pywencai

PyWenCai (Tonghuashun WenCai data query)

Use Chinese natural language via Python fromTonghuashun WenCaiquery A-shares and other market data.

⚠️ Requires Cookie: a valid Cookie for the WenCai website must be provided. See below for how to obtain it.

Environment Requirements

Python 3.7+
Node.js v16+(pywencai internally executes JS code)
pip package manager

Installation

pip install pywencai --upgrade

How to Obtain Cookie

If the environment variable WENCAI_COOKIEexists, it can be used directly; if the call fails, obtain it again as follows:

Open in your browser https://www.iwencai.com and log in.
Press F12 to open developer tools → switch to the Network tab.
Perform any query on the WenCai page.
Find the request sent toiwencai.comand copy from the request headers theCookievalue.
Use this string as thecookieparameter.

Cookie Management

Prefer using environment variables or files to manage the Cookie, avoiding hardcoding:

import os
import pywencai

# 方法1：从环境变量读取Cookie（推荐）
cookie = os.environ.get('WENCAI_COOKIE', '')

# 方法2：从文件读取Cookie
def load_cookie(path='~/.wencai_cookie'):
    path = os.path.expanduser(path)
    if os.path.exists(path):
        with open(path) as f:
            return f.read().strip()
    return ''

# 方法3：封装查询函数，统一管理Cookie和错误处理
def query(q, **kwargs):
    cookie = load_cookie()
    try:
        return pywencai.get(
            query=q, cookie=cookie,
            no_detail=True, retry=3, sleep=1,
            **kwargs
        )
    except Exception as e:
        print(f"查询失败: {e}")
        return None

# 使用
df = query('今日涨停的股票')

Basic Usage

import pywencai

# 查询今日涨幅前10的股票，需要有效cookie
res = pywencai.get(query='今日涨幅前10', cookie='your_cookie_here')
print(res)

API Reference:`pywencai.get(**kwargs)`

Required Parameters

query — Chinese natural language query string, e.g. '今日涨停股票'、'市盈率小于20的股票'
cookie — Cookie string obtained from the WenCai website (required)

Optional Parameters

sort_key — sort field name, e.g. '退市@退市日期'
sort_order — sort direction:'asc'(ascending) or 'desc'(descending)
page — page number (default:1）
perpage — number of results per page (default and maximum:100）
loop — set toTrueto get all pages; or set to an integernto get the first n pages
query_type — query category (default:'stock'), optional values:

Value	Description
`stock`	A-share stocks
`zhishu`	Indices
`fund`	Funds
`hkstock`	Hong Kong stocks
`usstock`	US stocks
`threeboard`	New Third Board
`conbond`	Convertible bonds
`insurance`	Insurance
`futures`	Futures
`lccp`	Wealth management products

retry — number of retries on failure (default:10）
sleep — delay in seconds between paginated requests (default:0）
log — set toTrueprint logs in the console
pro — set toTrueuse the paid version (requires the corresponding cookie)
no_detail — set toTruealways returnDataFrameorNone(do not return dict)
find — list of preferred stock codes, e.g. ['600519', '000010']
request_params — passed torequestsadditional parameters, e.g. {'proxies': proxies}

Return Values

List-type queries → return pandas.DataFrame
Detail-type queries → return dict(may contain text and DataFrame)

Usage Examples

Query by market type

import pywencai

# A股股票（默认）
res = pywencai.get(query='今日涨停的股票', cookie=cookie)

# 指数数据
res = pywencai.get(query='上证指数近5日涨跌幅', query_type='zhishu', cookie=cookie)

# 基金数据
res = pywencai.get(query='近一年收益率最高的前20只基金', query_type='fund', cookie=cookie)

# 可转债数据
res = pywencai.get(query='可转债溢价率小于10%', query_type='conbond', cookie=cookie)

# 港股数据
res = pywencai.get(query='港股市值最大的前20只股票', query_type='hkstock', cookie=cookie)

Fundamental stock selection

# 估值筛选
res = pywencai.get(query='市盈率小于20的股票', cookie=cookie)
res = pywencai.get(query='市盈率小于10且市净率小于1的股票', cookie=cookie)

