Compare commits

...

21 Commits

Author SHA1 Message Date
Junyan Qin
45e4096a12 Merge pull request #587 from RockChinQ/hotfix/openai-1.0-adaptation
Feat: 适配openai>=1.0.0
2023-11-10 11:49:20 +08:00
GitHub Actions
8a1d4fe287 Update override-all.json 2023-11-10 03:47:30 +00:00
RockChinQ
98f880ebc2 chore: 群内回复不再默认引用原消息 2023-11-10 11:47:10 +08:00
RockChinQ
2b852853f3 feat: 适配completion和chat_completions 2023-11-10 11:31:14 +08:00
RockChinQ
c7a9988033 feat: 以新的方式设置正向代理 2023-11-10 10:54:03 +08:00
RockChinQ
c475eebe1c chore: 不再限制openai版本 2023-11-10 10:14:11 +08:00
RockChinQ
0fe7355ae0 hotfix: 适配openai>=1.0.0 2023-11-10 10:13:50 +08:00
Junyan Qin
57de96e3a2 chore(requirements.txt): 锁定openai版本到0.28.1 2023-11-10 09:31:27 +08:00
Junyan Qin
70571cef50 Update README.md 2023-10-02 17:31:08 +08:00
Junyan Qin
0b6deb3340 Update README.md 2023-10-02 17:23:36 +08:00
Junyan Qin
dcda85a825 Merge pull request #580 from RockChinQ/dependabot/pip/openai-approx-eq-0.28.1
chore(deps): update openai requirement from ~=0.28.0 to ~=0.28.1
2023-10-02 16:10:37 +08:00
dependabot[bot]
9d3bff018b chore(deps): update openai requirement from ~=0.28.0 to ~=0.28.1
Updates the requirements on [openai](https://github.com/openai/openai-python) to permit the latest version.
- [Release notes](https://github.com/openai/openai-python/releases)
- [Commits](https://github.com/openai/openai-python/compare/v0.28.0...v0.28.1)

---
updated-dependencies:
- dependency-name: openai
  dependency-type: direct:production
...

Signed-off-by: dependabot[bot] <support@github.com>
2023-10-02 08:09:30 +00:00
RockChinQ
051376e0d2 Release v2.6.1 2023-09-28 12:18:46 +00:00
Junyan Qin
a113785211 Merge pull request #578 from RockChinQ/fix/blocked-audit-upload
[Fix] 阻塞地发送审计报告数据
2023-09-28 20:17:26 +08:00
RockChinQ
3f4ed4dc3c fix: 阻塞地发送审计报告数据 2023-09-28 12:16:30 +00:00
Junyan Qin
ac80764fae Merge pull request #577 from RockChinQ/doc/deadlinks-in-wiki
[Doc] 修复wiki中的死链
2023-09-28 20:04:01 +08:00
RockChinQ
e43afd4891 doc: 修复wiki中的死链 2023-09-28 12:03:27 +00:00
RockChinQ
f1aea1d495 doc: 统一改称指令为命令 2023-09-28 11:46:33 +00:00
GitHub Actions
0e2a5db104 Update override-all.json 2023-09-26 16:10:22 +00:00
Junyan Qin
3a4c9771fa feat(config): 默认超时时间改为两分钟 2023-09-27 00:09:58 +08:00
RockChinQ
f4f8ef9523 ci: 工作流统一双空格缩进 2023-09-13 08:27:47 +00:00
19 changed files with 197 additions and 179 deletions

