Initial commit

.gitattributes

@@ -0,0 +1,2 @@
+# Normalize EOL for all files that Git considers text files.
+* text=auto eol=lf

.gitignore

@@ -0,0 +1,4 @@
+# Godot 4+ specific ignores
+.godot/
+.build/
+.test/

README.md

@@ -0,0 +1,107 @@
+<h1 align="center">🐶 Godog 🐶</h1>
+<h2 align="center">AI Godot 桌宠</h2>
+
+简体中文 /  [English](README_EN.md)
+
+## 👋 介绍
+
+🚀 🚀 🚀 一款由大语言模型驱动、Godot 制作的 AI 桌宠，旨在提供一个全能的、丰富的桌面AI宠物。你可以将其作为基础，构建你自己独特的桌宠形象和角色行为
+
+- 使用Godot：轻量(内存占用90mb)、开发简单、提供动画、控件UI 等框架
+- 兼容大量的大语言模型
+- 有预制角色功能，多种形象
+- 支持多模态
+- 在此基础上，方便构建自己的 AI 桌宠
+
+![](/img/example.png)
+
+## 📙 使用手册
+
+### ⬇️ 安装
+
+下载地址：查看 `release`
+
+### 🔑 使用
+1. 点击 `编辑预设`，填写名称，API Key，点击 **保存(修改)**
+2. 点击 `存储预设`，选中刚刚添加的预设
+3. 开始对话
+
+> [!Caution]
+> 当前仅支持 OpenAI 的视觉能力(VISION)，建议使用 **gpt-4o** 模型
+> 你可以使用其他模型，但不支持多模态
+
+### ⌨️ 快捷键
+- **鼠标右键** 角色区域弹出对话框
+- 当焦点在角色窗口上，按  **ESC** 关闭软件 
+
+## 🚦开发计划
+
+- [ ] 添加多语言的支持
+- [ ] 添加更多个性化角色的动画，机制，事件
+- [ ] 添加更多模型支持
+- [ ] 添加对话框对 markdown 语法渲染的支持
+- [ ] 添加语音、文件上传等多模态的能力
+- [ ] 添加更多的预设
+- [ ] 添加更多角色主题设置，方便导入
+
+## 🐶 开发自己的桌宠
+
+基于 Godot 的轻便和易学，你可以很容易构建独属于你自己的桌面宠物形象。
+
+### 🐾 步骤
+
+1. 准备一套角色帧动： 你可以查看[itch.ion](https://itch.io/game-assets)等，选择一套自己喜欢的下载，**注意版权，仅自己用不用太注意**。
+2. Godot！启动！[下载](https://godotengine.org/download/windows/)
+3. 查看并学习官方文档：不需要害怕，Godot很简单，查看[Animation](https://docs.godotengine.org/en/4.2/tutorials/animation/index.html)。
+4. 开发（continue...）
+
+
+5. [导出](https://docs.godotengine.org/zh-cn/4.x/tutorials/export/index.html)
+
+### 🏗️ 项目架构
+
+```css
+- root(Window)
+  - Globals(Node)
+  - App(Node)
+	- Canvas(Node2D) - 角色部分
+	  - Grapic(Node2D) - 显示区域 和 角色管理
+	  - Dialogue(Control) - 对话显示
+	- Send(Window) - 发送消息
+	  - TabContainer(TabContainer)
+		- 对话栏(Control)
+		- 存储预设(Control)
+		- 编辑预设(Control)
+		  - Model(PanelContainer)
+		  - Parament(PanelContainer)
+```
+
+### 🔊 注意
+
+> [!Caution]
+> 1. 开发使用的godot版本号为 4.2.2 stable，Godot4 应该都是可以的(未测试) 
+> 2. 导入角色素材记得合适的缩放，同时不忘了调整 显示区域(ClickPolygon) 的大小
+> 3. Godot右上角 选择**兼容模式**
+
+## 🤝 贡献
+
+
+项目目前处于开发阶段，还存在许多 Bug，我正着手修复。
+欢迎任何贡献，即使细微。
+添加角色素材等等
+
+## ©️ 版权和授权
+
+你可以使用项目中的任何代码片段，但不能直接将完整项目打包上架市场。
+同时，需要注意项目中使用到的一些资产的版权信息。
+
+
+## ❤️ 鸣谢
+
+多亏了如下开源的项目和精彩绝伦的资产，才有了该项目：
+
+- [ChatGPT-stream-for-Godot-4](https://github.com/oceanbuilders/ChatGPT-stream-for-Godot-4) - stream输出的支持
+- [godot-click-through-transparent-window](https://github.com/atadenizoktay/godot-click-through-transparent-window) - 透明窗口
+- [tech-dungeon-roguelite](https://trevor-pupkin.itch.io/tech-dungeon-roguelite) - 角色资产
+- [tiny-swords](https://pixelfrog-assets.itch.io/tiny-swords) - 角色资产
+- [dino-characters](https://arks.itch.io/dino-characters) - 角色资产

