Financial Center: open financial data framework

• Financial Center: open financial data framework
• Overview
  • Running Shot
  • Logging Shot
  • Database Shot
  • Flask Demo Shot
• Description
• Installation guide
  • 1. Packages / Dependencies
  • 2. Database
  • 3. Redis
  • 4. Kafka
  • 5. Pip Packages
• Configure FinDy Settings
• Updates:
  • (25 Dec 2021)
  • (01 Mar 2021)
  • (16 Sep 2020):
• How to use
  • Fetch data
  • Read data
• Docker Usage
  • Docker Execute
  • View Log
• To Do

Overview

Running Shot

Logging Shot

Database Shot

Flask Demo Shot

Description

Based on registry framework, project can easily compat various of thrid-party data sources.
Below shows all data sources that project already supported.

Third Party Data Sources Support Table
China Market	US Market	HK Market
BaoStock	Yahoo
JoinQuant	alpaca
Tushare
EastMoney

To add more third party data sources, please reference (findy/database/plugins/***) modules for modification.

Installation guide

The FinDy installation consists of setting up the following components:

1. Packages / Dependencies

Command line tools

xcode-select --install #xcode command line tools

Homebrew

/bin/bash -c "$(curl -fsSL https://gitee.com/ineo6/homebrew-install/raw/master/install.sh)"
echo 'eval "$(/opt/homebrew/bin/brew shellenv)"' >> /Users/huangdon/.zprofile
eval "$(/opt/homebrew/bin/brew shellenv)"

brew install git cmake pkg-config openssl
brew link openssl --force

2. Database

FinDy recommends using a PostgreSQL database. But you can use MySQL too, see MySQL setup guide.

# Install package
brew install postgres
postgres -V

# Start PostgreSQL server
pg_ctl -D /usr/local/var/postgres start
#Or you can start the PostgreSQL server and have it start up at login automatically
brew services start postgresql

# Stop PostgreSQL server
pg_ctl -D /usr/local/var/postgres stop
#To make it stop starting up automatically at login
brew services stop postgresql

# Login to PostgreSQL
psql -U XXX(superuser role) -d postgres

# Create a user for FinDy
CREATE USER postgres with NOSUPERUSER CREATEDB CREATEROLE INHERIT LOGIN;

# Create the FinDy production database & grant all privileges on database
CREATE DATABASE findy OWNER postgres;

# Quit the database session
\q

# Try connecting to the new database with the new user
sudo -u git -H psql -d findy

# tuning postgres (optional)
cpan -i DBD::Pg
cd pgtune
chmod +x postgresqltuner.pl
./pg_tune.sh

For MacOS, if you wish to enlarge system share memory from 4MB to 512 MB, make following:

1. Disable SIP: Restart computer to Recovery mode, Select "Terminal" and run command csrutil disable, then restart computer
2. Create /Library/LaunchDaemons/com.startup.sysctl.plist

     <?xml version="1.0" encoding="UTF-8"?>
     <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
     <!-- DISABLE SIP TO USE: macOS Recovery > Utilities > Terminal > `csrutil disable` > Reboot -->
     <plist version="1.0">
         <dict>
             <key>Label</key>
             <string>com.startup.sysctl</string>
             <key>LaunchOnlyOnce</key>
             <true/>
             <key>StandardErrorPath</key>
                 <string>/private/tmp/sysctl.err</string>
             <key>StandardOutPath</key>
                 <string>/private/tmp/sysctl.out</string>
             <key>ProgramArguments</key>
             <array>
                 <string>/usr/sbin/sysctl</string>
                 <string>-w</string>
                 <string>kern.sysv.shmmax=16777216</string>
                 <string>kern.sysv.shmmin=1</string>
                 <string>kern.sysv.shmmni=128</string>
                 <string>kern.sysv.shmseg=512</string>
                 <string>kern.sysv.shmall=4096</string>
             </array>
             <key>RunAtLoad</key>
             <true/>
         </dict>
     </plist>

3. Load the PLIST after a few housekeeping items

   # sanity check
   sysctl -a | grep shm
   
   # set permissions
   sudo chown root:wheel /Library/LaunchDaemons/com.startup.sysctl.plist
   
   # validate key-value pairs
   plutil /Library/LaunchDaemons/com.startup.sysctl.plist
   
   # load plist
   sudo launchctl bootstrap system /Library/LaunchDaemons/com.startup.sysctl.plist
   
   # check logs
   tail -f /tmp/sysctl.out
   tail -f /tmp/sysctl.err
   
   # recheck sysctl values
   sysctl -a | grep shm

3. Redis

brew install redis
ln -sfv /usr/local/opt/redis/*.plist ~/Library/LaunchAgents

Redis config is located in /usr/local/etc/redis.conf. Make a copy:

cp /usr/local/etc/redis.conf /usr/local/etc/redis.conf.orig

Disable Redis listening on TCP by setting 'port' to 0

sed 's/^port .*/port 0/' /usr/local/etc/redis.conf.orig | sudo tee /usr/local/etc/redis.conf

