Waxsoosaarka qaabaysan

POST https://api.shannon-ai.com/v1/chat/completions

import json
from openai import OpenAI

client = OpenAI(api_key="YOUR_API_KEY", base_url="https://api.shannon-ai.com/v1")

response = client.chat.completions.create(
    model="Kimi-K3-3BIT-REAP",
    messages=[
        {"role": "user", "content": "Extract the person: John Doe, 30 years old, engineer."}
    ],
    response_format={
        "type": "json_schema",
        "json_schema": {
            "name": "person",
            "strict": True,
            "schema": {
                "type": "object",
                "properties": {
                    "name": {"type": "string"},
                    "age": {"type": "integer"},
                    "occupation": {"type": "string"},
                },
                "required": ["name", "age", "occupation"],
            },
        },
    },
)

# The JSON arrives as text in message.content: parse it, then check it.
try:
    person = json.loads(response.choices[0].message.content)
except json.JSONDecodeError:
    person = None  # ask again, or report the failure

print(person)

import OpenAI from "openai";

const client = new OpenAI({ apiKey: "YOUR_API_KEY", baseURL: "https://api.shannon-ai.com/v1" });

const response = await client.chat.completions.create({
  model: "Kimi-K3-3BIT-REAP",
  messages: [
    { role: "user", content: "Extract the person: John Doe, 30 years old, engineer." },
  ],
  response_format: {
    type: "json_schema",
    json_schema: {
      name: "person",
      strict: true,
      schema: {
        type: "object",
        properties: {
          name: { type: "string" },
          age: { type: "integer" },
          occupation: { type: "string" },
        },
        required: ["name", "age", "occupation"],
      },
    },
  },
});

// The JSON arrives as text in message.content: parse it, then check it.
let person = null;
try {
  person = JSON.parse(response.choices[0].message.content);
} catch {
  // ask again, or report the failure
}

console.log(person);

curl https://api.shannon-ai.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "Kimi-K3-3BIT-REAP",
    "messages": [
      {"role": "user", "content": "Extract the person: John Doe, 30 years old, engineer."}
    ],
    "response_format": {
      "type": "json_schema",
      "json_schema": {
        "name": "person",
        "strict": true,
        "schema": {
          "type": "object",
          "properties": {
            "name": {"type": "string"},
            "age": {"type": "integer"},
            "occupation": {"type": "string"}
          },
          "required": ["name", "age", "occupation"]
        }
      }
    }
  }'

JSON-ku waa qoraalka jawaabta, choices[0].message.content:

{
  "id": "chatcmpl-0c4e7a1b9d2f4e6a8b3c5d7e9f1a2b3c",
  "object": "chat.completion",
  "created": 1791590400,
  "model": "Kimi-K3-3BIT-REAP",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "{\"name\":\"John Doe\",\"age\":30,\"occupation\":\"engineer\"}",
        "reasoning_content": null
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 96,
    "completion_tokens": 19,
    "total_tokens": 115,
    "prompt_tokens_details": {"cached_tokens": 0},
    "completion_tokens_details": {"reasoning_tokens": 0}
  }
}

Qaababka response_format

response_format waa field ka mid ah /v1/chat/completions. La'aantiis jawaabtu waa qoraal caadi ah.

Qaabka shay JSON ah schema ma u baahna. Furayaasha ku magacaw prompt-ka:

{
  "model": "shannon-3",
  "messages": [
    {
      "role": "user",
      "content": "Name the capital of France. Answer as JSON with the keys city and country."
    }
  ],
  "response_format": {"type": "json_object"}
}

Taageerada model kasta

GET /v1/models wuxuu sheegaa isla shaygaas id kasta: capabilities.json_schema model kasta, iyo capabilities.response_format model-yada open-weight ee la martigeliyo. Model-yo & qiimo

Erayada schema ee la taageero

Waxa model-ku ka helo schema-gaaga wuxuu ku xiran yahay model-ka iyo endpoint-ka:

Qayb-yarta erayada model-yada Shannon /v1/chat/completions:

/v1/responses: text.format

/v1/responses wuxuu isla codsiga u qoraa text.format. Schema-hu wuxuu si toos ah ugu jiraa format, name iyo strict agtooda, ma aha furaha json_schema hoostiisa. Qaabka shay JSON ah dir {"format": {"type": "json_object"}}.

