量化交易学习-week1

1
---
2
title: "量化交易学习-week1"
3
description: "量化交易学习"
4
pubDate: "2025-10-23"
5
draft: false
6
---
7

8
## 学习 polars的`.groupby()`, `.rolling()`, `.resample()`
9

10
量化交易中最常见的是对于时间序列数据的处理
11
对于 polars 来说常见的 api 是 `group_by`, `rolling`, `groupby_dynamic`, `resample`
12

13
https://pola-rs.github.io/polars-book-cn/user-guide/dsl/groupby.html
14
https://pola-rs.github.io/polars-book-cn/user-guide/timeseries/intro.html
15

16
pandas 也是类似的逻辑，文档写的很详细我就不再重复了
17

18
来试试实践吧
19
baostock 提供了 5 分钟级别的数据
20

21
```python
22
import baostock as bs
23
import plotly.graph_objects as go
24
import polars as pl
25

26
lg = bs.login()
27

28
rs = bs.query_history_k_data_plus(
29
    "sh.600000",
30
    "date,time,code,open,high,low,close,volume,amount,adjustflag",
31
    start_date="2024-07-01",
32
    end_date="2024-12-31",
33
    frequency="5",
34
    adjustflag="3",
35
)
36
print("query_history_k_data_plus respond error_code:" + rs.error_code)
37
print("query_history_k_data_plus respond  error_msg:" + rs.error_msg)
38

39
data_list = []
40
while (rs.error_code == "0") & rs.next():
41
    # 获取一条记录，将记录合并在一起
42
    data_list.append(rs.get_row_data())
43
df = pl.DataFrame(data_list, schema=rs.fields, orient="row")
44

45

46
df = df.select(
47
    pl.col("date").str.to_date("%Y-%m-%d"),
48
    pl.col("time").str.to_datetime("%Y%m%d%H%M%S%3f"),
49
    pl.col("low").cast(pl.Float32),
50
    pl.col("high").cast(pl.Float32),
51
    pl.col("open").cast(pl.Float32),
52
    pl.col("close").cast(pl.Float32),
53
    pl.col("volume").cast(pl.Float32),
54
    pl.col("amount").cast(pl.Float32),
55
)
56

57
df = (
58
    df.group_by("date")
59
    .agg(
60
        pl.min("low"),
61
        pl.max("high"),
62
        pl.col("open").sort_by("time").first(),
63
        pl.col("close").sort_by("time").last(),
64
        pl.sum("volume"),
65
    )
66
    .sort("date")
67
)
68

69
df = df.with_columns(
70
    pl.col("close").rolling_mean(5).alias("MA5"),
71
    pl.col("close").rolling_mean(20).alias("MA20"),
72
)
73

74
fig = go.Figure()
75
fig.add_trace(
76
    go.Candlestick(
77
        x=df["date"],
78
        open=df["open"],
79
        high=df["high"],
80
        low=df["low"],
81
        close=df["close"],
82
        name="600000",
83
    )
84
)
85
fig.add_trace(
86
    go.Scatter(
87
        x=df["date"],
88
        y=df["MA5"],
89
        mode="lines",
90
        name="MA5",
91
        line=dict(width=1.2, color="orange"),
92
    )
93
)
94
fig.add_trace(
95
    go.Scatter(
96
        x=df["date"],
97
        y=df["MA20"],
98
        mode="lines",
99
        name="MA20",
100
        line=dict(width=1.2, color="blue"),
101
    )
102
)
103

104
fig.update_xaxes(type="category")
105
fig.show()