{ "version": "https://jsonfeed.org/version/1", "title": "Python", "description": "\u8fd9\u91cc\u8ba8\u8bba\u5404\u79cd Python \u8bed\u8a00\u7f16\u7a0b\u8bdd\u9898\uff0c\u4e5f\u5305\u62ec Django\uff0cTornado \u7b49\u6846\u67b6\u7684\u8ba8\u8bba\u3002\u8fd9\u91cc\u662f\u4e00\u4e2a\u80fd\u591f\u5e2e\u52a9\u4f60\u89e3\u51b3\u5b9e\u9645\u95ee\u9898\u7684\u5730\u65b9\u3002", "home_page_url": "go/python", "feed_url": "feed/python.json", "icon": "https://cdn.v2ex.com/navatar/8613/985e/90_large.png?m=1648339948", "favicon": "https://cdn.v2ex.com/navatar/8613/985e/90_normal.png?m=1648339948", "items": [ { "author": { "url": "member/guoguobaba", "name": "guoguobaba", "avatar": "https://cdn.v2ex.com/avatar/48a4/7fbb/307012_large.png?m=1742526918" }, "url": "t/1176823", "title": "\u95ee\u4e2a\u6c49\u5b57\u5904\u7406\u7684\u95ee\u9898", "id": "t/1176823", "date_published": "2025-12-04T02:14:43+00:00", "content_html": "

\u6709\u4e2a\u6587\u4ef6\uff0c\u5305\u542b'\u2edd' 11997 \u548c\u98df\u662f\u76f8\u540c\u7684\u5b57\uff0c\u4f46\u662f\u662f\u4e0d\u540c\u7684\u7f16\u7801\uff0c

\n

\u2edd (U+2EDD)\u548c \u98df (U+98DF)\u7684\u5173\u7cfb\u662f\uff1a\u5b83\u4eec\u662f\u540c\u4e00\u4e2a\u5b57\u7684\u4e0d\u540c\u89c6\u89c9\u8868\u73b0\u5f62\u5f0f\uff0c\u4f46 Unicode \u6307\u5b9a U+98DF \u4e3a\u6807\u51c6\u5f62\u5f0f\uff0cU+2EDD \u4e3a\u5176\u5f02\u4f53\u5f62\u5f0f\uff08\u7279\u522b\u662f\u4f5c\u4e3a\u90e8\u9996\u65f6\uff09\u3002

\n

\u95ee\u4e86\u4e00\u4e0b AI \uff0c\u90fd\u662f\u8ba9\u679a\u4e3e\u505a\u4e00\u4e2a map \uff0c\u6709\u6ca1\u6709\u7edf\u4e00\u7684\u8f6c\u5316\u8fd9\u6837\u6c49\u5b57\u7684\u65b9\u6848\uff0cunicodedata.normalize \u4e0d\u884c\u3002

\n" }, { "author": { "url": "member/Leon6868", "name": "Leon6868", "avatar": "https://cdn.v2ex.com/avatar/14ed/48e6/438369_large.png?m=1748955348" }, "url": "t/1175519", "title": "\u6709\u54ea\u4e9b Python 3.14 Free-Threading \u591a\u7ebf\u7a0b\u6027\u80fd\u5206\u6790\u65b9\u6848\uff1f", "id": "t/1175519", "date_published": "2025-11-27T15:07:31+00:00", "content_html": "

\u5982\u9898\uff0c\u6ca1\u627e\u5230\u5bf9\u4ee3\u7801\u5165\u4fb5\u5c0f\u3001\u80fd\u5904\u7406\u591a\u7ebf\u7a0b\u7684\u89e3\u51b3\u65b9\u6848\uff0c\u8bf7\u95ee\u5927\u5bb6\u90fd\u662f\u600e\u4e48\u89e3\u51b3\u7684\u5462\uff1f

\n" }, { "author": { "url": "member/Daming798", "name": "Daming798", "avatar": "https://cdn.v2ex.com/avatar/b2b5/a321/577749_large.png?m=1735891240" }, "url": "t/1175211", "date_modified": "2025-11-26T08:00:12+00:00", "content_html": "

\u6211\u4eec\u5728\u505a\u4ec0\u4e48

\n

\u6211\u4eec\u6b63\u9762\u5411\u4e2d\u5c0f\u4f01\u4e1a\u548c\u521d\u521b\u56e2\u961f\uff0c\u505a\u4e00\u6b3e\u4e13\u6ce8\u4e8e\u7528\u6237\u53cd\u9988\u7684\u7ba1\u7406\u5206\u6790\u5de5\u5177 \u2014\u2014 Feedalyze \u3002

\n

\u5b83\u80fd\u591a\u6e20\u9053\u7edf\u4e00\u6536\u96c6\u3001\u9ad8\u6548\u5904\u7406\u3001\u53ef\u89c6\u5316\u5206\u6790\u7528\u6237\u53cd\u9988\uff0c\u5e2e\u52a9\u4e2d\u5c0f\u4f01\u4e1a\u6d1e\u5bdf\u7528\u6237\u9700\u6c42\uff0c\u52a0\u901f\u4ea7\u54c1\u8fed\u4ee3\u3002\u5730\u5740\uff1a https://feedalyze.com/

Feedalyze \u5bf9\u6807\u56fd\u5916\u7684 UserVoice + Dovetail \uff0c\u5e0c\u671b\u6210\u4e3a\u56fd\u5185\u4f01\u4e1a\u9996\u9009\u7684\u7528\u6237\u53cd\u9988\u7ba1\u7406\u5206\u6790\u5e73\u53f0\u3002

\u89e3\u51b3\u7684\u95ee\u9898

1 \u3001\u4ea7\u54c1\u5f00\u53d1\u8fed\u4ee3\u5e38\u9760\u731c\u6d4b\u6216\u7ecf\u9a8c\uff0c\u800c\u975e\u7528\u6237\u771f\u5b9e\u9700\u6c42\u3002\uff08\u5982\u679c\u53c2\u4e0e\u8fc7\u4e2d\u5c0f\u4f01\u4e1a\u4ea7\u54c1\u7814\u53d1\uff0c\u5c31\u80fd\u7406\u89e3\u8fd9\u70b9\uff09

2 \u3001\u5728\u6536\u96c6\u3001\u5904\u7406\u548c\u5206\u6790\u7528\u6237\u53cd\u9988\uff0c\u8c03\u67e5\u7528\u6237\u9700\u6c42\u65f6\uff0c\u4f01\u4e1a\u6210\u672c\u9ad8\u6602

3 \u3001\u96be\u4ee5\u53ca\u65f6\u6d1e\u5bdf\u7528\u6237\u9700\u6c42\uff0c\u5f71\u54cd\u4ea7\u54c1\u8fed\u4ee3\u901f\u5ea6\u548c\u51b3\u7b56\u6b63\u786e\u7387

4 \u3001\u7f3a\u4e4f\u53ef\u91cf\u5316\u7684\u6570\u636e\u652f\u6491\uff0c\u4ea7\u54c1\u65b9\u5411\u5bb9\u6613\u504f\u79bb\u5e02\u573a\u9700\u6c42\u7b49

\u4e3a\u4ec0\u4e48\u503c\u5f97\u52a0\u5165\uff1f

\n
• \u76ee\u524d\u4ea7\u54c1\u548c demo \u672a\u4e0a\u7ebf\uff0c\u5df2\u6709 50 \u4f59\u4f4d\u8bd5\u7528\u8005\uff0c\u6570\u4f4d\u613f\u610f\u4ed8\u8d39\u7684\u4f01\u4e1a\u7528\u6237\u3002
\n
• \u5e02\u573a\u5de8\u5927\uff0c\u56fd\u5185\u505a\u6b64\u7c7b\u5de5\u5177\u7684\u5f88\u5c11\uff0c\u56fd\u5916\u4ea7\u54c1\u4e0d\u9002\u5e94\u672c\u571f
\n
• \u7c97\u653e\u65f6\u4ee3\u5df2\u8fc7\u53bb\uff0c\u4e14\u5927\u73af\u5883\u4e0d\u4f73\u7684\u60c5\u51b5\u4e0b\uff0c\u7528\u6237\u81f3\u4e0a\u7684\u7406\u5ff5\u6b63\u9010\u6b65\u6df1\u5165\u4f01\u4e1a
\n
• \u5ba2\u6237\u7fa4\u4f53\u6781\u4e3a\u5e7f\u6cdb\uff0c\u9664\u4e92\u8054\u7f51\u884c\u4e1a\u5916\uff0c\u8fd8\u9002\u5408\u5236\u9020\u4e1a\u3001\u96f6\u552e\u4e1a\u3001\u91d1\u878d\u4e1a\u3001\u533b\u836f\u4e1a\u7b49\u7b49
\n
• \u4f60\u5c06\u8d1f\u8d23\u6574\u4e2a\u540e\u7aef\u7684\u6280\u672f\u67b6\u6784\uff0c\u662f\u4ea7\u54c1\u7684\u6838\u5fc3\u6280\u672f\u521b\u5efa\u8005
\n

\u4f60\u5c06\u8d1f\u8d23

\n
• 0-1 \u5229\u7528 FastAPI \u6784\u5efa\u4ea7\u54c1\u540e\u7aef
\n
• \u8d1f\u8d23\u4ea7\u54c1\u7684\u529f\u80fd\u5b9e\u73b0\uff0c\u8c03\u8bd5\u4e0e\u4f18\u5316
\n
• \u6280\u672f\u8def\u7ebf\u89c4\u5212\u53ca\u67b6\u6784\u8bbe\u8ba1\uff0c\u4fdd\u8bc1\u7cfb\u7edf\u53ef\u6269\u5c55\u6027\u4e0e\u7a33\u5b9a\u6027
\n

\u5e0c\u671b\u4f60

\n
• \u5728\u676d\u5dde\uff0c\u6709\u4e00\u5230\u4e09\u5e74\u4ee5\u4e0a\u7684 FastAPI \u5f00\u53d1\u7ecf\u9a8c
\n
• \u719f\u6089 PostgreSQL \u3001Redis \u7b49\u5e38\u7528\u6570\u636e\u5e93
\n
• \u826f\u597d\u7684\u7f16\u7a0b\u4e60\u60ef\uff0c\u6ce8\u91cd\u6548\u7387\uff0c\u4e3a\u4eba\u8bda\u5b9e\u5f00\u653e
\n

\u6211\u4eec\u63d0\u4f9b

\n
• \u5206\u7ea2\u671f\u6743\uff1a1-3%\uff0c\u4ee5\u603b\u51c0\u5229\u6da6\uff0c\u6309\u6301\u6709\u6bd4\u4f8b\u6bcf\u5e74\u5e95\u5206\u7ea2
\n
• \u7ecf\u8425\uff1a\u4f60\u5c06\u53c2\u4e0e\u516c\u53f8\u7ecf\u8425\uff0c\u6709\u6295\u7968\u6743
\n
• \u9ad8\u5ea6\u81ea\u4e3b\u6743\uff0c\u8d1f\u8d23\u6574\u4e2a\u540e\u7aef\u6280\u672f\u65b9\u5411
\n
• \u85aa\u8d44\uff1a7k \u5de6\u53f3\uff0c\u534a\u5e74\u8c03\u6574\u4e00\u6b21
\n
• \u6210\u4e3a\u4ea7\u54c1\u7684\u6838\u5fc3\u521b\u5efa\u8005
\n

\u6211\u4eec\u6682\u672a\u63d0\u4f9b

\n
• \u4f18\u6e25\u7684\u529e\u516c\u73af\u5883
\n
• \u6781\u4f73\u7684\u786c\u4ef6\u8bbe\u5907\uff08\u4f46\u591f\u7528\uff09
\n

\u4f46\u6211\u4eec\u63d0\u4f9b\u66f4\u591a

\n
• \u4e00\u4e2a\u6301\u6709\u751f\u4ea7\u8d44\u6599\u7684\u673a\u4f1a
\n
• \u4e00\u4e2a\u98ce\u9669\u6781\u4f4e\u7684\u521b\u4e1a\u673a\u4f1a
\n
• \u4e00\u6b21\u521b\u4e1a\u6838\u5fc3\u6210\u5458\u7684\u5c65\u5386
\n

\u5907\u6ce8

\n
• \u4e3a\u4ec0\u4e48\u662f\u5206\u7ea2\u671f\u6743\u800c\u4e0d\u662f\u80a1\u4efd\uff0c\u56e0\u4e3a\u6301\u80a1\u610f\u5473\u7740\u4f60\u4e5f\u8981\u627f\u62c5\u7ecf\u8425\u5931\u8d25\u548c\u4e8f\u635f\u7684\u98ce\u9669\u3002\u5982\u679c\u4f60\u6709\u52c7\u6c14\uff0c\u6211\u5f88\u4e50\u610f\u5c06\u5206\u7ea2\u671f\u6743\u8f6c\u4e3a\u80a1\u4efd\u3002
\n
• 7k \u7684\u85aa\u8d44\u592a\u5783\u573e\u4e86\u3002\u85aa\u8d44\u786e\u5b9e\u4e0d\u9ad8\uff0c\u4f46\u4f60\u62e5\u6709\u5206\u7ea2\u671f\u6743\uff0c\u8fd9\u8fd8\u53ef\u80fd\u662f\u4e00\u6b21\u957f\u8fdc\u89e3\u51b3\u7ecf\u6d4e\u95ee\u9898\u7684\u673a\u4f1a\u3002\u5982\u679c\u671f\u671b\u9ad8\u85aa\u804c\u4f4d\uff0c\u6216\u6709\u7ecf\u6d4e\u538b\u529b\uff0c\u8bf7\u5bfb\u627e\u5176\u5b83\u673a\u4f1a\u3002
\n
• \u4e3a\u4ec0\u4e48\u662f 1-3% \u7684\u603b\u51c0\u5229\u6da6\u4f5c\u4e3a\u5206\u7ea2\u671f\u6743\uff1f\u521b\u4e1a\u521d\u671f\u4f1a\u9047\u5230\u5f88\u591a\u98ce\u9669\uff0c\u4e09\u5e74\u5185\u6211\u4f1a\u5c06 50% \u7684\u51c0\u5229\u6da6\u4f5c\u4e3a\u907f\u9669\u8d44\u91d1\uff0c\u975e\u5230\u5fc5\u8981\u65f6\u523b\u4e0d\u4f1a\u52a8\u300227% \u4f5c\u4e3a\u7ecf\u8425\u8d44\u91d1\uff0c23% \u4f5c\u4e3a\u5168\u4f53\u6210\u5458\u5206\u7ea2\u6c60\uff0c\u6211\u81ea\u5df1\u7684\u5206\u7ea2\u6bd4\u4f8b\u5728 3-5%\u3002\u5982\u679c\u5230\u4e86\u65e0\u53ef\u633d\u56de\u7684\u90a3\u4e00\u5929\uff0c\u6211\u5c06\u628a\u907f\u9669\u8d44\u91d1\u6309\u6bd4\u4f8b\u5206\u914d\u7ed9\u6240\u6709\u6210\u5458\u3002
\n
• \u6709\u610f\u4e49\u7684\u5de5\u4f5c\u548c\u4eba\u9645\u5173\u7cfb\u3002\u6211\u6240\u8ffd\u6c42\u7684\u4e00\u662f\u6709\u610f\u4e49\u7684\u5de5\u4f5c\uff0c\u4e8c\u662f\u6709\u610f\u4e49\u7684\u3001\u6709\u4ef7\u503c\u7684\u56e2\u961f\u4eba\u9645\u5173\u7cfb\uff0c\u8fd9\u662f\u56e2\u961f\u7684\u57fa\u672c\u539f\u5219\u3002
\n
• \u56e2\u961f\u89c4\u6a21\uff1a\u672a\u6765\u5c06\u5c3d\u53ef\u80fd\u514b\u5236\u56e2\u961f\u4eba\u5458\u5728 20 \u4eba\u4ee5\u5185\u3002
\n
• \u521b\u4e1a\u8270\u8f9b\uff1a\u8fd9\u662f\u4e00\u6b21\u82e6\u903c\u7684\u521b\u4e1a\u4e4b\u65c5\uff0c\u9700\u8981\u4e3b\u52a8\u51fa\u51fb\uff0c\u8fce\u63a5\u6311\u6218\u3002\u5728\u6b64\u4e4b\u524d\uff0c\u8bf7\u505a\u597d\u5fc3\u7406\u51c6\u5907\u3002
\n

