{"id":1162,"date":"2024-04-22T12:32:44","date_gmt":"2024-04-22T04:32:44","guid":{"rendered":"https:\/\/www.nicekj.com\/?p=1162"},"modified":"2024-04-22T12:32:44","modified_gmt":"2024-04-22T04:32:44","slug":"openaizhongbanggengxinhanshudiaoyonggongnengzuyigaibiankaifaliuchengdeyicigengxin","status":"publish","type":"post","link":"https:\/\/www.nicekj.com\/openaizhongbanggengxinhanshudiaoyonggongnengzuyigaibiankaifaliuchengdeyicigengxin.html","title":{"rendered":"OPENAI\u91cd\u78c5\u66f4\u65b0\uff1a\u51fd\u6570\u8c03\u7528\u529f\u80fd\uff0c\u8db3\u4ee5\u6539\u53d8\u5f00\u53d1\u6d41\u7a0b\u7684\u4e00\u6b21\u66f4\u65b0\uff01"},"content":{"rendered":"

\u5f15\u8a00<\/h2>\n

\u4eca\u5929OPENAI\u53d1\u5e03\u91cd\u78c5\u66f4\u65b0\uff0c\u5305\u62ec\uff1a
\n1.Chat Completions API \u4e2d\u7684\u65b0\u51fd\u6570\u8c03\u7528\u529f\u80fd
\n2.gpt-4\u548c\u7684\u66f4\u65b0\u548c\u66f4\u6613\u4e8e\u64cd\u7eb5\u7684\u7248\u672cgpt-3.5-turbo
\n3.\u65b0\u7684 16k \u4e0a\u4e0b\u6587\u7248\u672cgpt-3.5-turbo\uff08\u4e0e\u6807\u51c6 4k \u7248\u672c\u76f8\u6bd4\uff09
\n4.\u6211\u4eec\u6700\u5148\u8fdb\u7684\u5d4c\u5165\u6a21\u578b\u7684\u6210\u672c\u964d\u4f4e\u4e86 75%
\n5.\u8f93\u5165\u4ee4\u724c\u7684\u6210\u672c\u964d\u4f4e\uff0cgpt-3.5-turbo\u6a21\u578b\u964d\u4f4e\u4e8625%
\n6.gpt-3.5-turbo-0301\u548cgpt-4-0314\u6a21\u578b\u5ba3\u5e03\u5f03\u7528\u65f6\u95f4\u8868<\/p>\n

\u5f88\u591a\u4eba\u89c9\u5f97\u91cd\u78c5\u5728\u964d\u4f4e\u4e86\u6210\u672c\uff0c\u6269\u5927\u4e86token\u6570\uff0c\u4f46\u662f\u8fd9\u4e9b\u90fd\u662f\u5728\u4eba\u4eec\u9884\u6599\u4e4b\u4e2d\u7684\uff0c\u53ea\u6709\u65b0\u7684\u51fd\u6570\u8c03\u7528\u529f\u80fd\u624d\u662f\u5b9e\u81f3\u540d\u5f52\u7684\u201c\u91cd\u78c5\u201d\u3002<\/p>\n

\u5f88\u591a\u4eba\u4e0d\u7406\u89e3\uff0c\u4ec0\u4e48\u662f\u51fd\u6570\u8c03\u7528\u529f\u80fd\uff0c\u8fd9\u91cc\u4e3e\u4e00\u4e2a\u751f\u52a8\u7684\u4f8b\u5b50\uff1a
\n\u4ece\u524d\u6211\u4eec\u8981\u95eegpt\u4eca\u5929\u7684\u5929\u6c14\uff0cgpt\u4f1a\u8bf4\u4e0d\u77e5\u9053\uff0c\u56e0\u4e3a\u9700\u8981\u4f60\u6240\u5904\u7684\u5730\u5740\uff0c\u4ee5\u53ca\u4e00\u4e2a\u5929\u6c14\u9884\u62a5\u7684\u7f51\u7ad9\u7684API\u3002\u5f53\u6211\u4eec\u53ef\u4ee5\u4f7f\u7528\u51fd\u6570\u8c03\u7528\u529f\u80fd\uff0c\u90a3\u4e48\u5f53gpt\u53d1\u73b0\u81ea\u5df1\u4e0d\u77e5\u9053\u5730\u5740\u548cAPI\u7684\u65f6\u5019\uff0c\u5c31\u4f1a\u5728\u6211\u4eec\u63d0\u4f9b\u7684\u51fd\u6570\u5e93\u91cc\u9762\u627e\uff0c\u5982\u679c\u6709\u76f8\u5173\u7684\u51fd\u6570\u5c31\u4f1a\u8c03\u7528\uff0c\u4ece\u800c\u5b9e\u73b0\u529f\u80fd\u3002<\/p>\n

