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

meutils/apis/sunoai/suno.py

@@ -0,0 +1,373 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# @Project      : AI.  @by PyCharm
+# @File         : suno
+# @Time         : 2024/3/27 20:37
+# @Author       : betterme
+# @WeChat       : meutils
+# @Software     : PyCharm
+# @Description  :
+
+
+import jsonpath
+
+from meutils.pipe import *
+from meutils.schemas.task_types import Task
+from meutils.schemas.suno_types import SunoAIRequest
+from meutils.schemas.suno_types import MODELS, BASE_URL, CLIENT_BASE_URL, UPLOAD_BASE_UR, STUDIO_BASE_URL
+from meutils.schemas.suno_types import API_SESSION, API_FEED, API_BILLING_INFO, API_GENERATE_LYRICS, API_GENERATE_V2
+
+from meutils.decorators.retry import retrying
+from meutils.notice.feishu import send_message as _send_message
+from meutils.config_utils.lark_utils import get_next_token_for_polling
+from meutils.io.files_utils import to_bytes
+
+FEISHU_URL = "https://xchatllm.feishu.cn/sheets/Bmjtst2f6hfMqFttbhLcdfRJnNf?sheet=Jxlglo"
+FEISHU_URL_STEM = "https://xchatllm.feishu.cn/sheets/GYCHsvI4qhnDPNtI4VPcdw2knEd?sheet=rxldsA"
+
+send_message = partial(
+    _send_message,
+    title=__name__,
+    url="https://open.feishu.cn/open-apis/bot/v2/hook/dc1eda96-348e-4cb5-9c7c-2d87d584ca18"
+)
+
+
+@alru_cache(ttl=3600)
+# @retrying(max_retries=5, predicate=lambda r: not r)
+async def get_refresh_token(token: str):  # 定时更新一次就行
+    headers = {
+        "Cookie": f"__client={token}"
+    }
+    async with httpx.AsyncClient(base_url=CLIENT_BASE_URL, headers=headers, timeout=60) as client:
+        response = await client.get('')
+
+        # logger.debug(response.status_code)
+        # logger.debug(response.text)
+
+        if response.is_success:
+            data = response.json()  # {"response":null,"client":null}
+            if not data.get("response"):
+                return token, None
+
+            if ids := jsonpath.jsonpath(data, "$..last_active_session_id"):
+                return token, ids[0]  # last_active_session_id
+
+        send_message(f"未知错误：{response.status_code}\n\n{response.text}")
+
+
+@alru_cache(ttl=30 - 3)
+async def get_access_token(token: str):
+    _ = await get_refresh_token(token)
+    token, last_active_session_id = await get_refresh_token(token)  # last_active_token 没啥用
+
+    headers = {
+        "Cookie": f"__client={token}",
+    }
+
+    params = {
+        "__clerk_api_version": "2021-02-05",
+        "_clerk_js_version": "5.35.1"
+    }
+
+    async with httpx.AsyncClient(base_url=CLIENT_BASE_URL, headers=headers, timeout=60) as client:
+        response = await client.post(f"/sessions/{last_active_session_id}/tokens", params=params)
+        response.raise_for_status()
+        logger.debug(bjson(response.json()))
+        return response.json().get('jwt')
+
+# /tokens?__clerk_api_version=2021-02-05&_clerk_js_version=5.35.1
+
+@retrying(predicate=lambda r: not r)
+async def create_task(request: SunoAIRequest, token: Optional[str] = None):
+    token = token or await get_next_token_for_polling(FEISHU_URL, check_token=check_token, from_redis=True)
+
+    access_token = await get_access_token(token)
