MeUtils 2024.3.4.13.4.45__py3-none-any.whl → 2025.1.16.17.15.52__py3-none-any.whl

meutils/{fileparser/filetype.py → llm/completions/oi.py}

@@ -1,8 +1,8 @@
 #!/usr/bin/env python
 # -*- coding: utf-8 -*-
 # @Project      : AI.  @by PyCharm
-# @File         : filetype
-# @Time         : 2023/7/15 15:05
+# @File         : oi
+# @Time         : 2024/12/12 13:19
 # @Author       : betterme
 # @WeChat       : meutils
 # @Software     : PyCharm
@@ -11,4 +11,6 @@
 from meutils.pipe import *
 
 
-import filetype
+from openai import OpenAI
+
+

meutils/llm/completions/rag/__init__.py

@@ -0,0 +1,11 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# @Project      : AI.  @by PyCharm
+# @File         : __init__.py
+# @Time         : 2024/11/21 17:13
+# @Author       : betterme
+# @WeChat       : meutils
+# @Software     : PyCharm
+# @Description  : 
+
+from meutils.pipe import *

meutils/llm/completions/rag/fire.py

@@ -0,0 +1,157 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# @Project      : AI.  @by PyCharm
+# @File         : moonshot
+# @Time         : 2024/11/21 17:13
+# @Author       : betterme
+# @WeChat       : meutils
+# @Software     : PyCharm
+# @Description  : 文件url 或者 文件id
+import mimetypes
+
+from meutils.pipe import *
+from meutils.io.files_utils import to_bytes
+from meutils.llm.openai_utils import to_openai_params
+from meutils.str_utils.regular_expression import parse_url
+
+from openai import OpenAI, AsyncOpenAI
+from meutils.schemas.openai_types import chat_completion, chat_completion_chunk, ChatCompletionRequest, CompletionUsage
+
+moonshot_client = AsyncOpenAI(
+    api_key=os.getenv("MOONSHOT_API_KEY"),
+    base_url=os.getenv("MOONSHOT_BASE_URL")
+)
+
+
+# moonshot_client.files.
+
+# len(moonshot_client.files.list().data)
+
+class Completions(object):
+
+    def __init__(self, api_key: Optional[str] = None):
+        self.api_key = api_key
+
+    @staticmethod
+    @alru_cache(ttl=15 * 60)
+    async def file_extract(file):  # "https://oss.ffire.cc/files/招标文件备案表（第二次）.pdf"
+        """todo 定时删除文件"""
+        filename = Path(file).name
+        mime_type, _ = mimetypes.guess_type(filename)  # mime_type = "application/octet-stream"
+        file: bytes = await to_bytes(file)
+
+        file_object = await moonshot_client.files.create(
+            # file=file,
+            # file=("filename.pdf", file),
+            file=(filename, file, mime_type),
+            purpose="file-extract"
+        )
+        logger.debug(file_object)
+
+        response = await moonshot_client.files.content(file_id=file_object.id)
+        return response.text
+
+    async def create(self, request: ChatCompletionRequest):
+        """[{'role': 'user', 'content': 'hi'}]
+        {"type": "file_url", "file_url": {"url": "https://oss.ffire.cc/files/招标文件备案表（第二次）.pdf", "detai": "auto"}}"""
+
+        # logger.debug(request.urls)
+        if request.urls:
+            files = await asyncio.gather(*map(self.file_extract, request.urls))
+
+            for file_content in files:
+                request.system_messages.append({
+                    "role": "system",
+                    "content": file_content,
+                })
+
+            request.messages = request.system_messages + [{'role': 'user', 'content': request.last_content}]
+
+        elif isinstance(request.last_content, str) and (urls := parse_url(request.last_content)):
+            logger.debug(urls)
+
+            url = urls[-1]
+            request.last_content = request.last_content.replace(url, '')
+
+            files = await asyncio.gather(*map(self.file_extract, urls))
+
+            for file_content in files:
+                request.system_messages.append({
+                    "role": "system",
+                    "content": file_content,
+                })
+
+            request.messages = request.system_messages + [{'role': 'user', 'content': request.last_content}]
+
+        logger.debug(request)
+
+        data = to_openai_params(request)
+        return await AsyncOpenAI(api_key=self.api_key).chat.completions.create(**data)
+
+
+# data: {"event": "message", "task_id": "900bbd43-dc0b-4383-a372-aa6e6c414227", "id": "663c5084-a254-4040-8ad3-51f2a3c1a77c", "answer": "Hi", "created_at": 1705398420}\n\n
+if __name__ == '__main__':
+    c = Completions()
+
+    request = ChatCompletionRequest(
+        # model="qwen-turbo-2024-11-01",
+        # model="claude-3-5-sonnet-20241022",
+        model="gpt-4o-mini",
+
+        messages=[
+            {
+                'role': 'system',
+                'content': '你是一个文件问答助手'
+            },
+            # {
+            #     'role': 'user',
+            #     # 'content': {
+            #     #     "type": "file_url",
+            #     #     "file_url": {"url": "https://oss.ffire.cc/files/招标文件备案表（第二次）.pdf", "detai": "auto"}
+            #     # },
+            #     'content': [
+            #         {
+            #             "type": "text",
+            #             "text": "这个文件讲了什么？"
+            #         },
+            #         # 多轮的时候要剔除
+            #         {
+            #             "type": "file_url",
+            #             "file_url": {"url": "https://oss.ffire.cc/files/招标文件备案表（第二次）.pdf", "detai": "auto"}
+            #         }
+            #     ]
+            # },
+
+            {
+                'role': 'user',
+                # 'content': {
+                #     "type": "file_url",
+                #     "file_url": {"url": "https://oss.ffire.cc/files/招标文件备案表（第二次）.pdf", "detai": "auto"}
+                # },
+                'content': "https://oss.ffire.cc/files/%E6%8B%9B%E6%A0%87%E6%96%87%E4%BB%B6%E5%A4%87%E6%A1%88%E8%A1%A8%EF%BC%88%E7%AC%AC%E4%BA%8C%E6%AC%A1%EF%BC%89.pdf 这个文件讲了什么？",
+
+            },
+
+            # {'role': 'assistant', 'content': "好的"},
+            # {
+            #     'role': 'user',
+            #     # 'content': {
+            #     #     "type": "file_url",
+            #     #     "file_url": {"url": "https://oss.ffire.cc/files/招标文件备案表（第二次）.pdf", "detai": "auto"}
+            #     # },
+            #     'content': [
+            #         {
+            #             "type": "text",
+            #             "text": "错了 继续回答"
+            #         },
+            #         # {
+            #         #     "type": "file_url",
+            #         #     "file_url": {"url": "https://oss.ffire.cc/files/招标文件备案表（第二次）.pdf", "detai": "auto"}
+            #         # }
+            #     ]
+            # }
+        ]
+
+    )
+
+    arun(c.create(request))

