{"id":99081,"date":"2026-09-01T23:16:44","date_gmt":"2026-09-01T15:16:44","guid":{"rendered":"https:\/\/www.wsisp.com\/helps\/99081.html"},"modified":"2026-09-01T23:16:44","modified_gmt":"2026-09-01T15:16:44","slug":"chatgpt-plus-pro-codex-%e5%ae%9e%e6%88%98%e6%8c%87%e5%8d%97%ef%bc%882026%e5%b9%b49%e6%9c%881%e6%97%a5%ef%bc%89","status":"publish","type":"post","link":"https:\/\/www.wsisp.com\/helps\/99081.html","title":{"rendered":"ChatGPT Plus \/ Pro + Codex \u5b9e\u6218\u6307\u5357\uff082026\u5e749\u67081\u65e5\uff09"},"content":{"rendered":"

1. \u5f15\u8a00<\/h3>\n

2026 \u5e74，OpenAI \u7684 ChatGPT Plus \u4e0e Pro \u8ba2\u9605\u65b9\u6848，\u914d\u5408 Codex \u7f16\u7a0b\u52a9\u624b，\u5df2\u7ecf\u6210\u4e3a\u5f00\u53d1\u8005\u65e5\u5e38\u5de5\u4f5c\u4e2d\u4e0d\u53ef\u6216\u7f3a\u7684\u5de5\u5177\u7ec4\u5408\u3002ChatGPT \u8d1f\u8d23\u5bf9\u8bdd\u5f0f\u7406\u89e3\u4e0e\u65b9\u6848\u8bbe\u8ba1，Codex \u5219\u4e13\u6ce8\u4e8e\u4ee3\u7801\u751f\u6210\u3001\u5ba1\u67e5\u4e0e\u81ea\u52a8\u5316\u91cd\u6784\u3002\u4e24\u8005\u7ed3\u5408，\u53ef\u4ee5\u663e\u8457\u63d0\u5347\u4ece\u9700\u6c42\u5206\u6790\u5230\u4ee3\u7801\u843d\u5730\u7684\u5168\u6d41\u7a0b\u6548\u7387\u3002<\/p>\n

\u672c\u6587\u4e0d\u6d89\u53ca\u4efb\u4f55\u5145\u503c\u6216\u8d26\u53f7\u8d2d\u4e70\u76f8\u5173\u5185\u5bb9，\u7eaf\u7cb9\u4ece\u6280\u672f\u89d2\u5ea6\u51fa\u53d1，\u56f4\u7ed5 ChatGPT Plus \/ Pro \u4e0e Codex \u7684 API \u8c03\u7528\u3001\u672c\u5730\u96c6\u6210\u3001\u81ea\u52a8\u5316\u5de5\u4f5c\u6d41\u3001\u4ee3\u7801\u5ba1\u67e5\u3001\u6d4b\u8bd5\u751f\u6210\u7b49\u573a\u666f，\u7ed9\u51fa\u53ef\u8fd0\u884c\u7684\u4ee3\u7801\u793a\u4f8b\u4e0e\u5de5\u7a0b\u5b9e\u8df5\u5efa\u8bae\u3002\u5168\u6587\u7ea6 3000 \u5b57，\u6240\u6709\u793a\u4f8b\u5747\u57fa\u4e8e Python \u4e0e OpenAI \u5b98\u65b9 SDK \u7f16\u5199，\u8bfb\u8005\u53ef\u590d\u5236\u540e\u76f4\u63a5\u8fd0\u884c\u3002<\/p>\n

2. \u73af\u5883\u51c6\u5907\u4e0e\u4f9d\u8d56\u5b89\u88c5<\/h3>\n

\u5728\u5f00\u59cb\u4e4b\u524d，\u9700\u8981\u51c6\u5907 Python 3.10 \u6216\u66f4\u9ad8\u7248\u672c，\u5e76\u5b89\u88c5 OpenAI \u5b98\u65b9 SDK\u3002\u5efa\u8bae\u4f7f\u7528\u865a\u62df\u73af\u5883\u9694\u79bb\u4f9d\u8d56\u3002<\/p>\n

python -m<\/span> venv venv
\nsource<\/span> venv\/bin\/activate # Windows \u4e0b\u4f7f\u7528 venv\\\\Scripts\\\\activate<\/span>
\npip install<\/span> –upgrade<\/span> openai python-dotenv<\/p>\n

