fix: 更新模型配置和环境变量，调整版本号和模型参数

bot.py

@@ -74,36 +74,6 @@ def easter_egg():
     print(rainbow_text)
 
 
-def scan_provider(env_config: dict):
-    provider = {}
-
-    # 利用未初始化 env 时获取的 env_mask 来对新的环境变量集去重
-    # 避免 GPG_KEY 这样的变量干扰检查
-    env_config = dict(filter(lambda item: item[0] not in env_mask, env_config.items()))
-
-    # 遍历 env_config 的所有键
-    for key in env_config:
-        # 检查键是否符合 {provider}_BASE_URL 或 {provider}_KEY 的格式
-        if key.endswith("_BASE_URL") or key.endswith("_KEY"):
-            # 提取 provider 名称
-            provider_name = key.split("_", 1)[0]  # 从左分割一次，取第一部分
-
-            # 初始化 provider 的字典（如果尚未初始化）
-            if provider_name not in provider:
-                provider[provider_name] = {"url": None, "key": None}
-
-            # 根据键的类型填充 url 或 key
-            if key.endswith("_BASE_URL"):
-                provider[provider_name]["url"] = env_config[key]
-            elif key.endswith("_KEY"):
-                provider[provider_name]["key"] = env_config[key]
-
-    # 检查每个 provider 是否同时存在 url 和 key
-    for provider_name, config in provider.items():
-        if config["url"] is None or config["key"] is None:
-            logger.error(f"provider 内容：{config}\nenv_config 内容：{env_config}")
-            raise ValueError(f"请检查 '{provider_name}' 提供商配置是否丢失 BASE_URL 或 KEY 环境变量")
-
 
 async def graceful_shutdown():
     try:
@@ -229,9 +199,6 @@ def raw_main():
 
     easter_egg()
 
-    env_config = {key: os.getenv(key) for key in os.environ}
-    scan_provider(env_config)
-
     # 返回MainSystem实例
     return MainSystem()
 

src/config/api_ada_configs.py

@@ -3,7 +3,7 @@ from typing import List, Dict
 
 from packaging.version import Version
 
-NEWEST_VER = "0.1.0"  # 当前支持的最新版本
+NEWEST_VER = "0.1.1"  # 当前支持的最新版本
 
 @dataclass
 class APIProvider:

src/config/auto_update.py

@@ -1,162 +0,0 @@
-import shutil
-import tomlkit
-from tomlkit.items import Table, KeyType
-from pathlib import Path
-from datetime import datetime
-
-
-def get_key_comment(toml_table, key):
-    # 获取key的注释（如果有）
-    if hasattr(toml_table, "trivia") and hasattr(toml_table.trivia, "comment"):
-        return toml_table.trivia.comment
-    if hasattr(toml_table, "value") and isinstance(toml_table.value, dict):
-        item = toml_table.value.get(key)
-        if item is not None and hasattr(item, "trivia"):
-            return item.trivia.comment
-    if hasattr(toml_table, "keys"):
-        for k in toml_table.keys():
-            if isinstance(k, KeyType) and k.key == key:
-                return k.trivia.comment
-    return None
-
-
-def compare_dicts(new, old, path=None, new_comments=None, old_comments=None, logs=None):
-    # 递归比较两个dict，找出新增和删减项，收集注释
-    if path is None:
-        path = []
-    if logs is None:
-        logs = []
-    if new_comments is None:
-        new_comments = {}
-    if old_comments is None:
-        old_comments = {}
-    # 新增项
-    for key in new:
-        if key == "version":
-            continue
-        if key not in old:
-            comment = get_key_comment(new, key)
-            logs.append(f"新增: {'.'.join(path + [str(key)])}  注释: {comment if comment else '无'}")
-        elif isinstance(new[key], (dict, Table)) and isinstance(old.get(key), (dict, Table)):
-            compare_dicts(new[key], old[key], path + [str(key)], new_comments, old_comments, logs)
-    # 删减项
-    for key in old:
-        if key == "version":
-            continue
-        if key not in new:
-            comment = get_key_comment(old, key)
-            logs.append(f"删减: {'.'.join(path + [str(key)])}  注释: {comment if comment else '无'}")
-    return logs
-
-
-def update_config():
-    print("开始更新配置文件...")
-    # 获取根目录路径
-    root_dir = Path(__file__).parent.parent.parent.parent
-    template_dir = root_dir / "template"
-    config_dir = root_dir / "config"
-    old_config_dir = config_dir / "old"
-
-    # 创建old目录（如果不存在）
-    old_config_dir.mkdir(exist_ok=True)
-
-    # 定义文件路径
-    template_path = template_dir / "bot_config_template.toml"
-    old_config_path = config_dir / "bot_config.toml"
-    new_config_path = config_dir / "bot_config.toml"
-
-    # 读取旧配置文件
-    old_config = {}
-    if old_config_path.exists():
-        print(f"发现旧配置文件: {old_config_path}")
-        with open(old_config_path, "r", encoding="utf-8") as f:
-            old_config = tomlkit.load(f)
-
-        # 生成带时间戳的新文件名
-        timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
-        old_backup_path = old_config_dir / f"bot_config_{timestamp}.toml"
-
-        # 移动旧配置文件到old目录
-        shutil.move(old_config_path, old_backup_path)
-        print(f"已备份旧配置文件到: {old_backup_path}")
-
-    # 复制模板文件到配置目录
-    print(f"从模板文件创建新配置: {template_path}")
-    shutil.copy2(template_path, new_config_path)
-
-    # 读取新配置文件
-    with open(new_config_path, "r", encoding="utf-8") as f:
-        new_config = tomlkit.load(f)
-
-    # 检查version是否相同
-    if old_config and "inner" in old_config and "inner" in new_config:
-        old_version = old_config["inner"].get("version")  # type: ignore
-        new_version = new_config["inner"].get("version")  # type: ignore
-        if old_version and new_version and old_version == new_version:
-            print(f"检测到版本号相同 (v{old_version})，跳过更新")
-            # 如果version相同，恢复旧配置文件并返回
-            shutil.move(old_backup_path, old_config_path)  # type: ignore
-            return
-        else:
-            print(f"检测到版本号不同: 旧版本 v{old_version} -> 新版本 v{new_version}")
-
-    # 输出新增和删减项及注释
-    if old_config:
-        print("配置项变动如下：")
-        logs = compare_dicts(new_config, old_config)
-        if logs:
-            for log in logs:
-                print(log)
-        else:
-            print("无新增或删减项")
-
-    # 递归更新配置
-    def update_dict(target, source):
-        for key, value in source.items():
-            # 跳过version字段的更新
-            if key == "version":
-                continue
-            if key in target:
-                if isinstance(value, dict) and isinstance(target[key], (dict, Table)):
-                    update_dict(target[key], value)
-                else:
-                    try:
-                        # 对数组类型进行特殊处理
-                        if isinstance(value, list):
-                            # 如果是空数组，确保它保持为空数组
-                            if not value:
-                                target[key] = tomlkit.array()
-                            else:
-                                # 特殊处理正则表达式数组和包含正则表达式的结构
-                                if key == "ban_msgs_regex":
-                                    # 直接使用原始值，不进行额外处理
-                                    target[key] = value
-                                elif key == "regex_rules":
-                                    # 对于regex_rules，需要特殊处理其中的regex字段
-                                    target[key] = value
-                                else:
-                                    # 检查是否包含正则表达式相关的字典项
-                                    contains_regex = False
-                                    if value and isinstance(value[0], dict) and "regex" in value[0]:
-                                        contains_regex = True
-
-                                    target[key] = value if contains_regex else tomlkit.array(str(value))
-                        else:
-                            # 其他类型使用item方法创建新值
-                            target[key] = tomlkit.item(value)
-                    except (TypeError, ValueError):
-                        # 如果转换失败，直接赋值
-                        target[key] = value
-
-    # 将旧配置的值更新到新配置中
-    print("开始合并新旧配置...")
-    update_dict(new_config, old_config)
-
-    # 保存更新后的配置（保留注释和格式）
-    with open(new_config_path, "w", encoding="utf-8") as f:
-        f.write(tomlkit.dumps(new_config))
-    print("配置文件更新完成")
-
-
-if __name__ == "__main__":
-    update_config()