+
+    headers = {
+        "Authorization": f"Bearer {access_token}",
+    }
+    payload = request.model_dump(exclude_none=True)
+    async with httpx.AsyncClient(base_url=BASE_URL, headers=headers, timeout=60) as client:
+        response = await client.post(API_GENERATE_V2, json=payload)
+        if response.is_success:
+            data = response.json()
+            task_id, *clip_ids = jsonpath.jsonpath(data, "$..id")
+            clip_ids = [i for i in clip_ids if not str(i).startswith("m_")]
+
+            task_id = f"suno-{','.join(clip_ids)}"  # 需要返回的任务id
+            return Task(id=task_id, data=data, system_fingerprint=token)
+
+        response.raise_for_status()
+
+
+@alru_cache(ttl=15)
+@retrying(predicate=lambda r: not r)  # 触发重试
+async def get_task(task_id, token: str):  # task_id 实际是 clip_ids， 必须指定token获取任务
+    task_id = task_id.split("suno-")[-1]
+
+    access_token = await get_access_token(token)
+    headers = {
+        "Authorization": f"Bearer {access_token}",
+    }
+
+    params = {"ids": task_id}
+    async with httpx.AsyncClient(base_url=BASE_URL, headers=headers, timeout=60) as client:
+        response = await client.get(API_FEED, params=params)
+        if response.is_success:
+            return response.json()
+
+
+@alru_cache(ttl=15)
+@retrying(predicate=lambda r: not r)  # 触发重试
+async def generate_lyrics(prompt: str = '', token: Optional[str] = None):
+    token = token or await get_next_token_for_polling(FEISHU_URL)
+
+    access_token = await get_access_token(token)
+    headers = {
+        "Authorization": f"Bearer {access_token}",
+    }
+
+    payload = {"prompt": prompt}
+
+    async with httpx.AsyncClient(base_url=BASE_URL, headers=headers, timeout=60) as client:
+        response = await client.post(API_GENERATE_LYRICS, json=payload)
+        if response.is_success:
+            task_id = response.json().get("id")
+
+            for i in range(100):
+                response = await client.get(API_GENERATE_LYRICS + task_id)
+
+                logger.debug(response.text)
+
+                if response.is_success:
+                    if response.json().get("status") == "complete":  # "status": "complete"
+                        return response.json()  # todo: 大模型写歌词兜底
+                    elif response.json().get("status") == "error":  # 退出
+                        # send_message(f"生成歌词失败：{response.text}")
+                        return "可能触发内容审核，请检查提示词"
+
+                if i > 30:
+                    response.raise_for_status()
+
+                await asyncio.sleep(1 if i < 3 else 0.5)
+        response.raise_for_status()
+
+
+@alru_cache(ttl=3600)
+async def upload(file: bytes, title: str = 'xx.wav', token: Optional[str] = None):  # 必须指定token获取任务
+    token = token or await get_next_token_for_polling(FEISHU_URL)
+
+    access_token = await get_access_token(token)
+    headers = {
+        "Authorization": f"Bearer {access_token}",
+    }
+
+    async with httpx.AsyncClient(timeout=100) as client:
+        # payload = {"extension": "wav"} # "{\"extension\":\"wav\"}"
+        payload = "{\"extension\":\"wav\"}"
+        payload = "{\"extension\":\"\"}"