\u5982\u679c\u4f60\u5bf9\u8fd9\u4e2a\u673a\u4f1a\u611f\u5174\u8da3\uff0c\u6b22\u8fce\u52a0\u5fae\u4fe1\u4e00\u8d77\u804a\u804a\uff0c\u5403\u4e2a\u996d\u3002\u6dfb\u52a0\u65f6\u8bf7\u5907\u6ce8\uff1a\u59d3\u540d + FastAPI

\u6211\u7684\u5fae\u4fe1\uff1aM-ing2020

chatgpt \u4e5f\u6ca1\u7ed9\u4e00\u4e2a\u5b9e\u7528\u7684\u65b9\u6848\uff0c\u4f60\u4eec\u90fd\u662f\u600e\u4e48\u7ba1\u7406\u81ea\u5df1\u7684 python \u73af\u5883\u7684\uff1f

\u76ee\u524d\u662f fastapi+tortoise orm \uff0c\u8003\u5bdf\u4e86\u51e0\u4e2a

django admin \u6709\u70b9\u91cd

fastapi-admin\uff1a \u6298\u817e\u5f97\u5f88\u9ebb\u70e6\uff0c

fastadmin: \u8fd9\u4e2a\u662f\u6700\u7b80\u5355\u7684\uff0c\u4f46\u662f\u529f\u80fd\u6709\u70b9\u7b80\u964b\uff0c\u5f88\u591a\u524d\u7aef\u7684\u529f\u80fd\u6ca1\u6709\u5b9e\u73b0\u3002

\u6700\u8fd1\u521a\u5f00\u59cb\u5b66 python \u548c fastapi, \u6309\u7167 fastapi \u6559\u7a0b\u5b66\u4e60\u65f6\u9047\u5230\u4e2a\u95ee\u9898, \u56e0\u4e3a\u4f7f\u7528 sqlmodel \u7684 relationship, \u5bfc\u81f4\u4e24\u4e2a\u6a21\u578b\u6587\u4ef6\u76f8\u4e92\u5bfc\u5165, \u7136\u540e\u62a5\u9519\u4e86, \u6839\u636e\u5b98\u65b9\u6587\u6863\u7684\u89e3\u51b3\u529e\u6cd5, \u4f7f\u7528 TYPE_CHECKING \u548c\u5b57\u7b26\u4e32\u7248\u672c\u7c7b\u578b\u540e\u6682\u65f6\u89e3\u51b3\u4e86\u62a5\u9519.
\n\u4f46\u662f\u5f53\u63a5\u53e3\u901a\u8fc7 response_model \u6307\u5b9a\u4e86\u6570\u636e\u6a21\u578b(\u975e\u8868\u6a21\u578b)\u4f5c\u4e3a\u8fd4\u56de\u7c7b\u578b\u65f6, \u53c8\u51fa\u73b0\u4e86\u62a5\u9519, \u62a5\u9519\u63d0\u793a\u5982\u4e0b:

`TypeAdapter[typing.Annotated[app.models.teams.TeamPublicWithUser, FieldInfo(annotation=TeamPublicWithUser, required=True)]]` is not fully defined; you should define `typing.Annotated[app.models.teams.TeamPublicWithUser, FieldInfo(annotation=TeamPublicWithUser, required=True)]` and all referenced types, then call `.rebuild()` on the instance.\n

\u6c42\u5404\u4f4d\u5927\u4f6c\u5e2e\u5fd9\u770b\u770b, \u54ea\u91cc\u6709\u95ee\u9898.\npython \u7248\u672c\u7528\u7684\u662f 3.12, \u5177\u4f53\u4ee3\u7801\u5982\u4e0b
\nmodels/users.py

from typing import TYPE_CHECKING, Any, Optional\n\nfrom sqlmodel import Field, Relationship, SQLModel\n\nif TYPE_CHECKING:\n from .teams import Team, TeamPublic\n\n\nclass UserBase(SQLModel):\n username: str = Field(index=True, max_length=255, unique=True)\n email: str | NOne= Field(default=None, index=True, max_length=255)\n is_active: bool = True\n is_superuser: bool = False\n full_name: str | NOne= Field(default=None, max_length=255)\n team_id: int | NOne= Field(default=None, foreign_key=\"team.id\")\n\n\nclass User(UserBase, table=True):\n id: int | NOne= Field(default=None, primary_key=True)\n hashed_password: str\n team: Optional[\"Team\"] = Relationship(back_populates=\"members\")\n\n\nclass UserPublic(UserBase):\n id: int\n\n\nclass UserPublicWithTeam(UserPublic):\n team: Optional[\"TeamPublic\"] = None\n

models/teams.py

from typing import TYPE_CHECKING, Any, List\n\nfrom sqlmodel import Field, Relationship, SQLModel\n\nif TYPE_CHECKING:\n from .users import User, UserPublic\n\n\nclass TeamBase(SQLModel):\n name: str = Field(max_length=255)\n\n\nclass Team(TeamBase, table=True):\n id: int | NOne= Field(default=None, primary_key=True)\n members: List[\"User\"] = Relationship(back_populates=\"team\")\n\n\nclass TeamPublic(TeamBase):\n id: int\n\n\nclass TeamPublicWithUser(TeamPublic):\n members: List[\"UserPublic\"] = []\n

router/users.py

# \u7528\u6237\u76f8\u5173\u63a5\u53e3\nfrom sqlmodel import and_, select\n\nfrom app.api.deps import SessionDep\nfrom app.models.users import (\n User,\n UserPublicWithTeam,\n)\nfrom app.new_router import new_router\n\nrouter = new_router(prefix=\"/users\", tags=[\"\u7528\u6237\u7ba1\u7406\"])\n\n\n@router.get(\"/{id}\", summary=\"\u83b7\u53d6\u7528\u6237\u8be6\u60c5\", response_model=UserPublicWithTeam)\nasync def get_user_api(db: SessionDep, id: int):\n user = db.exec(select(User).where(and_(User.id == id, User.is_active))).first()\n return user\n

router/teams.py

# \u56e2\u961f\u76f8\u5173\u63a5\u53e3\nfrom sqlmodel import select\n\nfrom app.api.deps import SessionDep\nfrom app.models.teams import (\n Team,\n TeamPublicWithUser,\n)\nfrom app.new_router import new_router\n\nrouter = new_router(prefix=\"/teams\", tags=[\"\u56e2\u961f\u7ba1\u7406\"])\n\n\n@router.get(\"/{id}\", summary=\"\u83b7\u53d6\u56e2\u961f\u8be6\u60c5\", response_model=TeamPublicWithUser)\nasync def get_team_api(db: SessionDep, id: int):\n team = db.exec(select(Team).where(Team.id == id)).first()\n return team\n

\u4e3a\u4ec0\u4e48\u8fd8\u8981\u505a\u4e00\u4e2a MCP Filesystem \u7684 Python \u5b9e\u73b0\uff1f

\u76ee\u524d MCP \u751f\u6001\u91cc\uff0c\u5f88\u591a\u80fd\u529b\u662f Node.js \u5b9e\u73b0\u7684\uff0cPython \u8fd9\u8fb9\u8981\u4e48\u53ea\u662f\u5c01\u88c5\u4e00\u5c42\uff0c\u8981\u4e48\u504f\u540c\u6b65\u963b\u585e\u3002\u5bf9\u90a3\u79cd\u201c\u6211\u5c31\u60f3\u5199\u4e2a async \u5c0f\u5de5\u5177\u201d\u7684\u573a\u666f\uff0c\u5176\u5b9e\u4e0d\u592a\u53cb\u597d\uff1a

\n
• \u6574\u6761\u94fe\u8def\u6700\u597d\u90fd\u662f async I/O\uff0c\u8ddf\u4e3b\u6d41\u751f\u6001\u5bf9\u9f50
\n
• \u4e0d\u60f3\u4e3a\u4e86\u8fd9\u70b9\u80fd\u529b\u989d\u5916\u5f15\u4e00\u5806\u4f9d\u8d56
\n
• \u8fd4\u56de\u7ed3\u6784\u8981\u89c4\u6574\uff0c\u4eba\u80fd\u770b\u61c2\u8fd8\u4e0d\u591f\uff0cAI \u4e5f\u5f97\u8bfb\u5f97\u8212\u670d
\n

\u6240\u4ee5\u5728\u81ea\u5df1\u7684 MCP \u5de5\u5177\u91cc\uff0c\u57fa\u4e8e\u5b98\u65b9\u534f\u8bae\u641e\u4e86\u4e00\u4e2a\u300c\u7cbe\u7b80\u7248\u300d\u5b9e\u73b0\uff0c\u4e3b\u8981\u7279\u70b9\uff1a

\n
• \u5f02\u6b65 I/O \u5b9e\u73b0\uff1a\u63a5\u53e3\u5168\u90e8\u662f async\uff0c\u65b9\u4fbf\u76f4\u63a5\u63a5\u5230\u73b0\u6709\u7684 async Web \u6846\u67b6 / \u4efb\u52a1\u7cfb\u7edf
\n
• \u65e0\u7b2c\u4e09\u65b9\u4f9d\u8d56\uff08\u57fa\u7840\u64cd\u4f5c\uff09\uff1a\u53ea\u7528\u6807\u51c6\u5e93\uff0c\u9002\u5408\u53d7\u9650\u73af\u5883\u6216\u8005\u7b80\u5355\u811a\u672c\u9879\u76ee
\n
• \u7ed3\u6784\u4f53\u53cb\u597d\uff1a\u8fd4\u56de\u4f53\u8bbe\u8ba1\u5f97\u6bd4\u8f83\u89c4\u6574\uff0c\u65b9\u4fbf\u540e\u9762\u518d\u5c01\u88c5\u6210 dataclass / Pydantic \u7b49
\n
• \u4ee3\u7801\u7b80\u5355\u6709\u6ce8\u91ca\uff1a\u903b\u8f91\u523b\u610f\u5199\u5f97\u76f4\u767d\uff0c\u6ca1\u6709\u505a\u91cd\u5ea6\u4f18\u5316\uff0c\u66f4\u9002\u5408 fork \u4e0b\u6765\u6309\u81ea\u5df1\u9700\u6c42\u9b54\u6539
\n

\u53ef\u4ee5\u81ea\u884c TODO / \u4e8c\u6b21\u5f00\u53d1\u7684\u65b9\u5411\u5305\u62ec\uff1a

\n
• \u66f4\u4e25\u683c\u7684\u8bfb\u5199\u4fdd\u8bc1\u4e0e\u9519\u8bef\u5904\u7406\uff08\u91cd\u8bd5\u3001\u9650\u6d41\u3001\u5e42\u7b49\u7b49\uff09
\n
• \n\u76ee\u5f55\u8fb9\u754c\u7b56\u7565\u8c03\u6574\uff1a
  \n
  • \u76ee\u524d\u901a\u8fc7 DATA_DIR \u505a\u76ee\u5f55\u6c99\u7bb1\uff0c\u9632\u6b62\u8d8a\u754c\u8bbf\u95ee\uff0c\u907f\u514d\u4e00\u4e0a\u6765\u5c31\u6478\u6574\u4e2a\u78c1\u76d8
  \n
  • \u4e5f\u53ef\u4ee5\u6839\u636e\u4e1a\u52a1\u9700\u8981\u53bb\u6389\u6216\u66ff\u6362\u4e3a\u81ea\u5b9a\u4e49\u6839\u76ee\u5f55\u3001\u751a\u81f3\u5f00\u653e\u6574\u76d8\u8bbf\u95ee
  \n
  \n
\n

Github \u9879\u76ee\u5730\u5740\uff1a

mcp-filesystem-python

\u6280\u672f\u542b\u91cf\u4e0d\u591a\uff0c\u7eaf\u7cb9\u6709\u76f8\u540c\u9700\u6c42\u7684\u4eba\u53ef\u4ee5\u4e0d\u7528\u518d\u601d\u8003\u592a\u591a\uff0c\u76f4\u63a5\u7528\uff0c\u4e5f\u53ef\u4ee5\u5f53\u4f5c\u7ec4\u4ef6\u62fc\u63a5\u5230\u5176\u4ed6\u5730\u65b9\u3002

Neovim \u91c7\u7528 nvim-cmp + pyright \u8fdb\u884c Python \u4ee3\u7801\u68c0\u67e5\u3001\u8865\u5168\uff0c\u4f46\u662f\u5bf9\u4e8e\u590d\u6742\u7684\u4ee3\u7801\uff0c\u8b6c\u5982\u81ea\u5b9a\u4e49\u7c7b\uff0c\u7ee7\u627f\u81ea\u7236\u7c7b\u7684\u4e00\u4e9b\u5c5e\u6027\u3001\u65b9\u6cd5\u5c31\u68c0\u6d4b\u4e0d\u5230\uff0c\u63d0\u793a\u9519\u8bef\u3002\u5982\u56fe\uff1a\n\n\u81ea\u5b9a\u4e49\u7c7b MoneyAgent \u4e2d\u6709\u7ee7\u627f\u81ea\u7236\u7c7b CellAgent \u7684\u5bf9\u8c61 cell \uff0ccell \u5177\u6709\u5c5e\u6027 neighborhood \uff0cpyright \u5b8c\u5168\u610f\u8bc6\u4e0d\u5230\u3002\n\u8bf7\u6559\uff1a\n1 \u3001\u662f\u4e0d\u662f\u6709\u8bbe\u7f6e\u65b9\u6cd5\u53ef\u4ee5\u6539\u8fdb\u3001\u63d0\u5347\n2 \u3001\u9664\u4e86 pyright \u6216\u8005 nvim-cmp \uff0c\u662f\u4e0d\u662f\u8fd8\u6709\u66f4\u597d\u7684\u8865\u5168\u5de5\u5177\uff1f\n\u8c22\u8c22\uff0c\u8bf7\u5404\u4f4d V \u53cb\u4e0d\u541d\u8d50\u6559\u3002

pip \u9ed8\u8ba4\u8fde\u63a5\u7684\u662f\u4f4d\u4e8e\u6d77\u5916\u7684 PyPI \u5b98\u65b9\u6e90\uff0c\u7f51\u7edc\u5ef6\u8fdf\u7684\u65f6\u5019\u4e0b\u8f7d\u901f\u5ea6\u975e\u5e38\u6162\u3002