View File

@@ -7,32 +7,32 @@ on:
types: [published]
jobs:
publish-docker-image:
runs-on: ubuntu-latest
name: Build image
runs-on: ubuntu-latest
name: Build image
steps:
- name: Checkout
uses: actions/checkout@v2
- name: judge has env GITHUB_REF # 如果没有GITHUB_REF环境变量则把github.ref变量赋值给GITHUB_REF
run: |
if [ -z "$GITHUB_REF" ]; then
export GITHUB_REF=${{ github.ref }}
fi
- name: Check GITHUB_REF env
run: echo $GITHUB_REF
- name: Get version
id: get_version
if: (startsWith(env.GITHUB_REF, 'refs/tags/')||startsWith(github.ref, 'refs/tags/')) && startsWith(github.repository, 'RockChinQ/QChatGPT')
run: echo ::set-output name=VERSION::${GITHUB_REF/refs\/tags\//}
- name: Build # image name: rockchin/qchatgpt:<VERSION>
run: docker build --network=host -t rockchin/qchatgpt:${{ steps.get_version.outputs.VERSION }} -t rockchin/qchatgpt:latest .
- name: Login to Registry
run: docker login --username=${{ secrets.DOCKER_USERNAME }} --password ${{ secrets.DOCKER_PASSWORD }}
steps:
- name: Checkout
uses: actions/checkout@v2
- name: judge has env GITHUB_REF # 如果没有GITHUB_REF环境变量则把github.ref变量赋值给GITHUB_REF
run: |
if [ -z "$GITHUB_REF" ]; then
export GITHUB_REF=${{ github.ref }}
fi
- name: Check GITHUB_REF env
run: echo $GITHUB_REF
- name: Get version
id: get_version
if: (startsWith(env.GITHUB_REF, 'refs/tags/')||startsWith(github.ref, 'refs/tags/')) && startsWith(github.repository, 'RockChinQ/QChatGPT')
run: echo ::set-output name=VERSION::${GITHUB_REF/refs\/tags\//}
- name: Build # image name: rockchin/qchatgpt:<VERSION>
run: docker build --network=host -t rockchin/qchatgpt:${{ steps.get_version.outputs.VERSION }} -t rockchin/qchatgpt:latest .
- name: Login to Registry
run: docker login --username=${{ secrets.DOCKER_USERNAME }} --password ${{ secrets.DOCKER_PASSWORD }}
- name: Push image
if: (startsWith(env.GITHUB_REF, 'refs/tags/')||startsWith(github.ref, 'refs/tags/')) && startsWith(github.repository, 'RockChinQ/QChatGPT')
run: docker push rockchin/qchatgpt:${{ steps.get_version.outputs.VERSION }}
- name: Push image
if: (startsWith(env.GITHUB_REF, 'refs/tags/')||startsWith(github.ref, 'refs/tags/')) && startsWith(github.repository, 'RockChinQ/QChatGPT')
run: docker push rockchin/qchatgpt:${{ steps.get_version.outputs.VERSION }}
- name: Push image
if: (startsWith(env.GITHUB_REF, 'refs/tags/')||startsWith(github.ref, 'refs/tags/')) && startsWith(github.repository, 'RockChinQ/QChatGPT')
run: docker push rockchin/qchatgpt:latest
- name: Push latest image
if: (startsWith(env.GITHUB_REF, 'refs/tags/')||startsWith(github.ref, 'refs/tags/')) && startsWith(github.repository, 'RockChinQ/QChatGPT')
run: docker push rockchin/qchatgpt:latest