\u5b89\u88c5\u5b8c\u6210\u540e，\u5728\u9879\u76ee\u6839\u76ee\u5f55\u521b\u5efa .env \u6587\u4ef6，\u7528\u4e8e\u5b58\u653e API \u5bc6\u94a5\u3002\u6ce8\u610f：\u672c\u6587\u5047\u8bbe\u8bfb\u8005\u5df2\u7ecf\u62e5\u6709\u5408\u6cd5\u7684 API \u8bbf\u95ee\u6743\u9650，\u4e14\u4e0d\u4f1a\u8ba8\u8bba\u4efb\u4f55\u83b7\u53d6\u5bc6\u94a5\u7684\u975e\u5e38\u89c4\u9014\u5f84\u3002<\/p>\n

OPENAI_API_KEY<\/span>=<\/span>sk-your-key-here<\/p>\n

\u52a0\u8f7d\u73af\u5883\u53d8\u91cf\u5e76\u521d\u59cb\u5316\u5ba2\u6237\u7aef：<\/p>\n

import<\/span> os
\nfrom<\/span> dotenv import<\/span> load_dotenv
\nfrom<\/span> openai import<\/span> OpenAI<\/p>\n

load_dotenv(<\/span>)<\/span><\/p>\n

client =<\/span> OpenAI(<\/span>
\n api_key=<\/span>os.<\/span>getenv(<\/span>"OPENAI_API_KEY"<\/span>)<\/span>,<\/span>
\n)<\/span><\/p>\n

3. \u4f7f\u7528 ChatGPT \u6a21\u578b\u8fdb\u884c\u5bf9\u8bdd\u4e0e\u4ee3\u7801\u751f\u6210<\/h3>\n

ChatGPT Plus \u4e0e Pro \u8ba2\u9605\u7684\u6838\u5fc3\u4ef7\u503c\u4e4b\u4e00，\u662f\u53ef\u4ee5\u901a\u8fc7 API \u8c03\u7528 GPT \u7cfb\u5217\u6a21\u578b，\u5b8c\u6210\u81ea\u7136\u8bed\u8a00\u7406\u89e3\u3001\u4ee3\u7801\u751f\u6210\u3001\u6587\u6863\u64b0\u5199\u7b49\u4efb\u52a1\u3002\u4e0b\u9762\u662f\u4e00\u4e2a\u57fa\u7840\u7684\u5bf9\u8bdd\u8865\u5168\u793a\u4f8b\u3002<\/p>\n

response =<\/span> client.<\/span>chat.<\/span>completions.<\/span>create(<\/span>
\n model=<\/span>"gpt-4.1"<\/span>,<\/span>
\n messages=<\/span>[<\/span>
\n {<\/span>"role"<\/span>:<\/span> "system"<\/span>,<\/span> "content"<\/span>:<\/span> "\u4f60\u662f\u4e00\u540d\u8d44\u6df1 Python \u5de5\u7a0b\u5e08，\u64c5\u957f\u7f16\u5199\u9ad8\u8d28\u91cf\u3001\u53ef\u7ef4\u62a4\u7684\u4ee3\u7801\u3002"<\/span>}<\/span>,<\/span>
\n {<\/span>"role"<\/span>:<\/span> "user"<\/span>,<\/span> "content"<\/span>:<\/span> "\u8bf7\u7f16\u5199\u4e00\u4e2a\u51fd\u6570，\u7528\u4e8e\u8ba1\u7b97\u6590\u6ce2\u90a3\u5951\u6570\u5217\u7684\u7b2c n \u9879，\u8981\u6c42\u4f7f\u7528\u8fed\u4ee3\u65b9\u5f0f\u5b9e\u73b0\u3002"<\/span>}<\/span>,<\/span>
\n ]<\/span>,<\/span>
\n temperature=<\/span>0.3<\/span>,<\/span>
\n)<\/span><\/p>\n

print<\/span>(<\/span>response.<\/span>choices[<\/span>0<\/span>]<\/span>.<\/span>message.<\/span>content)<\/span><\/p>\n