template/bot_config_template.toml

@@ -229,120 +229,84 @@ show_prompt = false # 是否显示prompt
 
 
 [model]
-model_max_output_length = 1024 # 模型单次返回的最大token数
+model_max_output_length = 800 # 模型单次返回的最大token数
 
-#------------必填：组件模型------------
+#------------模型任务配置------------
+# 所有模型名称需要对应 model_config.toml 中配置的模型名称
 
 [model.utils] # 在麦麦的一些组件中使用的模型，例如表情包模块，取名模块，关系模块，是麦麦必须的模型
-name = "Pro/deepseek-ai/DeepSeek-V3"
+model_name = "siliconflow-deepseek-v3"  # 对应 model_config.toml 中的模型名称
-provider = "SILICONFLOW"
+temperature = 0.2  # 模型温度，新V3建议0.1-0.3
-pri_in = 2 #模型的输入价格（非必填，可以记录消耗）
+max_tokens = 800   # 最大输出token数
-pri_out = 8 #模型的输出价格（非必填，可以记录消耗）
-#默认temp 0.2 如果你使用的是老V3或者其他模型，请自己修改temp参数
-temp = 0.2 #模型的温度，新V3建议0.1-0.3
 
 [model.utils_small] # 在麦麦的一些组件中使用的小模型，消耗量较大，建议使用速度较快的小模型
-# 强烈建议使用免费的小模型
+model_name = "qwen3-8b"  # 对应 model_config.toml 中的模型名称
-name = "Qwen/Qwen3-8B"
+temperature = 0.7
-provider = "SILICONFLOW"
+max_tokens = 800
-pri_in = 0
-pri_out = 0
-temp = 0.7
 enable_thinking = false # 是否启用思考
 
 [model.replyer_1] # 首要回复模型，还用于表达器和表达方式学习
-name = "Pro/deepseek-ai/DeepSeek-V3"
+model_name = "siliconflow-deepseek-v3"  # 对应 model_config.toml 中的模型名称
-provider = "SILICONFLOW"
+temperature = 0.2  # 模型温度，新V3建议0.1-0.3
-pri_in = 2 #模型的输入价格（非必填，可以记录消耗）
+max_tokens = 800
-pri_out = 8 #模型的输出价格（非必填，可以记录消耗）
-#默认temp 0.2 如果你使用的是老V3或者其他模型，请自己修改temp参数
-temp = 0.2 #模型的温度，新V3建议0.1-0.3
 
 [model.replyer_2] # 次要回复模型
-name = "Pro/deepseek-ai/DeepSeek-V3"
+model_name = "siliconflow-deepseek-r1"  # 对应 model_config.toml 中的模型名称
-provider = "SILICONFLOW"
+temperature = 0.7  # 模型温度
-pri_in = 2 #模型的输入价格（非必填，可以记录消耗）
+max_tokens = 800
-pri_out = 8 #模型的输出价格（非必填，可以记录消耗）
-#默认temp 0.2 如果你使用的是老V3或者其他模型，请自己修改temp参数
-temp = 0.2 #模型的温度，新V3建议0.1-0.3
 