\u5982\u679c\u6709\u4e86\u89e3Langchain\u7684\u670b\u53cb\u4f1a\u53d1\u73b0\uff0c\u8fd9\u4e0d\u5c31\u662fagent\u7684\u529f\u80fd\u5417\uff1f\u662f\u7684\uff0cOPENAI\u5c06\u8fd9\u4e2a\u529f\u80fd\u201c\u6284\u201c\u5230\u4e86\u81ea\u5df1\u7684API\u63a5\u53e3\u670d\u52a1\u4e2d\u3002\u8ba9\u6574\u4e2a\u6d41\u7a0b\u66f4\u52a0\u4e1d\u6ed1\u5408\u7406\uff0c\u7b26\u5408\u903b\u8f91\u3002\u5177\u4f53\u7684\u5dee\u8ddd\u540e\u9762\u4f1a\u8bb2\u3002\u4e0b\u9762\u6211\u4eec\u5148\u4e86\u89e3\u4e00\u4e0bfunctions\u7684\u53c2\u6570\u3002<\/p>\n

functions\u5177\u4f53\u53c2\u6570\u4ecb\u7ecd<\/h2>\n

functions \u662f ChatCompletion API \u4e2d\u7684\u53ef\u9009\u53c2\u6570\uff0c\u53ef\u7528\u4e8e\u63d0\u4f9b\u51fd\u6570\u89c4\u8303\u3002\u8fd9\u6837\u505a\u7684\u76ee\u7684\u662f\u4f7f\u6a21\u578b\u80fd\u591f\u751f\u6210\u7b26\u5408\u51fd\u6570\u8f93\u5165\u6a21\u5f0f\u7684\u8f93\u51fa\u3002\u8bf7\u6ce8\u610f\uff0cAPI \u5b9e\u9645\u4e0a\u4e0d\u4f1a\u6267\u884c\u4efb\u4f55\u51fd\u6570\u8c03\u7528\u3002\u5f00\u53d1\u4eba\u5458\u53ef\u4ee5\u4f7f\u7528\u6a21\u578b\u8f93\u51fa\u6267\u884c\u51fd\u6570\u8c03\u7528\u3002<\/p>\n

\u5982\u679c\u63d0\u4f9b\u4e86\u51fd\u6570\u53c2\u6570\uff0c\u90a3\u4e48\u9ed8\u8ba4\u60c5\u51b5\u4e0b\uff0c\u6a21\u578b\u5c06\u51b3\u5b9a\u4f55\u65f6\u9002\u5408\u4f7f\u7528\u5176\u4e2d\u4e00\u4e2a\u51fd\u6570\u3002\u901a\u8fc7\u5c06 function_call \u53c2\u6570\u8bbe\u7f6e\u4e3a {“name”: “”}\uff0c\u4e5f\u53ef\u4ee5\u5f3a\u5236 API \u4f7f\u7528\u7279\u5b9a\u51fd\u6570\u3002\u5982\u679c\u4f7f\u7528\u51fd\u6570\uff0c\u8f93\u51fa\u5c06\u5728\u54cd\u5e94\u4e2d\u5305\u542b “finish_reason”: “function_call”\uff0c\u4ee5\u53ca\u5177\u6709\u51fd\u6570\u540d\u79f0\u548c\u751f\u6210\u7684\u51fd\u6570\u53c2\u6570\u7684 function_call \u5bf9\u8c61\u3002<\/p>\n

