+rBoAE02a_rcPnH7{{Q9*O5InkEbi9CN|vh?15CakhX>9KGfRA_{~gcw+iiG-eWDz zw)?rP!mlx!N-Mw6(vXt=NO{CrnYlAq2`jz0a7F0nXzbDr1(g~jW!1IM2*R3js@}e4 z)aA*)_`Y_wzQMn>BQsYsSBgT`s-1Rp}Jg>kVI7&fJAHf=&B@es_^iPh;AB%p=)u>#f_oIzKW`iiEol z>!4Uu<7afy&U0dVq`Hnc6`Q@OG@0^%=$``F}zWMkd z^m22Y*G*kyFnivP=ZPetZt{LgfJx@BfFK>CLRvS?EBTJzMwr(8@uee`g1LPjFVCta z&!vmx)L-o2Y?z$9do11PEW!Bfd1xBq6^6YWDN#0S=u54~({~37+0;LltJ@T$L^^E2 z&0`+Jg}>LSRdj#WPAJz?%}*{XbLpUw1B&SiB9fDHi*IO(a%DFiPoMW==_Q+7qwmN( zKVz{;CopOrtnh(1g4$2wfeaoR`jkWArjT2jt{(@SN$`@|z*WB3I1VKHW@g2^6eP^PFR} zCZEcABcLSDeA+}eUg*S*(TC^39pdH|Y?)TZuqa^LRwOr%y;{uqtggMgYPfuHVT^lb z$~7g#QjSl>ODmCAKrRX@1gGR>JBJ;8z5RlhQC;IK-94UvvO<7D9@}trv(_ojSGnJF z97*lYyAMW7BrLw$URM8z1hECcXo&%qNlI!y~O^gjn zIwOT>Wkk>0YKPl3!W&{qtj2ZN>|F^~XY8{tJGJef#7aBIuRgZ^P|GNEZ)!cUx>+%98emV3cf7SbP&j*=C`28wOHA~ah|<{a;7Iv{%DGaB!*WX z$ah)y$Pe7O@5phj#@k7lHLq*-g+}<>M$6u*dD=hPS!C9yQIov-@E4ce?geLeyNrEo zMvv9Ns(Z?(&A7g>rC7b$|B)CjZ(;(uDer{gO)srOvSLx z3-UoR3R$gBo$1GG%d;!(zoL@zV?H!&)2F>R&>Jmg zww~SgB6ZQ-I-@iTm6sib&n&2%)%B9Q#$d5n;IVM)YQt@eXsyqNfZb9NiX#A+KWF0+ z{npE|&1y7Y)b?YQ5cosOgi{H5=&CckpK8Kb+Vuc8MUGGWJ0Ig8ohMXNbrTaeX}f8f z0SdDiihKudEu9wIEvD>X3ZrwWfXp%~4Fk1Bt}r@^z;Wjf!}BAEmhP>gVB=_wH@aBU%FTyS_O zHGVgyNm<05*jd($?87vr^vbP>VsSLOv^+;wFg3BqWabKv!hnJzWDw}ojpk5R1Y zByQJoQGK}$IKOSs_2sbg2e2L_$-MCAHn>M5Wk@X&S{sol#hqM)exmg)fsG7t8y+e)xTNf8uW7XaY6yN6}==%q}%jx8C~h(f>M;QP+U`ZNaviDqML%>OQJuo zFejHwy}?nEZ3YW_p_o-__RkTSPY%>N`t-ifF$|;o0}PNO_k2hwjq}fUb#@L5yf;Kav|qJM9TW$Bo^pZVr;(~?R^zjrE?okq~Zf$YS=%r)9m?OhB@fBXpN15vTy zlY1bt-^0fLDP?LNyC>!KB^cX1APrtz9!lkZdy;lnKqvo?W=VNDJCJy4n*qU1ur5iO zs0O_O8x+#}!H@xR;OFF6p)swBz4|_~WN&RMRq8~CbUk*@7F05AK^Tf!^OR@0&Nfx+ z*tO~0kLV6C{%87eg&Fvi=p1xh=#vYn4!GT5b~U_skgktKfflh%-D%A2UJ9Q1d^xk!