julia-a beautiful language

To evaluate any Julia expression, we type ``` (a backtick) and Debugger gives us a Julia prompt. （windows键盘为中/英文+左上角. )(mac 为 option+左上角.)

在atom中，gui界面debugger

1	julia> Juno.@enter maxsum()

Below, square brackets denote optional arguments.····

Misc:

o: open the current line in an editor
q: quit the debugger, returning nothing
C: toggle compiled mode
L: toggle showing lowered code instead of source code
+/-: increase / decrease the number of lines of source code shown

Stepping (basic):

n: step to the next line
u [i::Int]: step until line i or the next line past the current line
s: step into the next call
so: step out of the current call
c: continue execution until a breakpoint is hit
f [i::Int]: go to the i-th function in the call stack (stepping is only possible in the function at the top of the call stack)
up/down [i::Int] go up or down one or i functions in the call stack

Stepping (advanced):

nc: step to the next call
se: step one expression step
si: same as se but step into a call if a call is the next expression
sg: step into a generated function

Querying:

st: show the “status” (current function, source code and current expression to run)
bt: show a backtrace
fr [i::Int]: show all variables in the current or ith frame

Evaluation:

w
- w add expr: add an expression to the watch list
- w: show all watch expressions evaluated in the current function’s context
- w rm [i::Int]: remove all or the i:th watch expression

Breakpoints:

bp add
- bp add “file.jl”:line [cond]: add a breakpoint att file file.jl on line line with condition cond
- bp add func [:line] [cond]: add a breakpoint to function func at line line (defaulting to first line) with condition cond
- bp add func(::Float64, Int)[:line] [cond]: add a breakpoint to methods matching the signature at line line (defaulting to first line) with condition cond
- bp add func(x, y)[:line] [cond]: add a breakpoint to the method matching the types of the local variable x, y etc with condition cond
- bp add line [cond] add a breakpoint to line of the file of the current function with condition cond
bp show all breakpoints
bp rm [i::Int]: remove all or the i:th breakpoint
bp toggle [i::Int]: toggle all or the i:th breakpoint
bp disable [i::Int]: disable all or the i:th breakpoint
bp enable [i::Int]: enable all or the i:th breakpoint
bp on/off
- bp on/off error - turn on or off break on error
- bp on/off throw - turn on or off break on throw

An empty command will execute the previous command.

Hit ` to enter “evaluation mode,” where any expression you type is executed in the debug context. Hit backspace as the first character of the line (or ^C anywhere) to return to “debug mode.”
1|de

Julia 安装

1	docker run -it --rm -v /Users/wjq:/Users/wjq/Documents/GitHub/Test/Test08 -w /Users/wjq/Documents/GitHub/Test/Test08 julia

—rm命令慎用，因为重启后之前保存的全部被删掉了。

1
2
3

(@v1.5) pkg>  add Pluto
julia> import Pluto
julia> Pluto.run()

https://techytok.com/from-zero-to-julia-using-docker/

docker in the cloud

docker pull Julia

linux安装

Add Julia 使用 jupyter环境—-》安装目录 ~/.Julia

Pluto快捷键

Which keys?	Action
✅	`F1` or `Ctrl+Shift+?`	List shortcuts
✅	`Shift+Enter`	Run cell
✅	`Ctrl+Enter`	Run cell if the code changed and add a new cell below
✅	`Ctrl+S`	Run all cells with modified code in a single batch
✅	`Tab`	Autocomplete
✅	`Backspace` or `Delete`	Delete empty cell (delete code manually first)
✅	`Page+Up` and `Page+Down`	Edit cell above/below
✅	`Ctrl+Q`	Interrupt all cells
✅	`Ctrl+P`	Export to PDF
⚪	`Ctrl+Z` but it’s tricky	Undo delete cell
⚪	TBD	Fold/unfold cell
⚪	`Ctrl+Up` and `Ctrl+Down`	Move cell(s) up/down
✅	`Alt+Up` and `Alt+Down`	Move line(s) up/down
⚪	`Ctrl+Shift+S`	Move notebook file
✅	`Ctrl+/` or `Cmd+/`	Comment/uncomment selection
⚪	TBD	Wrap in `begin ... end`
⚪	TBD	Wrap in `let ... end`
⚪	`Ctrl/Cmd+C` and `Ctrl/Cmd+V`	Cut, copy & paste cell(s)
⚪	TBD	Add cell without running
⚪	`Ctrl`+primary click	Jump to definition
⚪	TBD