\u6240\u4ee5\uff0c\u76f4\u63a5\u5199\u4e86\u4e00\u4e2a pip fast check pypi

\u76f4\u63a5\u5b89\u88c5

\n

pip install -U pip-fc

\n

\u76f4\u63a5\u8fd0\u884c

\n

pip-fc \u6216\u8005 python -m pip-fc

\n

\u4e3b\u8981\u529f\u80fd

\n
• \u81ea\u52a8\u9009\u62e9\u6700\u5feb\u7684\u955c\u50cf\u6e90\uff1a\u6d4b\u8bd5\u591a\u4e2a\u56fd\u5185\u955c\u50cf\u6e90\uff0c\u81ea\u52a8\u9009\u51fa\u54cd\u5e94\u6700\u5feb\u7684\u6e90\u3002
\n
• \u652f\u6301 Python 2.7 \u548c 3.x\uff1a\u65e0\u8bba\u4f60\u4f7f\u7528\u7684\u662f\u54ea\u4e2a\u7248\u672c\u7684 Python \uff0c\u90fd\u80fd\u517c\u5bb9\u3002
\n
• \u7b80\u5355\u7684\u547d\u4ee4\u884c\u64cd\u4f5c\uff1a\u65e0\u9700\u624b\u52a8\u914d\u7f6e\uff0c\u53ea\u9700\u6267\u884c\u7b80\u5355\u547d\u4ee4\uff0c\u5c31\u80fd\u5207\u6362\u5230\u6700\u5feb\u7684\u955c\u50cf\u6e90\u3002
\n
• \u652f\u6301\u5168\u5c40\u914d\u7f6e\uff1a\u955c\u50cf\u6e90\u8bbe\u7f6e\u652f\u6301\u5168\u5c40\u914d\u7f6e\uff0c\u751a\u81f3\u53ef\u4ee5\u5728 Docker \u5bb9\u5668\u4e2d\u4f7f\u7528\u3002
\n

\u4f7f\u7528 pdm \u7684\u9879\u76ee\u4eca\u5929\u7a81\u7136\u6784\u5efa\u5931\u8d25\u4e86\uff0c\u53bb\u5b98\u65b9 issue \u641c\u4e86\u4e0b\uff0c\u53d1\u73b0\u662f\u56e0\u4e3a hishel \u5347\u7ea7 1.0.0 \u5bfc\u81f4\u4e0d\u517c\u5bb9\u3002

\u6211\u7684 Dockerfile \u89e3\u51b3\u65b9\u6848\u5982\u4e0b

pip install --no-cache-dir -U \"pdm==2.25.9\" \"hishel<1.0.0\"

https://github.com/pdm-project/pdm/issues/3657

https://github.com/karpetrosyan/hishel

\u722c\u866b\u4e2d\u53ef\u80fd\u9047\u5230\u4f60\u4ec0\u4e48\u53c2\u6570\u90fd\u62fc\u5bf9\u4e86\u4f46\u662f\u8fd8\u662f\u6ca1\u6cd5\u8bf7\u6c42\u7684\u60c5\u51b5\uff0c\u6ca1\u9519\uff0c\u4f60\u53ef\u80fd\u662f\u9047\u5230\u4e86 JA3 \u6307\u7eb9\uff0c\u4ece TLS \u5c42\u9762\u5c31\u5df2\u7ecf\u628a\u4f60\u8bc6\u522b\u4e86\uff0c\u6240\u4ee5\u518d\u505a\u4ec0\u4e48\u4e5f\u662f\u65e0\u7528\uff0c\u653e\u5fc3\u4ed6\u80fd\u5e2e\u4f60\u3002\u867d\u7136\u53ea\u662f\u5957\u58f3\u4f46\u662f\u80fd\u8ba9 Python \u76f4\u63a5\u8c03\u7528\u4e5f\u662f\u5f88\u597d\u4e86\uff0c\u4f60\u4e5f\u53ef\u4ee5\u57fa\u4e8e\u6b64\u65b9\u6848\u5b9e\u73b0\u66f4\u9ad8\u7ea7\u7684\u8bf7\u6c42\u5e93\u3002

Github: https://github.com/rev1si0n/ja3

\u516c\u53f8\u4e00\u76f4\u7528 jetson orin nano \u548c NX \u8bbe\u5907\u5f00\u53d1\u6848\u4f8b\uff0c\u4f46\u662f\u9700\u8981\u4f7f\u7528 Python + Qt \uff0c\u76ee\u524d\u4e00\u76f4\u91c7\u7528 miniconda3 \u6765\u5b89\u88c5 PyQt5 \u8fd9\u4e2a\u5e93\uff0c\u65e0\u6cd5\u4f7f\u7528 pip \u76f4\u63a5\u7f16\u8bd1\u5b89\u88c5\uff0c\u6211\u8ba4\u4e3a\u6027\u80fd\u662f\u5b8c\u5168\u591f\u7684\uff0c8G \u5185\u5b58\uff0c\u4f46\u662f\u7f16\u8bd1\u5b89\u88c5 Qt \u8fd8\u662f\u5931\u8d25\u3002\u4f46\u662f miniconda \u6709\u70b9\u81c3\u80bf\uff0c\u80fd\u7528\u662f\u80fd\u7528\uff0c\u4f46\u662f\u4e0d\u597d\u7528\uff0cX86 \u8bbe\u5907\u90fd\u6539\u6210\u4e86 uv \u7edf\u4e00\u7ba1\u7406\uff0c\u56e0\u4e3a\u5f00\u53d1\u7684\u6848\u4f8b\u6d89\u53ca\u5230\u73af\u5883\u91cd\u7f6e\u7b49\uff0cuv \u548c poetry \u8fd9\u79cd\u6bd4\u8f83\u597d\uff0c\u53ef\u4ee5 100%\u8fd8\u539f\u4f9d\u8d56\u3002conda \u5bfc\u51fa\u7684 yaml \u6709\u65f6\u5019\u4e5f\u662f\u4f1a\u62bd\u98ce\uff0c\u603b\u662f\u6f0f\u4e2a\u5305\u4e4b\u7c7b\u7684\u3002\u4f46\u662f\u53c8\u79bb\u4e0d\u5f00 PyQt \u3002\u6240\u4ee5 \u8fd8\u6709\u62db\u5417\uff1f apt \u5b89\u88c5\u7684\u5168\u5c40\uff0c\u597d\u50cf\u4e0d\u592a\u517c\u5bb9\u3002

\u5728 arm64 \u4e0a\u4f7f\u7528 pip \u5b89\u88c5\u7f16\u8bd1 PyQt \u5e93\uff0c100%\u5931\u8d25\uff0cpip \u4e5f\u6ca1\u6709 arm64 \u8bbe\u5907\u9884\u7f16\u8bd1\u597d\u7684 wheel.\nconda \u5728 arm64 \u8bbe\u5907\u63d0\u4f9b\u9884\u7f16\u8bd1\u597d\u7684 pyqt \u5e93\uff0cconda install \u662f\u53ef\u4ee5\u76f4\u63a5\u57fa\u4e8e\u5f53\u524d Python \u7248\u672c\u5b89\u88c5\u7684\uff0c\u4e0d\u9700\u8981\u7f16\u8bd1.

\u80cc\u666f

\u8fd9\u4e24\u5929\u6784\u601d\u4e86\u4e00\u4e2a\u722c\u866b\u6846\u67b6\uff0c\u5bf9\u5916\u63d0\u4f9b API \u521b\u5efa\u722c\u866b\u4efb\u52a1\uff0c\u7136\u540e\u5185\u90e8\u7684\u961f\u5217\u4f1a\u8fdb\u884c\u722c\u866b\u7684\u6d88\u8d39\u3002\u53ea\u9700\u8981\u5b9e\u73b0\u6570\u636e\u7684\u89e3\u6790\u63a5\u53e3\u5c31\u80fd\u5feb\u901f\u7f16\u5199\u722c\u866b\u3002\u975e\u5e38\u9002\u5408\u9700\u8981\u5229\u7528 AI \u5feb\u901f\u751f\u6210\u722c\u866b\u4ee3\u7801\u7684\u56e2\u961f\u3002\n

\u8fd9\u4e2a\u6846\u67b6\u5bf9\u5916\u63d0\u4f9b\u4e86 API \u63a5\u53e3\u6765\u521b\u5efa\uff0c\u975e\u5e38\u4fbf\u5229\u3002\u76ee\u524d\u7684\u8bbe\u8ba1\u601d\u8def\u5c31\u662f\u53ea\u9700\u8981\u5b9e\u73b0\u4e00\u4e2a parse \u63a5\u53e3\uff0c\u5c31\u884c\u4e86\uff0c\u65b9\u4fbf\u540e\u7eed AI \u7684\u4ecb\u5165\u3002

\u540e\u7eed\u5f00\u53d1\u8ba1\u5212

\n
1. \u5f00\u653e AI \u63a5\u53e3\uff0c\u901a\u8fc7 AI \u81ea\u52a8\u751f\u6210\u722c\u866b\u4ee3\u7801
\n
2. \u96c6\u6210\u57fa\u4e8e redis \u7684\u4efb\u52a1\u961f\u5217
\n
3. \u5b9e\u73b0\u5bf9\u5916\u8f93\u51fa\u7684\u63a5\u53e3\u5c42\uff0c\u4f8b\u5982\u722c\u866b\u7ed3\u679c\u8f6c\u50a8\u5230 mysql \u7b49\u3002
\n

\u76ee\u524d\u8fd9\u662f\u4e00\u4e2a\u975e\u5e38\u7b80\u5355\u6e05\u6670\u7684\u9879\u76ee\uff0c\u5e0c\u671b\u548c\u611f\u5174\u8da3\u7684\u670b\u53cb\u5171\u5efa\u8fd9\u4e2a\u9879\u76ee\uff0c\u63d0\u5347\u5927\u5bb6\u7684\u6280\u672f\u5f71\u54cd\u529b\uff0c\u6216\u8bb8\u5bf9\u627e\u8fdc\u7a0b\u5de5\u4f5c\u4e5f\u662f\u6709\u5e2e\u52a9\u7684\u3002

\u9879\u76ee\u5730\u5740\uff1a\nhappytaoer/web-craft: A Python-based modular web scraping framework focused on efficient single URL crawling, supporting asynchronous processing, API services, and highly customizable spider modules.

\u6700\u8fd1\u5f00\u53d1\u4e00\u4e2a\u65b0\u9879\u76ee \u6280\u672f\u6808\u662f python \u670d\u52a1\u6253\u5305 docker \u955c\u50cf\u653e\u5230\u4e86\u4e91\u4e0a\u90e8\u7f72 \u90e8\u7f72\u4e0a\u4e4b\u540e\u6ca1\u6709\u95ee\u9898 \u4f46\u662f\u9694\u4e86\u4e00\u6bb5\u65f6\u95f4\u5c31\u4f1a\u6709\u8d85\u65f6\u672a\u54cd\u5e94\u7684\u95ee\u9898 \u53ea\u8981\u91cd\u542f\u5bb9\u5668\u5c31\u7acb\u9a6c\u80fd\u6b63\u5e38\u8bbf\u95ee\u4e86 \u8fd9\u4e2a\u670d\u52a1\u672c\u8eab\u4e5f\u6ca1\u6709\u591a\u5927\u8bbf\u95ee\u91cf \u8bf7\u95ee\u5404\u4f4d\u4e00\u4e0b\u6392\u67e5\u601d\u8def \u4ee5\u53ca\u5927\u6982\u54ea\u91cc\u4f1a\u6709\u95ee\u9898

\u6982\u8ff0

\u6700\u8fd1\u5728\u68b3\u7406 iterator \uff0c\u4e0d\u5f97\u4e0d\u8bf4\uff0c \u5373\u4f7f\u81ea\u5df1\u5199\u4e86\u5f88\u591a\u5e74\u7684\u4ee3\u7801\uff0c\u6211\u4ecd\u7136\u6ca1\u6709\u5728\u5b9e\u9645\u5e94\u7528\u4e2d\u770b\u5230\u81ea\u5b9a\u4e49\u7684\u8fed\u4ee3\u5668\u3002\u5373\u4f7f\u8bfb\u4e86\u5f88\u591a\u4e66\uff0c\u4f46\u662f\u8fd9\u4e9b\u4e66\u4e2d\u7684\u793a\u4f8b\u5927\u591a\u662f\u6ee5\u7afd\u5145\u6570\uff0c\u4e0d\u5177\u5907\u5b9e\u9645\u5e94\u7528\u610f\u4e49\u3002\u6240\u4ee5\u987a\u7740\u7f51\u7ebf\u722c\u4e0a V \u7ad9\u8bf7\u6559\u5404\u4f4d\u3002

\u53ef\u8fed\u4ee3\u5bf9\u8c61 & \u8fed\u4ee3\u5668\u5b9a\u4e49

\u53ef\u8fed\u4ee3\u5bf9\u8c61

\n

\u5982\u679c\u4e00\u4e2a\u5bf9\u8c61\u5b9a\u4e49\u4e86 __iter__() \u65b9\u6cd5\u6216\u5b9a\u4e49\u4e86 __getitem__() \u65b9\u6cd5\uff0c\u90a3\u4e48\u8fd9\u6837\u7684\u5bf9\u8c61\u79f0\u4e3a\u53ef\u8fed\u4ee3\u5bf9\u8c61(iterable)\u3002

\n

\u8fed\u4ee3\u5668

\n

\u5982\u679c\u4e00\u4e2a\u5bf9\u8c61\u5b9a\u4e49\u4e86 __iter__() \u65b9\u6cd5\u548c __next__() \u65b9\u6cd5\uff0c\u90a3\u4e48\u8fd9\u6837\u7684\u5bf9\u8c61\u79f0\u4e3a\u8fed\u4ee3\u5668(iterator)\u3002

\n

\u6ce8\uff1a

1.\u540e\u7eed\u7684\u8ba8\u8bba\u90fd\u662f\u57fa\u4e8e\u4ee5\u4e0a\u4e24\u4e2a\u5b9a\u4e49\u3002

2.\u56e0\u8fed\u4ee3\u5668\u5e38\u548c\u53ef\u8fed\u4ee3\u5bf9\u8c61\u7ed3\u5408\u4f7f\u7528\uff0c\u6545\u5f15\u5982\u53ef\u8fed\u4ee3\u5bf9\u8c61\u8fd9\u4e00\u6982\u5ff5\uff0c\u4f46\u8fed\u4ee3\u5668\u7684\u6982\u5ff5\u5148\u4e8e\u751f\u6210\u5668(generator)\uff0c\u5728\u540e\u7eed\u7684\u8ba8\u8bba\u4e2d\u8bf7\u52ff\u6d89\u53ca\u751f\u6210\u5668\u3002

\u4e00\u4e9b\u793a\u4f8b

\u793a\u4f8b 1

python 3 \u7684 range() \u662f\u4e00\u4e2a\u53ef\u8fed\u4ee3\u5bf9\u8c61\uff0c\u5176\u5b9e\u73b0\u4f7f\u7528\u4e86\u8fed\u4ee3\u5668\u3002\u4f7f\u7528\u8fed\u4ee3\u5668\u540e\u4e0d\u662f\u76f4\u63a5\u751f\u6210\u5217\u8868\uff0c\u8282\u7701\u4e86\u5185\u5b58\uff0c\u4f53\u73b0\u4e86\u8fed\u4ee3\u5668\u7684\u5e94\u7528\u610f\u4e49\u3002

