PyWorkflow(PyWF) - A Python Binding of C++ Workflow

Overview

PyWorkflow(PyWF) - A Python Binding of C++ Workflow

License Language PyPI - Python Version PyPI

概览

C++ Workflow是一个高性能的异步引擎,本项目着力于实现一个Python版的Workflow,让Python用户也能享受Workflow带来的绝佳体验。

快速上手

在用户深入了解Workflow相关概念之前,先来看几个简单的示例,可以对使用方法有一个初步印象。pywf是本项目Python包的名称,在文档中有时会直接使用wf作为其简称。

发起一个Http请求

import pywf as wf

def http_callback(http_task):
    resp = http_task.get_resp()
    print("Http status:{}\n{}".format(
        resp.get_status_code(), resp.get_body())) # body is bytes

http_task = wf.create_http_task("http://www.sogou.com/", redirect_max=4, retry_max=2, callback=http_callback)
http_task.start()
wf.wait_finish()

依次发起多个Http请求

import pywf as wf

def series_callback(s):
    print("All task in this series is done")

def http_callback(http_task):
    req = http_task.get_req()
    resp = http_task.get_resp()
    print("uri:{} status:{}".format(
        req.get_request_uri(),
        resp.get_status_code()))

def create_http_task(url):
    return wf.create_http_task(url, 4, 2, http_callback)

first_task = create_http_task("http://www.sogou.com")
series = wf.create_series_work(first_task, series_callback)
series.push_back(create_http_task("https://www.zhihu.com/people/kedixa"))
series.push_back(create_http_task("https://fanyi.sogou.com/document"))
series.start()
wf.wait_finish()

同时发起多个Http请求

import pywf as wf

def parallel_callback(p):
    print("All series in this parallel is done")

def http_callback(http_task):
    req = http_task.get_req()
    resp = http_task.get_resp()
    print("uri:{} status:{}".format(
        req.get_request_uri(),
        resp.get_status_code()))

url = [
    "http://www.sogou.com",
    "https://www.zhihu.com/people/kedixa",
    "https://fanyi.sogou.com/document"
]
parallel = wf.create_parallel_work(parallel_callback)
for u in url:
    task = wf.create_http_task(u, 4, 2, http_callback)
    series = wf.create_series_work(task, None) # without callback
    parallel.add_series(series)
parallel.start()

wf.wait_finish()

基本概念

任务

通过create_xxx_task等工厂函数创建的对象称作任务(task),例如create_http_task。一个任务被创建后,必须被启动或取消,通过执行http_task.start(),会自动以http_taskfirst_task创建一个串行并立即启动任务。如果用户指定了回调函数,当任务完成时回调函数会被调用,但在任务启动后且回调函数前,用户不能再操作该任务。当回调函数结束后,该任务被立即释放,用户也不能再操作该任务。

串行

通过create_series_work创建的对象称作串行(series),用户在创建时需要指定一个first_task来作为启动该series启动时应当执行的任务,用户可选地指定一个回调函数,当所有任务执行完成后,回调函数会被调用。

series的回调函数用于通知用户该串行中的任务均已完成,不能再继续添加新的任务,且回调函数结束后,该串行会立即被销毁。

并行

通过create_parallel_work创建的对象称作并行(parallel),用户可以创建一个空的并行,然后通过add_series接口向并行中添加串行,也可以在创建时指定一组串行。并行本身也是一种任务,所以并行也可以放到串行中。parallel.start()就会自动创建一个串行,并将parallel作为first_task立即开始执行。

parallel的回调函数用于通知用户该并行中的串行均已完成,不能再继续添加新的串行,且回调函数结束后,该并行会立即被销毁。

有了上述三个概念,就可以构建出各种复杂的任务结构,并在Workflow的管理下高效执行。

设计理念

Workflow认为,一个典型的后端程序由三个部分组成,并且完全独立开发。即:程序=协议+算法+任务流。

  • 协议
    • 大多数情况下,用户使用的是内置的通用网络协议,例如http,redis或各种rpc。
    • PyWF未支持用户自定义协议。
  • 算法
    • 算法是与协议对称的概念。
    • 如果说协议的调用是rpc,算法的调用就是一次apc(Async Procedure Call)。
    • 任何一次边界清晰的复杂计算,都应该包装成算法。
  • 任务流
    • 任务流就是实际的业务逻辑,就是把开发好的协议与算法放在流程图里使用起来。
    • 典型的任务流是一个闭合的串并联图。复杂的业务逻辑,可能是一个非闭合的DAG。
    • 任务流图可以直接构建,也可以根据每一步的结果动态生成。所有任务都是异步执行的。

Python Workflow将会逐步支持Workflow的六种基础任务:通讯,文件IO,CPU,GPU,定时器,计数器。

注意事项

  • 框架本身不抛出异常,也未处理任何异常,所以用户需要保证回调函数不会抛出异常,context的构造和析构不抛出异常。
  • 所有通过工厂函数创建出的task,必须start、dismiss或添加至一个series中。
  • 所有创建出的series必须start、dismiss或添加至一个parallel中。
  • 所有创建出的parallel必须start、dismiss或添加至一个series中。
  • 由PyWF工厂函数创建的对象的生命周期均由内部管理,在Python层面仅是一个引用,用户不能使用超出生命周期的对象。
  • 用户使用大部分get接口获取的对象可以自由使用,例如Http中的get_body