+        response = await client.post(f"{BASE_URL}/api/uploads/audio/", content=payload, headers=headers)
+
+        # logger.debug(response.text)
+
+        if response.is_success:
+            data = response.json()
+            logger.debug(data)
+
+            file_id = data.get("id")
+            # file_id = Path(data.get("fields").get('key')).stem
+
+            payload = data.get("fields")
+            files = {
+                'file': file
+                # 'file': ("xx", file, 'audio/mpeg')
+                # 'file': (title, file, 'audio/wav')
+            }
+
+            response = await client.post(url=UPLOAD_BASE_UR, data=payload, files=files)
+            logger.debug(response.status_code)
+            if response.is_success:
+                payload = {"upload_type": "file_upload", "upload_filename": title}
+                response = await client.post(
+                    f"{BASE_URL}/api/uploads/audio/{file_id}/upload-finish/",
+                    headers=headers,
+                    json=payload
+                )
+                for i in range(30):
+                    response = await client.get(f"{BASE_URL}/api/uploads/audio/{file_id}/", headers=headers)
+                    if response.is_success and response.json().get("s3_id"):
+                        logger.debug(response.json())
+                        break
+
+                        # {
+                        #     "id": "557f349e-2ce5-45f0-806c-efba18286599",
+                        #     "status": "complete",
+                        #     "error_message": null,
+                        #     "s3_id": "m_476a2e21-a0aa-4e33-92ff-2b5ddd587661",
+                        #     "title": "audio",
+                        #     "image_url": "https://cdn1.suno.ai/image_476a2e21-a0aa-4e33-92ff-2b5ddd587661.png"
+                        # }
+                        # {"id": "5d39c704-52c2-4d32-9bef-ac06c2c84dab", "status": "error",
+                        #  "error_message": "Uploaded audio is too short (currently 1.5 seconds). Minimum duration is 6 seconds.",
+                        #  "s3_id": null, "title": "", "image_url": null}
+
+                    elif response.json().get("status") == "error":
+                        logger.error(response.text)
+                        raise Exception(response.text)
+                    if i == 1:
+                        logger.debug(response.status_code)
+                        logger.debug(response.text)
+                        await asyncio.sleep(3)
+                    else:
+                        await asyncio.sleep(0.5)
+
+                response = await client.post(
+                    f"{BASE_URL}/api/uploads/audio/{file_id}/initialize-clip/",
+                    headers=headers
+                )
+                clip_id = response.json().get("clip_id")  # m_{clip_id} image_{clip_id}.png
+
+                payload = {"id": clip_id, "title": title}  # 好像不要也行，前端展示  title 传入文件名
+                resp = await client.post(f"{BASE_URL}/api/gen/{clip_id}/set_metadata/", json=payload, headers=headers)
+
+                data = await get_task(task_id=clip_id, token=token)  # 绑定token才能获取