\u8fd0\u884c\u4e0a\u8ff0\u4ee3\u7801，\u6a21\u578b\u4f1a\u8fd4\u56de\u4e00\u6bb5\u5b8c\u6574\u7684 Python \u5b9e\u73b0\u3002\u901a\u8fc7\u8c03\u6574 temperature \u53c2\u6570，\u53ef\u4ee5\u63a7\u5236\u8f93\u51fa\u7684\u968f\u673a\u6027：\u6570\u503c\u8d8a\u4f4e，\u8f93\u51fa\u8d8a\u786e\u5b9a；\u6570\u503c\u8d8a\u9ad8，\u8f93\u51fa\u8d8a\u591a\u6837\u3002<\/p>\n

4. \u4f7f\u7528 Codex \u6a21\u578b\u8fdb\u884c\u4ee3\u7801\u8865\u5168\u4e0e\u5ba1\u67e5<\/h3>\n

Codex \u662f OpenAI \u63a8\u51fa\u7684\u7f16\u7a0b\u4e13\u7528\u6a21\u578b，\u64c5\u957f\u4ee3\u7801\u8865\u5168\u3001\u4ee3\u7801\u5ba1\u67e5\u3001Bug \u4fee\u590d\u7b49\u4efb\u52a1\u3002\u4e0e ChatGPT \u7684\u5bf9\u8bdd\u5f0f\u63a5\u53e3\u4e0d\u540c，Codex \u66f4\u4fa7\u91cd\u4e8e\u4ee3\u7801\u4e0a\u4e0b\u6587\u7684\u7406\u89e3\u4e0e\u751f\u6210\u3002<\/p>\n

4.1 \u4ee3\u7801\u8865\u5168<\/h4>\n

response =<\/span> client.<\/span>completions.<\/span>create(<\/span>
\n model=<\/span>"codex-1"<\/span>,<\/span>
\n prompt=<\/span>"def quick_sort(arr):\\\\n if len(arr) <= 1:\\\\n return arr\\\\n pivot = arr[len(arr) \/\/ 2]\\\\n left = [x for x in arr if x < pivot]\\\\n middle = [x for x in arr if x == pivot]\\\\n right = [x for x in arr if x > pivot]\\\\n return quick_sort(left) + middle + quick_sort(right)\\\\n\\\\n# \u6d4b\u8bd5\u7528\u4f8b\\\\nprint(quick_sort([3, 6, 8, 10, 1, 2, 1]))"<\/span>,<\/span>
\n max_tokens=<\/span>200<\/span>,<\/span>
\n temperature=<\/span>0.2<\/span>,<\/span>
\n)<\/span><\/p>\n

print<\/span>(<\/span>response.<\/span>choices[<\/span>0<\/span>]<\/span>.<\/span>text)<\/span><\/p>\n

4.2 \u4ee3\u7801\u5ba1\u67e5<\/h4>\n

Codex \u4e5f\u53ef\u4ee5\u7528\u4e8e\u4ee3\u7801\u5ba1\u67e5\u3002\u5c06\u5f85\u5ba1\u67e5\u7684\u4ee3\u7801\u7247\u6bb5\u4f5c\u4e3a\u8f93\u5165，\u8ba9\u6a21\u578b\u6307\u51fa\u6f5c\u5728\u95ee\u9898\u5e76\u7ed9\u51fa\u6539\u8fdb\u5efa\u8bae\u3002<\/p>\n

code_snippet =<\/span> """
\ndef fetch_data(url):
\n import requests
\n response = requests.get(url)
\n return response.json()
\n"""<\/span><\/p>\n

response =<\/span> client.<\/span>chat.<\/span>completions.<\/span>create(<\/span>
\n model=<\/span>"gpt-4.1"<\/span>,<\/span>
\n messages=<\/span>[<\/span>
\n {<\/span>"role"<\/span>:<\/span> "system"<\/span>,<\/span> "content"<\/span>:<\/span> "\u4f60\u662f\u4e00\u540d\u4e25\u683c\u7684\u4ee3\u7801\u5ba1\u67e5\u4e13\u5bb6，\u8bf7\u6307\u51fa\u4ee3\u7801\u4e2d\u7684\u95ee\u9898\u5e76\u7ed9\u51fa\u6539\u8fdb\u5efa\u8bae\u3002"<\/span>}<\/span>,<\/span>
\n {<\/span>"role"<\/span>:<\/span> "user"<\/span>,<\/span> "content"<\/span>:<\/span> f"\u8bf7\u5ba1\u67e5\u4ee5\u4e0b\u4ee3\u7801：\\\\n<\/span>{<\/span>code_snippet}<\/span><\/span>"<\/span><\/span>}<\/span>,<\/span>
\n ]<\/span>,<\/span>
\n)<\/span><\/p>\n