View File

@@ -8,7 +8,7 @@
# QChatGPT
<!-- 高稳定性/持续迭代/架构清晰/支持插件/高可自定义的 ChatGPT QQ机器人框架 -->
“当然下面是一个使用Java编写的快速排序算法的示例代码”
<!-- “当然下面是一个使用Java编写的快速排序算法的示例代码” -->
[![GitHub release (latest by date)](https://img.shields.io/github/v/release/RockChinQ/QChatGPT)](https://github.com/RockChinQ/QChatGPT/releases/latest)
<a href="https://hub.docker.com/repository/docker/rockchin/qchatgpt">
@@ -25,7 +25,7 @@
<a href="http://qm.qq.com/cgi-bin/qm/qr?_wv=1027&k=nC80H57wmKPwRDLFeQrDDjVl81XuC21P&authKey=2wTUTfoQ5v%2BD4C5zfpuR%2BSPMDqdXgDXA%2FS2wHI1NxTfWIG%2B%2FqK08dgyjMMOzhXa9&noverify=0&group_code=738382634">
<img alt="Static Badge" src="https://img.shields.io/badge/%E7%A4%BE%E5%8C%BA%E7%BE%A4-738382634-purple">
</a>
<a href="https://lazyfree.top/2023/08/16/QChatGPT%E4%BD%BF%E7%94%A8%E6%89%8B%E5%86%8C/">
<a href="https://qchatgpt.rockchin.top">
<img alt="Static Badge" src="https://img.shields.io/badge/%E6%9F%A5%E7%9C%8B-%E7%A4%BE%E5%8C%BA%E7%BC%96%E5%86%99%E4%BD%BF%E7%94%A8%E6%89%8B%E5%86%8C-blue">
</a>
<a href="https://www.bilibili.com/video/BV14h4y1w7TC">
@@ -117,24 +117,24 @@
</details>
<details>
<summary>✅支持预设指令文字</summary>
<summary>✅支持预设文字</summary>
- 支持以自然语言预设文字,自定义机器人人格等信息
- 详见`config.py`中的`default_prompt`部分
- 支持设置多个预设情景,并通过!reset、!default等令控制,详细请查看[wiki](https://github.com/RockChinQ/QChatGPT/wiki/1-%E5%8A%9F%E8%83%BD%E4%BD%BF%E7%94%A8#%E6%9C%BA%E5%99%A8%E4%BA%BA%E6%8C%87%E4%BB%A4)
- 支持设置多个预设情景,并通过!reset、!default等令控制,详细请查看[wiki](https://github.com/RockChinQ/QChatGPT/wiki/1-%E5%8A%9F%E8%83%BD%E4%BD%BF%E7%94%A8#%E6%9C%BA%E5%99%A8%E4%BA%BA%E6%8C%87%E4%BB%A4)
</details>
<details>
<summary>✅支持对话、绘图等模型,可玩性更高</summary>
- 现已支持OpenAI的对话`Completion API`和绘图`Image API`
- 向机器人发送`!draw <prompt>`即可使用绘图模型
- 向机器人发送`!draw <prompt>`即可使用绘图模型
</details>
<details>
<summary>✅支持令控制热重载、热更新</summary>
<summary>✅支持令控制热重载、热更新</summary>
- 允许在运行期间修改`config.py`或其他代码后,以管理员账号向机器人发送`!reload`进行热重载,无需重启
- 运行期间允许以管理员账号向机器人发送`!update`进行热更新,拉取远程最新代码并执行热重载
- 允许在运行期间修改`config.py`或其他代码后,以管理员账号向机器人发送`!reload`进行热重载,无需重启
- 运行期间允许以管理员账号向机器人发送`!update`进行热更新,拉取远程最新代码并执行热重载
</details>
<details>
<summary>✅支持插件加载🧩</summary>
@@ -307,7 +307,7 @@ python3 main.py
</details>
**部署完成后必看: [令说明](https://github.com/RockChinQ/QChatGPT/wiki/1-%E5%8A%9F%E8%83%BD%E4%BD%BF%E7%94%A8#%E6%9C%BA%E5%99%A8%E4%BA%BA%E6%8C%87%E4%BB%A4)**
**部署完成后必看: [令说明](https://github.com/RockChinQ/QChatGPT/wiki/1-%E5%8A%9F%E8%83%BD%E4%BD%BF%E7%94%A8#%E6%9C%BA%E5%99%A8%E4%BA%BA%E6%8C%87%E4%BB%A4)**
所有功能查看[Wiki功能使用页](https://github.com/RockChinQ/QChatGPT/wiki/1-%E5%8A%9F%E8%83%BD%E4%BD%BF%E7%94%A8)
@@ -332,10 +332,8 @@ python3 main.py
- [WebwlkrPlugin](https://github.com/RockChinQ/WebwlkrPlugin) - 让机器人能联网!!
- [revLibs](https://github.com/RockChinQ/revLibs) - 将ChatGPT网页版、Claude、Bard、Hugging Chat等破解版接入此项目关于[官方接口和网页版有什么区别](https://github.com/RockChinQ/QChatGPT/wiki/8-%E5%AE%98%E6%96%B9%E6%8E%A5%E5%8F%A3%E3%80%81ChatGPT%E7%BD%91%E9%A1%B5%E7%89%88%E3%80%81ChatGPT-API%E5%8C%BA%E5%88%AB)
- [Switcher](https://github.com/RockChinQ/Switcher) - 支持通过令切换使用的模型
- [Switcher](https://github.com/RockChinQ/Switcher) - 支持通过令切换使用的模型
- [hello_plugin](https://github.com/RockChinQ/hello_plugin) - `hello_plugin` 的储存库形式,插件开发模板
- [dominoar/QChatPlugins](https://github.com/dominoar/QchatPlugins) - dominoar编写的诸多新功能插件语音输出、Ranimg、屏蔽词规则等
- [dominoar/QCP-NovelAi](https://github.com/dominoar/QCP-NovelAi) - NovelAI 故事叙述与绘画
- [oliverkirk-sudo/chat_voice](https://github.com/oliverkirk-sudo/chat_voice) - 文字转语音输出支持HuggingFace上的[VITS模型](https://huggingface.co/spaces/Plachta/VITS-Umamusume-voice-synthesizer),azure语音合成,vits本地语音合成,sovits语音合成
- [RockChinQ/WaitYiYan](https://github.com/RockChinQ/WaitYiYan) - 实时获取百度`文心一言`等待列表人数
- [chordfish-k/QChartGPT_Emoticon_Plugin](https://github.com/chordfish-k/QChartGPT_Emoticon_Plugin) - 使机器人根据回复内容发送表情包

View File

@@ -248,7 +248,7 @@ image_api_params = {
trace_function_calls = False
# 群内回复消息时是否引用原消息
quote_origin = True
quote_origin = False
# 群内回复消息时是否at发送者
at_sender = False
@@ -257,7 +257,7 @@ at_sender = False
include_image_description = True
# 消息处理的超时时间,单位为秒
process_message_timeout = 30
process_message_timeout = 120
# 回复消息时是否显示[GPT]前缀
show_prefix = False