\u51fd\u6570\u7531\u4ee5\u4e0b\u5b57\u6bb5\u6307\u5b9a\uff1a<\/p>\n

\u540d\u79f0Name<\/strong>\uff1a\u51fd\u6570\u7684\u540d\u79f0\u3002<\/p>\n

\u63cf\u8ff0Description<\/strong>\uff1a\u51fd\u6570\u4f5c\u7528\u7684\u63cf\u8ff0\u3002\u8be5\u6a21\u578b\u5c06\u4f7f\u7528\u5b83\u6765\u51b3\u5b9a\u4f55\u65f6\u8c03\u7528\u8be5\u51fd\u6570\u3002<\/p>\n

\u53c2\u6570Parameters<\/strong>\uff1a\u53c2\u6570\u5bf9\u8c61\u5305\u542b\u51fd\u6570\u9700\u8981\u7684\u6240\u6709\u8f93\u5165\u5b57\u6bb5\u3002\u8fd9\u4e9b\u8f93\u5165\u53ef\u4ee5\u662f\u4ee5\u4e0b\u7c7b\u578b\uff1aString\u3001Number\u3001Boolean\u3001Object\u3001Null\u3001AnyOf\u3002\u6709\u5173\u8be6\u7ec6\u4fe1\u606f\uff0c\u8bf7\u53c2\u9605 API \u53c2\u8003\u6587\u6863\u3002<\/p>\n

\u5fc5\u9700Required<\/strong>\uff1a\u8fdb\u884c\u67e5\u8be2\u9700\u8981\u54ea\u4e9b\u53c2\u6570\u3002\u5176\u4f59\u7684\u5c06\u88ab\u89c6\u4e3a\u53ef\u9009\u7684\u3002<\/p>\n

\u60a8\u53ef\u4ee5\u901a\u8fc7\u6267\u884c\u51fd\u6570\u5e76\u5c06\u51fd\u6570\u6267\u884c\u7684\u8f93\u51fa\u76f4\u63a5\u4f20\u56de\u52a9\u624b\u6765\u94fe\u63a5\u51fd\u6570\u8c03\u7528\u3002\u8fd9\u53ef\u80fd\u4f1a\u5bfc\u81f4\u6a21\u578b\u7ee7\u7eed\u65e0\u9650\u671f\u5730\u8c03\u7528\u51fd\u6570\u7684\u65e0\u9650\u5faa\u73af\u884c\u4e3a\uff0c\u4f46\u662f\u53ef\u4ee5\u8bbe\u7f6e\u9632\u62a4\u680f\u6765\u9632\u6b62\u8fd9\u79cd\u60c5\u51b5\u53d1\u751f\u3002<\/p>\n

\u5177\u4f53\u4f8b\u5b50<\/h2>\n

\u63a5\u4e0b\u6765\u7528\u4e00\u4e2a\u4f8b\u5b50\u5177\u4f53\u8bb2\u8bb2\u5982\u4f55\u8c03\u7528\u51fd\u6570\uff0c\u6ce8\u610f\uff0c\u8fd9\u91cc\u53ea\u4ecb\u7ecd\u5982\u4f55\u4f7f\u7528\u4ee5\u53ca\u6574\u4e2a\u7684\u903b\u8f91\uff0c\u8fc7\u591a\u7684\u4ee3\u7801\u7ec6\u8282\u4e0d\u4f1a\u8bb2\u89e3\u3002\uff08\u8fd9\u91cc\u7684\u4ee3\u7801\u662f\u4e00\u4e2a\u63a8\u53cb\u6839\u636e\u8fd9\u4e2a\u529f\u80fd\u5199\u7684\u4e00\u4e2a\u9879\u76eedemo\uff0c\u5730\u5740\u662f\uff1agithub.com\/hx23840\/ope\u2026<\/a><\/p>\n