meutils/llm/completions/rag/qwen.py

@@ -0,0 +1,11 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# @Project      : AI.  @by PyCharm
+# @File         : qwen
+# @Time         : 2024/11/21 17:13
+# @Author       : betterme
+# @WeChat       : meutils
+# @Software     : PyCharm
+# @Description  : 
+
+from meutils.pipe import *

meutils/llm/completions/rag/rag.py

@@ -0,0 +1,41 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# @Project      : AI.  @by PyCharm
+# @File         : rag
+# @Time         : 2024/11/21 16:58
+# @Author       : betterme
+# @WeChat       : meutils
+# @Software     : PyCharm
+# @Description  : 
+
+from meutils.pipe import *
+import os
+from pathlib import Path
+from openai import OpenAI
+
+client = OpenAI(
+    api_key=os.getenv("DASHSCOPE_API_KEY"),
+    base_url="https://dashscope.aliyuncs.com/compatible-mode/v1",
+)
+
+file_object = client.files.create(file=Path("百炼系列手机产品介绍.docx"), purpose="file-extract")
+print(file_object.id) # file-fe-fUEnuSjIEbuMnKWpWSPKWGVe
+
+# print(client.files.retrieve_content("file-fe-fUEnuSjIEbuMnKWpWSPKWGVe"))
+#
+# completion = client.chat.completions.create(
+#     model="qwen-turbo-2024-11-01",
+#     messages=[
+#         {'role': 'system', 'content': 'You are a helpful assistant.'},
+#         {'role': 'system', 'content': f'fileid://{file_object.id}'},
+#         {'role': 'user', 'content': '这篇文章讲了什么？'}
+#     ],
+#     stream=True,
+#     stream_options={"include_usage": True}
+# )
+#
+# for chunk in completion:
+#     print(chunk.model_dump())
+
+
+print(client.files.content(file_id=file_object.id).text)

