PL4Y

Introduction

This is the very first step of my attempt to break down our monolithic project code into micro-services. Opsworks is what we are using, but the deploy experience is far from comfortable. There is just too much magic behind "Stack Layer". A radical move gives us full control, which is awesome.

Steps

provision an ec2 instance. I use t2.small in my example; but it is recommended by Chef documentation that > Chef server should have at least 4 cores and 4 GB of RAM. It should also have a 64-bit operating system.

cd ~; wget https://web-dl.packagecloud.io/chef/stable/packages/ubuntu/trusty/chef-server-core_12.0.8-1_amd64.deb

sudo dpkg -i chef-server-core_12.0.8-1_amd64.deb

I ran into some error when hostname was not set. so i suggest add 127.0.0.1 HOST_NAME in /etc/hosts where HOST_NAME is from hostname -f

sudo chef-server-ctl reconfigure

Additionally, there's a UI the same as what we see on https://manage.chef.io/ sudo chef-server-ctl install opscode-manage; sudo opscode-manage-ctl reconfigure; sudo chef-server-ctl reconfigure. Done.

Conclusion

After only a few minutes, we have a fully-working chef server. Creating a user/organization can happen on the UI; so is downloading the ssh private key. Next let's see how knife helps us with deploy.

Consider the following scenario and note your intuitive answer to the question.

A  cab was involved in a hit-and-run accident at night.

Two cab companies, the Green and the Blue, operate in the city.

You are given the following data:

85% of the cabs in the city are Green and 15% are Blue.

A witness identified the cab as Blue. The court tested the reliability of the witness under the circumstances that existed on the night of the accident and concluded that the witness correctly identified each one of the two colors 80% of the time and failed 20% of the time.

What is the probability that the cab involved in the accident was Blue rather than Green?

This is a standard problem of Bayesian inference. The correct answer is 41%. Be away from keyboard and enjoy your time doing the math with a pen.

download GSL

install GSL by './configure && make && sudo make install'

gem install gsl

gem install statsample

gem install statsample-bivariate-extension

Afterwards, let's try some toy examples. I am interest in polychoric correlation between two variables/raters.

calculate the polychoric correlation of the following cross table:

------------------- | y=0 | y=1 | y=2 | ------------------- x = 0 | 1 | 10 | 20 | ------------------- x = 1 | 20 | 20 | 50 | -------------------

matrix=Matrix[[1,10,20],[20,20,50]] poly=Statsample::Bivariate::Polychoric.new(matrix, :method=>:joint) puts poly.r

or

poly=Statsample::Bivariate::Polychoric.new_with_vectors([1,10,20].to_scale, [20,20,50].to_scale) puts poly.r

Replace Polychoric with Tetrachoric if your data is dichotomous instead of ordinal.

After upgrading to Mavericks, I ran into the problem of being able to bundle install due to rmagick install failure. The solution of mine is: 1) brew update; 2) brew unlink imagemagick; 3) brew install imagemagick.