<\/div>
ini<\/span>
\u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>import requests<\/span>\nimport openai<\/span>\nimport json<\/span>\n<\/span>\n<\/span>\ndef send_request(method, url, data<\/span>=None, params=None):<\/span>\n    response<\/span> = requests.request(method, url, json=data, params=params)<\/span>\n    response.raise_for_status()<\/span>\n    return response.json()<\/span>\n<\/span>\n<\/span>\n# Base URL<\/span><\/span>\nBASE_URL<\/span> = \"https:\/\/plugin.askyourpdf.com\"<\/span>  # \u8bf7\u5c06\u8fd9\u4e2aURL\u66ff\u6362\u4e3a\u5b9e\u9645\u7684API URL<\/span><\/span>\n<\/span>\n<\/span>\ndef download_pdf(url):<\/span>\n    path<\/span> = \"\/api\/download_pdf\"<\/span><\/span>\n    full_url<\/span> = BASE_URL + path<\/span>\n    params<\/span> = {\"url\"<\/span>: url}<\/span>\n    response<\/span> = send_request(\"POST\"<\/span>, full_url, params=params)<\/span>\n    return response<\/span>\n<\/span>\n<\/span>\ndef perform_query(doc_id, query):<\/span>\n    path<\/span> = \"\/query\"<\/span><\/span>\n    full_url<\/span> = BASE_URL + path<\/span>\n    data<\/span> = {\"doc_id\"<\/span>: doc_id, \"query\"<\/span>: query}<\/span>\n    response<\/span> = send_request(\"POST\"<\/span>, full_url, data=data)<\/span>\n    return response<\/span>\n<\/span>\n<\/span>\nopenai.api_key<\/span> = \"YOUR_API_KEY\"<\/span><\/span>\n<\/span>\n<\/span>\n# Step 1, send model the user query and what functions it has access to<\/span><\/span>\ndef run_conversation():<\/span>\n    response<\/span> = openai.ChatCompletion.create(<\/span>\n        model<\/span>=\"gpt-3.5-turbo-0613\"<\/span>,<\/span>\n        messages<\/span>=[<\/span>\n            {\"role\"<\/span>: \"user\"<\/span>, \"content\"<\/span>: \"I want to upload pdf use this url:https:\/\/arxiv.org\/pdf\/1706.03762.pdf\"<\/span>}],<\/span>\n        functions<\/span>=[<\/span>\n            {<\/span>\n                \"name\"<\/span>: \"download_pdf\"<\/span>,<\/span>\n                \"description\"<\/span>: \"upload a pdf\"<\/span>,<\/span>\n                \"parameters\"<\/span>: {<\/span>\n                    \"type\"<\/span>: \"object\"<\/span>,<\/span>\n                    \"properties\"<\/span>: {<\/span>\n                        \"url\"<\/span>: {\"type\"<\/span>: \"string\"<\/span>, \"description\"<\/span>: \"The pdf url\"<\/span>},<\/span>\n                    },<\/span>\n                    \"required\"<\/span>: [\"url\"<\/span>],<\/span>\n                },<\/span>\n            }<\/span>\n        ],<\/span>\n        function_call<\/span>=\"auto\"<\/span>,<\/span>\n    )<\/span>\n<\/span>\n    message<\/span> = response[\"choices\"<\/span>][0<\/span>][\"message\"<\/span>]<\/span>\n<\/span>\n    # Step 2, check if the model wants to call a function<\/span><\/span>\n    if message.get(\"function_call\"):<\/span>\n        # Step 3, call the function<\/span><\/span>\n        function_call<\/span> = message.get('function_call'<\/span>)<\/span>\n        arguments<\/span> = json.loads(function_call.get('arguments'<\/span>))<\/span>\n        