+                return data, token  # clip_id: token 存一下
+
+            else:
+                response.raise_for_status()
+
+
+async def get_credits(token):
+    access_token = await get_access_token(token)
+    headers = {
+        "Authorization": f"Bearer {access_token}",
+    }
+    async with httpx.AsyncClient(base_url=BASE_URL, timeout=100, headers=headers) as client:
+        response = await client.get(API_BILLING_INFO)
+        response.raise_for_status()
+        if response.is_success:
+            data = response.json()
+            # logger.debug(bjson(data))
+            return data
+
+
+async def check_token(token, threshold=10):
+    try:
+        data = await get_credits(token)
+
+        logger.debug(data['total_credits_left'])
+
+        return data['total_credits_left'] >= threshold  # 视频
+    except Exception as e:
+        logger.error(f"{e}\n无效\n{token}")
+        return False
+
+
+@retrying()
+async def create_task_for_stems(stem_from_url):
+    token = await get_next_token_for_polling(FEISHU_URL_STEM)
+    access_token = await get_access_token(token)
+
+    file = await to_bytes(stem_from_url)
+    data, _ = await upload(file, token=token)
+    logger.debug(bjson(data))
+    stem_from_id = data[0]['id']  #
+
+    headers = {
+        "Authorization": f"Bearer {access_token}",
+    }
+    async with httpx.AsyncClient(base_url=BASE_URL, headers=headers, timeout=60) as client:
+        response = await client.post(f"/api/edit/stems/{stem_from_id}/")
+        response.raise_for_status()
+        if response.is_success:
+            data = response.json()
+            clip_ids = jsonpath.jsonpath(data, "$..id")
+
+            task_id = f"suno-{','.join(clip_ids)}"  # 需要返回的任务id
+            return Task(id=task_id, data=data, system_fingerprint=token)
+
+
+@retrying()
+async def create_task_for_cover(cover_from_url, lyrics):
+    token = await get_next_token_for_polling(FEISHU_URL_STEM)
+    access_token = await get_access_token(token)
+
+    file = await to_bytes(cover_from_url)
+    data, _ = await upload(file, token=token)
+    cover_from_id = data[0]['id']
+
+    headers = {
+        "Authorization": f"Bearer {access_token}",
+    }
+    payload = SunoAIRequest(
+        task="cover",
+        mv="chirp-v3-5-tau",
+        cover_clip_id=cover_from_id,
+        prompt=lyrics,
+    ).model_dump(exclude_none=True)
+
+    async with httpx.AsyncClient(base_url=BASE_URL, headers=headers, timeout=60) as client:
+        response = await client.post(API_GENERATE_V2, json=payload)
+        if response.is_success:
+            data = response.json()
+            task_id, *clip_ids = jsonpath.jsonpath(data, "$..id")
+            clip_ids = [i for i in clip_ids if not str(i).startswith("m_")]
+
+            task_id = f"suno-{','.join(clip_ids)}"  # 需要返回的任务id
+            return Task(id=task_id, data=data, system_fingerprint=token)
+
+        response.raise_for_status()
+
+
+if __name__ == '__main__':
+    # token = os.getenv("SUNO_API_KEY")