*3H)Wa^-{OXyv`kg6NyX z-q+r{Yfr55c5W_0cUx`?yVUbfjNiIf+=7Gonftxz9{4>lU+_E6nv0^H(%M+>UYB~W zL+=>FVb>B;bJ>kXSo!EN7BPfdazC^uk_f-j?TB29Iv-b4;6l<9$~5eg`psz?Q5?+KZYf? z9B%&7vrG!^uvGKl{!;U*XNxHm>gN8FPVNJGS)s;3B0;f%_Xz!|SF(j;6&&FnKaB=d zZ(4?@a^D}gD+0xB>iXe%tnoq;YOD%N@}s|N*_mZF>ckth1g2inV5N1#gE37hE z(&lj!SXda$$AwB;RF`rx6gy)cn4Bj}G9uo5!%SE!;+*O_^Y)HqpVM>P_BaNMkEKm3XRJJjUEfQ{|s{!W;`aL#R$+y&K*+BJbO?o@&1 zwxOnhxeB_}SHzGvIKjDkxwf&(BQz+{=hh$om5=$jz|&_1&kN)_^K!{*{)2i#&l{0x z_mKh1=;O>%emx13MtvF4s^YCym#xTR8w%N$VZcFWQ?sqnk-3;+Tf<-BZA zV$Fp7{)swYMUHfZ4@Zt|&lKk80+IWQE1F<+*CR*t0{9Z>C_zyoBLrNolF4U<f2T*BTpG(Q|Iw}R4W{z&KQj99!Rl2v8>#jdgR8C;jBW_uW2uLh1BCvg;QEW1Yd-w z1zTxQlouc2XziYIYHWAu-R6I?__(bg6C{6mOr%vGxH=z5+2x>dG!D^5W=u>&ll6yNt8gDpFijbD$U)Vx>)JBVKz zOJfxIVSofV@vl3)DM+j|ewKcE>J_b_p`wFogWPPDN-xQy3QI0TuJuLnST&)01Y{~h zmzrcB2~#cNr;0ioBhW`%TQ%(Y{*hqZa1i*lw?}e7qvL*l*O#FTsQ6{ef~EDQwkLL>ybpUrp%IB2bI2hhz!F=Ghp$s8!DU|C^-m@Nwd?<)B z@63l;f4ah#rG*6Q#$);A0Xg*3vV}_6;VsF|)zy(hFt69(nC;cV-v#&jrmOYXspd<6 z>>p6{Ed%V$S^fm1WQOoauguw~uXHbH5Xv1GU*tABtM0It z*7q}_r(#Di(cbU&x(~NJFo6%{%9$*pyb2F|>eb{k-^mEcZXDj%!MHo2lqmaab_%O&=TdeS+it)KfPks@4C|3v{9EwL8R_+UpLz zdw9FsVu;r`m%X!_4gN=EP^4_XEtUyGSt!TM4g`2Jatx zdP5?R4%VXW9f^GhP#*ticiHJ zEgkrLvT>Rc6WUKFr1SM}lJR|H?$H$=ZWAoLZECoZsp|fPh3wk_?4jlXcA8? z;jMY*QUp!=38tqUfz(UrKW%asn+C z-6AZd4T;O)I+3g%TL_HOdQ+@ETcIngF2Rh8(oRJX(F0^L) zjkyux4)^eU<FdW+J`&eF|U za?@eF@_#fkWyNZHV4WkkQxEonN*5oxuV*s@raqH#?gzEqBmN`w@y5*H-DGS&SSMWW zScDx_F!8qN)JArOs|X*suhzSBxA@x4uKoEqr0l9ln1moLM?aW^2Rk{59^ocp(mf zq#MQsTqaQecj6IW$_F@BB8~QCw%{u)-r2fudw0K9In&8s zeRq7^ZcQMK{p|VgH2*jGyn1wgzaiIvxAj`?KDw>6TjL&41GW=3@auM3m5dh4>Wqc8 ze?Ab@#0^Eo!n!83W6^%~P<~{0>1C~#=t{tOsT>=dr^7WUZK3oSF&Rot-ZZWq9MB9?)