function_response<\/span> = download_pdf(url=arguments.get(\"url\"<\/span>))<\/span>\n<\/span>\n        second_response<\/span> = openai.ChatCompletion.create(<\/span>\n            model<\/span>=\"gpt-3.5-turbo-0613\"<\/span>,<\/span>\n            messages<\/span>=[<\/span>\n                {\"role\"<\/span>: \"user\"<\/span>, \"content\"<\/span>: \"I want to search content with pdf, The query content is What is the \"<\/span><\/span>\n                                            \"Self-attention?,The doc id is \"<\/span> + function_response[\"docId\"<\/span>]}<\/span>\n            ],<\/span>\n            functions<\/span>=[<\/span>\n                {<\/span>\n                    \"name\"<\/span>: \"perform_query\"<\/span>,<\/span>\n                    \"description\"<\/span>: \"search content with a pdf\"<\/span>,<\/span>\n                    \"parameters\"<\/span>: {<\/span>\n                        \"type\"<\/span>: \"object\"<\/span>,<\/span>\n                        \"properties\"<\/span>: {<\/span>\n                            \"doc_id\"<\/span>: {\"type\"<\/span>: \"string\"<\/span>, \"description\"<\/span>: \"The doc id\"<\/span>},<\/span>\n                            \"query\"<\/span>: {\"type\"<\/span>: \"string\"<\/span>, \"description\"<\/span>: \"The query content\"<\/span>},<\/span>\n                        },<\/span>\n                        \"required\"<\/span>: [\"doc_id\"<\/span>, \"query\"<\/span>],<\/span>\n                    },<\/span>\n                }<\/span>\n            ],<\/span>\n            function_call<\/span>=\"auto\"<\/span>,<\/span>\n        )<\/span>\n<\/span>\n        message<\/span> = second_response[\"choices\"<\/span>][0<\/span>][\"message\"<\/span>]<\/span>\n        if message.get(\"function_call\"):<\/span>\n            # Step 3, call the function<\/span><\/span>\n            function_call<\/span> = message.get('function_call'<\/span>)<\/span>\n            arguments<\/span> = json.loads(function_call.get('arguments'<\/span>))<\/span>\n            function_response<\/span> = perform_query(doc_id=arguments.get(\"doc_id\"<\/span>), query=arguments.get(\"query\"<\/span>))<\/span>\n<\/span>\n            return function_response<\/span>\n<\/span>\n<\/span>\nprint(run_conversation())<\/span>\n<\/code><\/pre>\n
\u5177\u4f53\u8bb2\u4e00\u4e0b\u6574\u4e2a\u7684\u8c03\u7528\u903b\u8f91\uff1a
\n\u6211\u4eec\u5148\u4e0a\u4f20\u4e00\u4e2apdf\u6587\u4ef6\uff0c\u7136\u540e\u5411gpt\u63d0\u95ee\uff1a\u6211\u60f3\u8981\u4e0a\u4f20xxx\u6587\u4ef6\uff0c\u7136\u540egpt\u53d1\u73b0\u81ea\u5df1\u9700\u8981\u4e00\u4e2a\u4e0a\u4f20pdf\u6587\u4ef6\u7684\u51fd\u6570\uff0c\u68c0\u67e5\u4e86\u4f60\u5728functions\u91cc\u9762\u5199\u7684description\uff0c\u7136\u540e\u627e\u5230\u4e86\u90a3\u4e2a\u4e0a\u4f20\u6587\u4ef6\u7684\u51fd\u6570\uff0c\u7136\u540e\u8c03\u7528\u51fd\u6570\uff1b\u63a5\u7740\u6211\u4eec\u5411gpt\u63d0\u95ee\uff1a\u6211\u60f3\u8981\u67e5\u627exxx\u5185\u5bb9\u5728xx\u6587\u4ef6\u4e2d\uff08\u8fd9\u4e2a\u6587\u4ef6\u5c31\u662f\u6211\u4eec\u4e0a\u9762\u4e0a\u4f20\u7684\u6587\u4ef6\uff09\uff0c\u7136\u540egpt\u601d\u8003\u53d1\u73b0\u9700\u8981\u4e00\u4e2a\u67e5\u8be2\u6587\u4ef6\u7684\u51fd\u6570\uff0c\u91cd\u590d\u524d\u9762\u7684\u64cd\u4f5c\uff0c\u627e\u5230\u4e86\u6211\u4eec\u7ed9\u4ed6\u7684\u51fd\u6570\uff0c\u518d\u6b21\u5de5\u4f5c\u3002<\/p>\n