View File

@@ -191,13 +191,16 @@ def start(first_time_init=False):
# 配置OpenAI proxy
import openai
openai.proxy = None # 先重置因为重载后可能需要清除proxy
openai.proxies = None # 先重置因为重载后可能需要清除proxy
if "http_proxy" in config.openai_config and config.openai_config["http_proxy"] is not None:
openai.proxy = config.openai_config["http_proxy"]
openai.proxies = {
"http": config.openai_config["http_proxy"],
"https": config.openai_config["http_proxy"]
}
# 配置openai api_base
if "reverse_proxy" in config.openai_config and config.openai_config["reverse_proxy"] is not None:
openai.api_base = config.openai_config["reverse_proxy"]
openai.base_url = config.openai_config["reverse_proxy"]
# 主启动流程
database = pkg.database.manager.DatabaseManager()

View File

@@ -63,10 +63,10 @@
"size": "256x256"
},
"trace_function_calls": false,
"quote_origin": true,
"quote_origin": false,
"at_sender": false,
"include_image_description": true,
"process_message_timeout": 30,
"process_message_timeout": 120,
"show_prefix": false,
"force_delay_range": [
1.5,

View File

@@ -5,6 +5,7 @@
import hashlib
import json
import logging
import threading
import requests
@@ -42,15 +43,20 @@ class DataGatherer:
只会报告此次请求的使用量,不会报告总量。
不包含除版本号、使用类型、使用量以外的任何信息,仅供开发者分析使用情况。
"""
try:
config = pkg.utils.context.get_config()
if not config.report_usage:
def thread_func():
try:
config = pkg.utils.context.get_config()
if not config.report_usage:
return
res = requests.get("http://reports.rockchin.top:18989/usage?service_name=qchatgpt.{}&version={}&count={}&msg_source={}".format(subservice_name, self.version_str, count, config.msg_source_adapter))
if res.status_code != 200 or res.text != "ok":
logging.warning("report to server failed, status_code: {}, text: {}".format(res.status_code, res.text))
except:
return
res = requests.get("http://reports.rockchin.top:18989/usage?service_name=qchatgpt.{}&version={}&count={}&msg_source={}".format(subservice_name, self.version_str, count, config.msg_source_adapter))
if res.status_code != 200 or res.text != "ok":
logging.warning("report to server failed, status_code: {}, text: {}".format(res.status_code, res.text))
except:
return
threading.Thread(target=thread_func).start()
def get_usage(self, key_md5):
return self.usage[key_md5] if key_md5 in self.usage else {}

View File

@@ -1,4 +1,5 @@
import openai
from openai.types.chat import chat_completion_message
import json
import logging
@@ -13,13 +14,14 @@ class ChatCompletionRequest(RequestBase):
此类保证每一次返回的角色为assistant的信息的finish_reason一定为stop。
若有函数调用响应,本类的返回瀑布是:函数调用请求->函数调用结果->...->assistant的信息->stop。
"""
model: str
messages: list[dict[str, str]]
kwargs: dict
stopped: bool = False
pending_func_call: dict = None
pending_func_call: chat_completion_message.FunctionCall = None
pending_msg: str
@@ -46,16 +48,18 @@ class ChatCompletionRequest(RequestBase):
def __init__(
self,
client: openai.Client,
model: str,
messages: list[dict[str, str]],
**kwargs
):
self.client = client
self.model = model
self.messages = messages.copy()
self.kwargs = kwargs
self.req_func = openai.ChatCompletion.acreate
self.req_func = self.client.chat.completions.create
self.pending_func_call = None
@@ -84,39 +88,48 @@ class ChatCompletionRequest(RequestBase):
# 拼接kwargs
args = {**args, **self.kwargs}
from openai.types.chat import chat_completion
resp = self._req(**args)
resp: chat_completion.ChatCompletion = self._req(**args)
choice0 = resp["choices"][0]
choice0 = resp.choices[0]
# 如果不是函数调用且finish_reason为stop则停止迭代
if choice0['finish_reason'] == 'stop': # and choice0["finish_reason"] == "stop"
if choice0.finish_reason == 'stop': # and choice0["finish_reason"] == "stop"
self.stopped = True
if 'function_call' in choice0['message']:
self.pending_func_call = choice0['message']['function_call']
if hasattr(choice0.message, 'function_call') and choice0.message.function_call is not None:
self.pending_func_call = choice0.message.function_call
self.append_message(
role="assistant",
content=choice0['message']['content'],
function_call=choice0['message']['function_call']
content=choice0.message.content,
function_call=choice0.message.function_call
)
return {
"id": resp["id"],
"id": resp.id,
"choices": [
{
"index": choice0["index"],
"index": choice0.index,
"message": {
"role": "assistant",
"type": "function_call",
"content": choice0['message']['content'],
"function_call": choice0['message']['function_call']
"content": choice0.message.content,
"function_call": {
"name": choice0.message.function_call.name,
"arguments": choice0.message.function_call.arguments
}
},
"finish_reason": "function_call"
}
],
"usage": resp["usage"]
"usage": {
"prompt_tokens": resp.usage.prompt_tokens,
"completion_tokens": resp.usage.completion_tokens,
"total_tokens": resp.usage.total_tokens
}
}
else:
@@ -124,19 +137,23 @@ class ChatCompletionRequest(RequestBase):
# 普通回复一定处于最后方故不用再追加进内部messages
return {
"id": resp["id"],
"id": resp.id,
"choices": [
{
"index": choice0["index"],
"index": choice0.index,
"message": {
"role": "assistant",
"type": "text",
"content": choice0['message']['content']
"content": choice0.message.content
},
"finish_reason": choice0["finish_reason"]
"finish_reason": choice0.finish_reason
}
],
"usage": resp["usage"]
"usage": {
"prompt_tokens": resp.usage.prompt_tokens,
"completion_tokens": resp.usage.completion_tokens,
"total_tokens": resp.usage.total_tokens
}
}
else: # 处理函数调用请求
@@ -144,20 +161,20 @@ class ChatCompletionRequest(RequestBase):
self.pending_func_call = None
func_name = cp_pending_func_call['name']
func_name = cp_pending_func_call.name
arguments = {}
try:
try:
arguments = json.loads(cp_pending_func_call['arguments'])
arguments = json.loads(cp_pending_func_call.arguments)
# 若不是json格式的异常处理
except json.decoder.JSONDecodeError:
# 获取函数的参数列表
func_schema = get_func_schema(func_name)
arguments = {
func_schema['parameters']['required'][0]: cp_pending_func_call['arguments']
func_schema['parameters']['required'][0]: cp_pending_func_call.arguments
}
logging.info("执行函数调用: name={}, arguments={}".format(func_name, arguments))

