{"id":3497,"date":"2026-10-15T08:22:35","date_gmt":"2026-10-14T23:22:35","guid":{"rendered":"https:\/\/qol-kk.com\/wp2\/?p=3497"},"modified":"2026-10-15T08:22:35","modified_gmt":"2026-10-14T23:22:35","slug":"20261004-plan070","status":"publish","type":"post","link":"https:\/\/qol-kk.com\/wp2\/blog\/2026\/10\/15\/20261004-plan070\/","title":{"rendered":"OpenAI API \u3067\u69cb\u9020\u5316\u51fa\u529b\uff08JSON mode \/ Structured Outputs\uff09\u3092\u4f7f\u3046"},"content":{"rendered":"<h2>\u6982\u8981<\/h2>\n<p>LLM \u3092\u30a2\u30d7\u30ea\u306b\u7d44\u307f\u8fbc\u3080\u3068\u304d\u3001\u6700\u521d\u306b\u3076\u3064\u304b\u308b\u306e\u304c\u300c\u8fd4\u3063\u3066\u304f\u308b\u30c6\u30ad\u30b9\u30c8\u304c\u6bce\u56de\u5c11\u3057\u305a\u3064\u9055\u3046\u300d\u3068\u3044\u3046\u554f\u984c\u3067\u3059\u3002\u30d7\u30ed\u30f3\u30d7\u30c8\u3067\u300cJSON \u3067\u8fd4\u3057\u3066\u300d\u3068\u983c\u3093\u3067\u3082\u3001\u524d\u5f8c\u306b\u524d\u7f6e\u304d\u304c\u4ed8\u3044\u305f\u308a\u3001\u30b3\u30fc\u30c9\u30d5\u30a7\u30f3\u30b9\u3067\u56f2\u307e\u308c\u305f\u308a\u3001\u30ad\u30fc\u540d\u304c\u63fa\u308c\u305f\u308a\u3057\u307e\u3059\u3002\u305d\u306e\u305f\u3073\u306b\u5f8c\u51e6\u7406\u306e\u30d1\u30fc\u30b9\u3092\u9811\u5f35\u308b\u3053\u3068\u306b\u306a\u308a\u307e\u3059\u3002<\/p>\n<p>OpenAI API \u306b\u306f\u3001\u3053\u306e\u554f\u984c\u3092 API \u5074\u3067\u5438\u53ce\u3059\u308b\u4ed5\u7d44\u307f\u304c 2 \u6bb5\u968e\u3042\u308a\u307e\u3059\u3002<\/p>\n<ul>\n<li><strong>JSON mode<\/strong> \u2014 \u51fa\u529b\u304c\u5fc5\u305a\u59a5\u5f53\u306a JSON \u306b\u306a\u308b\u3053\u3068\u3092\u4fdd\u8a3c\u3059\u308b\uff08\u30b9\u30ad\u30fc\u30de\u306f\u4fdd\u8a3c\u3057\u306a\u3044\uff09<\/li>\n<li><strong>Structured Outputs<\/strong> \u2014 JSON Schema \u3092\u6e21\u3057\u3001\u51fa\u529b\u304c\u305d\u306e\u30b9\u30ad\u30fc\u30de\u306b\u53b3\u5bc6\u4e00\u81f4\u3059\u308b\u3053\u3068\u3092\u4fdd\u8a3c\u3059\u308b<\/li>\n<\/ul>\n<p>\u672c\u8a18\u4e8b\u3067\u306f\u3001\u3053\u306e 2 \u3064\u306e\u9055\u3044\u3068\u4f7f\u3044\u5206\u3051\u3001Python \u3067\u306e\u5b9f\u88c5\u3001\u30cf\u30de\u308a\u3069\u3053\u308d\u3092\u6574\u7406\u3057\u307e\u3059\u3002<\/p>\n<h2>\u74b0\u5883<\/h2>\n<ul>\n<li>Python 3.11<\/li>\n<li><code>openai<\/code>\uff08Python SDK v1 \u7cfb\uff09<\/li>\n<li><code>pydantic<\/code><\/li>\n<\/ul>\n<pre class=\"lang:default decode:true\">pip install openai pydantic<\/pre>\n<p>API\u30ad\u30fc\u306f\u74b0\u5883\u5909\u6570 <code>OPENAI_API_KEY<\/code> \u3067\u6e21\u3059\u524d\u63d0\u3067\u3059\uff08\u30b3\u30fc\u30c9\u306b\u76f4\u66f8\u304d\u3057\u306a\u3044\uff09\u3002<\/p>\n<pre class=\"lang:default decode:true\">from openai import OpenAI\nclient = OpenAI()  # OPENAI_API_KEY \u3092\u74b0\u5883\u5909\u6570\u304b\u3089\u8aad\u3080<\/pre>\n<h2>\u767a\u751f\u3057\u305f\u554f\u984c<\/h2>\n<p>\u7d20\u6734\u306b\u300cJSON \u3067\u8fd4\u3057\u3066\u300d\u3068\u3060\u3051\u983c\u3080\u3068\u3001\u5b9f\u904b\u7528\u3067\u306f\u6b21\u306e\u63fa\u308c\u304c\u8d77\u304d\u307e\u3059\u3002<\/p>\n<ul>\n<li>\u30b3\u30fc\u30c9\u30d5\u30a7\u30f3\u30b9\uff08```<code> <\/code>`<code>json <\/code>```\uff09\u3067\u56f2\u307e\u308c\u308b<\/li>\n<li>\u5148\u982d\u306b\u300c\u306f\u3044\u3001\u5206\u985e\u3057\u307e\u3057\u305f:\u300d\u306e\u3088\u3046\u306a\u524d\u7f6e\u304d\u304c\u4ed8\u304f<\/li>\n<li>\u30ad\u30fc\u540d\u304c <code>label<\/code> \u3060\u3063\u305f\u308a <code>sentiment<\/code> \u3060\u3063\u305f\u308a\u5b89\u5b9a\u3057\u306a\u3044<\/li>\n<li>\u307e\u308c\u306b JSON \u304c\u58ca\u308c\u3066\u3044\u308b<\/li>\n<\/ul>\n<p>\u3053\u308c\u3092 <code>json.loads()<\/code> \u306b\u305d\u306e\u307e\u307e\u6e21\u3059\u3068\u4f8b\u5916\u304c\u98db\u3073\u3001\u30ea\u30c8\u30e9\u30a4\u3084\u30b5\u30cb\u30bf\u30a4\u30ba\u3092\u66f8\u304f\u3053\u3068\u306b\u306a\u308a\u3001\u30b3\u30fc\u30c9\u304c\u9632\u5fa1\u7684\u306b\u306a\u3063\u3066\u3044\u304d\u307e\u3059\u3002<\/p>\n<h2>\u539f\u56e0<\/h2>\n<p>\u3053\u308c\u306f LLM \u306e\u78ba\u7387\u7684\u306a\u6027\u8cea\u305d\u306e\u3082\u306e\u304c\u539f\u56e0\u3067\u3059\u3002\u30d7\u30ed\u30f3\u30d7\u30c8\u306f\u300c\u51fa\u529b\u306e\u50be\u5411\u300d\u3092\u8a98\u5c0e\u3067\u304d\u307e\u3059\u304c\u3001<strong>\u69cb\u6587\u7684\u306a\u4fdd\u8a3c<\/strong>\u306f\u3067\u304d\u307e\u305b\u3093\u3002\u4fdd\u8a3c\u304c\u6b32\u3057\u3044\u306a\u3089\u3001\u30c8\u30fc\u30af\u30f3\u751f\u6210\u306e\u6bb5\u968e\u3067\u5236\u7d04\u3092\u304b\u3051\u308b\u4ed5\u7d44\u307f\u304c\u5fc5\u8981\u3067\u3059\u3002\u305d\u308c\u304c JSON mode \u3068 Structured Outputs \u3067\u3059\u3002<\/p>\n<h2>\u5bfe\u5fdc\u65b9\u6cd5<\/h2>\n<h3>1. JSON mode\uff08\u59a5\u5f53\u306a JSON \u3092\u4fdd\u8a3c\u3059\u308b\uff09<\/h3>\n<pre class=\"lang:default decode:true\">resp = client.chat.completions.create(\n    model=\"gpt-4o-mini\",\n    response_format={\"type\": \"json_object\"},\n    messages=[\n        {\"role\": \"system\", \"content\": \"\u3042\u306a\u305f\u306fJSON\u306e\u307f\u3092\u8fd4\u3059API\u3067\u3059\u3002\"},\n        {\"role\": \"user\", \"content\": \"\u300e\u601d\u3063\u305f\u3088\u308a\u826f\u304b\u3063\u305f\u300f\u306e\u611f\u60c5\u3092 label \u3068 score \u306eJSON\u3067\u8fd4\u3057\u3066\u3002\"},\n    ],\n)\nprint(resp.choices[0].message.content)  # {\"label\": \"positive\", \"score\": 0.82}<\/pre>\n<p>JSON \u3068\u3057\u3066\u59a5\u5f53\u3067\u3042\u308b\u3053\u3068\u306f\u4fdd\u8a3c\u3055\u308c\u307e\u3059\u3002\u305f\u3060\u3057\u6b21\u306e\u5236\u7d04\u304c\u3042\u308a\u307e\u3059\u3002<\/p>\n<ul>\n<li>\u30b9\u30ad\u30fc\u30de\uff08\u30ad\u30fc\u540d\u30fb\u578b\u30fb\u5fc5\u9808\uff09\u306f\u4fdd\u8a3c\u3055\u308c\u306a\u3044 \u2192 \u30d7\u30ed\u30f3\u30d7\u30c8\u983c\u307f<\/li>\n<li>\u30d7\u30ed\u30f3\u30d7\u30c8\u306b\u300cJSON\u300d\u3068\u3044\u3046\u5358\u8a9e\u3092\u542b\u3081\u308b\u5fc5\u8981\u304c\u3042\u308b<\/li>\n<li>\u51fa\u529b\u4e0a\u9650\u306b\u9054\u3059\u308b\u3068 JSON \u304c\u9014\u4e2d\u3067\u5207\u308c\u308b\u3053\u3068\u304c\u3042\u308b<\/li>\n<\/ul>\n<h3>2. Structured Outputs\uff08\u30b9\u30ad\u30fc\u30de\u307e\u3067\u4fdd\u8a3c\u3059\u308b\uff09<\/h3>\n<p><code>strict: true<\/code> \u3092\u6307\u5b9a\u3059\u308b\u3068\u3001\u51fa\u529b\u304c JSON Schema \u306b\u53b3\u5bc6\u4e00\u81f4\u3059\u308b\u3053\u3068\u304c\u4fdd\u8a3c\u3055\u308c\u307e\u3059\u3002<\/p>\n<pre class=\"lang:default decode:true\">schema = {\n    \"name\": \"sentiment\",\n    \"strict\": True,\n    \"schema\": {\n        \"type\": \"object\",\n        \"properties\": {\n            \"label\": {\"type\": \"string\", \"enum\": [\"positive\", \"negative\", \"neutral\"]},\n            \"score\": {\"type\": \"number\"},\n        },\n        \"required\": [\"label\", \"score\"],\n        \"additionalProperties\": False,\n    },\n}\n\nresp = client.chat.completions.create(\n    model=\"gpt-4o-mini\",\n    response_format={\"type\": \"json_schema\", \"json_schema\": schema},\n    messages=[{\"role\": \"user\", \"content\": \"\u300e\u601d\u3063\u305f\u3088\u308a\u826f\u304b\u3063\u305f\u300f\u306e\u611f\u60c5\u3092\u5206\u985e\u3057\u3066\u3002\"}],\n)\nprint(resp.choices[0].message.content)  # {\"label\":\"positive\",\"score\":0.82}<\/pre>\n<p><code>label<\/code> \u306f\u5fc5\u305a enum \u306e 3 \u5024\u3001<code>score<\/code> \u306f\u5fc5\u305a number\u3001\u4e21\u65b9 required\u3001\u4f59\u8a08\u306a\u30ad\u30fc\u306f\u751f\u3048\u307e\u305b\u3093\u3002<\/p>\n<h3>3. Pydantic \u3067\u66f8\u304f\uff08\u5b9f\u52d9\u3067\u306f\u3053\u308c\u304c\u697d\uff09<\/h3>\n<pre class=\"lang:default decode:true\">from pydantic import BaseModel\nfrom enum import Enum\n\nclass Label(str, Enum):\n    positive = \"positive\"\n    negative = \"negative\"\n    neutral = \"neutral\"\n\nclass Sentiment(BaseModel):\n    label: Label\n    score: float\n\nresp = client.beta.chat.completions.parse(\n    model=\"gpt-4o-mini\",\n    messages=[{\"role\": \"user\", \"content\": \"\u300e\u601d\u3063\u305f\u3088\u308a\u826f\u304b\u3063\u305f\u300f\u306e\u611f\u60c5\u3092\u5206\u985e\u3057\u3066\u3002\"}],\n    response_format=Sentiment,\n)\n\nresult = resp.choices[0].message.parsed  # Sentiment \u578b\u306e\u30a4\u30f3\u30b9\u30bf\u30f3\u30b9\nprint(result.label, result.score)        # =&gt; Label.positive 0.82<\/pre>\n<p><code>parsed<\/code> \u306f\u6700\u521d\u304b\u3089 Pydantic \u30e2\u30c7\u30eb\u306e\u30a4\u30f3\u30b9\u30bf\u30f3\u30b9\u306a\u306e\u3067\u3001\u578b\u88dc\u5b8c\u3082\u30d0\u30ea\u30c7\u30fc\u30b7\u30e7\u30f3\u3082\u52b9\u304d\u307e\u3059\u3002<\/p>\n<h2>\u5b9f\u88c5\u4f8b\uff1a\u60c5\u5831\u62bd\u51fa<\/h2>\n<pre class=\"lang:default decode:true\">from pydantic import BaseModel\n\nclass Inquiry(BaseModel):\n    category: str\n    urgency: int          # 1-5\n    needs_human: bool\n    summary: str\n\nresp = client.beta.chat.completions.parse(\n    model=\"gpt-4o-mini\",\n    messages=[\n        {\"role\": \"system\", \"content\": \"\u554f\u3044\u5408\u308f\u305b\u6587\u3092\u5206\u985e\u30fb\u8981\u7d04\u3059\u308b\u3002\"},\n        {\"role\": \"user\", \"content\": \"\u96fb\u6e90\u304c\u5165\u3089\u306a\u3044\u521d\u671f\u4e0d\u826f\u3002\u4eca\u65e5\u4e2d\u306b\u4ea4\u63db\u53ef\u5426\u3092\u77e5\u308a\u305f\u3044\u3002\"},\n    ],\n    response_format=Inquiry,\n)\n\ninq = resp.choices[0].message.parsed\nif inq.urgency &gt;= 4 and inq.needs_human:\n    print(\"\u2192 \u5373\u30a8\u30b9\u30ab\u30ec\u30fc\u30b7\u30e7\u30f3:\", inq.summary)<\/pre>\n<p>\u51fa\u529b\u304c\u578b\u306b\u53ce\u307e\u3063\u3066\u3044\u308b\u306e\u3067\u3001\u5206\u5c90\u3092\u305d\u306e\u307e\u307e\u66f8\u3051\u307e\u3059\u3002LLM \u306e\u51fa\u529b\u3092\u30d3\u30b8\u30cd\u30b9\u30ed\u30b8\u30c3\u30af\u306b\u76f4\u7d50\u3067\u304d\u308b\u306e\u304c\u69cb\u9020\u5316\u51fa\u529b\u306e\u4fa1\u5024\u3067\u3059\u3002<\/p>\n<h2>\u78ba\u8a8d\u65b9\u6cd5<\/h2>\n<pre class=\"lang:default decode:true\">msg = resp.choices[0].message\nif msg.refusal:              # \u5b89\u5168\u4e0a\u306e\u7406\u7531\u3067\u62d2\u5426\u3055\u308c\u305f\u5834\u5408\n    print(\"\u62d2\u5426\u3055\u308c\u307e\u3057\u305f:\", msg.refusal)\nelse:\n    print(msg.parsed)<\/pre>\n<ul>\n<li>JSON mode: <code>json.loads()<\/code> \u304c\u901a\u308b\u3053\u3068\uff0b\u4e2d\u8eab\u3092\u5225\u9014\u30a2\u30b5\u30fc\u30c8<\/li>\n<li>Structured Outputs: <code>parsed<\/code> \u304c <code>None<\/code> \u3067\u306a\u3044\u3053\u3068\uff0bPydantic \u30d0\u30ea\u30c7\u30fc\u30b7\u30e7\u30f3\u901a\u904e<\/li>\n<li><code>finish_reason<\/code> \u304c <code>length<\/code> \u306a\u3089\u9014\u4e2d\u5207\u65ad\u3092\u7591\u3046<\/li>\n<\/ul>\n<h2>\u6ce8\u610f\u70b9<\/h2>\n<ul>\n<li><strong>\u5bfe\u5fdc\u30e2\u30c7\u30eb\u3092\u78ba\u8a8d\u3059\u308b<\/strong>\u3002<code>json_schema<\/code> + <code>strict<\/code> \u306f\u6bd4\u8f03\u7684\u65b0\u3057\u3044\u4e16\u4ee3\u306e\u30e2\u30c7\u30eb\u3067\u5229\u7528\u53ef\u3002\u53e4\u3044\u30e2\u30c7\u30eb\u306f