\u793a\u4f8b 2

\u300a Learn Python Programming(4th)\u300b \u7b2c 246 \u9875\uff1a

class OddEven:\n def __init__(self, data):\n self._data = data\n self.indexes = list(range(0, len(data), 2)) + list(range(1, len(data), 2))\n\n def __iter__(self):\n return self\n\n def __next__(self):\n if self.indexes:\n return self._data[self.indexes.pop(0)]\n raise StopIteration\n\n\n# Testing the OddEven class\noddeven = OddEven(\"0123456789\")\nprint(\"\".join(c for c in oddeven)) # 0246813579\n\noddeven = OddEven(\"ABCD\") # or manually...\nit = iter(oddeven) # this calls oddeven.__iter__ internally\nprint(next(it)) # A\nprint(next(it)) # C\nprint(next(it)) # B\nprint(next(it)) # D\n\n

\u8be5\u793a\u4f8b\u867d\u7136\u521b\u5efa\u4e86\u4e00\u4e2a\u8fed\u4ee3\u5668\uff0c\u4f46\u5c31\u529f\u80fd\u800c\u8a00\u5176\u5b9e\u5c31\u662f\u201c\u5c06\u5947\u6570\u4f4d\u7f6e\u7684\u5b57\u7b26\u653e\u5728\u524d\u534a\u6bb5\uff0c\u5c06\u5076\u6570\u4f4d\u7f6e\u7684\u5b57\u7b26\u653e\u5728\u540e\u534a\u6bb5\u201d\uff0c\u5b8c\u5168\u6ca1\u6709\u5fc5\u8981\u4f7f\u7528\u8fed\u4ee3\u5668\u3002\u5173\u4e8e\u8fed\u4ee3\u5668\u7684\u5b9e\u529b\uff0c\u672c\u4eba\u770b\u5230\u7684\u5927\u591a\u662f\u8fd9\u6837\u7684\u2014\u2014\u6beb\u65e0\u5b9e\u9645\u5e94\u7528\u610f\u4e49\uff0c\u4ee4\u4eba\u6df1\u6076\u75db\u7edd\uff01

\u95ee\u9898

\u95ee\u9898 1

PEP 234 \u4e2d\u5199\u5230 iterator \u7684 virtues \u6709:

\n

\n
1. It provides an extensible iterator interface.
\n
2. It allows performance enhancements to list iteration.
\n
3. It allows big performance enhancements to dictionary iteration.
\n
4. It allows one to provide an interface for just iteration without pretending to provide random access to elements.
\n
5. It is backward-compatible with all existing user-defined classes and extension objects that emulate sequences and mappings, even mappings that only implement a subset of {__getitem__, keys, values, items}.
\n
6. It makes code iterating over non-sequence collections more concise and readable.
\n

\n

\u4e2d\u8bd1\u7248\uff1a

\n

\u5982\u679c\u5305\u542b\u8be5\u63d0\u6848\u7684\u6240\u6709\u90e8\u5206\uff0c\u5219\u4f1a\u4ee5\u4e00\u81f4\u4e14\u7075\u6d3b\u7684\u65b9\u5f0f\u89e3\u51b3\u8bb8\u591a\u95ee\u9898\u3002\u5176\u4e3b\u8981\u4f18\u70b9\u5305\u62ec\u4ee5\u4e0b\u56db\u70b9\u2014\u2014\u4e0d\uff0c\u4e94\u70b9\u2014\u2014\u4e0d\uff0c\u516d\u70b9

\n

\n
1. \u5b83\u63d0\u4f9b\u4e86\u4e00\u4e2a\u53ef\u6269\u5c55\u7684\u8fed\u4ee3\u5668\u63a5\u53e3\u3002
\n
2. \u5b83\u5141\u8bb8\u5bf9\u5217\u8868\u8fed\u4ee3\u8fdb\u884c\u6027\u80fd\u4f18\u5316\u3002
\n
3. \u5b83\u5141\u8bb8\u5bf9\u5b57\u5178\u8fed\u4ee3\u8fdb\u884c\u5927\u5e45\u5ea6\u6027\u80fd\u63d0\u5347\u3002
\n
4. \u5b83\u5141\u8bb8\u4e3a\u4ec5\u8fed\u4ee3\u63d0\u4f9b\u63a5\u53e3\uff0c\u800c\u65e0\u9700\u5047\u88c5\u63d0\u4f9b\u5bf9\u5143\u7d20\u7684\u968f\u673a\u8bbf\u95ee\u3002
\n
5. \u5b83\u4e0e\u6240\u6709\u73b0\u6709\u7684\u7528\u6237\u5b9a\u4e49\u7c7b\u548c\u6a21\u62df\u5e8f\u5217\u548c\u6620\u5c04\u7684\u6269\u5c55\u5bf9\u8c61\u5411\u540e\u517c\u5bb9\uff0c\u5373\u4f7f\u662f\u4ec5\u5b9e\u73b0\u4e86 {__getitem__, keys, values, items} \u5b50\u96c6\u7684\u6620\u5c04\u3002
\n
6. \u5b83\u4f7f\u904d\u5386\u975e\u5e8f\u5217\u96c6\u5408\u7684\u4ee3\u7801\u66f4\u52a0\u7b80\u6d01\u6613\u8bfb\u3002
\n

\n

\u4e0a\u9762\u6240\u5217\u51fa\u7684\u4f18\u70b9\u8f83\u62bd\u8c61\uff0c\u5404\u4f4d\u80fd\u5426\u63d0\u4f9b\u4e00\u4e9b\u5177\u4f53\u7684\u4f8b\u5b50\uff1f

\u95ee\u9898 2

\u5404\u4f4d\u5728\u5b9e\u9645\u5e94\u7528\u4e2d\u662f\u5426\u81ea\u5df1\u5b9e\u73b0\u8fc7\u8fed\u4ee3\u5668\uff1f\u5982\u679c\u6709\u9ebb\u70e6\u63d0\u4f9b\u4e00\u4e9b\u4f8b\u5b50\u3002

\u53c2\u8003\u8d44\u6599

[1] Python Document Glossary \uff0citerator\uff1a https://docs.python.org/3/glossary.html#term-iterator

[2] PEP 234 \u2013 Iterators\uff1a https://peps.python.org/pep-0234

[3] PEP 234 \u2013 \u8fed\u4ee3\u5668: https://peps.pythonlang.cn/pep-0234/

\u5982\u679c\u627e\u4eba\u4e00\u5bf9\u4e00\u6765\u5b66\u4e60\u662f\u4e0d\u662f\u66f4\u597d

\u7248\u672c\u53f7\u5341\u5206\u7279\u522b\uff0c\u521a\u597d\u5c31\u662f \ud835\udf0b

\u5df2\u7ecf\u6709\u4eba\u6539\u4e86\u4ee3\u7801\uff0c\u4e3a\u8fd9\u4e2a\u7248\u672c\u6dfb\u52a0\u4e13\u95e8\u7684\u540d\u79f0\uff1a\ud835\udf0bthon

gh-119535: Support \ud835\udf0bthon in Python 3.14 venvs

\u91cd\u5927\u66f4\u65b0\uff1a

\n
1. \u81ea\u7531\u7ebf\u7a0b\uff08 PEP 779 \uff09\u7279\u6027\u5df2\u7a33\u5b9a\u652f\u6301\u3002\u4e4b\u524d\u5728 Python 3.13 \u4e2d\uff0c\u8fd9\u4e00\u7279\u6027\u9700\u8981\u663e\u5f0f\u6307\u5b9a\u7f16\u8bd1\u9009\u9879\u3002\u4e0d\u8fc7\uff0c\u5168\u5c40\u89e3\u91ca\u5668\u9501\uff08 GIL \uff09\u4f9d\u7136\u662f\u53ef\u9009\u7684\uff0c\u800c\u662f\u5426\u8981\u5f7b\u5e95\u79fb\u9664 GIL \u4ecd\u5728\u8ba8\u8bba\u4e2d\uff08 PEP 703 \uff09\uff1b
\n
2. \u591a\u89e3\u91ca\u5668\uff08 PEP 734 \uff09\u7279\u6027\u5df2\u652f\u6301\u3002\u73b0\u5728\uff0c\u53ef\u4ee5\u5728\u540c\u4e00\u4e2a\u8fdb\u7a0b\u4e2d\u8fd0\u884c\u591a\u4e2a\u89e3\u91ca\u5668\uff0c\u6bcf\u4e2a\u89e3\u91ca\u5668\u90fd\u6709\u72ec\u7acb\u7684 GIL \u3002\u4f5c\u4e3a Python \u5e76\u884c\u7f16\u7a0b\u4e2d\u51cf\u5c11\u5bf9 multiprocessing \u4f9d\u8d56\u7684\u53c8\u4e00\u6b65\uff0c\u8fd8\u6dfb\u52a0\u4e86 InterpreterPoolExecutor \u6765\u5e2e\u52a9\u7ba1\u7406\u591a\u4e2a\u89e3\u91ca\u5668\u7684\u5e76\u884c\uff1b
\n
3. \u5b98\u65b9\u53d1\u5e03\u7684 macOS \u548c Windows \u7248\u4e8c\u8fdb\u5236\u73b0\u5df2\u5305\u542b\u00a0\u5b9e\u9a8c\u6027\u00a0\u7684\u5373\u65f6\u7f16\u8bd1\uff08 JIT \uff09\u5668\uff08 PEP 744 \uff09\u3002\u8be5 JIT \u7f16\u8bd1\u5668\u5c1a\u5904\u4e8e\u65e9\u671f\u5f00\u53d1\u9636\u6bb5\uff0c\u6027\u80fd\u8868\u73b0\u5b58\u5728\u6ce2\u52a8\uff1a\u542f\u7528\u540e\u6839\u636e\u5de5\u4f5c\u8d1f\u8f7d\u4e0d\u540c\uff0c\u53ef\u80fd\u4ea7\u751f 10% \u7684\u6027\u80fd\u4e0b\u964d\u81f3 20% \u7684\u6027\u80fd\u63d0\u5347\uff1b
\n
4. \u6a21\u677f\u5b57\u7b26\u4e32\uff08 t-string \uff0cPEP 750 \uff09\u652f\u6301\u3002\u8fd9\u4e00\u529f\u80fd\u4e3b\u8981\u662f\u5bf9 f-string \u7684\u8865\u5145\uff0c\u5141\u8bb8\u4ece\u7c7b\u4f3c\u5b57\u7b26\u4e32\u5b57\u9762\u91cf\u7684\u5199\u6cd5\u76f4\u63a5\u521b\u5efa\u4e00\u4e2a\u5b57\u7b26\u4e32\u6a21\u677f\u5bf9\u8c61\uff1b
\n
5. \u589e\u91cf\u5f0f\u5783\u573e\u56de\u6536\u3002\u5faa\u73af\u5783\u573e\u56de\u6536\u5668\u73b0\u5728\u91c7\u7528\u589e\u91cf\u5f0f\u5904\u7406\u3002\u8fd9\u610f\u5473\u7740\u5bf9\u4e8e\u8f83\u5927\u7684\u5806\u5185\u5b58\uff0c\u6700\u5927\u6682\u505c\uff08 Stop The World \uff09\u65f6\u95f4\u5c06\u51cf\u5c11\u4e00\u4e2a\u6570\u91cf\u7ea7\u6216\u66f4\u591a\uff1b
\n
6. \u4ea4\u4e92\u5f0f Shell \u652f\u6301\u8bed\u6cd5\u9ad8\u4eae\u3002\u9664\u975e\u663e\u5f0f\u7981\u7528\uff0c\u5426\u5219\u4ece Python 3.14 \u8d77\uff0cPython \u4ea4\u4e92\u5f0f\u73af\u5883\uff08 RHEL \uff09\u5c06\u9ed8\u8ba4\u5728\u7ec8\u7aef\u4e2d\u6e32\u67d3\u4ee3\u7801\u9ad8\u4eae\u3002\u6b64\u5916\uff0cShell \u8fd8\u589e\u52a0\u4e86\u5bf9 import ___ \u7684\u81ea\u52a8\u8865\u5168\uff1b
\n
7. asyncio \u5185\u7701\u80fd\u529b\u3002\u53ef\u4ee5\u4f7f\u7528 python\u00a0-m\u00a0asyncio\u00a0<ps|pstree>\u00a0<PID> \u529f\u80fd\u6765\u68c0\u67e5\u6b63\u5728\u8fd0\u884c\u7684\u5f02\u6b65 Python \u7a0b\u5e8f\u7684 async task \u6811\u3002
\n

\u66f4\u591a\u53ef\u5728\u8fd9\u91cc\u770b\u5230\uff1a https://docs.python.org/zh-cn/3.14/whatsnew/3.14.html

https://github.com/allmonday/fastapi-router-viz

\u5bf9\u4e8e\u9075\u5faa er \u6a21\u578b\u6765\u6784\u5efa\u89c6\u56fe\u6570\u636e\u7684\u9879\u76ee\uff0cfastapi-router-viz \u53ef\u4ee5\u4e3a\u4e86\u89e3 api \u8fd4\u56de\u7c7b\u578b\u95f4\u7684\u5173\u7cfb\u63d0\u4f9b\u5feb\u901f\uff0c\u76f4\u63a5\uff0c\u53ef\u4ea4\u4e92\u7684\u56fe\u6548\u679c\u3002

\u53ef\u4ee5\u901a\u8fc7\u70b9\u51fb\u8282\u70b9\u9ad8\u4eae\u5168\u90e8\u4e0a\u4e0b\u6e38\u94fe\u8def\uff0c \u4e86\u89e3 pydantic class \u7684\u4e0a\u4e0b\u6e38\u4f9d\u8d56\u60c5\u51b5

alt \u70b9\u51fb\u67e5\u770b\u8282\u70b9\u7684\u6e90\u4ee3\u7801\uff0c\u6216\u8005\u76f4\u63a5\u5728 vscode \u4e2d\u6253\u5f00

\u6839\u636e class + field name \u6765\u7cbe\u51c6\u5b9a\u4f4d\u6570\u636e\u88ab\u54ea\u4e9b\u9875\u9762/ \u63a5\u53e3\u4f7f\u7528

\u8fd9\u95ee\u9898\u90fd\u51fa\u73b0\u534a\u5e74\u4e86\u3002\u5c1d\u8bd5\u7528 AI \u89e3\u51b3\uff0c\u7ed9\u6211\u8bd5\u4e86\u534a\u5929\u8fd8\u90fd\u662f too many request.