View File

@@ -1,4 +1,5 @@
import openai
from openai.types import completion, completion_choice
from .model import RequestBase
@@ -17,10 +18,12 @@ class CompletionRequest(RequestBase):
def __init__(
self,
client: openai.Client,
model: str,
messages: list[dict[str, str]],
**kwargs
):
self.client = client
self.model = model
self.prompt = ""
@@ -31,7 +34,7 @@ class CompletionRequest(RequestBase):
self.kwargs = kwargs
self.req_func = openai.Completion.acreate
self.req_func = self.client.completions.create
def __iter__(self):
return self
@@ -63,49 +66,35 @@ class CompletionRequest(RequestBase):
if self.stopped:
raise StopIteration()
resp = self._req(
resp: completion.Completion = self._req(
model=self.model,
prompt=self.prompt,
**self.kwargs
)
if resp["choices"][0]["finish_reason"] == "stop":
if resp.choices[0].finish_reason == "stop":
self.stopped = True
choice0 = resp["choices"][0]
choice0: completion_choice.CompletionChoice = resp.choices[0]
self.prompt += choice0["text"]
self.prompt += choice0.text
return {
"id": resp["id"],
"id": resp.id,
"choices": [
{
"index": choice0["index"],
"index": choice0.index,
"message": {
"role": "assistant",
"type": "text",
"content": choice0["text"]
"content": choice0.text
},
"finish_reason": choice0["finish_reason"]
"finish_reason": choice0.finish_reason
}
],
"usage": resp["usage"]
}
if __name__ == "__main__":
import os
openai.api_key = os.environ["OPENAI_API_KEY"]
for resp in CompletionRequest(
model="text-davinci-003",
messages=[
{
"role": "user",
"content": "Hello, who are you?"
"usage": {
"prompt_tokens": resp.usage.prompt_tokens,
"completion_tokens": resp.usage.completion_tokens,
"total_tokens": resp.usage.total_tokens
}
]
):
print(resp)
if resp["choices"][0]["finish_reason"] == "stop":
break
}

View File

@@ -8,6 +8,8 @@ import openai
class RequestBase:
client: openai.Client
req_func: callable
def __init__(self, *args, **kwargs):
@@ -17,41 +19,17 @@ class RequestBase:
import pkg.utils.context as context
switched, name = context.get_openai_manager().key_mgr.auto_switch()
logging.debug("切换api-key: switched={}, name={}".format(switched, name))
openai.api_key = context.get_openai_manager().key_mgr.get_using_key()
self.client.api_key = context.get_openai_manager().key_mgr.get_using_key()
def _req(self, **kwargs):
"""处理代理问题"""
import config
ret: dict = {}
exception: Exception = None
ret = self.req_func(**kwargs)
logging.debug("接口请求返回:%s", str(ret))
async def awrapper(**kwargs):
nonlocal ret, exception
try:
ret = await self.req_func(**kwargs)
logging.debug("接口请求返回:%s", str(ret))
if config.switch_strategy == 'active':
self._next_key()
return ret
except Exception as e:
exception = e
loop = asyncio.new_event_loop()
thr = threading.Thread(
target=loop.run_until_complete,
args=(awrapper(**kwargs),)
)
thr.start()
thr.join()
if exception is not None:
raise exception
if config.switch_strategy == 'active':
self._next_key()
return ret