Edit file (nano /usr/local/etc/redis.conf) and uncomment:

unixsocket /tmp/redis.sock
unixsocketperm 777

Start Redis

launchctl load ~/Library/LaunchAgents/homebrew.mxcl.redis.plist

4. Kafka

# Install package
brew install zookeeper
brew install kafka

modify file /opt/homebrew/etc/kafka/server.properties, active line listeners=PLAINTEXT://:9092 as below:

############################# Socket Server Settings #############################
# The address the socket server listens on. It will get the value returned from 
# java.net.InetAddress.getCanonicalHostName() if not configured.
#   FORMAT:
#     listeners = listener_name://host_name:port
#   EXAMPLE:
#     listeners = PLAINTEXT://your.host.name:9092
listeners=PLAINTEXT://localhost:9092

# Start kafka server
brew services start zookeeper
brew services start kafka

pip install kafka_python

5. Pip Packages

pip install -r requirements_mac_arm64.txt

# optional, generate requirement.txt file from project imports
$ pipreqs ./ --force

Configure FinDy Settings

Default (config.json) setting

{
  "version": "0.0.2",

  "debug": 0,
  "processes": 1,
  "batch_size": 10000,
  
  "db_name": "findy",
  "db_host": "192.168.1.133",
  "db_port": "15432",
  "db_user": "",
  "db_pass": "",

  "redis_pass": "",
  "kafka": "localhost:9092",

  "jq_username": "",
  "jq_password": "",

  "tushare_token": "",

  "trade_host": "127.0.0.1",
  "trade_port": "11111",
  "trade_socket_port": "33333",
  "trade_socket_key": "",

  "http_proxy": "127.0.0.1:1087",
  "https_proxy": "127.0.0.1:1087",
  "smtp_host": "smtpdm.aliyun.com",
  "smtp_port": "80",
  "email_username": "",
  "email_password": "",
  "wechat_app_id": "",
  "wechat_app_secrect": "",
  "wechat_agent_id": ""
}

change database host address to following setting if you install your database locally:

  "db_host": "127.0.0.1",
  "db_port": "5432",

set database user and password to your custom settings:

  "db_user": "xxx",
  "db_pass": "xxx",

chinese stock market user are required to obtain joinquant and tushare authentication.

JoinQuant
TuShare

Updates:

(25 Dec 2021)

• rewrite yahoo finance lib (yfinance) to support aiohttp requests
• remove proxy pool
• add wechat notify (see utils/notify.py for more details)
• dev/async_db branch: add async sqlalchemy ORM
• many bugfix

(01 Mar 2021)

• database session decoupling
• entity provider removed
• database querying speed boost up
• asyncio logic framework (implementations requires rewrite request logic)
• multiprocessing logic rewrite, use processes to replace pool
• add proxy pool (see proxy_pool)
• add baostock thrid party data source

(16 Sep 2020):

• Switch Sql to Postgresql, which allow remote database access, separate fetching and analysis models to two individal project(see FinanceAnalysis project for more detail).
• Centralized http request interface，add Session and retry logic (Rxpy may be involved to rewrite request model pretty soon).
• multiprocessing is involved to boost up request speed.
• request progress display.
• auto joinquant account switching (banned)
• log to console -> log to file.
• US stock market data.
• global timezone support.
• tiny bug fix.

How to use

Fetch data

from findy.interface import Region
from findy.interface.fetch import fetching

fetching(Region.CHN)  # Region.US

Read data

from datetime import datetime, timedelta

from findy.interface import Region, Provider, EntityType, SP_500_TICKER
from findy.interface.reader import DataReader
from findy.database.schema.meta.stock_meta import Stock, StockDetail
from findy.database.schema.quotes.stock.stock_1d_kdata import Stock1dKdata
from findy.database.schema.register import get_entity_schema_by_type
from findy.database.context import get_db_session
from findy.database.quote import get_entities


end = datetime.now()
start = end - timedelta(days=365)

entity_schema = get_entity_schema_by_type(EntityType.StockDetail)
db_session = get_db_session(Region.US, Provider.Yahoo, entity_schema)
entities, column_names = get_entities(
        region=Region.US,
        provider=Provider.Yahoo,
        entity_schema=entity_schema,
        db_session=db_session,
        codes=SP_500_TICKER,
        columns=[StockDetail.code],
        filters=[StockDetail.market_cap > 10_000_000_000])

codes_cap = [entity.code for entity in entities]

stock_reader = DataReader(region=Region.US,
                          provider=Provider.Yahoo,
                          data_schema=Stock1dKdata,
                          entity_schema=Stock,
                          codes=codes_cap,
                          start_timestamp=start.strftime('%Y-%m-%d'),
                          end_timestamp=end.strftime('%Y-%m-%d'))

stock_reader.load_data()
df = stock_reader.data_df.reset_index(drop=True)
gb = df.groupby('code', sort=False)
stocks = [(code, gb.get_group(code)) for code in gb.groups]

Docker Usage

Docker Execute

docker-compose stop; docker-compose rm -f ; docker-compose build --no-cache
docker-compose up -d

View Log

docker-compose logs -f 

To Do

1. rewrite vendor/baostock package to support async logic
2. add more 3th-party data source packages
3. dev/async_db branch optimization