print<\/span>(<\/span>response.<\/span>choices[<\/span>0<\/span>]<\/span>.<\/span>message.<\/span>content)<\/span><\/p>\n

5. \u6784\u5efa\u81ea\u52a8\u5316\u4ee3\u7801\u751f\u6210\u5de5\u4f5c\u6d41<\/h3>\n

\u5c06 ChatGPT \u4e0e Codex \u7ed3\u5408，\u53ef\u4ee5\u6784\u5efa\u4e00\u4e2a\u81ea\u52a8\u5316\u7684\u4ee3\u7801\u751f\u6210\u6d41\u6c34\u7ebf：\u5148\u7531 ChatGPT \u7406\u89e3\u9700\u6c42\u5e76\u751f\u6210\u8bbe\u8ba1\u6587\u6863，\u518d\u7531 Codex \u6839\u636e\u8bbe\u8ba1\u6587\u6863\u751f\u6210\u5177\u4f53\u5b9e\u73b0\u3002<\/p>\n

def<\/span> generate_code_from_requirement<\/span>(<\/span>requirement:<\/span> str<\/span>)<\/span> –<\/span>><\/span> str<\/span>:<\/span>
\n # \u7b2c\u4e00\u6b65：\u4f7f\u7528 ChatGPT \u751f\u6210\u8bbe\u8ba1\u6587\u6863<\/span>
\n design_response =<\/span> client.<\/span>chat.<\/span>completions.<\/span>create(<\/span>
\n model=<\/span>"gpt-4.1"<\/span>,<\/span>
\n messages=<\/span>[<\/span>
\n {<\/span>"role"<\/span>:<\/span> "system"<\/span>,<\/span> "content"<\/span>:<\/span> "\u4f60\u662f\u4e00\u540d\u7cfb\u7edf\u67b6\u6784\u5e08，\u8bf7\u6839\u636e\u9700\u6c42\u8f93\u51fa\u8be6\u7ec6\u7684\u8bbe\u8ba1\u6587\u6863\u3002"<\/span>}<\/span>,<\/span>
\n {<\/span>"role"<\/span>:<\/span> "user"<\/span>,<\/span> "content"<\/span>:<\/span> requirement}<\/span>,<\/span>
\n ]<\/span>,<\/span>
\n )<\/span>
\n design_doc =<\/span> design_response.<\/span>choices[<\/span>0<\/span>]<\/span>.<\/span>message.<\/span>content<\/p>\n

# \u7b2c\u4e8c\u6b65：\u4f7f\u7528 Codex \u6839\u636e\u8bbe\u8ba1\u6587\u6863\u751f\u6210\u4ee3\u7801<\/span>
\n code_response =<\/span> client.<\/span>completions.<\/span>create(<\/span>
\n model=<\/span>"codex-1"<\/span>,<\/span>
\n prompt=<\/span>f"\u6839\u636e\u4ee5\u4e0b\u8bbe\u8ba1\u6587\u6863\u751f\u6210 Python \u4ee3\u7801：\\\\n<\/span>{<\/span>design_doc}<\/span><\/span>\\\\n\\\\n\u4ee3\u7801："<\/span><\/span>,<\/span>
\n max_tokens=<\/span>1500<\/span>,<\/span>
\n temperature=<\/span>0.2<\/span>,<\/span>
\n )<\/span>
\n return<\/span> code_response.<\/span>choices[<\/span>0<\/span>]<\/span>.<\/span>text<\/p>\n

requirement =<\/span> "\u5b9e\u73b0\u4e00\u4e2a\u7b80\u5355\u7684\u4efb\u52a1\u961f\u5217，\u652f\u6301\u6dfb\u52a0\u4efb\u52a1\u3001\u6267\u884c\u4efb\u52a1\u3001\u67e5\u8be2\u4efb\u52a1\u72b6\u6001\u3002"<\/span>
\ngenerated_code =<\/span> generate_code_from_requirement(<\/span>requirement)<\/span>
\nprint<\/span>(<\/span>generated_code)<\/span><\/p>\n

