From 232348fe7a8476134b31be64f57c502c01f4e197 Mon Sep 17 00:00:00 2001 From: sky2002 Date: Fri, 7 Mar 2025 13:37:44 +0000 Subject: [PATCH 1/6] =?UTF-8?q?=E4=BD=BF=E7=94=A8generate=5Fresponse=5Fasy?= =?UTF-8?q?nc?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- src/plugins/chat/llm_generator.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/plugins/chat/llm_generator.py b/src/plugins/chat/llm_generator.py index a19a222c2..04016f34b 100644 --- a/src/plugins/chat/llm_generator.py +++ b/src/plugins/chat/llm_generator.py @@ -194,6 +194,6 @@ class InitiativeMessageGenerate: prompt = prompt_builder._build_initiative_prompt( select_dot, prompt_template, memory ) - content, reasoning = self.model_r1.generate_response(prompt) + content, reasoning = self.model_r1.generate_response_async(prompt) print(f"[DEBUG] {content} {reasoning}") return content From cdb44d70b4cfc95827b9bff473fe8e7936c26fc2 Mon Sep 17 00:00:00 2001 From: sky2002 Date: Fri, 7 Mar 2025 13:39:14 +0000 Subject: [PATCH 2/6] =?UTF-8?q?=E4=BD=BF=E7=94=A8generate=5Fresponse=5Fasy?= =?UTF-8?q?nc?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- src/plugins/chat/llm_generator.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/plugins/chat/llm_generator.py b/src/plugins/chat/llm_generator.py index 04016f34b..4954c8494 100644 --- a/src/plugins/chat/llm_generator.py +++ b/src/plugins/chat/llm_generator.py @@ -21,7 +21,7 @@ config = driver.config class ResponseGenerator: def __init__(self): - self.model_r1 = LLM_request(model=global_config.llm_reasoning, temperature=0.7,max_tokens=1000) + self.model_r1 = LLM_request(model=global_config.llm_reasoning, temperature=0.7,max_tokens=1000,stream=True) self.model_v3 = LLM_request(model=global_config.llm_normal, temperature=0.7,max_tokens=1000) self.model_r1_distill = LLM_request(model=global_config.llm_reasoning_minor, temperature=0.7,max_tokens=1000) self.model_v25 = LLM_request(model=global_config.llm_normal_minor, temperature=0.7,max_tokens=1000) From 8497441507ed8b1a9bac9c580b058086c698e5fd Mon Sep 17 00:00:00 2001 From: sky2002 Date: Fri, 7 Mar 2025 14:17:03 +0000 Subject: [PATCH 3/6] =?UTF-8?q?=E4=BD=BF=E6=94=AF=E6=8C=81=E6=B5=81?= =?UTF-8?q?=E5=BC=8F=E8=BE=93=E5=87=BA?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- src/plugins/models/utils_model.py | 50 +++++++++++++++++++++++++++---- 1 file changed, 44 insertions(+), 6 deletions(-) diff --git a/src/plugins/models/utils_model.py b/src/plugins/models/utils_model.py index a471bd72d..7604389bf 100644 --- a/src/plugins/models/utils_model.py +++ b/src/plugins/models/utils_model.py @@ -1,5 +1,6 @@ import aiohttp import asyncio +import json import requests import time import re @@ -138,7 +139,12 @@ class LLM_request: } api_url = f"{self.base_url.rstrip('/')}/{endpoint.lstrip('/')}" - logger.info(f"发送请求到URL: {api_url}") + #判断是否为流式 + stream_mode = self.params.get("stream", False) + if self.params.get("stream", False) is True: + logger.info(f"进入流式输出模式，发送请求到URL: {api_url}") + else: + logger.info(f"发送请求到URL: {api_url}") logger.info(f"使用模型: {self.model_name}") # 构建请求体 @@ -151,6 +157,9 @@ class LLM_request: try: # 使用上下文管理器处理会话 headers = await self._build_headers() + #似乎是openai流式必须要的东西,不过阿里云的qwq-plus加了这个没有影响 + if stream_mode: + headers["Accept"] = "text/event-stream" async with aiohttp.ClientSession() as session: async with session.post(api_url, headers=headers, json=payload) as response: @@ -173,12 +182,41 @@ class LLM_request: elif response.status in policy["abort_codes"]: logger.error(f"错误码: {response.status} - {error_code_mapping.get(response.status)}") raise RuntimeError(f"请求被拒绝: {error_code_mapping.get(response.status)}") - + response.raise_for_status() - result = await response.json() - - # 使用自定义处理器或默认处理 - return response_handler(result) if response_handler else self._default_response_handler(result, user_id, request_type, endpoint) + + if stream_mode: + accumulated_content = "" + async for line_bytes in response.content: + line = line_bytes.decode("utf-8").strip() + if not line: + continue + if line.startswith("data:"): + data_str = line[5:].strip() + if data_str == "[DONE]": + break + try: + chunk = json.loads(data_str) + delta = chunk["choices"][0]["delta"] + delta_content = delta.get("content") + if delta_content is None: + delta_content = "" + accumulated_content += delta_content + except Exception as e: + logger.error(f"解析流式输出错误: {e}") + content = accumulated_content + reasoning_content = "" + think_match = re.search(r'(.*?)', content, re.DOTALL) + if think_match: + reasoning_content = think_match.group(1).strip() + content = re.sub(r'.