\u6211\u7684\u7f51\u7edc\u662f\u53ef\u4ee5\u7528 gemini cli \u4e4b\u7c7b\u7684\uff0c\u6309\u7406\u8bf4\u5e94\u8be5\u4e0d\u662f\u7f51\u7edc\u95ee\u9898\u554a\u3002

\u7b80\u5355\u6d4b\u8bd5\u4ee3\u7801\u5982\u4e0b\uff1a

import yfinance as yf\nimport pandas as pd\n\n# \u8bbe\u7f6e\u6c38\u8f89\u8d85\u5e02\u7684\u80a1\u7968\u4ee3\u7801\nticker = \"601933.SS\" # .SS \u8868\u793a\u4e0a\u6d77\u8bc1\u5238\u4ea4\u6613\u6240\n\n# \u521b\u5efa yfinance \u5bf9\u8c61\nstock = yf.Ticker(ticker)\n\n# \u83b7\u53d6\u5b9e\u65f6\u4ef7\u683c\uff08\u6700\u65b0\u6536\u76d8\u4ef7\u6216\u5f53\u524d\u4ef7\u683c\uff0c\u89c6\u5e02\u573a\u60c5\u51b5\u800c\u5b9a\uff09\ncurrent_price = stock.history(period=\"1d\")[\"Close\"].iloc[-1]\n\n# \u83b7\u53d6\u80a1\u7968\u7684\u8be6\u7ec6\u4fe1\u606f\uff08\u5305\u62ec\u516c\u53f8\u540d\u79f0\u7b49\uff09\ninfo = stock.info\n\n# \u8f93\u51fa\u7ed3\u679c\nprint(f\"\u6c38\u8f89\u8d85\u5e02 ({ticker}) \u7684\u5b9e\u65f6\u4ef7\u683c: {current_price:.2f} CNY\")\nprint(f\"\u516c\u53f8\u540d\u79f0: {info.get('shortName', '\u672a\u77e5')}\")\n

\u8fd0\u884c\u5c31\u76f4\u63a5\u62a5\u9519\uff1a

 line 11, in <module>\n current_price = stock.history(period=\"1d\")[\"Close\"].iloc[-1]\n ^^^^^^^^^^^^^^^^^^^^^^^^^^\n File \"/opt/anaconda3/lib/python3.12/site-packages/yfinance/utils.py\", line 103, in wrapper\n result = func(*args, **kwargs)\n ^^^^^^^^^^^^^^^^^^^^^\n File \"/opt/anaconda3/lib/python3.12/site-packages/yfinance/base.py\", line 91, in history\n return self._lazy_load_price_history().history(*args, **kwargs)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File \"/opt/anaconda3/lib/python3.12/site-packages/yfinance/utils.py\", line 103, in wrapper\n result = func(*args, **kwargs)\n ^^^^^^^^^^^^^^^^^^^^^\n File \"/opt/anaconda3/lib/python3.12/site-packages/yfinance/scrapers/history.py\", line 178, in history\n data = get_fn(\n ^^^^^^^\n File \"/opt/anaconda3/lib/python3.12/site-packages/yfinance/utils.py\", line 103, in wrapper\n result = func(*args, **kwargs)\n ^^^^^^^^^^^^^^^^^^^^^\n File \"/opt/anaconda3/lib/python3.12/site-packages/yfinance/data.py\", line 364, in get\n return self._make_request(url, request_method = self._session.get, user_agent_headers=user_agent_headers, params=\nparams, timeout=timeout)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n^^^^^^^^^^^^^^^^^^^^^^^^\n File \"/opt/anaconda3/lib/python3.12/site-packages/yfinance/utils.py\", line 103, in wrapper\n result = func(*args, **kwargs)\n ^^^^^^^^^^^^^^^^^^^^^\n File \"/opt/anaconda3/lib/python3.12/site-packages/yfinance/data.py\", line 424, in _make_request\n raise YFRateLimitError()\nyfinance.exceptions.YFRateLimitError: Too Many Requests. Rate limited. Try after a while.\n

\u8fd9\u8981\u600e\u4e48\u89e3\u51b3\u5462\uff0c\u96be\u9053 yahoo \u7684 API \u5c01\u8fd9\u4e48\u4e25\u91cd\u554a\u3002\u6709\u6ca1\u6709\u77e5\u9053\u600e\u4e48\u89e3\u51b3\u7684\u8001\u5144\u554a\uff0c\u611f\u8c22\uff01

\u53d1\u73b0 Celery 4.2.3 \u4e0d\u652f\u6301\u7c7b\u65b9\u6cd5

\u4f8b\u5b50\uff1a

class Demo:\n \u00a0@app.task\n \u00a0@classmethod\n \u00a0def test(cls):\n \u00a0 \u00a0 \u00a0pass\n \n\nschedule \u6ce8\u518c\u4efb\u52a1\uff1a\nDemo.test

\u76ee\u524d\u5df2\u6709 i18n \u5e93\u7684\u75db\u70b9

\n
1. Python \u81ea\u5e26\u7684 gettext \u4e0d\u652f\u6301\u52a8\u6001\u9009\u62e9\u8bed\u8a00 (\u6bd4\u5982 Bot \u9700\u8981\u6839\u636e\u7528\u6237\u9009\u62e9\u8bed\u8a00)
\n
2. \u5176\u4ed6\u5e93\u9700\u8981\u624b\u52a8\u7ef4\u62a4\u7ffb\u8bd1\u5b57\u5178, \u5f00\u53d1\u6548\u7387\u4f4e
\n
3. \u90fd\u4e0d\u652f\u6301 f-string \u8bed\u6cd5, \u5199\u8d77\u6765\u4e0d\u591f\u4f18\u96c5, \u4e0d\u591f\u65b9\u4fbf
\n

\u6240\u4ee5, Easy AI18n \u8bde\u751f\u4e86, \u89e3\u51b3\u4e86\u4ee5\u4e0a\u6240\u6709\u75db\u70b9

Easy AI18n \u662f\u4e00\u4e2a\u73b0\u4ee3\u5316\u7684 Python3 \u56fd\u9645\u5316\u5de5\u5177\u5e93\uff0c\u652f\u6301 AI \u7ffb\u8bd1\u3001\u591a\u7528\u6237\u573a\u666f\u4ee5\u53ca\u5b8c\u6574\u7684\u5b57\u7b26\u4e32\u683c\u5f0f\u5316\u8bed\u6cd5\uff0c\u8ba9\u9879\u76ee\u5168\u7403\u5316\u53d8\u5f97\u66f4\u52a0\u4f18\u96c5\u81ea\u7136\u3002

https://github.com/z-mio/easy-ai18n

\u7279\u70b9

\n
• \u81ea\u52a8\u63d0\u53d6\u5185\u5bb9\u751f\u6210\u7ffb\u8bd1\u5b57\u5178, \u63d0\u9ad8\u5f00\u53d1\u6548\u7387
\n
• \u652f\u6301\u52a8\u6001\u591a\u8bed\u8a00, \u9002\u914d\u591a\u7528\u6237\u573a\u666f
\n
• \u652f\u6301\u6240\u6709\u683c\u5f0f\u5316\u8bed\u6cd5, \u4ee3\u7801\u66f4\u4f18\u96c5
\n
• \u652f\u6301 AI \u7ffb\u8bd1, \u7ffb\u8bd1\u8d28\u91cf\u66f4\u597d
\n

\u5bf9\u6bd4

\u5176\u4ed6 i18n \u5de5\u5177	EasyAI18n
\u9700\u624b\u52a8\u7ef4\u62a4\u7ffb\u8bd1\u5b57\u5178, \u5f00\u53d1\u6548\u7387\u4f4e	\u81ea\u52a8\u63d0\u53d6\u7ffb\u8bd1\u5185\u5bb9\uff0c\u65e0\u9700\u624b\u52a8\u7ef4\u62a4\u6587\u4ef6
\u4ec5\u652f\u6301\u90e8\u5206\u683c\u5f0f\u5316\u8bed\u6cd5	\u5b8c\u5168\u652f\u6301\u6240\u6709\u683c\u5f0f\u5316\u8bed\u6cd5
\u4e0d\u652f\u6301\u5b9e\u65f6\u591a\u8bed\u8a00\u5207\u6362\uff0c\u4e0d\u9002\u7528\u4e8e\u591a\u7528\u6237\u573a\u666f	\u652f\u6301\u9ed8\u8ba4\u8bed\u8a00\u4e0e\u591a\u8bed\u8a00\u5207\u6362\uff0c\u9002\u914d\u591a\u7528\u6237\u73af\u5883

\ud83d\udc65 \u591a\u7528\u6237\u591a\u8bed\u8a00\u573a\u666f\uff08\u5982 Telegram Bot \uff09

\u901a\u8fc7\u81ea\u5b9a\u4e49\u8bed\u8a00\u9009\u62e9\u5668, \u5728\u591a\u7528\u6237\u73af\u5883\u4e2d\u5b9e\u73b0\u52a8\u6001\u8bed\u8a00\u9009\u62e9:

/i18n.py:

from pyrogram.types import Message\nfrom easy_ai18n import EasyAI18n, PostLanguageSelector\n\n\nclass MyPostLanguageSelector(PostLanguageSelector):\n def __getitem__(self, msg: Message):\n lang = msg.from_user.language_code\n return super().__getitem__(lang)\n\n\ni18n = EasyAI18n()\n\n_ = i18n.i18n(post_lang_selector=MyPostLanguageSelector)\n\nif __name__ == \"__main__\":\n i18n.build(target_lang=['en', 'ru'])\n

/bot.py:

@bot.on_message()\nasync def start(__, msg: Message):\n await msg.reply(_[msg](\"Hello, world!\"))\n

RT \uff0c\u8c22\u8c22\u5404\u4f4d\u6709\u7ecf\u9a8c\u7684\u5927\u4f6c

\u5982\u9898\uff0c\u4ee3\u7801\u6b63\u5e38\u80fd\u8dd1\u7684\u3002\u4f46\u662f debug \u6a21\u5f0f pycharm \u4f1a\u81ea\u52a8\u6253\u5370\u5bf9\u8c61\u3002\u7136\u540e\u5c31\u5806\u6808\u6ea2\u51fa\u4e86\uff0c\n\u8fd8\u770b\u4e0d\u5230\u5177\u4f53\u662f\u4ec0\u4e48\u8c03\u7528\u5bfc\u81f4\u7684\u6ea2\u51fa\uff0c\u53ea\u6709

\u8fdb\u7a0b\u5df2\u7ed3\u675f\uff0c\u9000\u51fa\u4ee3\u7801\u4e3a -1073741571 (0xC00000FD) \u7684\u9000\u51fa\u7801\n

\u559c\u6b22\u7684\u53ef\u4ee5 star \u4e00\u4e0b\u54c8\uff0c\u66f4\u5feb\u66f4\u5f3a

https://github.com/auv-sh/av

\u6709\u6ca1\u6709\u4e00\u8d77\u5b66\u6df1\u5ea6\u5b66\u4e60\u548c Transformer \u7684\uff0c\u6216\u8005\u9700\u8981\u5199\u8bba\u6587\u7684\uff0c\u6211\u4e70\u4e86\u5957\u6211\u89c9\u5f97\u8fd8\u4e0d\u9519\u7684\u8bfe\uff0c\u6211\u5df2\u7ecf\u57fa\u672c\u4e0a\u5b66\u5b8c\u4e86\uff0c\u4e5f\u624b\u64b8\u4e86\u51e0\u4e2a CNN \u3001LSTM \u3001Transformer \u7684\u6a21\u578b\u3002\u4f46\u662f\u8bfe 3K \u4e70\u7684\uff0c\u5c31\u60f3\u627e\u51e0\u4e2a\u60f3\u5b66\u7684\uff0c\u6211\u597d\u56de\u70b9\u672c\uff01

\u6211\u60f3\u4e86\u5f88\u4e45\u4e5f\u6ca1\u60f3\u660e\u767d\u4e3a\u4ec0\u4e48\u8981\u624b\u64b8\u4e00\u4e2a\u652f\u6301\u4f20\u5165 numpy \u8fdb\u884c\u8fd0\u7b97\u7684\u77e9\u9635\u7c7b\u800c\u4e0d\u662f\u76f4\u63a5\u7528 numpy \uff1f\n

>>> type({/})\n<class 'set'>\n>>> {/} == set()\nTrue\n

\u611f\u89c9\u597d\u5947\u602a\u7684\u8bed\u6cd5\uff0c\uff0c\u4e3a\u5565\u975e\u8981\u5f15\u5165\u4e00\u4e2a\u7a7a set \u8bed\u6cd5\uff1f\uff1f\u6709\u5565\u7279\u522b\u5927\u7684\u597d\u5904\u5417\uff1f\uff1f set() \u4e5f\u633a\u77ed\u7684\u554a\uff1f\uff1f

\u60f3\u8981\u722c\u7684\u7f51\u7ad9\uff1a https://crushon.ai/character/7e388767-63c7-45e3-b922-53458b4ae4f7/chat

\u4e00\u76f4\u8bf7\u6c42 403 \uff0c\u96be\u53d7\uff0c\u641e\u4e0d\u5b9a\uff0c\u7528\u65e0\u5934 playwrihgt \uff0c\u53c8\u4e0d\u8fc7 cloudflare \uff0c\u8bf7\u6c42\u53c8\u88ab\u62d2\u7edd\uff0c\u96be\u641e\uff01

\u60f3\u8bf7\u6559\u4e00\u4e0b v2 \u7684\u5404\u4f4d\u4f6c\uff0c\u600e\u4e48\u722c\u8fd9\u4e2a\u7f51\u7ad9\u804a\u5929\u5bf9\u8bdd\u3002

\n\u8ba9\u4ed6\u6539\u4e2a\u7b97\u6cd5\u4ee3\u7801\uff0c\u628a\u6211\u4ee3\u7801\u7ed9\u5220\u9519\u4e86\uff0c\u5bfc\u81f4\u6574\u4e2a\u5de5\u7a0b\u51fa\u73b0\u4e0d\u53ef\u63a7\u95ee\u9898

\u5982\u9898\uff0c\u9009\u6280\u672f\u6808 ing \uff0c\u4e0d\u77e5\u9053\u6709\u6ca1\u6709\u5751

\u65e5\u5e38\u9700\u8981\u628a\u8c37\u6b4c\u8ddf\u8e2a\u4ee3\u7801\u52a0\u5165\u5230 head \u5143\u7d20\u91cc\u9762\uff0c\u6240\u4ee5\u501f\u52a9 AI \u5199\u4e86\u4e00\u4e2a\u5c0f\u5de5\u5177\u3002

1 \u3001\u628a\u9700\u8981\u52a0\u5165\u7684\u5185\u5bb9\u4fdd\u5b58\u5230 GACode.txt \u6587\u4ef6\u4e2d

2 \u3001\u6267\u884c\u4ee3\u7801\uff0c\u8f93\u5165\u76ee\u5f55\u5373\u53ef\u3002\uff08\u53ea\u4f1a\u4fee\u6539 html \u6269\u5c55\u540d\u7684\u6587\u4ef6\uff09