meutils/llm/completions/rag.py

@@ -0,0 +1,38 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# @Project      : AI.  @by PyCharm
+# @File         : rag
+# @Time         : 2024/11/21 16:58
+# @Author       : betterme
+# @WeChat       : meutils
+# @Software     : PyCharm
+# @Description  : 
+
+from meutils.pipe import *
+import os
+from pathlib import Path
+from openai import OpenAI
+
+client = OpenAI(
+    api_key=os.getenv("DASHSCOPE_API_KEY"),
+    base_url="https://dashscope.aliyuncs.com/compatible-mode/v1",
+)
+
+file_object = client.files.create(file=Path("百炼系列手机产品介绍.docx"), purpose="file-extract")
+print(file_object.id) # file-fe-fUEnuSjIEbuMnKWpWSPKWGVe
+
+# print(client.files.retrieve_content("file-fe-fUEnuSjIEbuMnKWpWSPKWGVe"))
+
+completion = client.chat.completions.create(
+    model="qwen-long",
+    messages=[
+        {'role': 'system', 'content': 'You are a helpful assistant.'},
+        {'role': 'system', 'content': f'fileid://{file_object.id}'},
+        {'role': 'user', 'content': '这篇文章讲了什么？'}
+    ],
+    stream=True,
+    stream_options={"include_usage": True}
+)
+
+for chunk in completion:
+    print(chunk.model_dump())

