pip install quantdb-sdk

# 安装核心依赖与开发测试组件
pip install -e ".[dev]"

import os
from quantdb_sdk import QuantDBClient

# 从环境变量 QUANTDB_API_KEY 获取凭据
client = QuantDBClient(api_key=os.environ["QUANTDB_API_KEY"])

client = QuantDBClient(
    username="trader01",
    password="your_secure_password"
)

# 查询已用流量与当前订阅状态（套餐额度与流量包额度分开返回）
usage = client.get_usage()
print(f"套餐:   已用 {usage['plan_used_gb']:.2f} / {usage['plan_total_gb']:.2f} GB，剩余 {usage['plan_remaining_gb']:.2f} GB")
print(f"流量包: 已用 {usage['package_used_gb']:.2f} / {usage['package_total_gb']:.2f} GB，剩余 {usage['package_remaining_gb']:.2f} GB")
print(f"合计:   已用 {usage['used_gb']:.2f} / {usage['total_gb']:.2f} GB，剩余 {usage['remaining_gb']:.2f} GB")
print(f"订阅状态: {usage['subscription'].get('status')}")

# 列出已有 Key
keys = client.list_api_keys()

# 创建新 Key
new_key = client.create_api_key(description="Quant Server 01")
print(new_key["key"])

# 1. 列出可购买套餐
plans = client.list_plans()
for p in plans:
    print(p["id"], p["name"], p["amount_yuan"])

# 2. 创建专业版月付订单（¥99 / 月）
order = client.create_order(plan_id="pro-monthly")
print("支付链接:", order["pay_url"])

# 3. 校验订单状态
status = client.get_order(order_id=order["order_id"])
print(status["status"])

# 列出全部数据集（可按分类 / 访问级别 / 关键词过滤）
print(client.data.datasets().head())
print(client.data.datasets(category="6"))          # ML 数据集分类
print(client.data.datasets(keyword="复权"))

# 数据集详情 + 云端元数据（覆盖区间 / 文件数 / 文件大小）
print(client.data.info("l1_factors"))

# 在线查询：按数据集名自动路由（日线 / ETF / Tick / V2 / V1）
df = client.data.query(
    "daily_forward", symbol="600519.SH",
    start_date="2026-09-01", end_date="2026-09-05"
)
tick = client.data.query("tick_data", symbol="600519.SH", trade_date="20260924")

# 本地零流量读取（root 缺省读 QUANTDB_DATA_DIR，可显式指定）
local = client.data.read_local("daily_forward", start_date="2026-09-01")
paths = client.data.local_files("balance", symbol="600519.SH")

# 本地优先，缺失时才显式回退在线（会消耗流量）
df2 = client.data.read("valuation", symbol="600519.SH", prefer="local")

# 其余映射：preview / load / download / sync / mount
client.data.preview("balance", symbol="600519.SH", limit=5)

# 查询贵州茅台前复权日线（返回 DataFrame）
df = client.query_kline(
    symbol="600519.SH",
    adj_type="forward",
    start_date="2026-01-01",
    end_date="2026-07-22"
)
print(df.head())

# 仅查询收盘价与成交量字段
df_mini = client.query_kline("000001.SZ", fields="close,volume", limit=20)

# 查询沪深 300 ETF 日 K 线
df_etf = client.query_etf_kline(
    symbol="510300.SH",
    start_date="2026-09-01",
    end_date="2026-09-03"
)
print(df_etf.head())

# 查询全部子分类的分区大小
df_sizes = client.query_partition_size()
print(df_sizes[["dimension", "file_count", "file_size_gb"]].to_string(index=False))

# 只查 daily_forward 一个子分类的字节数（用于评估该数据集下载流量）
df_one = client.query_partition_size(sub_category="daily_forward")
print(df_one)

# 下载平安银行指定交易日五档 Tick 数据
df_tick = client.query_tick(
    symbol="000001.SZ",
    trade_date="20260720",
    start_ts="09:30:00",
    end_ts="11:30:00"
)
print(df_tick.head())

# 首次同步自动初始化本地 quantdb_sync.sqlite 元数据库
result = client.sync_dataset("daily_forward", save_dir="D:/quantdb-data")
print(f"已下载: {len(result['downloaded'])} 个, 对账跳过: {len(result['matched'])} 个")

# 财务表增量同步（支持断点重试）
financial = client.sync_dataset("balance", save_dir="D:/quantdb-data")

# ETF 日线增量同步（0.3.5 起支持）
etf = client.sync_dataset("etf_kline", save_dir="D:/quantdb-data")

from quantdb_sdk import QuantDBClient, DuckDBWarehouse

client = QuantDBClient(api_key=os.environ["QUANTDB_API_KEY"])
wh = client.get_local_warehouse()

# 挂载日线行情与财务利润表为本地 View
wh.mount_dataset("1", "daily_unadjusted", symbol="600519.SH", view_name="kline")
wh.mount_dataset("3", "income", symbol="600519.SH", view_name="income")

# 零流量本地 SQL JOIN 查询
df_join = wh.query("""
    SELECT k.trade_date, k.close, i.net_profit_excl_min_int_inc 
    FROM kline k 
    LEFT JOIN income i ON k.trade_date = i.m_anntime
    ORDER BY k.trade_date DESC
    LIMIT 20
""")
print(df_join)

# 一键将 300+ 维预计算模型训练宽表装入内存
df_factors = client.load_as_df(category_id="6", sub_category="l1_l2_factors", symbol="train_factors_2026q1")
print(df_factors.shape)  # (175217 行, 367 列)

# 直接喂入 LightGBM 训练
import lightgbm as lgb
feature_cols = [c for c in df_factors.columns if c not in ["wind_code", "date", "target_5d"]]
train_data = lgb.Dataset(df_factors[feature_cols], label=df_factors["target_5d"])
model = lgb.train({}, train_data, num_boost_round=100)

import asyncio
from quantdb_sdk import AsyncQuantDBClient

async def main():
    client = AsyncQuantDBClient(api_key=os.environ["QUANTDB_API_KEY"])

    # 异步并发请求多只股票日线
    tasks = [
        client.a_query_kline("600519.SH", adj_type="forward"),
        client.a_query_kline("000858.SZ", adj_type="forward"),
        client.a_query_stock_list(keyword="芯片"),
    ]
    results = await asyncio.gather(*tasks)
    for df in results:
        print(df.shape)

asyncio.run(main())

import pandas as pd

df = client.query_kline("600519.SH", adj_type="forward", start_date="2025-01-01")
df["ma5"] = df["close"].rolling(5).mean()
df["ma20"] = df["close"].rolling(20).mean()

# 信号生成：MA5 上穿 MA20 买入，下穿卖出
df["signal"] = 0
df.loc[df["ma5"] > df["ma20"], "signal"] = 1
df.loc[df["ma5"] < df["ma20"], "signal"] = -1

df["strat_ret"] = df["signal"].shift(1) * df["close"].pct_change()
print(f"累计收益: {(1 + df['strat_ret']).cumprod().iloc[-1]:.2%}")