View File

@@ -24,6 +24,8 @@ class OpenAIInteract:
"size": "256x256",
}
client: openai.Client = None
def __init__(self, api_key: str):
self.key_mgr = pkg.openai.keymgr.KeysManager(api_key)
@@ -31,7 +33,9 @@ class OpenAIInteract:
# logging.info("文字总使用量:%d", self.audit_mgr.get_total_text_length())
openai.api_key = self.key_mgr.get_using_key()
self.client = openai.Client(
api_key=self.key_mgr.get_using_key()
)
pkg.utils.context.set_openai_manager(self)
@@ -48,7 +52,7 @@ class OpenAIInteract:
cp_parmas = config.completion_api_params.copy()
del cp_parmas['model']
request = select_request_cls(model, messages, cp_parmas)
request = select_request_cls(self.client, model, messages, cp_parmas)
# 请求接口
for resp in request:

View File

@@ -5,9 +5,8 @@ ChatCompletion - gpt-3.5-turbo 等模型
Completion - text-davinci-003 等模型
此模块封装此两个接口的请求实现,为上层提供统一的调用方式
"""
import openai, logging, threading, asyncio
import openai.error as aiE
import tiktoken
import openai
from pkg.openai.api.model import RequestBase
from pkg.openai.api.completion import CompletionRequest
@@ -53,11 +52,11 @@ IMAGE_MODELS = {
}
def select_request_cls(model_name: str, messages: list, args: dict) -> RequestBase:
def select_request_cls(client: openai.Client, model_name: str, messages: list, args: dict) -> RequestBase:
if model_name in CHAT_COMPLETION_MODELS:
return ChatCompletionRequest(model_name, messages, **args)
return ChatCompletionRequest(client, model_name, messages, **args)
elif model_name in COMPLETION_MODELS:
return CompletionRequest(model_name, messages, **args)
return CompletionRequest(client, model_name, messages, **args)
raise ValueError("不支持模型[{}],请检查配置文件".format(model_name))

View File

@@ -278,7 +278,7 @@ class Session:
if resp['choices'][0]['message']['role'] == "assistant" and resp['choices'][0]['message']['content'] != None: # 包含纯文本响应
if not trace_func_calls:
res_text += resp['choices'][0]['message']['content'] + "\n"
res_text += resp['choices'][0]['message']['content']
else:
res_text = resp['choices'][0]['message']['content']
pending_res_text = resp['choices'][0]['message']['content']

View File

@@ -65,14 +65,14 @@ def process_normal_message(text_message: str, mgr, config, launcher_type: str,
if not event.is_prevented_default():
reply = [prefix + text]
except openai.error.APIConnectionError as e:
except openai.APIConnectionError as e:
err_msg = str(e)
if err_msg.__contains__('Error communicating with OpenAI'):
reply = handle_exception("{}会话调用API失败:{}\n您的网络无法访问OpenAI接口或网络代理不正常".format(session_name, e),
"[bot]err:调用API失败请重试或联系管理员或等待修复")
else:
reply = handle_exception("{}会话调用API失败:{}".format(session_name, e), "[bot]err:调用API失败请重试或联系管理员或等待修复")
except openai.error.RateLimitError as e:
except openai.RateLimitError as e:
logging.debug(type(e))
logging.debug(e.error['message'])
@@ -116,14 +116,14 @@ def process_normal_message(text_message: str, mgr, config, launcher_type: str,
else:
reply = handle_exception("{}会话调用API失败:{}".format(session_name, e),
"[bot]err:RateLimitError,请重试或联系作者,或等待修复")
except openai.error.InvalidRequestError as e:
except openai.BadRequestError as e:
if config.auto_reset and "This model's maximum context length is" in str(e):
session.reset(persist=True)
reply = [tips_custom.session_auto_reset_message]
else:
reply = handle_exception("{}API调用参数错误:{}\n".format(
session_name, e), "[bot]err:API调用参数错误请联系管理员或等待修复")
except openai.error.ServiceUnavailableError as e:
except openai.APIStatusError as e:
reply = handle_exception("{}API调用服务不可用:{}".format(session_name, e), "[bot]err:API调用服务不可用请重试或联系管理员或等待修复")
except Exception as e:
logging.exception(e)

File diff suppressed because one or more lines are too long

View File

@@ -1,5 +1,5 @@
requests
openai~=0.28.0
openai
dulwich~=0.21.6
colorlog~=6.6.0
yiri-mirai
@@ -10,4 +10,4 @@ Pillow
nakuru-project-idk
CallingGPT
tiktoken
PyYaml
PyYaml

