Nano Banana 系列開發指南

curl --location 'https://api.apiyi.com/v1beta/models/gemini-3-pro-image-preview:generateContent' \
  --header 'Authorization: Bearer sk-' \
  --header 'Content-Type: application/json' \
  --data '{
      "contents": [
          {
              "parts": [
                  {
                      "fileData": {
                          "fileUri": "https://raw.githubusercontent.com/apiyi-api/ai-api-code-samples/refs/heads/main/Vision-API-OpenAI/otter.png",
                          "mimeType": "image/png"
                      }
                  },
                  {
                      "text": "add five dogs"
                  }
              ],
              "role": "user"
          }
      ],
      "generationConfig": {"responseModalities": ["IMAGE"],
      "imageConfig": {
        "aspectRatio": "16:9",
        "imageSize": "2K"
      }},
      "safetySettings": []
  }'   > output.json

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Gemini 3 Pro Image - 圖片編輯（file_uri 最小化版本）
用途：僅用於快速驗證介面可用性
"""

import requests
import base64
import json
from pathlib import Path
from datetime import datetime

# ============================================================================
# 配置區域
# ============================================================================

API_KEY = "sk-"
API_URL = "https://api.apiyi.com/v1beta/models/gemini-3-pro-image-preview:generateContent"

# 圖片 URL
IMAGE_URL = "https://raw.githubusercontent.com/apiyi-api/ai-pics/refs/heads/main/1762260696217_dd0352c1f9604540.png"
IMAGE_MIME_TYPE = "image/png"

# 編輯指令
EDIT_PROMPT = "將照片中的人的衣服換成藍色夾克，頭髮換成紫色漸變色，人物的動作、眼睛朝向等其他結構不變"
SYSTEM_PROMPT = "您是一位專業的影像描述和生成專家。您的任務是根據使用者的請求，創作出細節豐富、藝術風格明確的高品質影像提示，或對現有影像進行準確、有創意的編輯。"

# 輸出引數
ASPECT_RATIO = "9:16"
RESOLUTION = "4K"
MAX_OUTPUT_TOKENS = 8000
OUTPUT_FILE = f"minimal_{datetime.now().strftime('%Y%m%d_%H%M%S')}.png"

# ============================================================================
# 核心程式碼
# ============================================================================

def main():
    print("=" * 60)
    print("開始測試 file_uri 格式介面")
    print("=" * 60)
    print(f"圖片 URL: {IMAGE_URL[:80]}...")
    print(f"編輯指令: {EDIT_PROMPT}")
    print(f"輸出引數: {RESOLUTION}, {ASPECT_RATIO}")
    print("-" * 60)

    # 構建請求體
    # 注意：fileData、mimeType、fileUri 必須使用駝峰命名
    payload = {
        "generationConfig": {
            "responseModalities": ["IMAGE", "TEXT"],
            "imageConfig": {
                "imageSize": RESOLUTION,
                "aspectRatio": ASPECT_RATIO
            },
            "maxOutputTokens": MAX_OUTPUT_TOKENS
        },
        "contents": [
            {
                "role": "model",
                "parts": [{"text": SYSTEM_PROMPT}]
            },
            {
                "role": "user",
                "parts": [
                    {
                        "fileData": {           # 駝峰命名：fileData（不是 file_data）
                            "mimeType": IMAGE_MIME_TYPE,  # 駝峰命名：mimeType
                            "fileUri": IMAGE_URL          # 駝峰命名：fileUri
                        }
                    },
                    {"text": EDIT_PROMPT}
                ]
            }
        ]
    }

    # 傳送請求
    print("\n正在傳送請求...")
    try:
        response = requests.post(
            API_URL,
            json=payload,
            headers={
                "Content-Type": "application/json",
                "Authorization": f"Bearer {API_KEY}"
            },
            timeout=300
        )

        print(f"響應狀態碼: {response.status_code}")

        if response.status_code != 200:
            print(f"❌ 錯誤: {response.text}")
            return

        # 解析響應
        data = response.json()
        print("✅ 成功獲取響應")

        # 儲存完整響應（方便除錯）
        with open(OUTPUT_FILE + ".response.json", "w", encoding="utf-8") as f:
            json.dump(data, f, indent=2, ensure_ascii=False)
        print(f"📄 響應已儲存: {OUTPUT_FILE}.response.json")

        # 提取並列印文本
        parts = data["candidates"][0]["content"]["parts"]
        for part in parts:
            if "text" in part:
                print(f"\n💬 文本響應: {part['text']}")

        # 儲存圖片
        for part in parts:
            if "inlineData" in part or "inline_data" in part:
                image_data = part.get("inlineData", part.get("inline_data", {})).get("data")
                if image_data:
                    image_bytes = base64.b64decode(image_data)
                    with open(OUTPUT_FILE, "wb") as f:
                        f.write(image_bytes)
                    print(f"\n✅ 圖片已儲存: {OUTPUT_FILE}")
                    print(f"📦 檔案大小: {len(image_bytes) / 1024:.1f} KB")
                    print(f"🔗 檔案路徑: {Path(OUTPUT_FILE).resolve()}")
                    return

        print("⚠️  響應中未找到圖片資料")

    except requests.Timeout:
        print("❌ 請求超時")
    except Exception as e:
        print(f"❌ 錯誤: {e}")

if __name__ == "__main__":
    main()
    print("\n" + "=" * 60)
    print("測試結束")
    print("=" * 60)

API Connection Error: HTTPSConnectionPool(host='api.apiyi.com', port=443): Read timed out. (read timeout=120)

timeout = {
    "1K": 300,  # 5 分鐘 - 快速預覽
    "2K": 300,  # 5 分鐘 - 推薦使用
    "4K": 600,  # 10 分鐘 - 超高畫質
}

import requests, base64

API_KEY = "sk-your-api-key"
URL = "https://api.apiyi.com/v1beta/models/gemini-3.1-flash-image-preview:generateContent"
H = {"Authorization": f"Bearer {API_KEY}", "Content-Type": "application/json"}
CFG = {"responseModalities": ["IMAGE"], "imageConfig": {"imageSize": "2K"}}

contents = []  # 全程維護同一個對話歷史

def turn(instruction, save_to):
    contents.append({"role": "user", "parts": [{"text": instruction}]})
    data = requests.post(URL, headers=H,
                         json={"contents": contents, "generationConfig": CFG}, timeout=300).json()
    part = next(p for p in data["candidates"][0]["content"]["parts"] if "inlineData" in p)
    contents.append({"role": "model", "parts": [part]})   # 關鍵：把產出圖回填進歷史
    with open(save_to, "wb") as f:
        f.write(base64.b64decode(part["inlineData"]["data"]))

turn("生成一隻橙色的貓，坐在藍色沙發上，簡筆畫風格", "step1.png")
turn("把沙發改成紅色，貓和構圖保持不變", "step2.png")     # 基於上一輪的圖
turn("給貓戴一頂黃色小帽子，其它保持不變", "step3.png")     # 繼續累積，紅沙發會保留

parts = response["candidates"][0]["content"]["parts"] or []   # 安全拒絕時 parts 為 null
images = [p["inlineData"]["data"] for p in parts if "inlineData" in p]

if images:
    final_image = images[-1]   # 最後一張 = 最終稿