\u4ee3\u7801

import os\nimport re\nimport sys\nfrom pathlib import Path\n\ndef get_resource_path(relative_path):\n \"\"\"\u83b7\u53d6\u8d44\u6e90\u6587\u4ef6\u7684\u7edd\u5bf9\u8def\u5f84\uff0c\u652f\u6301 PyInstaller \u6253\u5305\"\"\"\n try:\n # PyInstaller \u521b\u5efa\u4e34\u65f6\u6587\u4ef6\u5939\uff0c\u5e76\u5c06\u8def\u5f84\u5b58\u50a8\u5728_MEIPASS \u4e2d\n base_path = sys._MEIPASS\n except Exception:\n base_path = os.path.abspath(\".\")\n return os.path.join(base_path, relative_path)\n\ndef check_and_add_google_analytics():\n \"\"\"\n \u68c0\u67e5\u6240\u6709 HTML \u6587\u4ef6\u662f\u5426\u5305\u542b Google Analytics \u4ee3\u7801\uff0c\u5982\u679c\u4e0d\u5b58\u5728\u5219\u6dfb\u52a0\n \"\"\"\n # \u521d\u59cb\u5316\u7edf\u8ba1\u4fe1\u606f\n stats = {\n 'total': 0,\n 'modified': 0,\n 'skipped': 0,\n 'errors': 0\n }\n \n # \u4ece GACode.txt \u6587\u4ef6\u8bfb\u53d6 Google Analytics \u4ee3\u7801\n ga_code_file = get_resource_path('GACode.txt')\n \n try:\n with open(ga_code_file, 'r', encoding='utf-8') as f:\n ga_code = f.read().strip()\n except FileNotFoundError:\n print(f\"\u9519\u8bef\uff1a\u627e\u4e0d\u5230\u6587\u4ef6 {ga_code_file}\")\n stats['errors'] = 1\n return stats\n except Exception as e:\n print(f\"\u9519\u8bef\uff1a\u8bfb\u53d6 GA \u4ee3\u7801\u6587\u4ef6\u65f6\u51fa\u9519: {str(e)}\")\n stats['errors'] = 1\n return stats\n \n if not ga_code:\n print(\"\u9519\u8bef\uff1aGA \u4ee3\u7801\u6587\u4ef6\u4e3a\u7a7a\")\n stats['errors'] = 1\n return stats\n \n # \u83b7\u53d6\u5f53\u524d\u76ee\u5f55\u4e0b\u6240\u6709 HTML \u6587\u4ef6\n html_files = list(Path('.').glob('*.html'))\n stats['total'] = len(html_files)\n \n print(f\"\u627e\u5230 {len(html_files)} \u4e2a HTML \u6587\u4ef6\")\n \n for html_file in html_files:\n print(f\"\\n \u5904\u7406\u6587\u4ef6: {html_file}\")\n \n try:\n # \u8bfb\u53d6\u6587\u4ef6\u5185\u5bb9\n with open(html_file, 'r', encoding='utf-8') as f:\n cOntent= f.read()\n \n # \u68c0\u67e5\u662f\u5426\u5df2\u7ecf\u5305\u542b Google Analytics \u4ee3\u7801\n if 'googletagmanager.com/gtag/js' in content:\n print(f\" \u2713 {html_file} \u5df2\u5305\u542b Google Analytics \u4ee3\u7801\uff0c\u8df3\u8fc7\")\n stats['skipped'] += 1\n continue\n \n # \u67e5\u627e</head>\u6807\u7b7e\u7684\u4f4d\u7f6e\n head_end_pattern = r'</head>'\n match = re.search(head_end_pattern, content, re.IGNORECASE)\n \n if not match:\n print(f\" \u2717 {html_file} \u672a\u627e\u5230</head>\u6807\u7b7e\uff0c\u8df3\u8fc7\")\n stats['errors'] += 1\n continue\n \n # \u5728</head>\u6807\u7b7e\u524d\u63d2\u5165 Google Analytics \u4ee3\u7801\n head_end_pos = match.start()\n new_cOntent= (\n content[:head_end_pos] +\n ' ' + ga_code + '\\n' +\n content[head_end_pos:]\n )\n \n # \u5199\u5165\u4fee\u6539\u540e\u7684\u5185\u5bb9\n with open(html_file, 'w', encoding='utf-8') as f:\n f.write(new_content)\n \n print(f\" \u2713 {html_file} \u5df2\u6210\u529f\u6dfb\u52a0 Google Analytics \u4ee3\u7801\")\n stats['modified'] += 1\n \n except Exception as e:\n print(f\" \u2717 \u5904\u7406 {html_file} \u65f6\u51fa\u9519: {str(e)}\")\n stats['errors'] += 1\n \n print(\"\\n \u5904\u7406\u5b8c\u6210\uff01\")\n return stats\n\nif __name__ == \"__main__\":\n try:\n # \u83b7\u53d6\u7528\u6237\u8f93\u5165\u7684\u76ee\u5f55\u8def\u5f84\n current_dir = input(\"\u8bf7\u8f93\u5165\u8981\u5904\u7406\u7684\u76ee\u5f55\u8def\u5f84: \").strip()\n \n # \u5982\u679c\u7528\u6237\u6ca1\u6709\u8f93\u5165\uff0c\u5219\u4f7f\u7528\u5f53\u524d\u76ee\u5f55\n if not current_dir:\n current_dir = os.getcwd()\n \n print(f\"\u5f53\u524d\u5de5\u4f5c\u76ee\u5f55: {current_dir}\")\n \n # \u5207\u6362\u5230\u6307\u5b9a\u76ee\u5f55\n try:\n os.chdir(current_dir)\n except FileNotFoundError:\n print(f\"\u9519\u8bef\uff1a\u76ee\u5f55 '{current_dir}' \u4e0d\u5b58\u5728\")\n input(\"\\n \u6309\u56de\u8f66\u952e\u9000\u51fa...\")\n exit(1)\n except PermissionError:\n print(f\"\u9519\u8bef\uff1a\u6ca1\u6709\u6743\u9650\u8bbf\u95ee\u76ee\u5f55 '{current_dir}'\")\n input(\"\\n \u6309\u56de\u8f66\u952e\u9000\u51fa...\")\n exit(1)\n \n # \u68c0\u67e5\u662f\u5426\u5728\u5305\u542b HTML \u6587\u4ef6\u7684\u76ee\u5f55\u4e2d\n html_files = list(Path('.').glob('*.html'))\n if not html_files:\n print(\"\u5f53\u524d\u76ee\u5f55\u4e0b\u6ca1\u6709\u627e\u5230 HTML \u6587\u4ef6\uff0c\u8bf7\u786e\u4fdd\u5728\u6b63\u786e\u7684\u76ee\u5f55\u4e0b\u8fd0\u884c\u6b64\u811a\u672c\")\n input(\"\\n \u6309\u56de\u8f66\u952e\u9000\u51fa...\")\n exit(1)\n \n # \u6267\u884c\u6dfb\u52a0 Google Analytics \u4ee3\u7801\u7684\u64cd\u4f5c\n stats = check_and_add_google_analytics()\n \n # \u68c0\u67e5\u8fd4\u56de\u503c\u662f\u5426\u6709\u6548\n if stats is None:\n print(\"\\n\u26a0 \u51fd\u6570\u6267\u884c\u5f02\u5e38\uff0c\u65e0\u6cd5\u83b7\u53d6\u7edf\u8ba1\u4fe1\u606f\")\n stats = {'total': 0, 'modified': 0, 'skipped': 0, 'errors': 1}\n \n # \u663e\u793a\u7edf\u8ba1\u7ed3\u679c\n print(\"\\n=== \u6267\u884c\u7edf\u8ba1 ===\")\n print(f\"\u603b\u6587\u4ef6\u6570: {stats['total']}\")\n print(f\"\u6210\u529f\u4fee\u6539: {stats['modified']} \u4e2a\")\n print(f\"\u5df2\u5b58\u5728\u8df3\u8fc7: {stats['skipped']} \u4e2a\")\n print(f\"\u5904\u7406\u5931\u8d25: {stats['errors']} \u4e2a\")\n \n if stats['modified'] > 0:\n print(f\"\\n\u2713 \u6210\u529f\u4e3a {stats['modified']} \u4e2a HTML \u6587\u4ef6\u6dfb\u52a0\u4e86 Google Analytics \u4ee3\u7801\uff01\")\n elif stats['total'] > 0 and stats['skipped'] == stats['total']:\n print(\"\\n\u2713 \u6240\u6709 HTML \u6587\u4ef6\u90fd\u5df2\u5305\u542b Google Analytics \u4ee3\u7801\uff0c\u65e0\u9700\u4fee\u6539\u3002\")\n else:\n print(\"\\n\u26a0 \u6ca1\u6709\u6210\u529f\u4fee\u6539\u4efb\u4f55\u6587\u4ef6\uff0c\u8bf7\u68c0\u67e5\u6587\u4ef6\u683c\u5f0f\u6216\u6743\u9650\u3002\")\n \n print(\"\\n=== \u7a0b\u5e8f\u6267\u884c\u5b8c\u6210 ===\")\n input(\"\\n \u6309\u56de\u8f66\u952e\u9000\u51fa...\")\n \n except KeyboardInterrupt:\n print(\"\\n\\n \u7a0b\u5e8f\u88ab\u7528\u6237\u4e2d\u65ad\")\n input(\"\\n \u6309\u56de\u8f66\u952e\u9000\u51fa...\")\n except Exception as e:\n print(f\"\\n \u7a0b\u5e8f\u6267\u884c\u65f6\u53d1\u751f\u672a\u77e5\u9519\u8bef: {str(e)}\")\n input(\"\\n \u6309\u56de\u8f66\u952e\u9000\u51fa...\")\n

\u6211\u5c31\u4e0d\u53d1\u6210\u54c1\u4e86\uff0c\u81ea\u884c\u6253\u5305\u54c8

\u770b\u56fe

logger \uff0crequests \u90fd\u5c5e\u4e8e\u7b2c\u4e09\u65b9\u6a21\u5757\u3002\u4f46\u662f logger \u5c31\u662f\u7ea2\u7ebf\u3002\u6700\u521d\u4ee5\u4e3a\u662f\u73af\u5883\u6ca1\u8bc6\u522b\uff0c\u4f46\u662f\u8fd0\u884c\u90fd\u662f\u6b63\u5e38\u7684\uff0c\u8fd9\u4e8b\u60c5\u5f88\u8d39\u89e3\u554a\u3002\u6709\u89e3\u51b3\u529e\u6cd5\u4e48\uff1f

\u548b\u4e2a\u6ca1\u56fe\uff0c\u8d34\u4e2a\u5730\u5740\u5427 https://imgur.com/C7tWIvh

\u524d\u51e0\u5929\u5728\u77e5\u4e4e\u4e0a\u770b\u5230\u8fd9\u7bc7\u6587\u7ae0\uff0c\uff0c\u5f53\u65f6\u5b8c\u5168\u6ca1\u8003\u8651\u6587\u7ae0\u7684\u771f\u5b9e\u6027\uff0c\uff0c\u53ea\u662f\u89c9\u5f97 Python \u600e\u4e48\u52a0\u8fd9\u4e48\u591a\u5947\u8469\u8bed\u6cd5\uff0c\uff0c\u8981\u5b8c\u3002\u3002

\u4eca\u5929\u7a0d\u5fae\u641c\u4e86\u4e0b\uff0c\uff0c\u53d1\u73b0\u6587\u7ae0\u4e2d\u63d0\u5230\u7684 PEP 727 \u3001PEP 740 \u5e76\u4e0d\u662f\u6587\u7ae0\u4e2d\u63cf\u8ff0\u7684\u5185\u5bb9\u3002\u3002\u4f3c\u4e4e\u6574\u7bc7\u6587\u7ae0\u90fd\u662f\u80e1\u7f16\u4e71\u9020\u7684\u3002\u3002

\u641e\u4e0d\u61c2\u8c01\u4f1a\u5199\u8fd9\u79cd\u4e1c\u897f\uff1f\uff1f\u96be\u9053\u662f AI \u751f\u6210\u7684\uff1f\uff1f

\u6b63\u5728\u5904\u7406\u4e00\u5ea7\u5c4e\u5c71\uff0c\u5927\u6982\u60c5\u51b5\u662f\u8fd9\u6837\u7684\u3002\u6709\u4e00\u4e2a\u81ea\u5b9a\u4e49\u7684class X\uff0c\u9700\u8981\u6539\u6389\u7684\u4e1c\u897f\u6ce8\u91ca\u5199\u7684\u662fy=Union[X, str]\uff0c\u5b9e\u9645\u60c5\u51b5\u4e5f\u662f\u4e24\u79cd\u7c7b\u578b\u5230\u5904\u90fd\u6df7\u5728\u4e00\u8d77\u6839\u672c\u5206\u4e0d\u6e05\u3002\u540e\u9762\u6240\u6709\u8ddfX\u7c7b\u6709\u5173\u7684\u65b9\u6cd5\u4e00\u70b9\u6ce8\u91ca\u6ca1\u5199\uff0c\u751a\u81f3\u90fd\u4e0d\u505a\u7c7b\u578b\u5224\u65ad\uff0c\u800c\u662f\u5927\u91cf\u7684\u4f7f\u7528getattr(y, 'name', y)\u628a\u6c34\u6405\u5f97\u66f4\u6d51\uff08\u5982\u679cy\u662f\u5b57\u7b26\u4e32\uff0c\u6ca1\u6709name\uff0c\u8fd4\u56de\u5b57\u7b26\u4e32\u672c\u8eab\u3002\u5426\u5219\u8fd4\u56dey.name\u4e5f\u662f\u4e00\u4e2a\u5b57\u7b26\u4e32\uff09\u3002\u6700\u79bb\u8c31\u7684\u662f\u6574\u5ea7\u5c4e\u5c71\u8fd8\u914d\u4e86\u4e00\u5957 30 \u5206\u949f\u624d\u80fd\u8dd1\u5b8c\u7684 pytest \u3002

\u73b0\u5728\u9700\u8981\u628a\u6df7\u4e71\u7684y\u5206\u6210\u786e\u5b9a\u7c7b\u578b\u7684y\u548cy_str\uff0c\u4f7f\u7528 pydantic \u5f3a\u5236\u5b9a\u4e49\u6570\u636e\u7c7b\u578b\u6765\u4e00\u70b9\u70b9\u6392\u9519\u3002\u7136\u800c\u6539\u4e86\u51e0\u767e\u884c\u4ee5\u540e\uff0c\u6240\u6709\u8ddfX\u6709\u5173\u7684\u65b9\u6cd5\u90fd\u88ab\u8feb\u6539\u51fa\u4e24\u79cd\u7248\u672c\u7684\u3002pytest \u4e0d\u518d\u62a5\u7c7b\u578b\u9519\u8bef\u4e86\uff0c\u5374\u5f00\u59cb\u51fa\u73b0\u5404\u79cd\u79bb\u5947\u7684 bug \u6302\u6389\u3002

\u6709\u4ec0\u4e48\u66f4\u5f3a\u5927\u7684\u5de5\u5177\u53ef\u4ee5\u641e\u5b9a\u8fd9\u4ef6\u4e8b\u5417\uff1f\u6bd4\u5982\u540c\u6b65\u68c0\u6d4b\u4e24\u4e2a\u7248\u672c\u7684\u4ee3\u7801\uff0c\u770b\u4ece\u54ea\u4e00\u6b65\u5f00\u59cb\u4e24\u8fb9\u7684\u6570\u636e\u53d8\u5f97\u4e0d\u4e00\u81f4\u4e86\uff1f

