feat: custom placeholders

feat: add placeholder to chatbot

fix invalid_request_error in GPT models

feat: model description move to chatbot header

feat: Added description capability to display a piece of text in the chatbot. Added description for GPT models and Claude models for now.

feat: Added support for custom models

Added support for Groq, the super fast inference service.

fix system prompt issues.

Merged GPT4-V into GPT4-Turbo to accommodate OpenAI's new API change.

fix [Bug]: 配置"hide_history_when_not_logged_in": true后，界面报错 #1101

openai besed RAG uses text-embedding-3-large now.

feat: improved Chuanhu Assistant behaviour when handling file uploads

since available tools can be infered from config, Chuanhu Assistant now uses GPT 3.5 by default, and uses GPT4-Turbo (preview) in Pro mode

code clean up, bumped openai version

beautified tool usage logging

bumped ddg version, renamed tool in Chuanhu Agent

Improved Chuanhu Agent with OpenAI Tools support

gradio 4 update

added tokens per second logging

fix file uploading error

ChuanhuChat now runs on gradio4 (sort of)

bugfix: bulk response from Claude 3

fixed multimodal warning

feat: added logging of tokens per second

removed abundant warnings

fix #1071

fix #1066

Fix stop sequence initialization as empty string (#1063)

bugfix: Improvements on GPT4V

bugfix: retry doesn't work when Claude API errors out (such as when overloaded)

feat: Claude 3 multimmodal improvement, ChuanhuChat now converts unsupported image formats such as .jpg into jpeg

feat: Added Claude 3 Opus and Claude 3 Sonnet support, with multimodal capabilities! resolve #1062

chore: RAG code quality improvements

chore: compability fix

feat: Ollama后端支持 (#1051)

bugfix: Gemma generation too short

feat: Added Gemma support (needs hf auth token)

feat: 加入Gemini Pro (Vision) 支持 #1039

feat: 打开网页时创建新对话，而不是加载最晚的那一条

feat: 增加对OpenAiVison的图片token计算 (#1031)

feat: 搜索功能现在可以使用config中设置的http_proxy和https_proxy了

fix [其他]: 川虎助手问题 #1001

bugfix: fix #971

feat: Qwen支持加载本地/自定义模型

bugfix: 修复Qwen bug

bugfix: 修复Vision模型回复中断的问题

bugfix: 适配GPT4-V最新对话格式

feat: 适配新版openai API

修复了api-host在DALLE3中的问题 (#994)

feat: OpenAI模型现在会检查API Key是否为空