{
  "model": "Kimi-K3-3BIT-REAP",
  "input": "Extract the person: John Doe, 30 years old, engineer.",
  "text": {
    "format": {
      "type": "json_schema",
      "name": "person",
      "strict": true,
      "schema": {
        "type": "object",
        "properties": {
          "name": {"type": "string"},
          "age": {"type": "integer"},
          "occupation": {"type": "string"}
        },
        "required": ["name", "age", "occupation"]
      }
    }
  }
}

JSON-ku waa text ee qaybta output_text ee shayga message:

{
  "output": [
    {
      "id": "msg_7e1d3c5b9a2f4d6e8c0b1a3f5e7d9c2b",
      "type": "message",
      "role": "assistant",
      "status": "completed",
      "content": [
        {
          "type": "output_text",
          "text": "{\"name\":\"John Doe\",\"age\":30,\"occupation\":\"engineer\"}",
          "annotations": []
        }
      ]
    }
  ]
}

Taageerada model kasta iyo khaladaadku waxay la mid yihiin /v1/chat/completions. Responses API

/v1/messages: tool sida schema

Qaabka Messages ma laha field structured-output. JSON ka timaada isla model-yada, wac /v1/chat/completions ama /v1/responses isla furahaaga.

Si aad ugu sii jirto /v1/messages, shayga u sharax sida tool: geli schema-gaaga input_schema oo tool-ka magacaw tool_choice. input ee block-ga tool_use kaddib waa shayga, horeba la parse-gareeyay. tool_choice la magacaabay waxaa dabaqa model-yada open-weight ee la martigeliyo.

{
  "model": "Kimi-K3-3BIT-REAP",
  "max_tokens": 1024,
  "messages": [
    {
      "role": "user",
      "content": "Extract the person: John Doe, 30 years old, engineer."
    }
  ],
  "tools": [
    {
      "name": "record_person",
      "description": "Record the extracted person.",
      "input_schema": {
        "type": "object",
        "properties": {
          "name": {"type": "string"},
          "age": {"type": "integer"},
          "occupation": {"type": "string"}
        },
        "required": ["name", "age", "occupation"]
      }
    }
  ],
  "tool_choice": {"type": "tool", "name": "record_person"}
}

{
  "stop_reason": "tool_use",
  "content": [
    {
      "type": "tool_use",
      "id": "toolu_4b8d2f6a1c3e4a5b9d7f0e2c4a6b8d1f",
      "name": "record_person",
      "input": {"name": "John Doe", "age": 30, "occupation": "engineer"}
    }
  ]
}

Hubi in stop_reason yahay tool_use ka hor inta aadan akhrin block-ga. Wacyigelinta Shaqada

Sida JSON-ku u yimaado

• JSON-ku waa string. Jawaabtu shay la parse-gareeyay agteeda ma haysato: JSON parser-kaaga ku orodsii message.content.
• Marka stream: true JSON-ku wuxuu u yimaadaa gabal-gabal delta.content. Isku xir oo parse-garee marka stream-ku dhammaado; gabal keligiis ah ma aha JSON sax ah.
• Model open-weight oo la martigeliyay wuxuu ku jawaabaa codsi response_format oo qoraal reasoning la'aan ah: reasoning_content waa null.

Hubi waxa aad heshay

Schema-hu wuxuu hagaa sida model-ku u qoro. API-gu jawaabta dhammaystiran schema-gaaga kuma barbar dhigo, sidaas darteed code-kaagu wuxuu sameeyaa hubinta ugu dambeysa:

• Parse-garee gudaha error handler, sida tusaalaha kore.
• Shayga la parse-gareeyay ku xaqiiji isla schema code-kaaga, tusaale ahaan JSON Schema library, Pydantic ama Zod.
• Jawaab dheer waxay ku dhammaan kartaa ka hor inta JSON-ku dhammaystirmin. Jawaab aan stream ahayn waxay sheegtaa finish_reason sida stop ama tool_calls oo keliya, sidaas darteed khalad parse waa calaamadda. Mar kale codso, ama codso shay ka yar.

Khaladaad

Labada jawaab ee 400 waxaa la diraa ka hor inta wax laga qaadin haraaggaaga. Khaladaadka Maareynta