学习资源:

Julia Package

Julia debugger - https://www.youtube.com/watch?v=SU0SmQnnGys

详细介绍:http://webpages.csus.edu/fitzgerald/julia-debugger-tutorial/

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
1
2
3
4
5
6
7
8
9
10
11
12
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

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    # /src/JuliaExampleTest.jl
    # 主函数
    module JuliaExampleTest

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

    include("subfunction.jl")
    export subfunction

    end
    1
    2
    3
    4
    # /src/subfunction.jl
    # 子函数
    subfunction(x,y) = 3x+2y

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    # /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

dxptkzwlaW5Lscj

  • 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

YwlsbuQGBHNtpik

D7HPZ5VKsYz6jTQ

NolSPVKd5YtqMF8

  • 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

aJDvRPc2wqxlgEu

7RzIvD3iXfLneCB

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:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
### 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
1
2
3
4
5
6
# 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

opB5HJWlvTC3eYn

@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

1
2
3
4
5
6
7
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

  1. 首先,通过PkgTemplates.jl创建package

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    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")


  2. 在github官网建立仓库,仓库名为Swirlflow.jl

  3. git status, git add ., git commit -m “Changes: Version, Test; Message: commenting extra block…”, git push origin master
  4. git tag v0.1.0, git push —tag
  5. atom 进入 folder, activate . , add package
  6. Add function in src folder, 注意名字不要写错
  7. using Swirlflow 如果成功,表示function之间的关联都没有问题了。
  8. 通过ci actions 自动生成document,检查代码在不同环境的兼容性问题
  9. add documents in docs
  10. 创建issue, @JuliaRegistrator register 等几天
  11. 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