+    # token = "eyJhbGciOiJSUzI1NiIsInR5cCI6IkpXVCJ9.eyJpZCI6ImNsaWVudF8yaGlHSTlCZFVwOUdZcUlGM3ZmTU1IT25SNFAiLCJyb3RhdGluZ190b2tlbiI6ImltOWMzOGJ4bnV2OThiZXplMW8yOG1zd2Y2c3lrdzd6YnM2ejJubHkifQ.SnC8-G2LVQztTiA2davFS413mQIaBmRFDzIw1JmvHg4UOMXq95z0CgbfK8Gx8Zv-FXdpKVqkamiNTzZP9qsLOSgREqCSSq5bmA6SPIWx-R6dj1PMDFRX-qv5qGyyPe4sadF6wnr45MS9859148gRmr_Go8rAT_7Hu0DKySextl-Xbs6ClDaYYUyyV3HudWQh4F8jwvxkyer05AgN6smQH5eZI-NRKVgZn_i6Mtl8IJz8R1fzD2YNIcvH4QC4qGhrg9n74ljIeORCMsoJzW2SBZa4QWWDx_0VYs-tA_Z43bqwN_2ojMGM63fm2hLOZmwf6S1LQy9_O6UdcUQiEs__OA"
+    token = "eyJhbGciOiJSUzI1NiIsInR5cCI6IkpXVCJ9.eyJpZCI6ImNsaWVudF8ybEIwRFlXTWYya3NxNWlxazl4S2dZbmQ0eVciLCJyb3RhdGluZ190b2tlbiI6Im5vejdpMzlvY2YzMTEzaHp6aDZwY2N5MzltZm5xNmZvdnhtOG9oNngifQ.qUeBLl-NxKzhpoUBpo_EwkHH0qwIsG0nMSD_yvHv5EK7YfybupoPWw8SpSKDhmZu5K_KsgdOF0RQH22jll4U-x0BwfVu1ze-GBxjnNEoSerUB7hu1cfvmg7xMH8rHJQig2TWE2h0hzP6dMajPHQWRTltbb5MMkKHgBFj0CiAFqaGwzSSvAtERwwHBIK3KalbaV1oyd6DJYG4FrVgQLubkp7VXj11LszxD6qXklRhsc9h55kvYASDPHhnZJi9u2QfIbiKVkraXb6ShqDmNtqXbj22p6g2R9fwMEB-m68S7QSZyAWRArWzsSZujzhxmNuMGbuVX1v7op7F3hA2zFphYw"
+    # print(arun(get_refresh_token(token)))
+    print(arun(get_access_token(token)))
+
+    # arun(generate_lyrics(prompt=''))
+
+    # ids = "ee6d4369-3c75-4526-b6f1-b5f2f271cf30"
+    # print(api_feed(api_key, ids))
+
+    # for i in range(100):  # 测试过期时间
+    #     print(api_billing_info(api_key))
+    #     time.sleep(60)
+
+    # print(arun(get_api_key()))
+    # task_id = music_ids = 1
+    # send_message(f"""
+    #     https://api.chatfire.cn/task/suno/v1/tasks/{task_id}
+    #     https://api.chatfire.cn/task/suno/v1/music/{music_ids}
+    #     """)
+    # file = open("/Users/betterme/PycharmProjects/AI/test.mp3", 'rb').read()
+    # file = open("/Users/betterme/Downloads/audio.wav", 'rb').read()
+    # arun(upload(file=file))
+    # token = "eyJhbGciOiJSUzI1NiIsInR5cCI6IkpXVCJ9.eyJpZCI6ImNsaWVudF8yajM1dkNXaTE1b0VQWnlUU0RMMnBsU3RiMVQiLCJyb3RhdGluZ190b2tlbiI6Ijd5MDV0cG1nMWJpbjBmaGIxZW1kYzZoazQzbDRuaG84bmJ4Yzc1dzgifQ.LsEfuPgwXu33f_UD2pRY4HjHIwU_rPG2rvG46BVDKqXcPmhRTWl5LjKgFrSzU51tgxfG-wMopVJhRxgS6YZUMtKVojDFtV_ZImyJ30u6LYA5nSbrkhqUrBdU4P5WmkL9irvh4sGPtvv8ML_pyXzsittsDrNnDCtm_isacOD-Fy3VKOCOjWj4W3qUUdnmBvTeeUbrnepQqurAjYPg6Ug-WkR43xZ5tWtxxTQ4ebtZglgRyCbyF3TM9XMKa67FTwHQsja8cLo2CyGRzb89e3uuF8Na_CY17ZlxJyQ2p_FZmL0egWr0EveZFeVzIUUs704pkhKd-RC9Q47Jqcg8qdiRYA"
+    # arun(generate_lyrics('hi', token))
+    # arun(generate_lyrics())
+    # task_id = "0b017d4e-c559-4cc6-9339-9cd53aa25af4"
+    # token = "eyJhbGciOiJSUzI1NiIsInR5cCI6IkpXVCJ9.eyJpZCI6ImNsaWVudF8ybEIwRFlXTWYya3NxNWlxazl4S2dZbmQ0eVciLCJyb3RhdGluZ190b2tlbiI6Im5vejdpMzlvY2YzMTEzaHp6aDZwY2N5MzltZm5xNmZvdnhtOG9oNngifQ.qUeBLl-NxKzhpoUBpo_EwkHH0qwIsG0nMSD_yvHv5EK7YfybupoPWw8SpSKDhmZu5K_KsgdOF0RQH22jll4U-x0BwfVu1ze-GBxjnNEoSerUB7hu1cfvmg7xMH8rHJQig2TWE2h0hzP6dMajPHQWRTltbb5MMkKHgBFj0CiAFqaGwzSSvAtERwwHBIK3KalbaV1oyd6DJYG4FrVgQLubkp7VXj11LszxD6qXklRhsc9h55kvYASDPHhnZJi9u2QfIbiKVkraXb6ShqDmNtqXbj22p6g2R9fwMEB-m68S7QSZyAWRArWzsSZujzhxmNuMGbuVX1v7op7F3hA2zFphYw"
+    # arun(get_task(task_id=task_id, token=token))
+
+    # arun(get_credits(token))
+    # arun(check_token(token))
+
+    # arun(upload(Path('cover.mp3').read_bytes(), title="翻唱"))
+    url = "https://cdn1.suno.ai/14d5afd8-057e-4d97-bc13-b8521dde568a.mp3"
+