6. \u4f7f\u7528 Codex \u81ea\u52a8\u751f\u6210\u5355\u5143\u6d4b\u8bd5<\/h3>\n

\u5355\u5143\u6d4b\u8bd5\u662f\u4fdd\u8bc1\u4ee3\u7801\u8d28\u91cf\u7684\u91cd\u8981\u624b\u6bb5\u3002Codex \u53ef\u4ee5\u6839\u636e\u5df2\u6709\u4ee3\u7801\u81ea\u52a8\u751f\u6210\u6d4b\u8bd5\u7528\u4f8b，\u51cf\u5c11\u624b\u52a8\u7f16\u5199\u6d4b\u8bd5\u7684\u5de5\u4f5c\u91cf\u3002<\/p>\n

def<\/span> generate_unit_tests<\/span>(<\/span>source_code:<\/span> str<\/span>)<\/span> –<\/span>><\/span> str<\/span>:<\/span>
\n response =<\/span> client.<\/span>chat.<\/span>completions.<\/span>create(<\/span>
\n model=<\/span>"gpt-4.1"<\/span>,<\/span>
\n messages=<\/span>[<\/span>
\n {<\/span>"role"<\/span>:<\/span> "system"<\/span>,<\/span> "content"<\/span>:<\/span> "\u4f60\u662f\u4e00\u540d\u6d4b\u8bd5\u5de5\u7a0b\u5e08，\u8bf7\u4e3a\u7ed9\u5b9a\u7684\u4ee3\u7801\u751f\u6210 pytest \u5355\u5143\u6d4b\u8bd5\u3002"<\/span>}<\/span>,<\/span>
\n {<\/span>"role"<\/span>:<\/span> "user"<\/span>,<\/span> "content"<\/span>:<\/span> f"\u8bf7\u4e3a\u4ee5\u4e0b\u4ee3\u7801\u751f\u6210\u5355\u5143\u6d4b\u8bd5：\\\\n<\/span>{<\/span>source_code}<\/span><\/span>"<\/span><\/span>}<\/span>,<\/span>
\n ]<\/span>,<\/span>
\n temperature=<\/span>0.2<\/span>,<\/span>
\n )<\/span>
\n return<\/span> response.<\/span>choices[<\/span>0<\/span>]<\/span>.<\/span>message.<\/span>content<\/p>\n

source =<\/span> """
\ndef add(a, b):
\n return a + b<\/p>\n

def subtract(a, b):
\n return a – b
\n"""<\/span><\/p>\n

tests =<\/span> generate_unit_tests(<\/span>source)<\/span>
\nprint<\/span>(<\/span>tests)<\/span><\/p>\n

7. \u5c06 ChatGPT \u4e0e Codex \u96c6\u6210\u5230 CI\/CD \u6d41\u6c34\u7ebf<\/h3>\n

\u5728\u5b9e\u9645\u5de5\u7a0b\u4e2d，\u53ef\u4ee5\u5c06 ChatGPT \u4e0e Codex \u96c6\u6210\u5230 CI\/CD \u6d41\u6c34\u7ebf\u4e2d，\u5b9e\u73b0\u81ea\u52a8\u5316\u7684\u4ee3\u7801\u5ba1\u67e5\u4e0e\u6587\u6863\u751f\u6210\u3002\u4ee5\u4e0b\u662f\u4e00\u4e2a\u57fa\u4e8e GitHub Actions \u7684\u793a\u4f8b\u3002<\/p>\n

name<\/span>:<\/span> AI Code Review<\/p>\n

on<\/span>:<\/span>
\n pull_request<\/span>:<\/span>
\n types<\/span>:<\/span> [<\/span>opened,<\/span> synchronize]<\/span><\/p>\n

jobs<\/span>:<\/span>
\n ai-review<\/span>:<\/span>
\n runs-on<\/span>:<\/span> ubuntu–<\/span>latest
\n steps<\/span>:<\/span>
\n –<\/span> name<\/span>:<\/span> Checkout code
\n uses<\/span>:<\/span> actions\/checkout@v4<\/p>\n