\u5f53 Django \u7ed3\u5408\u5f02\u6b65\u4efb\u52a1(\u5982 Celery \u6216\u8005 MQ \u7b49)\uff0c\u6267\u884c\u957f\u8017\u65f6\u4efb\u52a1\uff0cDjango \u4f1a\u4e3a\u8be5\u4efb\u52a1\u7684\u51fd\u6570\u7b2c\u4e00\u6b21\u8c03\u7528 ORM \u65f6\u5206\u914d\u4e00\u6761\u6570\u636e\u5e93\u8fde\u63a5\uff0c\u800c\u4e14\u8fd9\u4e2a\u8fde\u63a5\u5728 ORM \u8c03\u7528\u7ed3\u675f\u65f6\u4e0d\u4f1a\u7acb\u5373\u91ca\u653e\uff0c\u5f53\u4f60\u5728\u540c\u4e00\u4e2a\u4efb\u52a1(\u957f\u8017\u65f6)\u91cc\u5176\u4ed6\u5730\u65b9\u518d\u6b21\u8c03\u7528 ORM \u65f6\uff0c\u5c31\u4f1a\u51fa\u73b0\"MySQL server has gone away\"\u7684\u62a5\u9519(\u539f\u56e0\u662f MySQL \u4e3b\u52a8\u65ad\u5f00\u7684)

\u5b98\u65b9\u63a8\u8350\u89e3\u51b3\u65b9\u6848\uff1a\n\u5982\u679c\u5728 Django \u7684\u8bf7\u6c42-\u54cd\u5e94\u5468\u671f\u4e4b\u5916\u7684\u957f\u8fd0\u884c\u8fdb\u7a0b\u4e2d\u521b\u5efa\u4e86\u8fde\u63a5\uff0c\u8be5\u8fde\u63a5\u5c06\u4fdd\u6301\u6253\u5f00\u72b6\u6001\uff0c\u76f4\u5230\u663e\u5f0f\u5173\u95ed\u6216\u8d85\u65f6\u53d1\u751f\u3002\u4f60\u53ef\u4ee5\u4f7f\u7528 django.db.close_old_connections() \u6765\u5173\u95ed\u6240\u6709\u65e7\u7684\u6216\u4e0d\u53ef\u7528\u7684\u8fde\u63a5\u3002

\u6211\u89c9\u5f97\u592a low \u4e86\u3002\u7406\u8bba\u4e0a\uff0c\u53ea\u9700\u8981\u8bbe\u7f6e\u6bcf\u6b21\u8c03\u7528 ORM \u5f00\u59cb\u524d\u83b7\u53d6\u4e00\u6761\u8fde\u63a5\uff0c\u8c03\u7528\u7ed3\u675f\u540e\u5173\u95ed\u8fde\u63a5\u5373\u53ef\u89e3\u51b3\u8fd9\u4e2a\u95ee\u9898\u3002

\u6709\u6ca1\u6709\u5927\u4f6c\u9047\u5230\u8fc7\u8fd9\u4e2a\u95ee\u9898\uff1f\u5982\u4f55\u66f4\u52a0\u4f18\u96c5\u7684\u89e3\u51b3\uff1f

\u6700\u8fd1\u53d1\u73b0\u4e86\u8fd9\u4e2a\u4e8b\u60c5\u3002\u6709\u70b9\u60ca\u8bb6\uff01

pytest \u7684\u539f\u751f\u5c55\u793a\u4fe1\u606f\u8fd9\u4e48\u591a\u5e74\u4e86\u4e5f\u6ca1\u600e\u4e48\u53d8, \u6211\u53c2\u8003 nextest \u5199\u4e86\u63d2\u4ef6 pytest-modern, \u63d0\u4f9b\u66f4\u73b0\u4ee3\u5316\u7684\u8f93\u51fa

V \u53cb\u4eec\uff0c\u5e2e\u5fd9\u7ed9\u70b9\u6392\u67e5\u610f\u89c1

\u95ee\u9898\u4e0d\u662f\u5fc5\u73b0\u7684\uff0c\u5076\u5c14\u5c31\u51fa\u73b0\u4e00\u6b21

\u95ee\u4e86 AI \u7ed9\u7684\u4e5f\u662f\u6a21\u7cca\u7684\u7684\u56de\u7b54

\u8fde\u63a5 pg \u6570\u636e\u5e93\u7528\u7684\u662f pool = PooledDB()

\u6267\u884c\u4ee3\u7801\u5982\u4e0b

 def dbQuery(self, sql):\n cOnn= pool.connection()\n cur = conn.cursor()\n cur.execute(sql)\n result = cur.fetchall()\n cur.close()\n conn.close()\n return result\n

\u62a5\u9519\u5982\u4e0b\uff1a

could not receive data from server: Software caused connection abort (0x00002745/10053)\n\nTraceback (most recent call last):\n File \"aaTestV5.py\", line 844, in process\n input_values = self.getInputValuesInfo()\n File \"aa_process.py\", line 117, in getInputValuesInfo\n order_list = super(tf, self).getInputValuesInfo()\n File \"aaTestV5.py\", line 555, in getInputValuesInfo\n results_fromno, no_list = self.getFromnoResult()\n File \"aa_process.py\", line 366, in getFromnoResult\n return self.getFromnoResultInner(False)\n File \"aa_process.py\", line 320, in getFromnoResultInner\n results_fromno, no_list = super(tf, self).getFromnoResult(True)\n File \"aaTestV5.py\", line 119, in getFromnoResult\n result_fromno1 = self.dbQuery(query_sql)\n File \"aa_process.py\", line 391, in dbQuery\n return super(tf, self).dbQuery(sql)\n File \"aaTestV5.py\", line 823, in dbQuery\n cur.execute(sql)\n File \"\\python\\lib\\site-packages\\dbutils\\steady_db.py\", line 598, in tough_method\n result = method(*args, **kwargs) # try to execute\npsycopg2.DatabaseError: could not receive data from server: Software caused connection abort (0x00002745/10053)\n

https://github.com/python/pymanager

\u597d\u50cf\u662f\u521a\u53d1\u5e03\u7684, \u6ca1\u770b\u5230\u6709\u4eba\u8ba8\u8bba, \u6211\u4e5f\u662f\u6628\u5929\u770b\u5230\u7684

\u76ee\u524d\u53ea\u652f\u6301 windows , \u4f46\u5b98\u65b9\u4ea7\u54c1\u5e94\u8be5\u4ee5\u540e\u4f1a\u652f\u6301 linux, mac \u628a

\u5b98\u65b9\u603b\u7b97\u8ba4\u8bc6\u5230 python \u7684\u7248\u672c\u7ba1\u7406\u662f\u4e2a\u95ee\u9898\u4e86

\u6c34\u5370\u662f\u672a\u77e5\u7684\uff0c\u6709\u6ca1\u6709\u76f8\u5173\u6a21\u578b\u6216\u8005\u5de5\u5177\u53ef\u4ee5\u8bc6\u522b\u5462

win11 \u4e2d\u6709\u4e00\u4e2a\u6587\u4ef6\u76ee\u5f55\uff0c\u8fd9\u4e2a\u6587\u4ef6\u76ee\u5f55\u4e0b\u6709\u5f88\u591a\u5b50\u76ee\u5f55\u4ee5\u53ca\u5b50\u76ee\u5f55\u7684\u5b50\u76ee\u5f55\u7b49\u7b49\uff0c\u5728\u8fd9\u4e9b\u76ee\u5f55\u4e2d\u6709\u8bb8\u591a\u5e26 vba \u4ee3\u7801\u7684 excel \u6587\u4ef6\uff0c\u4e5f\u6709\u8bb8\u591a\u4e0d\u5e26 vba \u7684 excel \u6587\u4ef6\uff0c\u5927\u7ea6\u51e0\u5341\u4e07\u4e2a\uff0c\u8fd9\u4e9b excel \u6587\u4ef6\u7684\u7248\u672c\u4e0d\u4e00\u6837\u3002\u73b0\u5728\u9700\u8981\u7f16\u5199\u4e00\u4e2a\u7a0b\u5e8f\uff0c\u5728\u6781\u77ed\u65f6\u95f4\u5185\u627e\u51fa\u76ee\u5f55\u4e0b\u6240\u6709\u5e26 vba \u4ee3\u7801\u7684 excel \u6587\u4ef6\uff0c\u5e76\u628a\u5b83\u4eec\u8f6c\u6362\u6210\u6ca1\u6709 vba \u4ee3\u7801\u7684 excel \u6587\u4ef6\uff0c\u4fdd\u6301\u539f\u6765\u7684\u6587\u4ef6\u7248\u672c\u548c\u539f\u6765\u7684\u6587\u4ef6\u540d\uff0c\u5927\u5bb6\u6709\u4ec0\u4e48\u597d\u7684\u65b9\u6848\uff1f

\u4f7f\u7528 SVHN \u7684\u6570\u636e\u96c6\u8fdb\u884c\u6a21\u578b\u7684\u8bad\u7ec3\uff0c\u4f46\u662f\u6574\u4e2a\u6a21\u578b\u5728\u8bad\u7ec3\u96c6\u4e0a\u7684\u51c6\u786e\u7387\u662f\u4e00\u76f4\u5728\u4e0a\u5347\u7684\uff0c\u4f46\u662f\u5230\u4e86\u6d4b\u8bd5\u96c6\u4e0a\u5c31\u4e00\u76f4\u5361\u5728 95%\uff0c\u90fd 3 \u5929\u4e86\uff0c\u6c42\u5404\u4f4d\u5927\u4f6c\u5e2e\u6211\u770b\u4e0b\u6709\u6ca1\u6709\u4f18\u5316\u7684\u65b9\u6848\uff01\u8dea\u8c22!