_*+STR~Y@n5dPTp4s?1TTxD z67~W%!x=uFzxuSjtm;YJ-Gr6ataqfH<9;%BvI=)PpCs*Xk9{xe){NI}Xj3=8(v?3f zY`Pd7EL^fb8B|s`BV{KtiBj?93Jfgo`o_|v(5H7Cu0*`#r_9PVZVpqEE)LEOh9ul( zuslbczLRUGRIxUGFDLd{Pr16PSf3^vk+zznT9-fua`oYg3fJh*v**s)hYzS|R(C-) zD_s)IgWEp1igi~yZ$;CnleYGIx=(hDemv_r==OOVOZ)p9ruOb8Ai=1pA&%C+0$pOG zFN#qyAIxM3ty%32`f@Ir?TdmrCHGPL7MJdI9!_uR4d{WRfBu*h@Wn2WQ|}O|`bVHHn|gm(sM~;-NP(5RYRy2Hs88_jh(y zNdHP0&WY_J@?DTMc;#2JuiZxM@(6RdV0b)6h$+`jJ`9FpnseHpNA#3aNOPI^Dm1M# z&1Lk!O^ls(Ve#l2+SVelmkM`GYt3@Hf;^JxB%?PxW)@X5{bxWte*1-d&l^0E$Ci2L z-EU(WF0!KX#*&e4v^GwPRvf%Co+@!!jHu4o>u7oK)$PF(t3r=e!UBmm}Neuh*6Ao$7tbK<42wPA;$8t0gN@O`wdd zfbqb{?~2vhA0C_d@7r3@g^hQCrCGH`r^{Ffj!MMWw<#>tfVmd~GG+y*q?C&C zhS^NLgi&k5XO@hjJDy()W|8E))sPDAX{4f<=X`LhdDO15zaWqD}^Sr7i%XjsS7_vBI z0wqmSln~zWi{vwXZ;hoM=VFEQe!scdlE*fU8pv~(Bq%gVtxaCJN)^+sW{gh%Ztv2; z?IqmVPK9udW&F{%>zH!bX{G5(CWtD%BvAju@nmYX<8ppdMJ=DSd4BFg?EQgFE;QbK zQUuhQxX^teFUH_1uba5KWQWdds`+@hGPU?c7OmQgLXb_w%_XGDcUlG)PJD=jhjLgh z!#=`&w?iF1rIp8^->P%^7jgo9Me#qgE5GNB>`T3%fE_F_tOW5@Ux>N<%qQ-av znrRM0;<0&p$9}z`TS(o?luq}YYkr9yysM-BMrX&JnXuxadQ9*o8tkrIi|Uy%?DO;ffLS?h%EojLU$h#$MD#b;*;^8jb8(KiDNm0v7stVqriZb`8;Fd+ z*Ybr2PPP0?vJlE}`uc5NI~Kxgr4X0R&n3>uk87~lMR`qq$>09lquE>&t{Dz*u+{Z} zK{P5KonM;{Dde=VM}>MzFDNbM6y(V@C;d%hF-DfiC+-V{Y#CYmt$yUOiq4OzszxE` zq$&r#x$OBAxBV_RwpB}*jx(0D%cz1Pc63^e`+>Wh_yyHN$O!jz+z>s8HVcMDf3wwX z&{GfHndqFTz|BR(^980|6xyDvc*#!j-#d`6S2GN18(W1ul>1{f2?b-d$%|R}P`Hma z`(2sQ@2QqcMz_^Ie7>qh*L;}u-_KE{X#SE*FKrgq%Ud6N_qHWjm6o@!YWMgmqs|z= zMcN_dt2!h|>GkC;GgO~*SVHGzoFjkQ(btn@Wi@}&XerU&CrooL_`MPb{*@l&LCY`X zgKiJwb4p)^a?)IgPz1yx^@jrwt| z<+Q)fLkMfz$z(NGJz@GSN*|BSm>{;Z^XJC)ole^QXRs=WeK3T*UHb0b8&>W8=X7=4 zeVLd!$i+#JBzg44_K$sbBstk!kfqVJs6nK+%igYh+MEPKM`pj?m^W;l4|}AuHDP_l z&dh#p)@03AqfX`RZ+<_t$pdteu@JLuglO~6gs!(fQ%|a_DEBR4w`>=UW_{n=^1J$- z0gDvI+c%(27h6Y08r?jcJDxDXs;{(Y!q~GXa@Zkq+dV=BS4;D}s?