atom环境- Julia 项目管理

atom安装
插件安装-Julia-client & uber-juno
enter to start Julia
add PkgTemplates
init a github repo

t = Template(;
        user = "jiaqi-knight",
        license="MIT",
        authors=["Jiaqi"],
        plugins=[
                TravisCI(), 
                Codecov(),
                Coveralls(),
        ],
)

t("JuliaExampleTest")

using Revise -更新时候不用重启julia #初学者暂时不加

具体代码示例参考https://github.com/Jiaqi-knight/JuliaExampleTest.jl

# /src/JuliaExampleTest.jl
# 主函数
module JuliaExampleTest

# Write your package code here.
greet() = print("hello word!")

include("subfunction.jl")
export subfunction

end

# /src/subfunction.jl
# 子函数
subfunction(x,y) = 3x+2y

# /test/runtests.jl
# 用来检查code
using JuliaExampleTest
using Test

@testset "JuliaExampleTest.jl" begin
    # Write your tests here.
    @test subfunction(2,1) == 7
    @test subfunction(2,3) == 13

end

github desktop — Git push to the github if necessary
Travis CI —测试你的代码是否适用

建立branch，然后修改代码做一些测试，pull request会更新这些，以及和原代码对比，同时Travis CI会对branch代码做测试。这些操作都不会影响主

如果解决了问题，就可以Merge Pull request, 此时该oull request也被关闭掉了。新的改动将合并到

Pluto的使用

Pluto是一款时尚的IDE编辑软件，类似于jupyter，但更加注重markdown和一些互动功能的扩展。其本质是Markdown和InteractiveUtils环境的jupyter集成。

因此，在写该jl文件时，抬头包含如下代码行：

### A Pluto.jl noterbook###
# v0.12.12
using Markdown
using InteractiveUtils

# ╔═╡ c8413e10-41bf-11eb-3f32-f557907ba7aa #表示分割列，可被IDE识别
begin
	import Pkg
	Pkg.activate(mktempdir()) #设置路径为当前目录
end


#正式代码开始
。。。
。。。


#最后会集中显示 cell order
# ╔═╡ Cell order:
# ╠═c8413e10-41bf-11eb-3f32-f557907ba7aa
# ╠═0dd17fb0-41bd-11eb-22a8-8db1993ad00e
# ╠═ab3a7d10-41c2-11eb-0645-ab0cd0535a86

# 插入markdown语法于Pluto也非常简单
md”“”
#Title
###subTitile
This is Pluto markdown exampele.
“”“

atom 安装weave

可以用来生成网页。

Julia图像处理 by 3blue1brown

1
2
3

pkg>  add Plots
pkg>  add Images

抽象Abstraction

数组array是一个容器，包含元素，元素可以是任何类型的，这是matlab所没有的。就这个功能非常牛逼，拓展性非常强。

julia不仅将python的array语法囊括了，更将java的一些网站语言包含其中，因此对于一些符号、图像、任何元素的矩阵排练，也轻而易举实现，公式符号的优美性不会因为语言而流失。这是相当难得的，是基于web的语言。

Arrays 基础

@view 浅复制- SubArray —》指向同一个地址

改变复制的结果，也会改变父本

= copy 深复制 —-〉开销比较大，要新开内存

示例代码

链接: https://pan.baidu.com/s/1WnW3cfygz7WKOKzDemGzsA 密码: 7acb

Julia性能提升技巧总结

@doc 宏来查看或者编辑文档。

最后一行作为返回值，或者nothing

一切都是对象，但julia的对象都是一个类型的实例。

类型在 Julia 中主要有两个作用：一是用来派发方法（method），二是用来包装数据结构。

julia的类型主要分为抽象类型（Abstract Type）和实体类型（Concret Type），实体类型主要分为可变类型（Mutable Type）和不可变类型（Immutable Type）。前者是成员的值在定义之后可变的类型，后者是成员的值在定义之后不可变的类型。