 [model.planner] #决策：负责决定麦麦该做什么的模型
-name = "Pro/deepseek-ai/DeepSeek-V3"
+model_name = "siliconflow-deepseek-v3"  # 对应 model_config.toml 中的模型名称
-provider = "SILICONFLOW"
+temperature = 0.3
-pri_in = 2
+max_tokens = 800
-pri_out = 8
-temp = 0.3
 
 [model.emotion] #负责麦麦的情绪变化
-name = "Pro/deepseek-ai/DeepSeek-V3"
+model_name = "siliconflow-deepseek-v3"  # 对应 model_config.toml 中的模型名称
-provider = "SILICONFLOW"
+temperature = 0.3
-pri_in = 2
+max_tokens = 800
-pri_out = 8
-temp = 0.3
-
 
 [model.memory] # 记忆模型
-name = "Qwen/Qwen3-30B-A3B"
+model_name = "qwen3-30b"  # 对应 model_config.toml 中的模型名称
-provider = "SILICONFLOW"
+temperature = 0.7
-pri_in = 0.7
+max_tokens = 800
-pri_out = 2.8
-temp = 0.7
 enable_thinking = false # 是否启用思考
 
 [model.vlm] # 图像识别模型
-name = "Pro/Qwen/Qwen2.5-VL-7B-Instruct"
+model_name = "qwen2.5-vl-72b"  # 对应 model_config.toml 中的模型名称
-provider = "SILICONFLOW"
+max_tokens = 800
-pri_in = 0.35
-pri_out = 0.35
 
 [model.voice] # 语音识别模型
-name = "FunAudioLLM/SenseVoiceSmall"
+model_name = "sensevoice-small"  # 对应 model_config.toml 中的模型名称
-provider = "SILICONFLOW"
-pri_in = 0
-pri_out = 0
 
 [model.tool_use] #工具调用模型，需要使用支持工具调用的模型
-name = "Qwen/Qwen3-14B"
+model_name = "qwen3-14b"  # 对应 model_config.toml 中的模型名称
-provider = "SILICONFLOW"
+temperature = 0.7
-pri_in = 0.5
+max_tokens = 800
-pri_out = 2
-temp = 0.7
 enable_thinking = false # 是否启用思考（qwen3 only）
 
 #嵌入模型
 [model.embedding]
-name = "BAAI/bge-m3"
+model_name = "bge-m3"  # 对应 model_config.toml 中的模型名称
-provider = "SILICONFLOW"
-pri_in = 0
-pri_out = 0
-
 
 #------------LPMM知识库模型------------
 
 [model.lpmm_entity_extract] # 实体提取模型
-name = "Pro/deepseek-ai/DeepSeek-V3"
+model_name = "siliconflow-deepseek-v3"  # 对应 model_config.toml 中的模型名称
-provider = "SILICONFLOW"
+temperature = 0.2
-pri_in = 2
+max_tokens = 800
-pri_out = 8
-temp = 0.2
-
 
 [model.lpmm_rdf_build] # RDF构建模型
-name = "Pro/deepseek-ai/DeepSeek-V3"
+model_name = "siliconflow-deepseek-v3"  # 对应 model_config.toml 中的模型名称
-provider = "SILICONFLOW"
+temperature = 0.2
-pri_in = 2
+max_tokens = 800
-pri_out = 8
-temp = 0.2
-
 
 [model.lpmm_qa] # 问答模型
-name = "Qwen/Qwen3-30B-A3B"
+model_name = "deepseek-r1-distill-qwen-32b"  # 对应 model_config.toml 中的模型名称
-provider = "SILICONFLOW"
+temperature = 0.7
-pri_in = 0.7
+max_tokens = 800
-pri_out = 2.8
-temp = 0.7
 enable_thinking = false # 是否启用思考
 
+
 [maim_message]
 auth_token = [] # 认证令牌，用于API验证，为空则不启用验证
 # 以下项目若要使用需要打开use_custom，并单独配置maim_message的服务器

template/compare/model_config_template.toml

@@ -1,5 +1,5 @@
 [inner]
-version = "0.1.0"
+version = "0.1.1"
 