-@c5Dc`Pr+3q0 zax2oIw!bAGJQOcn$DXlx6gGgi)`7?zTPq~w8U(oJXn8;%lX-kRib8SK zVe66z7s_dcSV!EUUK~##^mTB+GDFv2*}p84`71?xh{!hpa7Xzy2B=2fMh_!yr}guAqQ3>&mE67Fxdow+V6(ppJ)x6D@)$Yr7gD zp4Yuua?zmdGCk%5#J-8t!opt&)U(=hF%od%Fj9tF#{eNJO5jo~TIg}r=8T;-K~LoO zLZ5zu@=2lI{`D0+1X<((x9`vcsq3B7)0V-@qXLsTpIxxN)mA$Y@9&WHn3aE0!k8i_ zhlk?32bh-+EN7sYw6<$ZV)J_1cP#Dur_*z}vfw2%u=YAdER>vJ-h0nrtq7f!P5Ie@ zM8~4P7nLl)3=Lc{JFX)ida$&YI}1fwblft~GHR)gRIWmWXt)Y$Kz;!y$Ho?NC#%@z zQ!iee!b6U=#q++tV`_+q>E{u&m|MIcjk%{7f@Psd7C1m~r zpOEZZ^LQ2rq`<@2t+Ch`ktRkxnE5OP>!9Ev1TJX?K^r}Fyl_351R>b2ymeW-hwn;0 zK6(_KhqzU@FUQ$fLRq{E1=tZb_E?4(w>ujn1`9UVJNwn6~DtE&-KNm|DZ zZW4?i(ip6hAZ-4ZERP>!{Jgb_^Pgc>)~Ap3Z*F-!J^sz+AfwPz;CYwJNcEZdy^-tB zOP76(=*MsIBjCM**;jO~C3I4*T~m>->uN0RoFo@<0U_v?j2z{tfGeTA9}+C(hg56x zm|w&$d^Q(4YPA37z-WKUcY@+P16|EEC{oDWb%Bq)XlpHK2(!BWmkA{)M)aB8_qYwp z7Of!ge@*Zfi%9`BKi}dQnd~Y_Q^B=k20|*ikw}ZEv3jkv94SG|nX#+7o;l0)g1NQ$ zH^jZkSn9h1&nz0DAEGsBLJY&vorRJLZABv&1;Ik>VnDF6xq`e&>Dhmg{O+IJIE>k5 zUMI^fhkHKI+mf^xe!%j_F}Z#K;sK`xUY(sl)ZMXBJLrZGYpO4Z{x67seohkocXnrX zAG*4p=n}Ogn>j(81y%BO$t%BKXS6~bU}`C(g?|9hxm{DV8sZIAy}r7HNq5GFs#&<^ zyu7&=Dr@>;+FeZ*0pS;5?PC+$3G88)D8Fon`wGo0Xd%0la^1@@4lDSwz1^>SGz?5R zI5i>)N;HjY7p{BhR-md2>{qvUbOVd)o?Nc%O1l=lO(4UpA3_?m6DzQ?`T>>l>-Khb z_6d`E3_D29_%Wm5gadu#BW7@3+6;N5?%+ClVa8&rdk_L-R_x4v_6LcH*MZI*-mh7l zQn+iNHh!v}wjl;fgvel03b_Jm2P=;N#&|vJJPjiXndaDyzNMvz4$4;u^|=l~-s^yr z4P>T@AA%K2KP>DNP-?6_`X}|Zye>@8{_sHWP^W_$m(j>q*!byF9qs6#$44+P4?8Mn zg$)ium+lIyE~y^k`M$piY0Ywen%CHO6aiK$%j@bZ-6!>c$a9E?cGG-B!AnUHp|2s zO(tM)3Zk??EVr+Zn9{EZ5;n424ym>fbI&>VigjQ4aD#g~0XM4^J#Fx{61K(pzf+a2$ev&AOOhJP8LnjY^|^V_Mr=fN&lnk;YYO$+}*X z8*{|XLyYhD*m6oi6*?O3FeKA5qZu5uSQj~ORn$1GAy(i}$1^+bzS{;?08?9r3iPcSvno-=BtQ(nTD4*I zGsf>nBhx}og0{{C2&T9EYZ>7NJwGrqcg*095NW$AF(=UD(yrB^IR^IU07T`*FE8kM2#+osLQb^7~opU{7q|# z`V0t%JU#Q6QgOkM%5~DSXQ|hoM?Vt;XxTgm-ohOn2?