\u6574\u4e2a\u8fc7\u7a0b\u4f60\u4f1a\u53d1\u73b0\u975e\u5e38\u7684\u4e1d\u6ed1\uff0c\u8fc7\u53bb\u6211\u4eec\u662f\u7528langchain\u628a\u8fc7\u7a0b\u5199\u6b7b\uff0c\u5148\u8c03\u7528xxx\u7684api\u62ff\u5230xxx\uff0c\u518d\u7ed9gpt\u4f7f\u7528\uff0c\u4e5f\u5c31\u662f\u6211\u4eec\u4eba\u4e3a\u786e\u5b9a\u6574\u4e2a\u94fe\u8def\u3002\u4f46\u662f\u8fd9\u91cc\u6211\u4eec\u662f\u5c06gpt\u4f5c\u4e3a\u7cfb\u7edf\u7684\u5927\u8111\uff0c\u9700\u8981\u4ec0\u4e48\u529f\u80fd\uff0c\u5728functions\u91cc\u9762\u627e\u3002<\/p>\n
\u603b\u7ed3<\/h2>\n
\u6211\u76f8\u4fe1\u672a\u6765\u8fd9\u4e2a\u751f\u6001\u4e5f\u662f\u5982\u6b64\uff0cgpt\u94fe\u63a5\u7740\u51e0\u4e4e\u6240\u6709\u7684API\u5de5\u5177\uff0c\u5728\u601d\u8003\u7684\u8fc7\u7a0b\u4e2d\u8c03\u7528\u5373\u53ef\uff0c\u4e0d\u9700\u8981\u6211\u4eec\u4eba\u4e3a\u9650\u5236\u5b83\u601d\u8003\u7684\u8fc7\u7a0b\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"
\u5f15\u8a00 \u4eca\u5929OPENAI\u53d1\u5e03\u91cd\u78c5\u66f4\u65b0\uff0c\u5305\u62ec\uff1a 1.Chat Completions API \u4e2d\u7684\u65b0\u51fd\u6570\u8c03\u7528\u529f\u80fd 2.gpt-4\u548c\u7684\u66f4\u65b0\u548c\u66f4\u6613\u4e8e\u64cd\u7eb5\u7684\u7248\u672cgpt-3.5-turbo 3.\u65b0\u7684 16k \u4e0a\u4e0b<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"rank_math_title":"OpenAI\u51fd\u6570\u8c03\u7528\uff1aGPT\u81ea\u4e3b\u8c03\u7528API\uff0c\u5f00\u542fAI\u65b0\u65f6\u4ee3 - \u5948\u65af\u79d1\u6280\u793e\u533a","rank_math_description":"OpenAI\u91cd\u78c5\u66f4\u65b0Chat Completions API\uff0c\u5f15\u5165\u51fd\u6570\u8c03\u7528\u529f\u80fd\uff0c\u8ba9GPT\u66f4\u5f3a\u5927\uff01\u65e0\u9700\u4eba\u4e3a\u5e72\u9884\uff0cGPT\u53ef\u81ea\u4e3b\u8c03\u7528API\u5b8c\u6210\u4efb\u52a1\uff0c\u5f00\u542fAI\u65b0\u65f6\u4ee3\uff01","rank_math_focus_keyword":"OpenAI, GPT, \u51fd\u6570\u8c03\u7528, Langchain, API","views":"12","footnotes":""},"categories":[3],"tags":[128,129,136,126,127],"collection":[],"class_list":["post-1162","post","type-post","status-publish","format-standard","hentry","category-fenlei2","tag-128","tag-129","tag-136","tag-gpt","tag-ai"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/posts\/1162","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/comments?post=1162"}],"version-history":[{"count":0,"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/posts\/1162\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/media?parent=1162"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/categories?post=1162"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/tags?post=1162"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/collection?post=1162"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}