 # 配置文件版本号迭代规则同bot_config.toml
 
@@ -50,21 +50,79 @@ price_out = 8.0
 #（可选，若无该字段，默认值为false）
 #force_stream_mode = true
 
-#[[models]]
+[[models]]
-#model_identifier = "deepseek-reasoner"
+model_identifier = "deepseek-reasoner"
-#name = "deepseek-r1"
+name = "deepseek-r1"
-#api_provider = "DeepSeek"
+api_provider = "DeepSeek"
-#model_flags = ["text", "tool_calling", "reasoning"]
+model_flags = [ "text", "tool_calling", "reasoning",]
-#price_in = 4.0
+price_in = 4.0
-#price_out = 16.0
+price_out = 16.0
-#
+
-#[[models]]
+[[models]]
-#model_identifier = "BAAI/bge-m3"
+model_identifier = "Pro/deepseek-ai/DeepSeek-V3"
-#name = "siliconflow-bge-m3"
+name = "siliconflow-deepseek-v3"
-#api_provider = "SiliconFlow"
+api_provider = "SiliconFlow"
-#model_flags = ["text", "embedding"]
+price_in = 2.0
-#price_in = 0
+price_out = 8.0
-#price_out = 0
+
+[[models]]
+model_identifier = "Pro/deepseek-ai/DeepSeek-R1"
+name = "siliconflow-deepseek-r1"
+api_provider = "SiliconFlow"
+price_in = 4.0
+price_out = 16.0
+
+[[models]]
+model_identifier = "Pro/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B"
+name = "deepseek-r1-distill-qwen-32b"
+api_provider = "SiliconFlow"
+price_in = 4.0
+price_out = 16.0
+
+[[models]]
+model_identifier = "Qwen/Qwen3-8B"
+name = "qwen3-8b"
+api_provider = "SiliconFlow"
+price_in = 0
+price_out = 0
+
+[[models]]
+model_identifier = "Qwen/Qwen3-14B"
+name = "qwen3-14b"
+api_provider = "SiliconFlow"
+price_in = 0.5
+price_out = 2.0
+
+[[models]]
+model_identifier = "Qwen/Qwen3-30B-A3B"
+name = "qwen3-30b"
+api_provider = "SiliconFlow"
+price_in = 0.7
+price_out = 2.8
+
+[[models]]
+model_identifier = "Qwen/Qwen2.5-VL-72B-Instruct"
+name = "qwen2.5-vl-72b"
+api_provider = "SiliconFlow"
+model_flags = [ "vision", "text",]
+price_in = 4.13
+price_out = 4.13
+
+[[models]]
+model_identifier = "FunAudioLLM/SenseVoiceSmall"
+name = "sensevoice-small"
+api_provider = "SiliconFlow"
+model_flags = [ "audio",]
+price_in = 0
+price_out = 0
+
+[[models]]
+model_identifier = "BAAI/bge-m3"
+name = "bge-m3"
+api_provider = "SiliconFlow"
+model_flags = [ "text", "embedding",]
+price_in = 0
+price_out = 0
 
 
 [task_model_usage]

template/model_config_template.toml

@@ -1,5 +1,5 @@
 [inner]
-version = "0.1.0"
+version = "0.1.1"
 
 # 配置文件版本号迭代规则同bot_config.toml
 
@@ -50,27 +50,85 @@ price_out = 8.0
 #（可选，若无该字段，默认值为false）
 #force_stream_mode = true
 