*vJ7cD8nh-aM^^uNMH1@%LRE^&f z?Nni*{ek7WFqO80+MSyxs+b2y5(n>w6OV!q(A?dG{w*>NL8Y>5e={BG59X#j>!dJa zhg7JNdMc{z&OM~!YOp&8gWJ?#ew_hPAR_WGVPrHr6i1;zT=7RNhjTonjY6OQg9wM1 zO<-4;PbL|&v8k}Vw>t$|W*!7tq8MNO46tnY!UU{&A2-mQ!W{Tp!t8#WI-LB!{M!G@ za>M^up5>QY4#h2)h5N`&b;wej1cZD4zKcL97sUl@d*j9y;!so}hakut1>>fyU&fCj znbrYB&VTN?l34@ulYjvw(p|t06<0*SP_B;8_)%1!NmD4-aX=_JZO~x*2KW6c?DX}+ zdK*S#;&01Un4}=U*gpx`#5IDHYMdOyUk9i}`^m73A)1%Cc^710&3uz z>nQ%EKXmtBVD@;}5$F{H5PJY&P~Z=_g0Qxamw=Y>)ukNES@wen-7>|QS#+i2smX{}W%c>l~pLo#tPo~qZKuwKksgIIiEc9lI=AvI_ z=&=zfIeVEM5H4#y!>0N8j^)WR+QEC(F6Dn}+TMtGU8fu@Vg0bTY(zF{{%W>hs+T*Z z6{_%s4Xo@9(ycrQQ4&>DJb*GbdOY7ER@YUJaKDsYEyOx5FUz^`(p&eQl5@?@2d-2f zH0=R&+g~As$Pw;j&LDa1M;!Q-s)?m}}<(R+r0?C#(LCK9U897PVmxsXgTI;9Sq zg(xofUbGj)(t3M?$0lnyk5}kbOUJP|(i%?dL>G zls}M&gx<>0hUHz90%|Qru<_IZIl%mh@$mlmKhLuK528IZz5kiTg!~KkE6@`s3jwnR ztNf*$iSL>Zebs+`(#Sz5Y^oqv*Fyy=h{TJck40On9paQjHLE<_zCKZ}G`reqm7HBw zva_BR4JW*SJgB>-1#7O(K?V^kHYU0|H~>-;o9%!2d?9nlqgNoBo)QB*Vzm`-3E>Us z2yL@a{MG^)c|;$xS*`mbJ&po4#KK`_bEi2}Llh5DXCN|?848^p2Jpkb4=$B1mH|Mg zeM1dNw5o&Zpe|-eB9ac+X23z)F!UB298`F&m)>O*l!1JHGb6etdnDk7!9Yy@)L&J$ zyk0GBfMFHK!Jb3!+<_c^G4T$nE)~Wds&SUk?p_N1TYFy~57qm(e=Ln9+b1b1TMLD< z?_?{It&bW)whXc_*_URNNXV4JFi1#dEZN4sgzSb$6C-5Fo-I2w&pmyg=lSFL`}gPX zpI$l4J?B2>KKJ|mzTVe$9pEGbS>PmXDb3S@sFJ%1%%k6{_#qi!3{ffzKrj~2hJ>K0 zQd#misSMVU1mFSVd)IfBpeANG+@EX)37La3U)*KWc3uN=s0fR)uK8mRb)doDJs(0s zZAE}fMz0wKYJG+OF$dwT$EEgia1avieacY+T4KT0HzNa-o%%( zcYw9up+hEu)HBBUG@6g@|E&5Z3;57kG*j^rs7?m_4*M@pvl{won;RWe3^o|Z*>9Av zp(%h{@C5r`3^+i10NPOc&LXtkAGmiw&fp56WFL^XYBkMpT#Pb6a1F96ZX^riJrelXvF7^n;Uprei#{;zR(V?