JSON mode \u307e\u3067\u3002<\/li>\n<li><strong>strict \u306e\u524d\u63d0<\/strong>: \u5168\u30d7\u30ed\u30d1\u30c6\u30a3 <code>required<\/code> \uff0b <code>additionalProperties: false<\/code>\u3002\u624b\u66f8\u304d\u3067\u7701\u304f\u3068\u30a8\u30e9\u30fc\u3002Pydantic \u7d4c\u7531\u306a\u3089\u81ea\u52d5\u3067\u6e80\u305f\u3055\u308c\u308b\u3002<\/li>\n<li><strong>\u30b9\u30ad\u30fc\u30de\u5236\u7d04<\/strong>: \u4e00\u90e8\u306e JSON Schema \u30ad\u30fc\u30ef\u30fc\u30c9\u3084\u904e\u5ea6\u306a\u518d\u5e30\u306f\u672a\u5bfe\u5fdc\u306e\u3053\u3068\u304c\u3042\u308b\u3002<\/li>\n<li><strong>\u51fa\u529b\u4e0a\u9650\u3067\u306e\u5207\u65ad\u306b\u6ce8\u610f<\/strong>\u3002\u69cb\u9020\u5316\u51fa\u529b\u3067\u3082 <code>max_tokens<\/code> \u306b\u9054\u3059\u308b\u3068\u9014\u4e2d\u3067\u6b62\u307e\u308b\u3002<\/li>\n<li><strong>\u4fdd\u8a3c\u3055\u308c\u308b\u306e\u306f\u300c\u5f62\u300d\u3067\u3042\u3063\u3066\u300c\u4e2d\u8eab\u306e\u6b63\u3057\u3055\u300d\u3067\u306f\u306a\u3044<\/strong>\u3002\u578b\u306f\u5b88\u3089\u308c\u3066\u3082\u5024\u306e\u59a5\u5f53\u6027\u306f\u5225\u554f\u984c\u3002<\/li>\n<\/ul>\n<h2>\u307e\u3068\u3081<\/h2>\n<ul>\n<li>\u3068\u306b\u304b\u304f\u58ca\u308c\u306a\u3044 JSON \u304c\u6b32\u3057\u3044\u306a\u3089 <strong>JSON mode<\/strong>\u3002<\/li>\n<li>\u30ad\u30fc\u30fb\u578b\u30fb\u5fc5\u9808\u307e\u3067\u4fdd\u8a3c\u3057\u3066\u30a2\u30d7\u30ea\u306e\u578b\u306b\u76f4\u7d50\u3057\u305f\u3044\u306a\u3089 <strong>Structured Outputs<\/strong>\uff08\u5b9f\u52d9\u306f Pydantic \u7d4c\u7531\uff09\u3002<\/li>\n<li>\u69cb\u9020\u5316\u51fa\u529b\u3092\u4f7f\u3046\u3068\u5f8c\u51e6\u7406\u306e\u30d1\u30fc\u30b9\u304b\u3089\u89e3\u653e\u3055\u308c\u3001\u51fa\u529b\u3092\u305d\u306e\u307e\u307e\u5206\u5c90\u30fb\u4fdd\u5b58\u30fb\u9023\u643a\u306b\u56de\u305b\u307e\u3059\u3002<\/li>\n<li>\u305f\u3060\u3057\u4fdd\u8a3c\u3055\u308c\u308b\u306e\u306f\u300c\u5f62\u300d\u3060\u3051\u3002\u4e2d\u8eab\u306e\u59a5\u5f53\u6027\u306f\u5f93\u6765\u3069\u304a\u308a\u81ea\u5206\u3067\u691c\u8a3c\u3059\u308b\u5fc5\u8981\u304c\u3042\u308a\u307e\u3059\u3002<\/li>\n<\/ul>\n<p>LLM \u3092\u30a2\u30d7\u30ea\u306e\u90e8\u54c1\u3068\u3057\u3066\u6271\u3046\u7b2c\u4e00\u6b69\u306f\u3001\u51fa\u529b\u3092\u300c\u6587\u5b57\u5217\u300d\u304b\u3089\u300c\u578b\u300d\u306b\u5909\u3048\u308b\u3053\u3068\u3067\u3059\u3002\u69cb\u9020\u5316\u51fa\u529b\u306f\u305d\u306e\u305f\u3081\u306e\u6700\u3082\u5b9f\u7528\u7684\u306a\u5165\u53e3\u3067\u3059\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"<p>LLM \u3092\u30a2\u30d7\u30ea\u306b\u7d44\u307f\u8fbc\u3080\u3068\u304d\u3001\u6700\u521d\u306b\u3076\u3064\u304b\u308b\u306e\u304c\u300c\u8fd4\u3063\u3066\u304f\u308b\u30c6\u30ad\u30b9\u30c8\u304c\u6bce\u56de\u5c11\u3057\u305a\u3064\u9055\u3046\u300d\u3068\u3044\u3046\u554f\u984c\u3067\u3059\u3002\u30d7\u30ed\u30f3\u30d7\u30c8\u3067\u300cJSON \u3067\u8fd4\u3057\u3066\u300d\u3068\u983c\u3093\u3067\u3082\u3001\u524d\u5f8c\u306b\u524d\u7f6e\u304d\u304c\u4ed8\u3044\u305f\u308a\u3001\u30b3\u30fc\u30c9\u30d5\u30a7\u30f3\u30b9\u3067\u56f2\u307e\u308c\u305f\u308a\u3001\u30ad\u30fc\u540d\u304c\u63fa\u308c\u305f\u308a\u3057\u307e\u3059\u3002\u305d\u306e\u305f\u3073\u306b\u5f8c\u2026<\/p>\n","protected":false},"author":1,"featured_media":3496,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"sns_share_botton_hide":"","vkExUnit_sns_title":"","_vk_print_noindex":"","sitemap_hide":"","_veu_custom_css":"","veu_display_promotion_alert":"","vkexunit_cta_each_option":"","_jetpack_feature_clip_id":0,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_publicize_message":"","jetpack_publicize_feature_enabled":true,"jetpack_social_post_already_shared":true,"jetpack_social_options":{"image_generator_settings":{"template":"highway","default_image_id":0,"font":"","enabled":false},"version":2},"jetpack_post_was_ever_published":false},"categories":[38],"tags":[64,166,988,147,285],"class_list":["post-3497","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-python-memorandum","tag-api","tag-model","tag-object","tag-python","tag-role"],"veu_head_title_object":{"title":"","add_site_title":""},"jetpack_publicize_connections":[],"jetpack-related-posts":[],"jetpack_sharing_enabled":true,"jetpack_featured_media_url":"https:\/\/qol-kk.com\/wp2\/wp-content\/uploads\/2026\/10\/20261004-plan070-thumbnail.jpg","_links":{"self":[{"href":"https:\/\/qol-kk.com\/wp2\/wp-json\/wp\/v2\/posts\/3497","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/qol-kk.com\/wp2\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/qol-kk.com\/wp2\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/qol-kk.com\/wp2\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/qol-kk.com\/wp2\/wp-json\/wp\/v2\/comments?post=3497"}],"version-history":[{"count":1,"href":"https:\/\/qol-kk.com\/wp2\/wp-json\/wp\/v2\/posts\/3497\/revisions"}],"predecessor-version":[{"id":3535,"href":"https:\/\/qol-kk.com\/wp2\/wp-json\/wp\/v2\/posts\/3497\/revisions\/3535"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/qol-kk.com\/wp2\/wp-json\/wp\/v2\/media\/3496"}],"wp:attachment":[{"href":"https:\/\/qol-kk.com\/wp2\/wp-json\/wp\/v2\/media?parent=3497"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/qol-kk.com\/wp2\/wp-json\/wp\/v2\/categories?post=3497"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/qol-kk.com\/wp2\/wp-json\/wp\/v2\/tags?post=3497"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}