–<\/span> name<\/span>:<\/span> Set up Python
\n uses<\/span>:<\/span> actions\/setup–<\/span>python@v5
\n with<\/span>:<\/span>
\n python-version<\/span>:<\/span> "3.10"<\/span><\/p>\n

–<\/span> name<\/span>:<\/span> Install dependencies
\n run<\/span>:<\/span> pip install openai python–<\/span>dotenv<\/p>\n

–<\/span> name<\/span>:<\/span> Run AI code review
\n env<\/span>:<\/span>
\n OPENAI_API_KEY<\/span>:<\/span> ${<\/span>{<\/span> secrets.OPENAI_API_KEY }<\/span>}<\/span>
\n run<\/span>:<\/span> |<\/span>
\n python scripts\/ai_review.py<\/span><\/p>\n

\u5bf9\u5e94\u7684 scripts\/ai_review.py \u811a\u672c：<\/p>\n

import<\/span> os
\nimport<\/span> subprocess
\nfrom<\/span> openai import<\/span> OpenAI<\/p>\n

client =<\/span> OpenAI(<\/span>api_key=<\/span>os.<\/span>getenv(<\/span>"OPENAI_API_KEY"<\/span>)<\/span>)<\/span><\/p>\n

def<\/span> get_changed_files<\/span>(<\/span>)<\/span>:<\/span>
\n result =<\/span> subprocess.<\/span>run(<\/span>
\n [<\/span>"git"<\/span>,<\/span> "diff"<\/span>,<\/span> "–name-only"<\/span>,<\/span> "origin\/main…HEAD"<\/span>]<\/span>,<\/span>
\n capture_output=<\/span>True<\/span>,<\/span>
\n text=<\/span>True<\/span>,<\/span>
\n )<\/span>
\n return<\/span> result.<\/span>stdout.<\/span>strip(<\/span>)<\/span>.<\/span>split(<\/span>"\\\\n"<\/span>)<\/span><\/p>\n

def<\/span> review_file<\/span>(<\/span>file_path:<\/span> str<\/span>)<\/span> –<\/span>><\/span> str<\/span>:<\/span>
\n with<\/span> open<\/span>(<\/span>file_path,<\/span> "r"<\/span>,<\/span> encoding=<\/span>"utf-8"<\/span>)<\/span> as<\/span> f:<\/span>
\n content =<\/span> f.<\/span>read(<\/span>)<\/span>
\n response =<\/span> client.<\/span>chat.<\/span>completions.<\/span>create(<\/span>
\n model=<\/span>"gpt-4.1"<\/span>,<\/span>
\n messages=<\/span>[<\/span>
\n {<\/span>"role"<\/span>:<\/span> "system"<\/span>,<\/span> "content"<\/span>:<\/span> "\u4f60\u662f\u4e00\u540d\u8d44\u6df1\u4ee3\u7801\u5ba1\u67e5\u4e13\u5bb6，\u8bf7\u8f93\u51fa\u5ba1\u67e5\u610f\u89c1\u3002"<\/span>}<\/span>,<\/span>
\n {<\/span>"role"<\/span>:<\/span> "user"<\/span>,<\/span> "content"<\/span>:<\/span> f"\u8bf7\u5ba1\u67e5\u4ee5\u4e0b\u6587\u4ef6：\\\\n<\/span>{<\/span>file_path}<\/span><\/span>\\\\n\\\\n<\/span>{<\/span>content}<\/span><\/span>"<\/span><\/span>}<\/span>,<\/span>
\n ]<\/span>,<\/span>
\n )<\/span>
\n return<\/span> response.<\/span>choices[<\/span>0<\/span>]<\/span>.<\/span>message.<\/span>content<\/p>\n

if<\/span> __name__ ==<\/span> "__main__"<\/span>:<\/span>
\n files =<\/span> get_changed_files(<\/span>)<\/span>
\n for<\/span> file<\/span> in<\/span> files:<\/span>
\n if<\/span> file<\/span>.<\/span>endswith(<\/span>".py"<\/span>)<\/span>:<\/span>
\n print<\/span>(<\/span>f"\u5ba1\u67e5\u6587\u4ef6：<\/span>{<\/span>file<\/span>}<\/span><\/span>"<\/span><\/span>)<\/span>
\n print<\/span>(<\/span>review_file(<\/span>file<\/span>)<\/span>)<\/span><\/p>\n