`bOi`XAuaW!;sViBtIAk%)7}zPB`pXhiRBc z12fcw0fr9~p9?6??wccFe8Bhx+^43E-a^W3@$l$qtl%#?C`cHbxSC!WVJl;ZdXQ@a z%~!BOO-F%)$%Mto*4|`x0#;Vz*nEX6uqXmS0{~2@yqr>Mrsrt?Jx>IfNc5*9c8C_E zfb+`P!n5r$(>8XmA3ZO&9wfM^Hh~1#1b9?v-l}RA|YWzPk%m5=@M1g<)|A7C;bmZd+ z2u4fTpLQoxHQ>xttB#Ie)ID2r8g-v)^JBk}x>8FAlcU*pZ4Kmd`qt`dlZli5>(pA{ zB~{HV_0})`V1as|(fFLyVb*cOiWJ*R-QgrLqvOROw)5=Y7r=IK3yx&tbJIzwG}`C{ zI)WBR{U+%~MFYhr?5>UApk7LQ zS@=zjB*V=7lb=E9iY06NcEx6 z&tn<75t_4LcFs;TnY+qWiqm~Or7s7YNRqCGQ)l&uAdgUMDx1ycuME@e!w9 zIo`wxgr3mi>%Tw;dD{=*oB*8K%czo)*{8ozwml_AOotLI{CBB@uQ9!swD|f!l%P6T z!=ocw0PL~2M-QDe0%h#80dx7xTo-aWV7T)OVGCTpwYQH)<*#gOoD+sjKFTu5P_H}EJjL)6dsRr{HOY4S% ze&j$6{i(NyVMSVzr( zFtr~W-IA!AU^oO(M|c(eOvR8Z#K$GJ7S=22QFBb^AWN|6x*X%BX}?H+iDYY=1Urq7NHgcRj<2y z8280y86hiM#jJ37)x9og7VN|}m{+}|GVDb8vZ9&t(T~-;C7o2ff!;Hl-6dJ^W^PfT z&~TrL21WAt$p(gEa1rZX&9D-{N0Pdf7L}CQ49!?O76^w@t09vlj`|DV3JP}VF%yQwQt9`orRJPRP+8)m zbB3*ALW2x5?{ZvApFCOq8v~m#bKk<&lDM(7#$a|O?$`ID)o$cQ^|i|nebDPAH!&_Z za;fyCvE~(^*q?ZFhi>9R5%!E~`N{Q?XHPO33Bsc>iY2=~@E_LTInLZXdVBowW_&z; zjpSM3kw(NmQU!s*gM6Ll6%borf8DNpOi5`+xU94HY(pjGn8sM))aII$ji_X;ZYrT3 zpy&YyfjMsH)8HdP%L0cEf!pa#4=yJhbQdmC^_c_maErC=Zf=1s8dPQESa=Yv6F^Ex za5frrH(xvDVv=!g7!Zg|KnDN&2uc=&7s=N!PamG&(dY=JdFmC8hAn}yM&z~{_b<5= zJdzSD`gOBb_z7M9Cvw$iH$)!MLACvC1E!&9agR727?{wkd;kKPDvVxLTFQXbF=4OC z-K-k>kdnr0eD6@BCuhwz#50U<+$6g1Xe;sTQzy66r%9754=;1%03kE?d4^(5uo7;; z7%CI#$cOv3e$rLD9@x6iJZ*P=e&m52XL{w8;jG!^Epzv9W$BbY-2|v5bZpP9^<`4Q z#Cx0Orn$~$MzteXgsxdq_D2G|R#Dmil$gsDpFz$(H$E>;SJPn0i2#MGd*Kxlf?siW;K$plx&>b zgHTc9XBtcRoHbgn$O$NFcjlArc~xta6HMg$Bb5ptw0)vjt(_WPTs)MQ;4d z3-a6-38yhMsALPpJ$d{^{pNNC;YCi&HsEF~wYqK8@B9en)l|3BWQvOh&o>5k?~tm( zF>>{cujCYJ2*N?Sh>t^5LcM_KDceXob~yW6R=ayJ%fl6qA!=|`0qhn?`9Oj_fr+|% zn81*1GRYX^Jg2MM?TX?uC9MU|k&Ioh@bEsRGozmD?n;RerFNE{nn2{1k=!b9i=6b? zAJrO4W?8(+guuc2{S^kqIJH99GXN`wm?`!~(@h)!dx#{Pw0GkLGoCF7rS~%@F=L%ROeTYhp=@`k`8c%GE@D5&#)p?8bCQ+ zzL);E^iB!)-to*Ij}a5sX+yX2M5gy=4^30005W!