Contribution Guide

Before You Start:

• You should have some idea about the project and its purpose. If you don't, start here: http://madlib.net
• You should be familiar with the project contents. If you are not, look at below links:
  • Existing modules: http://doc.madlib.net/master/modules.html
  • Module Development Status
  • Proposed modules: Proposed Methods
• You should have some knowledge on relational databases, SQL, Python, C/C++, and analytical methods that you want to implement.

Step 1: Make friends with GitHub

• Create an account on GitHub.com.
• Get a git client from git-scm.com.
• If you are not familiar with this version control software follow bootcamp guides on help.github.com to gain some confidence.

Step 2: Find your task (new module, feature request or bug fix)

• Create an account on MADlib JIRA and review the open issues.
• Check again the Module Development Status page.
• Get in touch with us on MADlib Dev Forum
  • Tell us what you'd like to work on OR,
  • We can discuss the most pressing needs and suggest a task.
• Once we have a consensus we'll open a JIRA ticket (bug, task, or new feature) to track the progress and update the Dev Status on Wiki.

Step 3: Fork MADlib project

• Go to GitHub MADlib repo and fork the project using the FORK button (top right).
  • This step will create a "connected" copy of madlib/madlib repo under yourGitHubAccount/madlib.
  • Need some help with forking? Check here

Step 4: Develop away...

by using your own copy/fork of the MADlib repository.

Make sure you have all the module components in place: MADlib Module Anatomy

And remember about:

• Proper SQL API: SQL API Guide
• C++ coding practices: The C++ Abstraction Layer
• Comprehensive documentation: Documentation Guide (Doxygen)
• Complete unit tests: Unit Test Guide

Step 5: Send pull-request

Done with coding? Follow this to get your code checked-in:

• Test build/install/execution on your side.
• Create a pull-request from your forked repository.
  • Use different topic branches to separate your commits into few pull-requests.
  • If this sounds like Greek: read more about pull-requests here: http://help.github.com/send-pull-requests/.
• Create a JIRA Task: http://jira.madlib.net/
  • Use subject "pull-request" and explain the details.
  • Put the github pull-request link into your JIRA.
  • If you know who would be the best reviewer assigne the JIRA to that person. Otherwise don't worry.
• Code comments are easy to add on Github. JIRA better serves as a high level log tracker.
• Once the code is merged:
  • Github pull request closes automagically
  • JIRA must be closed manually