构建和安装

通过pip安装

本项目仅支持Python3.6以上,正在准备发布一组manylinux2014版本,用户即将可以通过较高版本的pip直接安装。

# We are working on it
pip3 install pywf

编译安装

用户可以下载本项目源码进行编译安装。

# CentOS 7
yum install cmake3 ninja-build python36 python36-devel python36-pip
yum install gcc-c++ # if needed
git clone https://github.com/sogou/pyworkflow --recursive
cd pyworkflow
pip3 install wheel
python3 setup.py bdist_wheel
pip3 install dist/*.whl --user
# CentOS 8
yum install cmake ninja-build python36 python36-devel python3-pip
git clone https://github.com/sogou/pyworkflow --recursive
cd pyworkflow
pip3 install wheel
python3 setup.py bdist_wheel
pip3 install dist/*.whl --user
Owner
Sogou-inc
Sogou-inc
Service for working with open data of the State Duma of the Russian Federation

Сервис для работы с открытыми данными Госдумы РФ Исходные данные из API Госдумы РФ извлекаются с помощью Apache Nifi и приземляются в хранилище Clickh

Aleksandr Sergeenko 2 Feb 14, 2022
One line Brainfuck interpreter in Python

One line Brainfuck interpreter in Python

16 Dec 21, 2022
IST-Website - IST Tutoring Portal for python

IST Tutoring Portal This portal is a web based interface to handle student help

Jean 3 Jan 03, 2022
Unfinished Python library based on ndspy, for Zelda: Phantom Hourglass and Spirit Tracks.

zed An unfinished library and toolset by me, for viewing and editing files from The Legend of Zelda: Phantom Hourglass and The Legend of Zelda: Spirit

4 Oct 13, 2022
Automatically skip sponsor segments in YouTube videos playing on Apple TV.

iSponsorBlockTV Skip sponsor segments in YouTube videos playing on an Apple TV. This project is written in asycronous python and should be pretty quic

David 64 Dec 17, 2022
Automates the fixing of problems reported by yamllint by parsing its output

yamlfixer yamlfixer automates the fixing of problems reported by yamllint by parsing its output. Usage This software automatically fixes some errors a

OPT Nouvelle Caledonie 26 Dec 26, 2022
RCCで開催する『バックエンド勉強会』の資料

RCC バックエンド勉強会 開発環境 Python 3.9 Pipenv 使い方 1. インストール pipenv install 2. アプリケーションを起動 pipenv run start 本コマンドを実行するとlocalhost:8000へアクセスできるようになります。 3. テストを実行

Averak 7 Nov 14, 2021
30DaysOfCode-PhoenixClub - Solution of everyday coding problem given in 30DaysofCode contest held on Hackerrank

30DaysOfCode-PhoenixClub 👨‍💻 Every day problems solution given in 30DaysOfCode

Urveshkumar 8 Jan 30, 2022
Return-Parity-MDP - Towards Return Parity in Markov Decision Processes

Towards Return Parity in Markov Decision Processes Code for the AISTATS 2022 pap

Jianfeng Chi 3 Nov 27, 2022
Beancount: Double-Entry Accounting from Text Files.

beancount: Double-Entry Accounting from Text Files Contents Description Documentation Download & Installation Versions Filing Bugs Copyright and Licen

2.3k Dec 28, 2022
FBChecker Account using python , package requests and web old facebook

fbcek FBChecker Account using python , package requests and web old facebook using python 3.x apt upgrade -y apt update -y pkg install bash -y pkg ins

XnuxersXploitXen 5 Dec 24, 2022
An After Effects render queue for ShotGrid Toolkit.

AEQueue An After Effects render queue for ShotGrid Toolkit. Features Render multiple comps to locations defined by templates in your Toolkit config. C

Brand New School 5 Nov 20, 2022
Simple module with some functions such as generate password (get_random_string)

Simple module with some functions such as generate password (get_random_string), fix unicode strings, size converter, dynamic console, read/write speed checker, etc.

Dmitry 2 Dec 03, 2022
Blender addon that simplifies access to useful operators and adds missing functionality

Quick Menu is a Blender addon that simplifies common tasks Compatible with Blender 3.x.x Install through Edit - Preferences - Addons - Install... -

passivestar 94 Dec 27, 2022
A basic python project which replicates the functionalities on an 8 Ball.

Magic-8-Ball To the people who wish to make decisions using a Magic 8 Ball but can't get one? I gotchu. This is a basic python project which replicate

3 Jun 24, 2021
A Brainfuck interpreter written in Python.

A Brainfuck interpreter written in Python.

Ethan Evans 1 Dec 05, 2021
Collections of python projects

nppy, mostly contains projects written in Python. Some projects are very simple while some are a bit lenghty and difficult(for beginners) Requirements

ghanteyyy 75 Dec 20, 2022
TrackGen - The simplest tropical cyclone track map generator

TrackGen - The simplest tropical cyclone track map generator Usage Each line is a point to be plotted on the map Each field gives information about th

TrackGen 6 Jul 20, 2022
Solutions for the Advent of Code 2021 event.

About 📋 This repository holds all of the solution code for the Advent of Code 2021 event. All solutions are done in Python 3.9.9 and done in non-real

robert yin 0 Mar 21, 2022
LinuxHelper - A collection of utilities for non-technical Linux users accessible via a GUI

Linux Helper A collection of utilities for non-technical Linux users accessible via a GUI This app is still in very early development, expect bugs and

Seth 7 Oct 03, 2022