-+}msOSNkg!)IDy=91cTl3`qxV zpfQ7QCH|bI|2~VYZqS+5OQfM79VLJGj z`cw#+^xCvV~nwgLJ+#3K4A#T;4)ti(B>SF zc2Zv6MPlSw3=&KF8N{){Zg^bsP{ncdO!+L;(F~70H1sa;b`pno79^Plso99%)E z5OnwX6SOhYfY0V1DZ`$HyG@=64@*N5pAJpGA0z=rR_C1#?DHpa#yau1=&lPyr$Zzq zFM-0VXMg{`5+%7mixBN#5a9!AFawYF#56z|&iK0U%R$BZCCARt8ae0w@)$gDEU6?d z-`v%#vH2M%+fvvx0m@xt2Gk`pa6%+%V3rQ8ImQGd1&m({cqQyRA~v|DfVyneiE0PH z!pAYWbPkS%7>?0bMU`|H$9t{;xe^C|v{XKYHvzx6chT=dp;%%XTPtZ-F-vtLe-j;i zk3@!=a}9uv5iRR$s8OEjG#Z#;&0FtPYb6DV`#BFJrV&FO{eECy%FdV43)vu%?A&JN z@H`ULa(jrc=^4lf$`AT~uv(I@yjQE%$Kh01e+;v-6?G5w=9_rF!nJToJvtm?W)qX| zWaIZdx40vqS_P&E{?muCea+n+l=`S@prGD#r`jbxJ)|KZ z^pF%V-0aVqwn)X_vt+I5kpW!wyI!j+lKz}^L3k*PSh)ZZhz5vaQr!iLP++I9^?^MPc8%j!e zM2ZFH#Bj-e3t@Gj^Z-X+ei-? z|H{OwV7t5gnfA8L(g|>8Bs2fK%rsLQn3!J(#k-vUeGyQRCO@JC3-Fqe15UiE5WdIG zO(HA@%so5o<@>N`tkd;Zz7{WM9fdHOIz=ZyyWYiP?nd>Uz;utCCGJuCu`xI(-7_(N zk!VIK{o;7xR^`krR*OVNydVb*&}ok7n)&lVey7(N&H)@qd&obrIzC>XbF4<X zkVIG)$(Z~2%FdwO<*VnH1a1w}qoc1BzpQzrYB)^eI;%{BuN8%$W}m@Pc5BzvWvPE% zdiuyl~ulj9!f13KkEy(6^-&%vW5NCt6gwi@~%2?*17U9*&?(_MeM_&wbnwJ>ds~#X1y!}8pr=%YZK4r!~LR{XZ-5A)P~UR zH2W9e>?3neunBlnH<^-wFYPfJiJ6-Q$UPTtS&8Df^rNQ**J%+LUp~p|9H!wB9fMh3 zv0V)El_i~*fSvtPk(}UDj?Y>|Ci@VAPVxW?bW^eH z%Vt2_&OY*th*ROz;#FhM8gbRxA-}pe5>FAEGu^88+DYmNl_K=~-Y`d80@!JBwO8a# z#Icp0%aJVeA4N1Wp*}hac;|V!<#1~gr>C;Uo&hxiRiDSH#zLq4Ig9nwv&2v}9^2aOa!`)} zr||I5C671rbqjhte5-6=7A_SbgZ?oa-`tUngNHA|=d%G9O++$067hw290w&)ql(Pk zK)egwzKD*$s$Ckiawull^kzq5N#&X+fa@UI9=&?wjKskOKZ!!`7xga({|ed!s-Auw z?U@!42P1VJ7g1UpUp1Ob=`?y2l7Bu;6J0kdFsNgTwfxUXfJi$zdT1vs?>jPj#sjnA z>K`EKw{j7k4!+P)5_Nd^MIe|kA~3j%Sr$>;ofkiE1?xOl^je?n5W@Vj#m8Fk2r3Wk zw0cQ;=@lN^da#P+Xonx|d`uqWQ37|=4dsIUdjZ*vH-t-&XdCbIIFUkJTY04VXJThI z^9U5Yr{Z@R*6C3ATA$8hE(>Fi1H5!w)>TF@uN;CpoB9V`rBCzue>5=EpsA=%qo}<( zzXrJh7NVHYVpyyRSq=#pNB%z&B1qnK)tKMb>5U5b+p@ZFqoj=BaQ;gxQU363&*XST9?EbR*=Bm< z1VgILo^0#(r?I<)?i^=p){eaz-PcrxKJWx*pl`zP86s~1hEk}gP?