View File

@@ -48,12 +48,12 @@
</details>
<details>
<summary>✅支持预设指令文字</summary>
<summary>✅支持预设文字</summary>
- 支持以自然语言预设文字,自定义机器人人格等信息
- 详见`config.py`中的`default_prompt`部分
- 支持设置多个预设情景,并通过!reset、!default等令控制,详细请查看[wiki](https://github.com/RockChinQ/QChatGPT/wiki/%E5%8A%9F%E8%83%BD%E4%BD%BF%E7%94%A8#%E6%9C%BA%E5%99%A8%E4%BA%BA%E6%8C%87%E4%BB%A4)
- 支持使用文件存储情景预设文字,并加载: 在`prompts/`目录新建文件写入预设文字,即可通过`!reset <文件名>`令加载
- 支持设置多个预设情景,并通过!reset、!default等令控制,详细请查看[wiki](https://github.com/RockChinQ/QChatGPT/wiki/1-%E5%8A%9F%E8%83%BD%E4%BD%BF%E7%94%A8#%E6%9C%BA%E5%99%A8%E4%BA%BA%E5%91%BD%E4%BB%A4)
- 支持使用文件存储情景预设文字,并加载: 在`prompts/`目录新建文件写入预设文字,即可通过`!reset <文件名>`令加载
</details>
<details>
@@ -61,25 +61,25 @@
- 使用SQLite进行会话内容持久化
- 最后一次对话一定时间后自动保存,请到`config.py`中修改`session_expire_time`的值以自定义时间
- 运行期间可使用`!reset` `!list` `!last` `!next` `!prompt`令管理会话
- 运行期间可使用`!reset` `!list` `!last` `!next` `!prompt`令管理会话
</details>
<details>
<summary>✅支持对话、绘图等模型,可玩性更高</summary>
- 现已支持OpenAI的对话`Completion API`和绘图`Image API`
- 向机器人发送`!draw <prompt>`即可使用绘图模型
- 向机器人发送`!draw <prompt>`即可使用绘图模型
</details>
<details>
<summary>✅支持令控制热重载、热更新</summary>
<summary>✅支持令控制热重载、热更新</summary>
- 允许在运行期间修改`config.py`或其他代码后,以管理员账号向机器人发送`!reload`进行热重载,无需重启
- 运行期间允许以管理员账号向机器人发送`!update`进行热更新,拉取远程最新代码并执行热重载
- 允许在运行期间修改`config.py`或其他代码后,以管理员账号向机器人发送`!reload`进行热重载,无需重启
- 运行期间允许以管理员账号向机器人发送`!update`进行热更新,拉取远程最新代码并执行热重载
</details>
<details>
<summary>✅支持插件加载🧩</summary>
- 自行实现插件加载器及相关支持
- 详细查看[插件使用页](https://github.com/RockChinQ/QChatGPT/wiki/%E6%8F%92%E4%BB%B6%E4%BD%BF%E7%94%A8)
- 详细查看[插件使用页](https://github.com/RockChinQ/QChatGPT/wiki/5-%E6%8F%92%E4%BB%B6%E4%BD%BF%E7%94%A8)
</details>
<details>
<summary>✅私聊、群聊黑名单机制</summary>
@@ -153,14 +153,14 @@
<img alt="绘图功能" src="https://github.com/RockChinQ/QChatGPT/blob/master/res/屏幕截图%202022-12-29%20194948.png" width="550" height="348"/>
### 机器人
### 机器人
目前支持的
目前支持的
> `<>` 中的为必填参数,使用时请不要包含`<>`
> `[]` 中的为可选参数,使用时请不要包含`[]`
#### 用户级别
#### 用户级别
> 可以使用`!help`命令来查看命令说明
@@ -174,7 +174,7 @@
!del all 删除本会话对象的所有历史记录
!last 切换到前一次会话
!next 切换到后一次会话
!reset [使用预设] 重置对象的当前会话,可指定使用的情景预设值(通过!default令查看可用的)
!reset [使用预设] 重置对象的当前会话,可指定使用的情景预设值(通过!default令查看可用的)
!prompt 查看对象当前会话的所有记录
!usage 查看api-key的使用量
!draw <提示语> 进行绘图
@@ -184,7 +184,7 @@
!default 查看可用的情景预设值
```
#### 管理员
#### 管理员
仅管理员私聊机器人时可使用,必须先在`config.py`中的`admin_qq`设置管理员QQ
@@ -197,9 +197,9 @@
!delhst all 删除所有会话的所有历史记录
```
<details>
<summary>⚙ !cfg 令及其简化形式详解</summary>
<summary>⚙ !cfg 令及其简化形式详解</summary>
令可以在运行期间由管理员通过QQ私聊窗口修改配置信息**重启之后会失效**。
令可以在运行期间由管理员通过QQ私聊窗口修改配置信息**重启之后会失效**。
用法:
1. 查看所有配置项及其值
@@ -239,7 +239,7 @@
格式:`!~<配置项名称>`
其中`!~`等价于`!cfg `
则前述三个令分别可以简化为:
则前述三个令分别可以简化为:
```
!~all
!~default_prompt