import torch\nimport torch.nn as nn\nimport torch.optim as optim\nfrom torch.distributed.checkpoint import load_state_dict\nfrom torch.hub import load_state_dict_from_url\nfrom torch.nn.modules.loss import _Loss\nfrom torch.optim import Optimizer\nfrom torch.utils.data import Dataset, random_split, DataLoader\nimport torchvision\nfrom torchvision.transforms import transforms\nimport torchvision.models as m\nimport matplotlib.pyplot as plt\nimport random\nimport gc # \u7528\u4e8e\u5783\u573e\u56de\u6536\nfrom torchinfo import summary\nimport numpy as np\nimport random\n\nimport gc\n\n# \u8bbe\u7f6e\u968f\u673a\u6570\u79cd\u5b50\nSEED = 420\nrandom.seed(SEED)\nnp.random.seed(SEED)\ntorch.manual_seed(SEED)\ntorch.cuda.manual_seed(SEED)\ntorch.cuda.manual_seed_all(SEED)\ntorch.backends.cudnn.deterministic = True\ntorch.backends.cudnn.benchmark = False\n\n# \u8bbe\u7f6e\u4f7f\u7528 gpu \u8fd8\u662f cpu \u8fdb\u884c\u8bad\u7ec3\ndevice = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n\n# \u5b9a\u4e49\u8bad\u7ec3\u7684\u8bba\u8ff0\nepochs = 100\nlr = 0.0001\n\n# \u5b9a\u4e49\u6570\u636e\u96c6\u9700\u8981\u7684\u53c2\u6570\nbatchSize = 64\n\n# \u52a0\u8f7d\u8bad\u7ec3\u96c6\u9700\u8981\u7684\u6570\u636e\u8f6c\u6362\u5668\ntrainT = transforms.Compose([\n transforms.RandomCrop(28),\n transforms.RandomRotation(degrees=[-15, 15]),\n transforms.ToTensor(),\n transforms.Normalize(mean = [0.4377, 0.4438, 0.4728], std = [0.1980, 0.2010, 0.1970])\n])\n\n# \u52a0\u8f7d\u6d4b\u8bd5\u96c6\u9700\u8981\u7684\u6570\u636e\u8f6c\u6362\u5668\ntestT = transforms.Compose([\n transforms.CenterCrop(28),\n transforms.ToTensor(),\n transforms.Normalize(mean = [0.4377, 0.4438, 0.4728], std = [0.1980, 0.2010, 0.1970])\n])\n\n# \u52a0\u8f7d\u8bad\u7ec3\u96c6\u6570\u636e\nsvhn_train = torchvision.datasets.SVHN(root='C:\\\\FashionMNIST'\n , split=\"train\"\n , download=True\n , transform=trainT\n )\n# \u52a0\u8f7d\u6d4b\u8bd5\u96c6\u6570\u636e\nsvhn_test = torchvision.datasets.SVHN(root='C:\\\\FashionMNIST'\n , split=\"test\"\n , download=True\n , transform=testT\n )\n\n# \u5b9a\u4e49\u795e\u7ecf\u7f51\u7edc\uff0c\u56e0\u4e3a\u6211\u4eec\u7684\u56fe\u7247\u7684\u5c3a\u5bf8\u548c\u6837\u672c\u6570\u91cf\u90fd\u4e0d\u662f\u5f88\u5927\uff0c\u6240\u4ee5\u9009\u62e9\u4ece ResNet18 \u548c Vgg16 \u4e2d\u62bd\u53d6\u5c42\u6765\u6784\u5efa\u7f51\u7edc\nresnet18_ = m.resnet18()\n\n\nclass MyResNet(nn.Module): # \u8fd9\u4e2a\u662f\u57fa\u4e8e ResNet18 \u6784\u5efa\u7684\u7f51\u7edc\n def __init__(self):\n super(MyResNet, self).__init__()\n self.block1 = nn.Sequential(\n nn.Conv2d(3, 64, 3, 1, 1),\n resnet18_.bn1,\n resnet18_.relu\n )\n\n self.block2 = resnet18_.layer2 # \u8fde\u6743\u91cd\u90fd\u4f1a\u590d\u7528\u8fc7\u6765\uff0c\u5728 resnet18_ = m.resnet18() \u8fd9\u513f\u5c31\u5df2\u7ecf\u521d\u59cb\u5316\u597d\u4e86\u6743\u91cd\u6570\u636e\uff01\n self.block3 = resnet18_.layer3\n self.block4 = resnet18_.layer4 # \u4ece Resnet18 \u4e2d\u54ea layer \u65b0\u589e\u5230\u81ea\u5df1\u7684\u6a21\u578b\u4e2d\n self.avgpool = resnet18_.avgpool\n\n self.fc = nn.Linear(512, 10, True)\n\n def forward(self, x):\n x = self.block1(x)\n x = self.block2(x)\n x = self.block3(x)\n x = self.block4(x) # \u8fd9\u513f\u65b0\u589e\u4e00\u6761\u5904\u7406\u4ee3\u7801\n x = self.avgpool(x)\n x = x.view(-1, 512)\n return self.fc(x)\n\n\nvgg16_ = m.vgg16()\n\n\nclass MyVgg(nn.Module): # \u8fd9\u4e2a\u662f\u57fa\u4e8e Vgg16 \u6784\u5efa\u7684\u7f51\u7edc\n def __init__(self):\n super(MyVgg, self).__init__()\n\n self.features = nn.Sequential(\n *vgg16_.features[0:9],\n # \u4f7f\u7528*\u662f \u5c06 .features[0:9]\u63d0\u53d6\u51fa\u6765\u7684\u5c42\uff0c\u5168\u90e8\u53d6\u51fa\u6765\uff0c\u4e00\u4e2a\u4e2a\u653e\u5230\u5f53\u524d\u7684 Sequential \u4e2d\uff0c\u800c\u4e0d\u662f\u7ec4\u6210\u4e00\u4e2a Sequential \u653e\u5230\u5f53\u524d\u7684 Sequential \u4e2d\uff01\n nn.Conv2d(128, 128, 3, 1, 1),\n nn.ReLU(inplace=True),\n nn.MaxPool2d(2, 2, padding=0, dilation=1, ceil_mode=False)\n )\n self.avgpool = vgg16_.avgpool\n\n self.fc = nn.Sequential(\n nn.Linear(6272, 4096, True),\n *vgg16_.classifier[1:6],\n nn.Linear(4096, 10, True)\n )\n\n def forward(self, x):\n x = self.features(x)\n x = self.avgpool(x)\n x = x.view(-1, 6272)\n x = self.fc(x)\n return x\n\n\n# summary(MyVgg(), input_size=(10, 3, 28, 28)) # \u4e00\u5b9a\u8981\uff0c\u5b9e\u4f8b\u5316\u8dd1\u4e00\u4e0b\uff0c\u770b\u770b\u6709\u6ca1\u6709\u95ee\u9898\uff01\n\n\nclass earlyStopping():\n\n def __init__(self, patience=5, tol=0.0005):\n # \u5f53\u8fde\u7eed patience=5 \u6b21\uff0c\u672c\u8f6e\u6b21\u7684\u8fed\u4ee3\u7684\u635f\u5931\u4e0e\u5386\u53f2\u6700\u5c0f\u7684\u635f\u5931\u7684\u5dee\u503c\u5927\u4e8e 0.0005 \u8fd9\u4e2a\u9608\u503c\uff0c\u5c31\u4f1a\u505c\u6b62\u8bad\u7ec3\n self.patience = patience\n self.tol = tol\n self.counter = 0 # \u8ba1\u6570\u5668\n self.lowest_loss = None # \u8bb0\u5f55\u5386\u53f2\u6700\u5c0f\u635f\u5931\n self.early_stop = False # \u9700\u8981\u8fd4\u56de\u662f\u5426\u9700\u8981\u63d0\u524d\u505c\u6b62\n\n def __call__(self, val_loss): # val_loss \u662f\u8bb0\u5f55\u6d4b\u8bd5\u96c6\u6216\u8bad\u7ec3\u96c6\u4e0a\u4e00\u6b21 epoch \u7684\u635f\u5931\n if self.lowest_loss is None:\n self.lowest_loss = val_loss\n elif self.lowest_loss - val_loss > self.tol:\n self.lowest_loss = val_loss\n self.counter = 0\n elif self.lowest_loss - val_loss < self.tol:\n self.counter += 1\n print('Notice: Early stopping counter {} of {}'.format(self.counter, self.patience))\n\n if self.counter >= self.patience:\n print('Notice: Early stopping counter Active')\n self.early_stop = True\n return self.early_stop\n\n\n# \u5b9a\u4e49\u8bad\u7ec3\u51fd\u6570\ndef fit(net: nn.Module, lossFunc: _Loss, op: Optimizer, trainData: DataLoader, testData: DataLoader, epochs: int):\n transLost = [] # \u7528\u4e8e\u6536\u96c6\u6bcf\u8f6e\u8bad\u7ec3\u548c\u6d4b\u8bd5\u7684\u7ed3\u679c\uff0c\u7528\u4e8e\u540e\u9762\u753b\u56fe\u8868\u4f7f\u7528\n trainCorrect = []\n testLost = []\n testCorrect = []\n trainedSampleNum = 0\n\n # \u521d\u59cb\u5316 earlystopping \u7c7b\n early_stopping = earlyStopping(patience=15, tol=0.00000005)\n # \u521d\u59cb\u5316\u6d4b\u8bd5\u96c6\u7684\u5386\u53f2\u6700\u9ad8\u51c6\u786e\u7387\n test_highest_correct = None\n test_lowest_loss = None\n\n # \u83b7\u53d6\u5230\u6574\u4e2a\u8bad\u7ec3\u96c6\u4e2d\u7684\u6837\u672c\u6570\u91cf\n trainTotalNum = trainData.dataset.__len__()\n # \u83b7\u53d6\u5230\u6574\u4e2a\u6d4b\u8bd5\u96c6\u4e2d\u7684\u6837\u672c\u6570\u91cf\n testTotalNum = testData.dataset.__len__()\n\n for epoch in range(epochs):\n net.train()\n train_loss = 0\n train_correct = 0\n for batch_index, (data, target) in enumerate(trainData):\n data = data.to(device, non_blocking=True)\n target = target.to(device, non_blocking=True).view(data.shape[0]) # \u786e\u4fdd\u6807\u7b7e\u662f 1 \u7ef4\u7684\u7ed3\u6784\n trainRes = net(data) # \u7ecf\u8fc7\u5b66\u4e60\uff0c\u8fd9\u513f\u6bcf\u4e2a\u6837\u672c\u4f1a\u8f93\u51fa 10 \u4e2a\u7279\u5f81\u7ed3\u679c\u5bf9\u5e94\u7684\u6570\u636e(\u5982\u679c\u6a21\u578b\u4e2d\u6709 softmax \uff0c\u5c31\u662f\u6982\u7387)\uff0c\u53ef\u4ee5\u7528\u4e8e\u540e\u7eed\u8ba1\u7b97\u51c6\u786e\u7387\n loss = lossFunc(trainRes, target)\n op.zero_grad() # \u6e05\u7a7a\u4f18\u5316\u5668\u4e0a\u7684\u68af\u5ea6\n loss.backward()\n op.step()\n # \u5f00\u59cb\u8ba1\u7b97\u51c6\u786e\u6570\uff0c\u5e76\u7d2f\u52a0\n yhat = torch.max(trainRes, 1)[1] # \u4ece trainRes \u4e00\u4e2a\u77e9\u9635\u4e2d\uff0c\u53d6\u51fa\u6bcf\u4e2a\u6837\u672c\u7684\u6700\u5927\u503c\u548c\u6700\u5927\u503c\u6240\u5728\u7684\u7d22\u5f15\uff0c\u5f97\u5230[1,2,1,4]\u8fd9\u79cd\u7c7b\u578b\u7684\u7ed3\u679c\n correct_num = torch.sum(\n yhat == target) # yhat \u3001target \u90fd\u662f\u4e00\u7ef4\u5f20\u91cf\uff0c\u4f7f\u7528 == \u4f1a\u6328\u4e2a\u5bf9\u6bd4\u5f20\u91cf\u4e2d\u7684\u5143\u7d20\u662f\u5426\u76f8\u7b49\uff0c\u6700\u7ec8\u5f97\u5230[False, True, Flase]\u8fd9\u6837\u7684\u6570\u636e\uff0c\u7136\u540e\u4f7f\u7528 torch.sum \u5c31\u53ef\u4ee5\u5f97\u5230\u4e00\u4e2a\u6570\u5b57\uff0c\u56e0\u4e3a True \u4e3a 1 \uff0cFalse \u4e3a 0 \u3002\n train_correct += correct_num # \u5c06\u51c6\u5907\u6570\u7d2f\u52a0\n # \u8ba1\u7b97\u635f\u5931\uff0c\u5e76\u7d2f\u52a0\n train_loss += loss.item() # \u8fd9\u513f\u9700\u8981\u5f97\u5230\u6240\u6709\u6837\u672c\u7684\u635f\u5931\u7684\u548c\n trainedSampleNum += data.shape[0]\n # print(\"\u672c\u6279\u6b21\u8bad\u7ec3\u635f\u5931\u4e3a\uff1a\", loss.item() / data.shape[0])\n if (batch_index + 1) % 125 == 0:\n # \u73b0\u5728\u8fdb\u884c\u5230\u4e86\u54ea\u4e2a epoch \u3001\u603b\u5171\u8981\u8bad\u7ec3\u591a\u5c11\u4e2a\u6837\u672c\u3001\u5df2\u7ecf\u8bad\u7ec3\u4e86\u591a\u5c11\u4e2a\u6837\u672c\u3001\u5df2\u8bad\u7ec3\u7684\u6837\u672c\u7684\u767e\u5206\u6bd4\n print(\"Epoch{}:{} / {} = ({:.0f}%)\".format(\n epoch + 1,\n trainedSampleNum,\n epochs * len(trainData) * batchSize,\n 100 * trainedSampleNum / (epochs * len(trainData) * batchSize)\n ))\n\n print(\"-------------------------------\")\n avg_correct = (float(train_correct) / trainTotalNum) * 100\n # print(\"\u672c\u8f6e\u8bad\u7ec3\u5e73\u5747\u51c6\u786e\u7387\uff1a\", avg_correct)\n trainCorrect.append(avg_correct)\n avg_loss = (float(train_loss) / trainTotalNum) * 100\n # print(\"\u672c\u8f6e\u8bad\u7ec3\u5e73\u5747\u635f\u5931\u7387\uff1a\", avg_loss)\n transLost.append(avg_loss)\n\n del data, target, train_loss, train_correct\n gc.collect()\n torch.cuda.empty_cache()\n\n # \u4e00\u8f6e\u8bad\u7ec3\u7ed3\u675f\uff0c\u5c31\u4f7f\u7528\u6d4b\u8bd5\u96c6\u8fdb\u884c\u6d4b\u8bd5\n net.eval()\n test_loss = 0\n test_correct = 0\n for batch_index, (test_data, test_target) in enumerate(testData):\n with torch.no_grad():\n test_data = test_data.to(device, non_blocking=True)\n test_target = test_target.to(device, non_blocking=True).view(test_data.shape[0]) # \u786e\u4fdd\u6807\u7b7e\u662f 1 \u7ef4\u7684\u7ed3\u6784\n testRes = net(test_data)\n loss = lossFunc(testRes, test_target)\n # \u8ba1\u7b97\u635f\u5931\uff0c\u5e76\u7d2f\u52a0\n test_loss += loss.item()\n # \u8ba1\u7b97\u51c6\u5907\u6570\uff0c\u5e76\u7d2f\u52a0\n yhat = torch.max(testRes, 1)[1] # \u4ece trainRes \u4e00\u4e2a\u77e9\u9635\u4e2d\uff0c\u53d6\u51fa\u6bcf\u4e2a\u6837\u672c\u7684\u6700\u5927\u503c\u548c\u6700\u5927\u503c\u6240\u5728\u7684\u7d22\u5f15\uff0c\u5f97\u5230[1,2,1,4]\u8fd9\u79cd\u7c7b\u578b\u7684\u7ed3\u679c\n correct_num = torch.sum(\n yhat == test_target) # yhat \u3001target \u90fd\u662f\u4e00\u7ef4\u5f20\u91cf\uff0c\u4f7f\u7528 == \u4f1a\u6328\u4e2a\u5bf9\u6bd4\u5f20\u91cf\u4e2d\u7684\u5143\u7d20\u662f\u5426\u76f8\u7b49\uff0c\u6700\u7ec8\u5f97\u5230[False, True, Flase]\u8fd9\u6837\u7684\u6570\u636e\uff0c\u7136\u540e\u4f7f\u7528 torch.sum \u5c31\u53ef\u4ee5\u5f97\u5230\u4e00\u4e2a\u6570\u5b57\uff0c\u56e0\u4e3a True \u4e3a 1 \uff0cFalse \u4e3a 0 \u3002\n test_correct += correct_num # \u5c06\u51c6\u5907\u6570\u7d2f\u52a0\n\n avg_test_correct = (float(test_correct) / testTotalNum) * 100\n # print(\"\u672c\u8f6e\u6d4b\u8bd5\u5e73\u5747\u51c6\u786e\u7387\uff1a\", avg_test_correct)\n testCorrect.append(avg_test_correct)\n avg_test_loss = (float(test_loss) / testTotalNum) * 100\n # print(\"\u672c\u8f6e\u6d4b\u8bd5\u5e73\u5747\u635f\u5931\u7387\uff1a\", avg_test_loss)\n testLost.append(avg_test_loss)\n\n print(\"\u672c\u8f6e\u8bad\u7ec3\u5e73\u5747\u51c6\u786e\u7387\uff1a{}, \u672c\u8f6e\u8bad\u7ec3\u5e73\u5747\u635f\u5931\u7387: {}, \u672c\u8f6e\u6d4b\u8bd5\u5e73\u5747\u51c6\u786e\u7387\uff1a{}, \u672c\u8f6e\u6d4b\u8bd5\u5e73\u5747\u635f\u5931\u7387\uff1a{}\".format(\n avg_correct, avg_loss, avg_test_correct, avg_test_loss))\n\n del test_data, test_target, test_loss, test_correct\n gc.collect()\n torch.cuda.empty_cache()\n\n # \u5982\u679c\u6d4b\u8bd5\u96c6\u635f\u5931\u51fa\u73b0\u65b0\u4f4e\u6216\u8005\u51c6\u786e\u7387\u51fa\u73b0\u65b0\u9ad8\uff0c\u5c31\u4fdd\u5b58\u5728\u6a21\u578b\u7684\u6743\u91cd\uff0c\u9632\u6b62\u4e2d\u9014\u65ad\u7535\u7b49\u539f\u56e0\u9700\u8981\u4ece\u5934\u518d\u6765\n if test_highest_correct is None:\n test_highest_correct = avg_test_correct\n if test_highest_correct < avg_test_correct:\n test_highest_correct = avg_test_correct\n torch.save(net.state_dict(), './v6/model-' + str(epoch + 1) + '.pth')\n print(\"model saved\")\n\n # \u6700\u597d\u5728\u6d4b\u8bd5\u96c6\u4e0a\u4f7f\u7528\u63d0\u524d\u505c\u6b62\uff0c\u5982\u679c\u4f7f\u7528\u8bad\u7ec3\u96c6\u65e0\u6cd5\u9884\u6d4b\u8fc7\u62df\u5408\u8fd9\u79cd\u60c5\u51b5\n early_stop = early_stopping(avg_test_loss) # \u8fd9\u513f\u4f7f\u7528\u63d0\u524d\u505c\u6b62\uff01\n if early_stop:\n break\n print(\"mission completed\")\n return transLost, trainCorrect, testLost, testCorrect\n\n\nmodel = MyResNet().to(device)\n# model.load_state_dict(torch.load(\"./v4/model-49.pth\"))\nloss_func = nn.CrossEntropyLoss(reduction='sum') # \u56e0\u4e3a\u6211\u4eec\u5728\u8bad\u7ec3\u51fd\u6570\u4e2d\uff0c\u5728\u8ba1\u7b97\u635f\u5931\u7684\u65f6\u5019\u662f\u8ba1\u7b97\u7684\u6bcf\u4e2a\u6837\u672c\u7684\u635f\u5931\u7684\u548c\uff0c\u6240\u4ee5\u8fd9\u513f\u9700\u8981\u4f7f\u7528 reduction='sum'\nopt = optim.RMSprop(model.parameters(), lr=lr, weight_decay=0.00005, momentum=0.0001)\n\ntrain_data = DataLoader(svhn_train, batch_size=batchSize, shuffle=True, drop_last=False, pin_memory=True)\ntest_data = DataLoader(svhn_test, batch_size=batchSize, shuffle=False, drop_last=False, pin_memory=True)\n\n# \u5f00\u59cb\u8bad\u7ec3\ntransLost, trainCorrect, testLost, testCorrect = fit(model, loss_func, opt, train_data, test_data, epochs)\n\n# \u8bad\u7ec3\u7ed3\u679c\u53ef\u89c6\u5316\nplt.plot(transLost, label='train loss')\nplt.plot(testLost, label='test loss')\nplt.plot(trainCorrect, label='train correcct')\nplt.plot(testCorrect, label='test correcct')\nplt.xlabel('Epoch')\nplt.ylabel('CrossEntropy Loss')\nplt.title('Training Loss')\nplt.legend()\nplt.grid(True)\nplt.show()