README_EN.md

@@ -0,0 +1,104 @@
+<h1 align="center">🐶 Godog 🐶</h1>
+<h2 align="center">AI Godot Desktop Pet</h2>
+
+[简体中文](README.md) / English
+
+## 👋 Introduction
+
+🚀 🚀 🚀 A desktop AI pet powered by a large language model and made with Godot. This project aims to provide a versatile and rich desktop AI pet. You can use it as a foundation to create your own unique pet image and character behavior.
+
+- Using Godot: lightweight (90MB memory usage), simple development, provides animation, control UI framework, etc.
+- Compatible with many large language models
+- Pre-made character feature with multiple appearances
+- Supports multi-modal capabilities
+- Easy to build your own AI desktop pet on this basis
+
+![](/img/example.png)
+
+## 📙 User Manual
+
+### ⬇️ Installation
+
+Download: Check `release`
+
+### 🔑 Usage
+1. Click `Edit Preset`, fill in the name and API Key, and click **Save (Modify)**
+2. Click `Store Preset` and select the preset you just added
+3. Start the conversation
+
+> [!Caution]
+> Currently, only OpenAI's vision capability (VISION) is supported. It is recommended to use the **gpt-4o** model.
+> You can use other models, but multi-modal capabilities are not supported.
+
+### ⌨️ Shortcuts
+- **Right-click** on the character area to pop up the dialog box
+- When the focus is on the character window, press **ESC** to close the software
+
+## 🚦 Development Plan
+
+- [ ] Add support for multiple languages
+- [ ] Add more personalized character animations, mechanisms, events
+- [ ] Add support for more models
+- [ ] Add support for Markdown syntax rendering in the dialog box
+- [ ] Add multi-modal capabilities such as voice, file upload, etc.
+- [ ] Add more presets
+- [ ] Add more character theme settings for easy import
+
+## 🐶 Developing Your Own Desktop Pet
+
+With Godot's lightweight and easy-to-learn nature, you can easily create your own unique desktop pet image.
+
+### 🐾 Steps
+
+1. Prepare a set of character frames: You can check [itch.io](https://itch.io/game-assets), choose a set you like and download it, **be mindful of copyright if not for personal use**.
+2. Godot! Start! [Download](https://godotengine.org/download/windows/)
+3. Review and learn the official documentation: Don’t be afraid, Godot is simple. Check out [Animation](https://docs.godotengine.org/en/4.2/tutorials/animation/index.html).
+4. Development (continue...)
+
+5. [Export](https://docs.godotengine.org/en/4.x/tutorials/export/index.html)
+
+### 🏗️ Project Structure
+
+```css
+- root(Window)
+  - Globals(Node)
+  - App(Node)
+	- Canvas(Node2D) - Character part
+	  - Graphic(Node2D) - Display area and character management
+	  - Dialogue(Control) - Dialog display
+	- Send(Window) - Send message
+	  - TabContainer(TabContainer)
+		- Dialog tab(Control)
+		- Store preset(Control)
+		- Edit preset(Control)
+		  - Model(PanelContainer)
+		  - Parameters(PanelContainer)
+```
+
+### 🔊 Notes
+
+> [!Caution]
+> 1. The Godot version used for development is 4.2.2 stable, Godot4 should work as well (not tested)
+> 2. Remember to scale the imported character assets appropriately and adjust the display area (ClickPolygon) size
+> 3. In the upper right corner of Godot, select **Compatibility Mode**
+
+## 🤝 Contributions
+
+The project is currently under development and has many bugs that I am working to fix.
+Any contributions are welcome, even small ones.
+Adding character assets, etc.
+
+## ©️ License
+
+You can use any code snippet from the project, but you cannot directly package and sell the complete project.
+Additionally, pay attention to the copyright information of some assets used in the project.
+
+## ❤️ Acknowledgements
+
+Thanks to the following open-source projects and amazing assets, without which this project would not exist:
+
+- [ChatGPT-stream-for-Godot-4](https://github.com/oceanbuilders/ChatGPT-stream-for-Godot-4) - Support for stream output
+- [godot-click-through-transparent-window](https://github.com/atadenizoktay/godot-click-through-transparent-window) - Transparent window
+- [tech-dungeon-roguelite](https://trevor-pupkin.itch.io/tech-dungeon-roguelite) - Character assets
+- [tiny-swords](https://pixelfrog-assets.itch.io/tiny-swords) - Character assets
+- [dino-characters](https://arks.itch.io/dino-characters) - Character assets