8. \u4f7f\u7528 Codex \u8fdb\u884c\u4ee3\u7801\u91cd\u6784<\/h3>\n

\u4ee3\u7801\u91cd\u6784\u662f\u65e5\u5e38\u5f00\u53d1\u4e2d\u5e38\u89c1\u7684\u9700\u6c42\u3002Codex \u53ef\u4ee5\u8bc6\u522b\u4ee3\u7801\u4e2d\u7684\u574f\u5473\u9053，\u5e76\u7ed9\u51fa\u91cd\u6784\u5efa\u8bae\u6216\u76f4\u63a5\u751f\u6210\u91cd\u6784\u540e\u7684\u4ee3\u7801\u3002<\/p>\n

def<\/span> refactor_code<\/span>(<\/span>source_code:<\/span> str<\/span>)<\/span> –<\/span>><\/span> str<\/span>:<\/span>
\n response =<\/span> client.<\/span>chat.<\/span>completions.<\/span>create(<\/span>
\n model=<\/span>"gpt-4.1"<\/span>,<\/span>
\n messages=<\/span>[<\/span>
\n {<\/span>"role"<\/span>:<\/span> "system"<\/span>,<\/span> "content"<\/span>:<\/span> "\u4f60\u662f\u4e00\u540d\u4ee3\u7801\u91cd\u6784\u4e13\u5bb6，\u8bf7\u4f18\u5316\u4ee3\u7801\u7ed3\u6784，\u63d0\u5347\u53ef\u8bfb\u6027\u4e0e\u53ef\u7ef4\u62a4\u6027\u3002"<\/span>}<\/span>,<\/span>
\n {<\/span>"role"<\/span>:<\/span> "user"<\/span>,<\/span> "content"<\/span>:<\/span> f"\u8bf7\u91cd\u6784\u4ee5\u4e0b\u4ee3\u7801：\\\\n<\/span>{<\/span>source_code}<\/span><\/span>"<\/span><\/span>}<\/span>,<\/span>
\n ]<\/span>,<\/span>
\n temperature=<\/span>0.1<\/span>,<\/span>
\n )<\/span>
\n return<\/span> response.<\/span>choices[<\/span>0<\/span>]<\/span>.<\/span>message.<\/span>content<\/p>\n

messy_code =<\/span> """
\ndef process(data):
\n result = []
\n for item in data:
\n if item['type'] == 'a':
\n result.append(item['value'] * 2)
\n elif item['type'] == 'b':
\n result.append(item['value'] * 3)
\n return result
\n"""<\/span><\/p>\n

refactored =<\/span> refactor_code(<\/span>messy_code)<\/span>
\nprint<\/span>(<\/span>refactored)<\/span><\/p>\n

9. \u7ed3\u5408\u5411\u91cf\u6570\u636e\u5e93\u6784\u5efa\u4ee3\u7801\u77e5\u8bc6\u5e93<\/h3>\n

\u5bf9\u4e8e\u5927\u578b\u9879\u76ee，\u53ef\u4ee5\u5c06\u4ee3\u7801\u5e93\u5411\u91cf\u5316，\u6784\u5efa\u4e00\u4e2a\u8bed\u4e49\u68c0\u7d22\u7cfb\u7edf\u3002\u5f53\u5f00\u53d1\u8005\u63d0\u95ee\u65f6，\u5148\u68c0\u7d22\u76f8\u5173\u4ee3\u7801\u7247\u6bb5，\u518d\u4ea4\u7ed9 ChatGPT \u751f\u6210\u56de\u7b54\u3002\u8fd9\u6837\u53ef\u4ee5\u663e\u8457\u63d0\u5347\u56de\u7b54\u7684\u51c6\u786e\u6027\u3002<\/p>\n

