开源软件名称：

开源软件地址：

开源编程语言：

开源软件介绍：

ExoJulia

A Julia package for fast exoplanet modeling

This code is being developed by Astronomy graduate students at the University of Washington for the course ASTR 598A: Exoplanets. The course and this project are led by Professor Eric Agol.

ASTR 598A Collaboration Instructions

Getting started

You should only need to perform the following once:

1. Fork this repository (click button in upper right of this page) to create a working branch on your github account. If you need help with git/github, check out this 'Getting Started' page by [Phil Marshall:] (https://github.com/drphilmarshall/GettingStarted)

2. Clone your fork onto your computer:

git clone https://github.com/YOUR_USERNAME/ExoJulia.git

3. Add this repo as an upstream remote for your fork:

git remote add upstream https://github.com/jlustigy/ExoJulia.git

3. Do HW assignments in the Homework/ subdirectories (see below)

Now you can code on your machine and commit/push changes to your remote github repo without worrying about messing up or being messed up by other peoples work-in-progress code.

e.g. if you write a module over the course of a few days, it's nice to push to github a few times over those days to ensure you're backing everything up remotely. With git+github the dogs can hang around our homework, but they can never eat it. Good dog.

Each week

1. Sync your fork with the upstream branch to ensure that we are all writing code that uses the same ExoJulia package:

git fetch upstream  
git checkout master  
git merge upstream/master  

This brings your fork's master branch into sync with the upstream repository, without losing your local changes.

2. Add, commit, push to your online fork (if you've made changes on your fork):

git add .
git commit -am "Message"
git push origin master

2. Create a unique homework directory: Homework\hw#\Name1_Name2 where "#" is the homework number, and "Name1" and "Name2" (and so forth) are the names of the group members. So each homework assignment will have a directory, within which each coding group will have a directory. This will ensure everyone's code is saved for future reference and will (hopefully) minimize merge conflicts that can occur when people have worked on the same files.
3. Write code and work in Jupyter Notebooks within the directory you just created. Make sure to reference the "official" (previously selected) code if you are building on previous work. DO NOT LOAD MODULES FROM OTHER HOMEWORK DIRECTORIES.
4. Add, commit, push to your remote branch:

git add .
git commit -am "Message"
git push origin master

5. Submit pull request by navigating to your fork on github.com and pressing the green button for "New Pull Request". This will request that your fork be "pulled" onto the Master branch. Since you've done all your work in a compeletly new directory (didn't you?) there won't be any issues merging :)
6. Each week the fastest code (and then cleanist in the case of ties?) will be selected to become the "offical" code in the ExoJulia package.

Finding the fastest functions for the ExoJulia package

Owen wrote an awesome script that will evaluate the runtime for all of our code.

• In order for the code to run on your functions, put the following line in a comment at the top of a *.jl file that contains the function you want to test:

  #@stest func(params)

  For instance,

  # Suppose this is 'demo.jl'
  
  # The following line tells the stest script to run the function, func() with an argument of 0.5
  #@stest func(0.5)
  
  # You can have multiple stests!
  #@stest func(0.9)
  
  function func(x)
    10.0^x  
  end

• We will need to discuss as a class standard parameters to use to that we are truely comparing apples to apples.

• To test the runtime of a function using this script (always good to check before submitting a pull request), navigate to the Homework/ directory and execute the bash script:

  ./stest hw1

  or hw2, hw3, etc.

Coding Conventions

• See the Julia progamming style guide

• Write functions within *.jl files:

  # Suppose this is 'function.jl'
  function func()
    # Epic code here!
  end

• include functions in other/test *.jl files:

    include("function.jl")
    func()

• For using functions that have been merged into the official ExoJulia package:

  # Add ExoJulia/ to path
  push!(LOAD_PATH, "/PATH/TO/ExoJulia/ExoJulia")
  
  # import ExoJulia
  using ExoJulia
  
  # Run code!
  x = ExoJulia.Orbit.kepler_solve(0.5, 0.5)

• Remember to add the line to test the runtime of your functions! #@stest func(params)

• Working on a new branch: See Astropy example of workflow

  • On the command line in your ExoJulia directory, get the latest version of ExoJulia:

    git fetch upstream

  • Now, create a new branch:

    git checkout -b newbranch

  • Do work on newbranch

  • Add, commit, and then push the newbranch to origin:

    git push origin newbranch

  • Now you can submit pull request on the github webpage from newbranch to master