|YG<6^iTbZ>2K zyYF4hnyDxx(ETaf3oEr2*@qMF@N=7=*>~+xpsoQBNR#jl*#7k=&IDkrDpgjLfikJB zO)=bvlFTF1MU9vR#aZK5f^8IKskKvwkV|w~w-prJE?q%h-=_BD@(}wSZj{}d=}}^= z@?8%kA3aoNMJ%s~Zhw5FUHbJd0oTy0%y^@(>dqJLkYkMkAdjadvyuf{ncLNk3f)|G zG?^8?6X``h4ut3D(e(`m{SVpcU?~TAtbC*nUO6MAmhXFEO|gCC)P#__PJlM@51P*A z2FUNMGM`j(t9mW1ATtddn#yMzZERQk*9BiIBtKz1wL)YYor0iu|d?RUv1qmc3>Qr5>i2hpJACJ_?ZS_w53sjoZiajdENX`p)bNJUY zwttVHB|z}yN94-Fh1ldu^b05nJzQGrrsPK{zh}ZTxzO6tey4QVgd126%@QtK95NB# zu>#X^khqzR*KX_Z+6g6=`ig=A8dAqdt^p}EC>+5RmwxOl1FgSc<2olZI~$|b^aBlh z=&?@wejoA^-UjUZe7}7i`@yV*?;VY zI8pr4B8`3(ID9MZ?gOKM0{LDmsz1$1)*A-ah#`$H<~4b(mK1gK!pBl&_G);S;j=fr zF&*+ZG9a=8%6#%T_^Xllc)6yz#4ju#eHofUQ?MEiPben?e zZ0QK|3m`Q*Z{(dykpEPz^w6<)@c!e6C5(u{L3hvJpE)<0o<8fFUM%`&0uGX5Snh>1;%q zJ`FEyQU((w5yYm-lQ(9PE&zZ_oh-u2t%FLhY7RaggejgPEwpIE*VWNr1(9~G}y6Rn<(~se`vmmFeJ^q^UdHDiHm@$aX$^jL79$ac8 z=GA*5R0(tC+xcZ&6PSjpKu}`&_JKBP7QD~lN~#E|>FJk0ODrN(wZA@sHApJHYSN=Q zbka?ns^$s&l;N3j$v;@ZYY5W8BXM$phNNfVJfj~2&Q)djYEuR>;}Xo#MR0{F%T7c= zErHAK(x&c>c6$LB>L^&&S4J0}H5F@P7^)=YWv3jU7bO8W}(2X0=CW|?=o-Ji=cKl!}6TAF<|ZbroDq>w^ud%?t35Lh4SS4LGZtC`v6W~-0uZn=L;}|0DM73 zaqTjlt-|)o3g)}F05?b1i7bzFTVgeE-vSufJ1hD-ZBRXc$m{R^p05k^k~GZQ#b}*A zDQ4sgLcu%SVqgp7Mdx4d4r%>;)j|*1%vNakD4k5dD9G*cj-3z50&QBOWP5DEgMoZN zAzdmL9fGXvv5^jbu5jr@937#VnWY^i3~s$vocF6lvj+|3mO8e&LGeBROZM_}k0a9$ zv^L^!>OXm<&I8`zhp&QfpPd(#2uYmh}+Qvu|$0ZH*0+Eph+Ra~w+p9py+a_BP=gs8~G7QWyZ6|j3(_KehF zJ5>=(vlQ-^*Df-P@C}JtYF(iS|8f2Y5NvCTR+!sR_rNR*E}N_Wx8Pt{pP@STC?lx9 z*2w8q&x5yjD2J*da&f)xUS5QNUEdMi#G;q@lV_N*hCAdf&(wZ3ARH)ii!4!Q?-`Ik z#t?!3d7mLdFYB%~JkV$_LK-Z`Y?=QR_qe2Kl0D&~3^n}?c4W=EFTGQJeN)A@5~BP6 z4f!f!?4Ju*pYsTnHx439?jyxv3}rhMWI&J^Jk~F~HO5fo=Y4_Vm%%l%TLdc_h;H!8m~}j{X)$o3f1{v-qTsE#8b4N^{*RW JE7G=m{9n8>3@ZQt