多重派发

Julia 通过结合 JIT 解决了这个问题，Julia 的编译过程是动态的，每次编译发生在函数被调用之前。这就意味着当编译器能够正确推导出类型的时候，我们就可以进行静态派发，当编译器无法推导类型的时候，我们自动进行动态派发。这表现为 Julia 语言的类型稳定性（type stability）。而所有的编译都以函数为界。这也就是说，如果函数 f1 是类型不稳定的，但是实际的程序中它的输出输入到一个类型稳定的函数 f2 里，Julia 依然能够被静态编译 f2，使得 f2 内部的函数派发变成静态的。

打开julia ：

1	exec '/Volumes/Julia-1.4.2/Julia-1.4.app/Contents/Resources/julia/bin/julia'

更换julia服务器

1	ENV["JULIA_PKG_SERVER"] = "https://mirrors.bfsu.edu.cn/julia"

cmake error

1
2
3

ENV["CMAKE_JL_BUILD_FROM_SOURCE"]=1

Pkg.build("cmake")

(@v1.5) pkg> build ADCME

Building Conda ─→ ~/.julia/packages/Conda/x5ml4/deps/build.log

Building PyCall → ~/.julia/packages/PyCall/tqyST/deps/build.log

Building CMake ─→ ~/.julia/packages/CMake/ULbyn/deps/build.log

Building HDF5 ──→ ~/.julia/packages/HDF5/YX0jU/deps/build.log

Building FFTW ──→ ~/.julia/packages/FFTW/eADNB/deps/build.log

Building ADCME ─→ ~/.julia/packages/ADCME/hCmEo/deps/build.log

┌ Error: Error building ADCME:

│ Collecting package metadata (repodata.json): …working… done

│ Solving environment: …working… done

│

│ Downloading and Extracting Packages

c-ares-1.16.1 | 91 KB | ########## | 100%

keras-preprocessing- | 36 KB | ########## | 100%

make-4.3 | 249 KB | ########## | 100%

python_abi-3.7 | 4 KB | ########## | 100%

scipy-1.5.1 | 19.0 MB | ########## | 100%

libblas-3.8.0 | 11 KB | ########## | 100%

werkzeug-0.16.1 | 258 KB | ########## | 100%

libopenblas-0.3.10 | 8.2 MB | ########## | 100%

llvm-openmp-10.0.1 | 265 KB | ########## | 100%

ld64-530 | 14 KB | ########## | 100%

libcblas-3.8.0 | 11 KB | ########## | 100%

clang-10.0.1 | 12.1 MB | ########## | 100%

ld64_osx-64-530 | 1.3 MB | ########## | 100%

tapi-1000.10.8 | 4.9 MB | ########## | 100%

openblas-0.3.10 | 9.1 MB | ########## | 100%

ninja-1.10.0 | 108 KB | ########## | 100%

hdf5-1.10.6 | 3.0 MB | ########## | 100%

cloudpickle-1.5.0 | 22 KB | ########## | 100%

google-pasta-0.2.0 | 42 KB | ########## | 100%

tensorflow-1.15.0 | 4 KB | ########## | 100%

_tflow_select-2.3.0 | 3 KB | ########## | 100%

certifi-2020.6.20 | 151 KB | ########## | 100%

numpy-1.19.1 | 5.1 MB | ########## | 100%

conda-4.8.4 | 3.0 MB | ########## | 100%

clangxx-10.0.1 | 123 KB | ########## | 100%

zipp-3.1.0 | 13 KB | ########## | 100%

absl-py-0.9.0 | 162 KB | ########## | 100%

importlib-metadata-1 | 44 KB | ########## | 100%

protobuf-3.12.3 | 688 KB | ########## | 100%

libclang-cpp10-10.0. | 11.7 MB | ########## | 100%

liblapack-3.8.0 | 11 KB | ########## | 100%

termcolor-1.1.0 | 6 KB | ########## | 100%

astor-0.8.1 | 25 KB | ########## | 100%

libllvm10-10.0.1 | 20.8 MB | ########## | 100%

h5py-2.10.0 | 925 KB | ########## | 100%

libgfortran-4.0.0 | 716 KB | ########## | 100%

decorator-4.4.2 | 14 KB | ########## | 100%

libgcc-4.8.5 | 785 KB | ########## | 100%

tensorflow-estimator | 271 KB | ########## | 100%

ca-certificates-2020 | 146 KB | ########## | 100%

tensorflow-probabili | 1.2 MB | ########## | 100%

unzip-6.0 | 149 KB | ########## | 100%

tensorflow-base-1.15 | 75.8 MB | ########## | 100%

grpcio-1.30.0 | 1.9 MB | ########## | 100%

markdown-3.2.2 | 61 KB | ########## | 100%

tensorboard-1.15.0 | 3.8 MB | ########## | 100%

wrapt-1.12.1 | 42 KB | ########## | 100%

libprotobuf-3.12.3 | 2.1 MB | ########## | 100%

openssl-1.1.1g | 1.9 MB | ########## | 100%

gast-0.2.2 | 10 KB | ########## | 100%

opt_einsum-3.3.0 | 51 KB | ########## | 100%

keras-applications-1 | 29 KB | ########## | 100%

lapack-3.6.1 | 2.1 MB | ########## | 100%

│ Preparing transaction: …working… done

│ Verifying transaction: …working… done

│ Executing transaction: …working… done

│

│ ==> WARNING: A newer version of conda exists. <==

│ current version: 4.8.3

│ latest version: 4.9.2

│

│ Please update conda by running

│

│ $ conda update -n base -c defaults conda

│

│ #

│ # To activate this environment, use

│ #

│ # $ conda activate base

│ #

│ # To deactivate an active environment, use

│ #

│ # $ conda deactivate

└ @ Pkg.Operations /Users/julia/buildbot/worker/package_macos64/build/usr/share/julia/stdlib/v1.5/Pkg/src/Operations.jl:949

Julia package

创建一个完整的software教程

然后将其分发到github上面

echo "# swirlflow.jl" >> README.md
git init
git add README.md
git commit -m "first commit"
git branch -M main
git remote add origin https://github.com/Jiaqi-knight/swirlflow.jl.git
git push -u origin main

C:\Users\wjq>C:\Users\wjq.julia\adcme\python.exe -m pip install \Mac\Home\Desktop\tensorflow_cpu-2.3.0-cp37-cp37m-win_amd64.whl -i http://pypi.douban.com/simple —trusted-host pypi.douban.com

首先，通过PkgTemplates.jl创建package

using PkgTemplates
t=Template(;
    user="Jiaqiwang969",
    dir="./",
    authors="Jiaqi-knight",
    julia=v"1.5.3",
    plugins=[
        License(; name="MIT"),
        Git(; manifest=true, ssh=true),
        GitHubActions(; x86=true),
        Documenter{GitHubActions}(),

    ],
)

t("Swirlflow.jl")

在github官网建立仓库，仓库名为Swirlflow.jl
git status, git add ., git commit -m “Changes: Version, Test; Message: commenting extra block…”, git push origin master
git tag v0.1.0, git push —tag
atom 进入 folder， activate . , add package
Add function in src folder, 注意名字不要写错
using Swirlflow 如果成功，表示function之间的关联都没有问题了。
通过ci actions 自动生成document，检查代码在不同环境的兼容性问题
add documents in docs
创建issue， @JuliaRegistrator register 等几天
然后就是下载documenterTools，注册账号，自动上传，遗憾电脑的原因没能安装成功，后续再想办法是解压的问题