addons/HTTPSSEClient/HTTPSSEClient_modified.gd

@@ -0,0 +1,144 @@
+class_name HTTPSSEClient
+extends Node
+
+signal new_sse_event
+signal connected
+signal connection_error(error)
+
+
+var httpclient = HTTPClient.new()
+var is_connected = false
+
+var domain
+var url_after_domain
+var headers : PackedStringArray
+var body: String
+var port
+var told_to_connect = false
+var connection_in_progress = false
+var request_in_progress = false
+var is_requested = false
+var response_body = PackedByteArray()
+
+var ai_status_message
+
+func connect_to_host(domain : String, url_after_domain : String, headers: PackedStringArray, body: String, ai_message: ChatMessageAI, port : int = -1):
+	self.domain = domain
+	self.url_after_domain = url_after_domain
+	self.port = port
+	self.headers = headers
+	self.body = body
+	told_to_connect = true
+	ai_status_message = ai_message
+
+func close_connection():
+	httpclient.close()
+	is_connected = false
+	told_to_connect = false
+	connection_in_progress = false
+	request_in_progress = false
+	is_requested = false
+
+func attempt_to_connect():
+	var err = httpclient.connect_to_host(domain, port)
+	if err == OK:
+		emit_signal("connected")
+		is_connected = true
+	else:
+		emit_signal("connection_error", str(err))
+
+func attempt_to_request(httpclient_status):
+	if httpclient_status == HTTPClient.STATUS_CONNECTING or httpclient_status == HTTPClient.STATUS_RESOLVING:
+		return
+
+	if httpclient_status == HTTPClient.STATUS_CONNECTED:
+		var err = httpclient.request(HTTPClient.METHOD_POST, url_after_domain, headers, body)
+		if err == OK:
+			is_requested = true
+
+func _process(delta):
+	if !told_to_connect:
+		return
+
+	if !is_connected:
+		if !connection_in_progress:
+			attempt_to_connect()
+			connection_in_progress = true
+		return
+
+	httpclient.poll()
+	var httpclient_status = httpclient.get_status()
+	if !is_requested:
+		if !request_in_progress:
+			attempt_to_request(httpclient_status)
+		return
+
+	var httpclient_has_response = httpclient.has_response()
+
+	if httpclient_has_response or httpclient_status == HTTPClient.STATUS_BODY:
+		var response_headers = httpclient.get_response_headers_as_dictionary()
+
+		httpclient.poll()
+		var chunk = httpclient.read_response_body_chunk()
+		if(chunk.size() == 0):
+			return
+		else:
+			# each chunk sometimes can contain more than one delta of data.
+			response_body = chunk 
+
+		var string_body = response_body.get_string_from_utf8()
+		if string_body:
+			var partial_reply = get_open_ai_events_data(string_body)
+			emit_signal("new_sse_event", partial_reply, ai_status_message)
+
+
+# Since each chunk sometimes can contain more than one delta of data this
+# function will iterate through each "data" block found and extract its
+# content concatenating it in an array.
+func get_open_ai_events_data(string_body: String) -> Array:
+	var results = []
+	string_body = string_body.strip_edges()  # Remove leading and trailing whitespaces, including new lines
+	var data_entries = string_body.split("\n\n")  # Split the body into individual data entries
+
+	for entry in data_entries:
+		if entry == "data: [DONE]":
+			results.append("[DONE]")
+			continue
+
+		entry = entry.replace("data: ", "")
+		print("Processing entry: ", entry)  # Debugging line
+		var parsed_data = JSON.parse_string(entry)
+		if parsed_data == null:
+			printerr("Error: Received data is not a valid JSON object")
+			results.append("[ERROR]")
+		else:
+			if "choices" in parsed_data and parsed_data["choices"].size() > 0:
+				var choices = parsed_data["choices"][0]
+				if "delta" in choices:
+					var delta = choices["delta"]
+					if "content" in delta:
+						results.append(delta["content"])
+					elif delta.is_empty():
+						results.append("[EMPTY DELTA]")
+					else:
+						printerr("Error: 'content' field not found in the 'delta'")
+						results.append("[ERROR]")
+				else:
+					printerr("Error: 'delta' field not found in the first choice")
+					results.append("[ERROR]")
+			else:
+				printerr("Error: 'choices' field not found in the received data")
+				results.append("[ERROR]")
+
+	return results
+
+
+func _exit_tree():
+	if httpclient:
+		httpclient.close()
+
+func _notification(what):
+	if what == NOTIFICATION_WM_CLOSE_REQUEST:
+		if httpclient:
+			httpclient.close()
+		get_tree().quit()