+    # arun(create_task_for_stems(url))
+    #
+    # arun(get_task("suno-644b4e39-abd5-44c0-95cb-25ee56b0f6ed,ee0443b2-af84-4fb5-b29d-d23d41ce8142", token=token))
+    # lyrics = """
+    # [Verse]\n每天天还未亮\n我们已经在路上\n辛苦工作拼命干\n工资却总是拖欠\n\n[Verse 2]\n加班加点不休息\n老板总是不提起\n口袋越来越空虚\n哪里是我们的权益\n\n[Chorus]\n一起团结呼吁\n要把话语权夺回\n为我们的权利\n我们绝不喊后退\n\n[Bridge]\n不怕辛劳与压力\n只怕公正被歪曲\n强忍泪水背负艰辛\n梦想不可被抹去\n\n[Verse 3]\n工资迟迟不见影\n梦想渐渐被耗尽\n但我们心不认命\n勇敢站出来喊停\n\n[Chorus]\n一起团结呼吁\n要把话语权夺回\n为我们的权利\n我们绝不喊后退
+    # """
+    # arun(create_task_for_cover(url, lyrics))
+
+    # token = "eyJhbGciOiJSUzI1NiIsInR5cCI6IkpXVCJ9.eyJpZCI6ImNsaWVudF8yazY3OG5pWWI5M2xjWHQ2a1dZMVNSalcxNVQiLCJyb3RhdGluZ190b2tlbiI6Imk2bzE0eXkxMW5mNTduaXFpcG9kOWx2cHBrbXBvbG16bTdqMDQ3bXQifQ.mD5dwo6c1xyg4kS_Z6Nwz4N_UonetHBFLcvHRTGTsqd4Pc-QYTEJ2IiedILX7nrA1b3K-mhN-M9isyzUxl3Pa9DCqQ4p9snsib_Rz-v_ou2V-rOORPHLwzIQSZbkE4gDyH4DF_UT6TqZRp_KkOyPEmyGQHQWjiUDuWv7CjDXM0AbqhV0gafCKA6cQ0ZdL9yKNq_0rFgIWP9mRrDGP6r237gXqLm84SmlhtMpob5T_QByCpunmTKe2Gi9qNEkKm_8EHAKSw3vj9XHBH3cMsiVPO47cT_rs98tIUq20sqk4C6hlL7mhLomm9L33k4wVTgyH7ln8UK9xxQ3lGm-W-mjKA"
+    #
+    # arun(get_credits(token))
+
+    # arun(generate_lyrics('hi'))
+    # arun(create_task(SunoAIRequest(prompt='写首中国风的歌曲')))

meutils/apis/textcard/__init__.py

@@ -0,0 +1,11 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# @Project      : AI.  @by PyCharm
+# @File         : __init__.py
+# @Time         : 2024/9/18 10:01
+# @Author       : betterme
+# @WeChat       : meutils
+# @Software     : PyCharm
+# @Description  : 
+
+from meutils.pipe import *

meutils/apis/textcard/demo.py

@@ -0,0 +1,25 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# @Project      : AI.  @by PyCharm
+# @File         : demo
+# @Time         : 2024/9/18 10:39
+# @Author       : betterme
+# @WeChat       : meutils
+# @Software     : PyCharm
+# @Description  :
+import re
+
+from meutils.pipe import *
+
+HTML_PARSER = re.compile(r'```html(.*?)```', re.DOTALL)
+s = """
+这是一堆文本
+```html
+这是一段html
+```
+这是一堆文本
+"""
+
+print(HTML_PARSER.findall(s))
+
+# {"messages":[{"role":"user","content":"早饭"}],"chat_id":"ssZZTRs","model":"Qwen/Qwen2-Math-72B-Instruct"}

meutils/apis/textcard/hanyuxinjie.py

@@ -0,0 +1,81 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# @Project      : AI.  @by PyCharm
+# @File         : hanyuxinjie
+# @Time         : 2024/9/18 13:30
+# @Author       : betterme
+# @WeChat       : meutils
+# @Software     : PyCharm
+# @Description  :
+
+from meutils.pipe import *
+
+BASE_URL = 'https://xy.siliconflow.cn'
+
+HTML_PARSER = re.compile(r'```html(.*?)```', re.DOTALL)
+
+
+# s = """
+# 这是一堆文本
+# ```html
+# 这是一段html
+# ```
+# 这是一堆文本
+# """
+#
+# print(HTML_PARSER.findall(s))
+
+
+async def create(
+        prompt: str = '996',
+        model: str = "meta-llama/Meta-Llama-3.1-405B-Instruct",
+        title: str = "🔥汉语新解",
+):
+    """
+    "Pro/THUDM/glm-4-9b-chat"
+    "Qwen/Qwen2-Math-72B-Instruct"
+    “deepseek-ai/DeepSeek-V2.5”
+    "meta-llama/Meta-Llama-3.1-405B-Instruct"
+    """
+    payload = {
+        "messages": [
+            {
+                "role": "user",
+                "content": prompt
+            }
+        ],
+        # "chat_id": "i8yw46k",