meutils/llm/completions/tryblend.py

@@ -0,0 +1,201 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# @Project      : AI.  @by PyCharm
+# @File         : tryblend
+# @Time         : 2024/9/4 09:22
+# @Author       : betterme
+# @WeChat       : meutils
+# @Software     : PyCharm
+# @Description  :
+
+
+from meutils.pipe import *
+from meutils.decorators.retry import retrying
+
+from meutils.notice.feishu import send_message as _send_message
+from meutils.db.redis_db import redis_client, redis_aclient
+from meutils.config_utils.lark_utils import aget_spreadsheet_values, get_next_token_for_polling
+from meutils.schemas.tryblend_types import BASE_URL, FEISHU_URL, FEISHU_URL_VIP
+from meutils.schemas.tryblend_types import GPT_4O_MINI, GPT_4O, CHECK_PAYLOAD
+from meutils.schemas.tryblend_types import CLAUDE_3_HAIKU, CLAUDE_3_SONNET, CLAUDE_3_OPUS, CLAUDE_35_SONNET
+from meutils.schemas.tryblend_types import PERPLEXITY_SONAR_SMALL, PERPLEXITY_SONAR_LARGE, PERPLEXITY_SONAR_HUGE
+
+from meutils.llm.openai_utils import to_openai_completion_params, token_encoder, token_encoder_with_cache
+from meutils.schemas.openai_types import chat_completion, chat_completion_chunk, ChatCompletionRequest, CompletionUsage
+from openai import OpenAI, AsyncOpenAI, APIStatusError
+
+# from meutils.llm.openai_utils import create_chat_completion_chunk, create_chat_completion
+
+send_message = partial(
+    _send_message,
+    url="https://open.feishu.cn/open-apis/bot/v2/hook/e0db85db-0daf-4250-9131-a98d19b909a9",
+    title=__name__
+)
+
+
+class Completions(object):
+
+    def __init__(self, api_key: Optional[str] = None, vip: bool = False):
+        self.api_key = api_key
+        self.vip = vip
+
+    async def create(self, request: ChatCompletionRequest):
+        if request.stream:
+            return self.stream(request)
+        else:
+            content = ""
+            completion_tokens = 0
+            prompt_tokens = len(str(request.messages))
+            async for chunk in self.stream(request):
+                content += chunk
+                completion_tokens += 1
+            # logger.debug(content)
+            chat_completion.choices[0].message.content = content
+            chat_completion.usage = CompletionUsage(
+                prompt_tokens=prompt_tokens,
+                completion_tokens=completion_tokens,
+                total_tokens=prompt_tokens + completion_tokens
+            )
+            return chat_completion
+
+    async def stream(self, request: ChatCompletionRequest):
+        token = self.api_key or await get_next_token_for_polling(
+            feishu_url=FEISHU_URL_VIP if self.vip else FEISHU_URL,
+            check_token=check_token
+        )
+
+        if request.model.startswith("gpt-4o-mini"):
+            payload = GPT_4O_MINI
+        elif request.model.startswith("gpt-4"):
+            payload = GPT_4O
+
+        elif request.model.startswith("claude-3-5-sonnet"):
+            payload = CLAUDE_35_SONNET
+        elif request.model.__contains__("haiku"):
+            logger.debug(request.model)
+            payload = CLAUDE_3_HAIKU
+        elif request.model.__contains__("sonnet"):
+            payload = CLAUDE_3_SONNET
+        elif request.model.__contains__("opus"):
+            payload = CLAUDE_3_OPUS
+
+        elif request.model.startswith(("perplexity-sonar-huge",)):
+            payload = PERPLEXITY_SONAR_HUGE
+        elif request.model.startswith(("perplexity-sonar-large", "net-gpt-4",)):
+            payload = PERPLEXITY_SONAR_LARGE
+        elif request.model.startswith(("perplexity-sonar-small", "perplexity", "net", "meta",)):
+            payload = PERPLEXITY_SONAR_SMALL
+
+        else:
+            payload = GPT_4O_MINI
+
+        payload[0]['messages'] = request.messages
+
+        headers = {
+            'Cookie': token,
+            'next-action': 'ca5ce500ddee37bddc3c986bee81b599f41e3efb',
+        }
+
+        async with httpx.AsyncClient(base_url=BASE_URL, headers=headers, timeout=300) as client:
+            async with client.stream(method="POST", url="", json=payload) as response:
+                logger.debug(response.status_code)
+
+                yield "\n"  # 提升首字速度
+
+                async for chunk in response.aiter_lines():
+                    # logger.debug(response.status_code)
+                    logger.debug(chunk)  # {"error":{"message":"User not found","status":500}}
+
+                    if 'Please add more credits to use this model' in chunk:
+                        # send_message("tryblend 余额不足")