# 财务指标
res = pywencai.get(query='ROE大于15%且营收同比增长率大于20%的股票', cookie=cookie)

# 多条件综合筛选
res = pywencai.get(query='市盈率小于20且营收同比增长大于30%且机构持仓比例大于10%的股票', cookie=cookie)
res = pywencai.get(query='今日站上20日均线且市盈率小于30且ROE大于10%的股票', cookie=cookie)

Technical and capital flow

# 技术信号
res = pywencai.get(query='今日MACD金叉的股票', cookie=cookie)
res = pywencai.get(query='KDJ的J值小于0的股票', cookie=cookie)
res = pywencai.get(query='今日成交量是5日均量2倍以上且涨幅大于5%的股票', cookie=cookie)

# 资金流向
res = pywencai.get(query='今日主力资金净流入前20的股票', cookie=cookie)
res = pywencai.get(query='北向资金持股比例最高的前20只股票', cookie=cookie)

Sorting, pagination, and proxy

# 按指定字段排序
res = pywencai.get(
    query='退市股票',
    sort_key='退市@退市日期',
    sort_order='asc',
    cookie=cookie
)

# 自动分页获取全部数据（使用代理）
proxies = {'http': 'http://proxy:8080', 'https': 'http://proxy:8080'}
res = pywencai.get(
    query='昨日涨幅',
    loop=True,
    log=True,
    cookie=cookie,
    request_params={'proxies': proxies}
)

Historical data query

# 指定日期
res = pywencai.get(query='2024年1月2日涨幅前10的股票', cookie=cookie)

# 日期范围
res = pywencai.get(query='2024年上半年涨幅最大的前20只股票', cookie=cookie)

Complete example: multi-strategy automated stock selection and export

import pywencai
import pandas as pd
import time

cookie = os.environ.get('WENCAI_COOKIE', '')

# 定义多个筛选策略
strategies = {
    "低估值高分红": "市盈率小于15且股息率大于3%的股票",
    "高成长": "营收同比增长大于30%且净利润同比增长大于30%的股票",
    "技术突破": "今日放量突破20日均线且涨幅大于3%的股票",
    "机构关注": "近一个月机构调研次数大于3次的股票",
    "北向资金": "北向资金今日净买入前20的股票",
}

results = {}
for name, query in strategies.items():
    try:
        res = pywencai.get(query=query, cookie=cookie, no_detail=True)
        if res is not None and not res.empty:
            results[name] = res
            print(f"策略 [{name}] 选出 {len(res)} 只股票")
        else:
            print(f"策略 [{name}] 无结果")
    except Exception as e:
        print(f"策略 [{name}] 查询失败: {e}")
    time.sleep(2)  # 每次查询间隔2秒，避免被封禁

# 保存结果到Excel（每个策略一个工作表）
if results:
    with pd.ExcelWriter("选股结果.xlsx") as writer:
        for name, df in results.items():
            df.to_excel(writer, sheet_name=name, index=False)
    print("筛选结果已保存到 选股结果.xlsx")

Usage tips

Use in moderation — high-frequency calls may be banned by the WenCai server; it is recommended to have an interval of ≥ 2 seconds between each query.
Always usethe latest version：pip install pywencai --upgrade
Query string usageChinese natural language — write queries as if searching on the WenCai website.
Whenloop=Trueand setfindwhen,loopis ignored, and only the first 100 results are returned.
When using paid data, setpro=Trueand provide a validcookie。

Common error handling

Error	Cause	Solution
`Cookie expired`	Cookie expired	Log in to the WenCai website again to obtain a new Cookie
Return`None`	Query returned no results or was rate-limited	Check the query statement and reduce the call frequency
`Node.js not found`	Node.js not installed	Install Node.js v16+
`JSONDecodeError`	Server returned an exception	Increase`retry`parameter and retry later
Returns a dict instead of a DataFrame	The query is of detail type	Set`no_detail=True`force return of a DataFrame

Rules

Before using this Skill, confirm that what the user needs isWenCai natural language data query, rather than a pure market data API (such as baostock) or a backtesting framework (such as backtrader/rqalpha).
In the examples, the cookie parameter should uniformly use the environment variable WENCAI_COOKIE or the Cookie file management method,do not hard-code Cookie strings。
Always set no_detail=True to ensure returning DataFrame(unless the user explicitly needs a detail dict).
For batch queries, the sleep parameter must be set (recommended ≥ 2 seconds) to avoid being banned by the server.
Query strings should use Chinese natural language, consistent with the input method in the WenCai website search box.