-#[[models]]
+[[models]]
-#model_identifier = "deepseek-reasoner"
+model_identifier = "deepseek-reasoner"
-#name = "deepseek-r1"
+name = "deepseek-r1"
-#api_provider = "DeepSeek"
+api_provider = "DeepSeek"
-#model_flags = ["text", "tool_calling", "reasoning"]
+model_flags = [ "text", "tool_calling", "reasoning",]
-#price_in = 4.0
+price_in = 4.0
-#price_out = 16.0
+price_out = 16.0
-#
+
-#[[models]]
+[[models]]
-#model_identifier = "BAAI/bge-m3"
+model_identifier = "Pro/deepseek-ai/DeepSeek-V3"
-#name = "siliconflow-bge-m3"
+name = "siliconflow-deepseek-v3"
-#api_provider = "SiliconFlow"
+api_provider = "SiliconFlow"
-#model_flags = ["text", "embedding"]
+price_in = 2.0
-#price_in = 0
+price_out = 8.0
-#price_out = 0
+
+[[models]]
+model_identifier = "Pro/deepseek-ai/DeepSeek-R1"
+name = "siliconflow-deepseek-r1"
+api_provider = "SiliconFlow"
+price_in = 4.0
+price_out = 16.0
+
+[[models]]
+model_identifier = "Pro/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B"
+name = "deepseek-r1-distill-qwen-32b"
+api_provider = "SiliconFlow"
+price_in = 4.0
+price_out = 16.0
+
+[[models]]
+model_identifier = "Qwen/Qwen3-8B"
+name = "qwen3-8b"
+api_provider = "SiliconFlow"
+price_in = 0
+price_out = 0
+
+[[models]]
+model_identifier = "Qwen/Qwen3-14B"
+name = "qwen3-14b"
+api_provider = "SiliconFlow"
+price_in = 0.5
+price_out = 2.0
+
+[[models]]
+model_identifier = "Qwen/Qwen3-30B-A3B"
+name = "qwen3-30b"
+api_provider = "SiliconFlow"
+price_in = 0.7
+price_out = 2.8
+
+[[models]]
+model_identifier = "Qwen/Qwen2.5-VL-72B-Instruct"
+name = "qwen2.5-vl-72b"
+api_provider = "SiliconFlow"
+model_flags = [ "vision", "text",]
+price_in = 4.13
+price_out = 4.13
+
+[[models]]
+model_identifier = "FunAudioLLM/SenseVoiceSmall"
+name = "sensevoice-small"
+api_provider = "SiliconFlow"
+model_flags = [ "audio",]
+price_in = 0
+price_out = 0
+
+[[models]]
+model_identifier = "BAAI/bge-m3"
+name = "bge-m3"
+api_provider = "SiliconFlow"
+model_flags = [ "text", "embedding",]
+price_in = 0
+price_out = 0
 
 
 [task_model_usage]
-#llm_reasoning = {model="deepseek-r1", temperature=0.8, max_tokens=1024, max_retry=0}
+llm_reasoning = {model="deepseek-r1", temperature=0.8, max_tokens=1024, max_retry=0}
-#llm_normal = {model="deepseek-r1", max_tokens=1024, max_retry=0}
+llm_normal = {model="deepseek-r1", max_tokens=1024, max_retry=0}
-#embedding = "siliconflow-bge-m3"
+embedding = "siliconflow-bge-m3"
 #schedule = [
 #    "deepseek-v3",
 #    "deepseek-r1",

template/template.env

@@ -1,16 +1,2 @@
 HOST=127.0.0.1
 PORT=8000
-
-#key and url
-SILICONFLOW_BASE_URL=https://api.siliconflow.cn/v1/
-DEEP_SEEK_BASE_URL=https://api.deepseek.com/v1
-CHAT_ANY_WHERE_BASE_URL=https://api.chatanywhere.tech/v1
-BAILIAN_BASE_URL = https://dashscope.aliyuncs.com/compatible-mode/v1
-xxxxxxx_BASE_URL=https://xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
-
-# 定义你要用的api的key(需要去对应网站申请哦)
-DEEP_SEEK_KEY=
-CHAT_ANY_WHERE_KEY=
-SILICONFLOW_KEY=
-BAILIAN_KEY = 
-xxxxxxx_KEY=