+        "model": model
+    }
+    async with httpx.AsyncClient(base_url=BASE_URL, timeout=300) as client:
+        async with client.stream(method="POST", url="/api/chat", json=payload) as response:
+            # logger.debug(response.status_code)
+            async for chunk in response.aiter_lines():
+                # for chunk in "response.aiter_lines()":
+                yield chunk.replace("智说新语", title)
+
+
+if __name__ == '__main__':
+    pass
+    # model = "Pro/THUDM/glm-4-9b-chat"
+    model = "Qwen/Qwen2.5-72B-Instruct"
+    model = "deepseek-ai/DeepSeek-V2.5"
+
+    # arun(create(text="火宝", model="meta-llama/Meta-Llama-3.1-405B-Instruct"))
+
+    # arun(create(text="火宝", model="google/gemma-2-27b-it"))
+
+    # arun(create(prompt="火宝", model=model))
+
+    # arun(main(create(text="火宝", model="Qwen/Qwen2-Math-72B-Instruct", stream=True)))
+    s = """
+    这是一堆文本
+    ``html
+    这是一段html
+    ```
+    这是一堆文本
+    """
+
+    print(HTML_PARSER.findall(s))
+
+    arun(create(prompt="996", model=model))

meutils/apis/textin.py

@@ -0,0 +1,159 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# @Project      : AI.  @by PyCharm
+# @File         : textin
+# @Time         : 2024/6/26 08:22
+# @Author       : betterme
+# @WeChat       : meutils
+# @Software     : PyCharm
+# @Description  : todo: 重构 https://tools.textin.com/
+# https://www.textin.com/console/recognition/robot_enhance?service=watermark-remove
+
+# proxies 取消了 https://github.com/encode/httpx/pull/2879
+
+from meutils.pipe import *
+from meutils.decorators.retry import retrying
+from meutils.notice.feishu import send_message
+from meutils.io.files_utils import to_url
+from meutils.apis.proxy.ips import get_one_proxy
+
+from fake_useragent import UserAgent
+
+ua = UserAgent()
+
+BASE_URL = "https://api.textin.com/home"
+
+proxy = "http://110.42.51.201:38443"
+
+
+@alru_cache(ttl=3600 * 24)
+async def document_process(data: bytes,
+                           service: str = "pdf_to_markdown",
+                           response_format: str = "url",
+                           token: Optional[str] = None,  # 821480df15b637852f7e4a0dabb8a156
+                           **kwargs):
+    """
+
+    :param data:
+    :param service: pdf_to_markdown watermark-remove dewarp
+    :param response_format:
+    :param kwargs:
+    :return:
+    """
+
+    params = {
+        "service": service,
+        **kwargs
+
+        # "page_count": page_count,
+        # "get_image": "objects"
+        # "apply_document_tree": 0,
+    }
+
+    request_kwargs = {}
+    # request_kwargs = {"proxies": proxies}
+
+    for i in range(3):
+
+        headers = {
+            "token": token or str(uuid.uuid4()).replace("-", ""),
+            'User-Agent': ua.random,
+        }
+        try:
+            async with httpx.AsyncClient(base_url=BASE_URL, timeout=60, headers=headers, **request_kwargs) as client:
+                response = await client.post('/user_trial_ocr', content=data, params=params)
+                response.raise_for_status()
+
+                logger.debug(response.status_code)
+                # logger.info(response.json())  # {'msg': '今日请求超过限制次数', 'code': 431}