*?', '', content, flags=re.DOTALL).strip() + # 构造一个伪result以便调用自定义响应处理器或默认处理器 + result = {"choices": [{"message": {"content": content, "reasoning_content": reasoning_content}}]} + return response_handler(result) if response_handler else self._default_response_handler(result, user_id, request_type, endpoint) + else: + result = await response.json() + # 使用自定义处理器或默认处理 + return response_handler(result) if response_handler else self._default_response_handler(result, user_id, request_type, endpoint) except Exception as e: if retry < policy["max_retries"] - 1: From 7f35e70483f7b611481af31504992badc5c0dfe4 Mon Sep 17 00:00:00 2001 From: sky2002 Date: Fri, 7 Mar 2025 14:39:16 +0000 Subject: [PATCH 4/6] =?UTF-8?q?=E5=AF=B9openai=E7=9A=84CoT=E6=A8=A1?= =?UTF-8?q?=E5=9E=8B=E8=BF=9B=E8=A1=8C=E7=89=B9=E5=88=AB=E4=BF=AE=E6=94=B9?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- src/plugins/models/utils_model.py | 23 +++++++++++++++++++---- 1 file changed, 19 insertions(+), 4 deletions(-) diff --git a/src/plugins/models/utils_model.py b/src/plugins/models/utils_model.py index 7604389bf..680230343 100644 --- a/src/plugins/models/utils_model.py +++ b/src/plugins/models/utils_model.py @@ -233,8 +233,18 @@ class LLM_request: async def _build_payload(self, prompt: str, image_base64: str = None) -> dict: """构建请求体""" + # 复制一份参数，避免直接修改 self.params + params_copy = dict(self.params) + if self.model_name.lower() == "o3-mini" or "o1-mini" or "o1" or "o1-2024-12-17" or "o1-preview-2024-09-12" or "o3-mini-2025-01-31" or "o1-mini-2024-09-12": + # 删除可能存在的 'temprature' 参数 + params_copy.pop("temprature", None) + # 如果存在 'max_tokens' 参数，则将其替换为 'max_completion_tokens' + if "max_tokens" in params_copy: + params_copy["max_completion_tokens"] = params_copy.pop("max_tokens") + # 构造基础请求体，注意这里依然使用 global_config.max_response_length 填充 'max_tokens' + # 如果需要统一改为 max_completion_tokens，也可以在下面做相同的调整 if image_base64: - return { + payload = { "model": self.model_name, "messages": [ { @@ -246,15 +256,20 @@ class LLM_request: } ], "max_tokens": global_config.max_response_length, - **self.params + **params_copy } else: - return { + payload = { "model": self.model_name, "messages": [{"role": "user", "content": prompt}], "max_tokens": global_config.max_response_length, - **self.params + **params_copy } + # 如果是 o3-mini 模型，也将基础请求体中的 max_tokens 改为 max_completion_tokens + if self.model_name.lower() == "o3-mini" and "max_tokens" in payload: + payload["max_completion_tokens"] = payload.pop("max_tokens") + return payload + def _default_response_handler(self, result: dict, user_id: str = "system", request_type: str = "chat", endpoint: str = "/chat/completions") -> Tuple: From 527b505712a28a409c7ef19dfe37b557f9863411 Mon Sep 17 00:00:00 2001 From: KawaiiYusora Date: Sat, 8 Mar 2025 00:23:56 +0800 Subject: [PATCH 5/6] =?UTF-8?q?typo:=20template=E6=8B=BC=E5=86=99=E9=94=99?= =?UTF-8?q?=E8=AF=AF?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- {templete => template}/auto_format.py | 0 {templete => template}/bot_config_template.toml | 0 2 files changed, 0 insertions(+), 0 deletions(-) rename {templete => template}/auto_format.py (100%) rename {templete => template}/bot_config_template.toml (100%) diff --git a/templete/auto_format.py b/template/auto_format.py similarity index 100% rename from templete/auto_format.py rename to template/auto_format.py diff --git a/templete/bot_config_template.toml b/template/bot_config_template.toml similarity index 100% rename from templete/bot_config_template.toml rename to template/bot_config_template.toml From 5247edc755d224b7cc3439a8d0a44a24546a7839 Mon Sep 17 00:00:00 2001 From: KawaiiYusora Date: Sat, 8 Mar 2025 00:24:18 +0800 Subject: [PATCH 6/6] =?UTF-8?q?fix:=20=E5=88=9D=E5=A7=8B=E5=8C=96=E6=97=B6?= =?UTF-8?q?=E6=B2=A1=E6=9C=89config=E6=96=87=E4=BB=B6=E5=A4=B9=E4=BC=9A?= =?UTF-8?q?=E6=8A=A5=E9=94=99?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- bot.py | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/bot.py b/bot.py index 1bebfe553..715fe9aa7 100644 --- a/bot.py +++ b/bot.py @@ -20,8 +20,12 @@ print(rainbow_text) if not os.path.exists("config/bot_config.toml"): logger.warning("检测到bot_config.toml不存在，正在从模板复制") import shutil + # 检查config目录是否存在 + if not os.path.exists("config"): + os.makedirs("config") + logger.info("创建config目录") - shutil.copy("templete/bot_config_template.toml", "config/bot_config.toml") + shutil.copy("template/bot_config_template.toml", "config/bot_config.toml") logger.info("复制完成，请修改config/bot_config.toml和.env.prod中的配置后重新启动") # 初始化.env 默认ENVIRONMENT=prod