from<\/span> openai import<\/span> OpenAI
\nimport<\/span> numpy as<\/span> np<\/p>\n

client =<\/span> OpenAI(<\/span>api_key=<\/span>os.<\/span>getenv(<\/span>"OPENAI_API_KEY"<\/span>)<\/span>)<\/span><\/p>\n

def<\/span> embed_text<\/span>(<\/span>text:<\/span> str<\/span>)<\/span> –<\/span>><\/span> list<\/span>:<\/span>
\n response =<\/span> client.<\/span>embeddings.<\/span>create(<\/span>
\n model=<\/span>"text-embedding-3-small"<\/span>,<\/span>
\n input<\/span>=<\/span>text,<\/span>
\n )<\/span>
\n return<\/span> response.<\/span>data[<\/span>0<\/span>]<\/span>.<\/span>embedding<\/p>\n

def<\/span> cosine_similarity<\/span>(<\/span>a:<\/span> list<\/span>,<\/span> b:<\/span> list<\/span>)<\/span> –<\/span>><\/span> float<\/span>:<\/span>
\n a =<\/span> np.<\/span>array(<\/span>a)<\/span>
\n b =<\/span> np.<\/span>array(<\/span>b)<\/span>
\n return<\/span> np.<\/span>dot(<\/span>a,<\/span> b)<\/span> \/<\/span> (<\/span>np.<\/span>linalg.<\/span>norm(<\/span>a)<\/span> *<\/span> np.<\/span>linalg.<\/span>norm(<\/span>b)<\/span>)<\/span><\/p>\n

# \u793a\u4f8b：\u5c06\u4ee3\u7801\u7247\u6bb5\u5411\u91cf\u5316\u5e76\u68c0\u7d22<\/span>
\ncode_snippets =<\/span> [<\/span>
\n "def add(a, b): return a + b"<\/span>,<\/span>
\n "def quick_sort(arr): …"<\/span>,<\/span>
\n "class DatabaseConnection: …"<\/span>,<\/span>
\n]<\/span><\/p>\n

embeddings =<\/span> [<\/span>embed_text(<\/span>snippet)<\/span> for<\/span> snippet in<\/span> code_snippets]<\/span><\/p>\n

query =<\/span> "\u5982\u4f55\u5b9e\u73b0\u4e24\u4e2a\u6570\u76f8\u52a0？"<\/span>
\nquery_embedding =<\/span> embed_text(<\/span>query)<\/span><\/p>\n

similarities =<\/span> [<\/span>cosine_similarity(<\/span>query_embedding,<\/span> emb)<\/span> for<\/span> emb in<\/span> embeddings]<\/span>
\nbest_match_index =<\/span> np.<\/span>argmax(<\/span>similarities)<\/span>
\nprint<\/span>(<\/span>f"\u6700\u76f8\u5173\u7684\u4ee3\u7801\u7247\u6bb5：<\/span>{<\/span>code_snippets[<\/span>best_match_index]<\/span>}<\/span><\/span>"<\/span><\/span>)<\/span><\/p>\n

10. \u603b\u7ed3\u4e0e\u6700\u4f73\u5b9e\u8df5<\/h3>\n

ChatGPT Plus \/ Pro \u4e0e Codex \u7684\u7ec4\u5408，\u4e3a\u5f00\u53d1\u8005\u63d0\u4f9b\u4e86\u4e00\u5957\u5b8c\u6574\u7684 AI \u8f85\u52a9\u5f00\u53d1\u89e3\u51b3\u65b9\u6848\u3002\u4ece\u9700\u6c42\u5206\u6790\u3001\u4ee3\u7801\u751f\u6210\u3001\u6d4b\u8bd5\u7f16\u5199\u5230\u4ee3\u7801\u5ba1\u67e5\u4e0e\u91cd\u6784，\u6bcf\u4e00\u4e2a\u73af\u8282\u90fd\u53ef\u4ee5\u501f\u52a9 AI \u63d0\u5347\u6548\u7387\u3002<\/p>\n

\u5728\u5b9e\u9645\u4f7f\u7528\u4e2d，\u5efa\u8bae\u9075\u5faa\u4ee5\u4e0b\u6700\u4f73\u5b9e\u8df5：<\/p>\n