+                        # logger.debug(chunk)
+                        break
+
+                    if chunk and (chunk := chunk.split(":", maxsplit=1)[-1]).strip() and chunk.startswith("{"):
+
+                        try:
+                            chunk = json.loads(chunk)
+                            chunk = chunk.get('diff', [""])[-1] or chunk.get('curr', "")
+                            yield chunk
+                            # break
+
+                        except Exception as e:
+                            _ = f"{e}\n{chunk}"
+                            logger.error(_)
+                            send_message(_)
+                            yield ""
+                    else:
+                        yield ""
+
+
+# 使用断点续传：
+# 实现一个机制来记录已接收的数据量，在连接中断时从断点处继续：
+# async def resumable_stream(client, url, payload, start_byte=0):
+#     headers = {"Range": f"bytes={start_byte}-"}
+#     while True:
+#         try:
+#             async with client.stream("POST", url, json=payload, headers=headers) as response:
+#                 async for chunk in response.aiter_bytes():
+#                     yield chunk
+#                     start_byte += len(chunk)
+#         except httpx.ReadTimeout:
+#             print(f"Connection lost, resuming from byte {start_byte}")
+#             continue
+#         break
+
+@alru_cache()
+@retrying(title=__name__)
+async def check_token(token):  ##########################
+    headers = {
+        'Cookie': token,
+        'next-action': '9447c867550adf321952367efe04d997e4c7d9a5',
+        # 'next-action': 'ca5ce500ddee37bddc3c986bee81b599f41e3efb'
+    }
+    # 0:["$@1",["EpTP5QUHhypCPwmq05kN6",null]]
+    async with httpx.AsyncClient(base_url=BASE_URL, headers=headers, timeout=60) as client:
+        response = await client.post("", json=CHECK_PAYLOAD)
+        logger.debug(response.status_code)
+        logger.debug(response.text)
+        return response.is_success or "500" not in response.text  # 1:{"error":{"message":"User not found","status":500}}
+
+
+if __name__ == '__main__':
+    token = "HMACCOUNT_BFESS=2A2062EF2FE81046; io=4gPrU0FUQr5w60EmAdXS; Hm_lvt_a3c8711bce1795293b1793d35916c067=1726115406; Hm_lpvt_a3c8711bce1795293b1793d35916c067=1726115406; HMACCOUNT=2A2062EF2FE81046; _gcl_au=1.1.112315428.1726115409.331920663.1726115409.1726115409; id=221988a1d5f300f3||t=1726115409|et=730|cs=002213fd481785fd77a9cb7651; session=eyJhbGciOiJIUzI1NiJ9.eyJ1c2VyIjp7ImVtYWlsIjoiZjk5Y241ZGJhQG1haWwueGl1dmkuY24ifSwiZXhwaXJlcyI6IjIwMjQtMTItMjFUMDQ6MzA6MTcuMTQ5WiIsImlhdCI6MTcyNjExNTQxNywiZXhwIjoxNzM0NzU1NDE3fQ.G7MvMcx_96aHeVZV7I28Fssl14MqbhTyAkHIM8ZgwOk; __cf_bm=rAWQ3zYw5tpAvJSeP3uD8Sd0mZir3NbcJdqiLg3YHlE-1726115419-1.0.1.1-pPoQZFTpwqS1eGcSfxmSoGYOCHkxEVhPl8LVig05cFf.Ri3smhoG5KCzCcupGTtQOM0To5WwGbPLX24mP3rEpw"
+    df = arun(aget_spreadsheet_values(feishu_url=FEISHU_URL_VIP, to_dataframe=True), debug=False)
+    # df = arun(aget_spreadsheet_values(feishu_url=FEISHU_URL, to_dataframe=True), debug=False)
+
+    api_keys = df[0]
+    # api_keys = [token]
+    for i, _ in enumerate(filter(None, api_keys)):
+        # logger.debug(i)
+        if token:
+            # token = "__Host-next-auth.csrf-token=5aacaeb122d54d7ee0aea2b5f34b61425e259696c0677aeb5a337887d1ca1f78%7Cc5e3861a0501a66449f97b1a4b0128d4f740f334a69111c480ad037f129c1f60; __Secure-next-auth.callback-url=https%3A%2F%2Fwww.tryblend.ai%2F; _gcl_au=1.1.1117087196.1725410224.646576272.1726110797.1726110798; session=eyJhbGciOiJIUzI1NiJ9.eyJ1c2VyIjp7ImVtYWlsIjoiNzY4M0BuZXNjLmNuIn0sImV4cGlyZXMiOiIyMDI0LTEyLTIxVDAzOjEzOjQ4LjgyM1oiLCJpYXQiOjE3MjYxMTA4MjgsImV4cCI6MTczNDc1MDgyOH0.r3FIs9T_ai9geOEH8iVnLa8ouQQUs8_ifURHWu5XEUg"
+            # arun(check_token(token))
+            # break
+            # token = None
+
+            c = Completions(token, vip=True)
+            request = ChatCompletionRequest(
+                # model="claude-3-haiku-20240307",
+                # model="claude-3-5-sonnet",
+                model="perplexity-sonar-small",
+                # model="gpt-4o-mini",
+
+                # messages=[{'role': 'user', 'content': '南京天气怎么样'}],
+                messages=[{'role': 'user', 'content': '你是谁'}],
+                stream=False,
+
+            )
+
+
+            async def main():
+                if request.stream:
+                    async for i in await c.create(request):
+                        print(i, end='')
+
+                else:
+                    for i in await c.create(request):
+                        print(i, end='')
+
+
+            arun(main())