+
+                _ = response.json()
+
+                if '今日请求超过限制次数' in str(_):
+                    # logger.info(response.json())
+                    raise Exception("更新请求头或者加代理")
+                else:
+                    break
+
+
+        except Exception as e:  # 加代理
+            logger.error(e)
+
+            request_kwargs = {
+                "proxy": await get_one_proxy(),
+                # "proxy": proxy,
+
+            }
+
+    if response_format == "url" and service in {"watermark-remove", "crop_enhance_image"}:
+        if "data" not in _:
+            send_message(_, title=__name__)
+
+        # _ => {'msg': '今日请求超过限制次数', 'code': 431}
+        logger.debug(bjson(_))
+        # logger.debug(list(_['data']['result']['image']))
+
+        url = await to_url(_['data']['result']['image'])
+        _ = {"data": [{"url": url}]}
+
+        # _['data']['result']['image'] = await to_url(_['data']['result']['image'])
+
+    return _
+
+
+if __name__ == '__main__':
+    from meutils.io.files_utils import to_bytes, to_url
+
+    # data = open("/Users/betterme/PycharmProjects/AI/11.jpg", 'rb').read()
+    # # data = open("/Users/betterme/PycharmProjects/AI/蚂蚁集团招股书.pdf", 'rb').read()
+    # with timer("解析"):
+    #     # arun(textin_fileparser(data))
+    #     print(arun(textin_fileparser(data, service="pdf_to_markdown")))
+
+    # response = requests.request("POST", url, data=data)
+    data = open("/Users/betterme/PycharmProjects/AI/qun.png", 'rb').read()
+    data = open("img.png", 'rb').read()
+    # data = arun(to_bytes("https://cdn.meimeiqushuiyin.cn/ori/tmp_e8d6329e1b2c1bc541ca530fcbae14e3ec12f65d8d4ec97d.jpg"))
+
+    from meutils.schemas.task_types import Purpose
+
+    service = Purpose.watermark_remove.value
+
+    # service = "pdf_to_markdown"
+
+    service = 'bill_recognize_v2'
+
+
+    async def main(n=1):
+        for i in tqdm(range(n)):
+            try:
+                _ = await document_process(data, service=service)
+                logger.debug(bjson(_))
+            except Exception as e:
+                logger.error(f"{i}: {e}")
+                break
+
+
+    # with timer("解析"):
+    #
+    #     # arun(textin_fileparser(data))
+    #     data = arun(document_process(data, service=service))
+    #
+    #     # b64 = data['data']['result']['image']
+    #
+    #     # base64_to_file(b64, "demo.png")
+    #
+    #     # data['data']['result']['image'] = arun(to_url(b64))
+    #
+    #     logger.debug(data)
+    #
+    #     # {'msg': 'success',
+    #     #  'data': {
+    #     #      'result': {
+    #     #          'image': 'https://sfile.chatglm.cn/chatglm-videoserver/image/e5/e5d4011c.png'
+    #     #      },
+    #     #      'file_type': '', 'file_data': ''
+    #     #  }, 'code': 200
+    #     #  }
+
+    arun(main(1))

meutils/apis/to_image/__init__.py

@@ -0,0 +1,11 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# @Project      : AI.  @by PyCharm
+# @File         : __init__.py
+# @Time         : 2024/9/19 08:50
+# @Author       : betterme
+# @WeChat       : meutils
+# @Software     : PyCharm
+# @Description  : 
+
+from meutils.pipe import *