@@ -290,11 +290,11 @@ sensitive_word_filter = True
### 预设文字(default模式)
编辑`config.py`中的`default_prompt`字段,预设文字不宜过长(建议1000字以内),目前所有会话都会射到预设文字的影响。
或将情景预设文字写入到`prompts/`目录下,运行期间即可使用`!reset <文件名>`令加载,或使用`!default <文件名>`令将其设为默认
或将情景预设文字写入到`prompts/`目录下,运行期间即可使用`!reset <文件名>`令加载,或使用`!default <文件名>`令将其设为默认
### 预设文字(full_scenario模式)
将JSON情景写入到`scenario/`目录下,运行期间即可使用`!reset <文件名>`令加载,或使用`!default <文件名>`令将其设为默认.
将JSON情景写入到`scenario/`目录下,运行期间即可使用`!reset <文件名>`令加载,或使用`!default <文件名>`令将其设为默认.
JSON情景模板参考`scenario/default_template.json`
@@ -367,7 +367,7 @@ prompt_submit_length = <模型单次请求token数上限> - 情景预设中token
在运行期间使用管理员QQ账号私聊机器人发送`!reload`加载修改后的`config.py`的值或编辑后的代码,无需重启
使用管理员账号私聊机器人,发送`!update`拉取最新代码并进行热更新,无需重启
详见前述`管理员令`段落
详见前述`管理员令`段落
### 群内无需@响应规则

View File

@@ -4,7 +4,7 @@
#### 自动更新
由管理员QQ私聊机器人QQ发送`!update`
由管理员QQ私聊机器人QQ发送`!update`
#### 手动更新

View File

@@ -1,7 +1,7 @@
QChatGPT 插件开发Wiki
> 请先阅读[插件使用页](https://github.com/RockChinQ/QChatGPT/wiki/%E6%8F%92%E4%BB%B6%E4%BD%BF%E7%94%A8)
> 请先阅读[技术信息页](https://github.com/RockChinQ/QChatGPT/wiki/%E6%8A%80%E6%9C%AF%E4%BF%A1%E6%81%AF)
> 请先阅读[插件使用页](https://github.com/RockChinQ/QChatGPT/wiki/5-%E6%8F%92%E4%BB%B6%E4%BD%BF%E7%94%A8)
> 请先阅读[技术信息页](https://github.com/RockChinQ/QChatGPT/wiki/4-%E6%8A%80%E6%9C%AF%E4%BF%A1%E6%81%AF)
> 建议先阅读本项目源码,了解项目架构
> 问题、需求请到仓库issue发起
@@ -354,18 +354,18 @@ PersonNormalMessageReceived = "person_normal_message_received"
"""
PersonCommandSent = "person_command_sent"
"""判断为应该处理的私聊令时触发
"""判断为应该处理的私聊令时触发
kwargs:
launcher_type: str 发起对象类型(group/person)
launcher_id: int 发起对象ID(群号/QQ号)
sender_id: int 发送者ID(QQ号)
command: str
command: str
params: list[str] 参数列表
text_message: str 完整令文本
text_message: str 完整令文本
is_admin: bool 是否为管理员
returns (optional):
alter: str 修改后的完整令文本
alter: str 修改后的完整令文本
reply: list 回复消息组件列表元素为YiriMirai支持的消息组件
"""
@@ -383,18 +383,18 @@ GroupNormalMessageReceived = "group_normal_message_received"
"""
GroupCommandSent = "group_command_sent"
"""判断为应该处理的群聊令时触发
"""判断为应该处理的群聊令时触发
kwargs:
launcher_type: str 发起对象类型(group/person)
launcher_id: int 发起对象ID(群号/QQ号)
sender_id: int 发送者ID(QQ号)
command: str
command: str
params: list[str] 参数列表
text_message: str 完整令文本
text_message: str 完整令文本
is_admin: bool 是否为管理员
returns (optional):
alter: str 修改后的完整令文本
alter: str 修改后的完整令文本
reply: list 回复消息组件列表元素为YiriMirai支持的消息组件
"""

View File

@@ -0,0 +1,24 @@
import os
import openai
client = openai.Client(
api_key=os.environ["OPENAI_API_KEY"],
)
openai.proxies = {
'http': 'http://127.0.0.1:7890',
'https': 'http://127.0.0.1:7890',
}
resp = client.chat.completions.create(
model="gpt-3.5-turbo",
messages=[
{
"role": "user",
"content": "Hello, how are you?",
}
]
)
print(resp)