pyetras

pyetras

def chunkUntil[I](emit: I => Boolean): Process1[I, Vector[I]] = {
  def go(acc: Vector[I]): Process1[I, Vector[I]] =
    Process.receive1Or[I,Vector[I]](Process.emit(acc)) { i =>
      val chunk = acc :+ i
      if (emit(i)) Process.emit(chunk) ++ go(Vector())
      else go(chunk)
    }
  go(Vector())
}

pyetras

Install openssl with brew, add the following to ~/.R/Makevars

pyetras

Add

CC=gcc-5
CXX=g++-5
PKG_CFLAGS=  -I/Users/Pietras/Downloads/nloptr-master/nlopt-2.4.2/include
PKG_LIBS=  -lm /Users/Pietras/Downloads/nloptr-master/nlopt-2.4.2/lib/libnlopt_cxx.a

to ~/.R/Makevars

pyetras

• Virtualenv

Create a new virtualenv with conda. Install digits' requirements:

conda create -n digits --file requirements.txt -c https://conda.binstar.org/slobodan -c https://conda.binstar.org/dougal -c https://conda.binstar.org/pyetras

• Dependencies

First apply brew-boost.patch to brew repository (patches boost to version 1.5.5)

pyetras

object UnliftType {
  def apply[L <: HList, M[_]](implicit unlift: UnliftType[L, M]) = unlift
  type Aux[L <: HList, M[_], R <: HList] = UnliftType[L, M] { type Out = R }

  implicit def unliftType[H, L <: HList, R <: HList, M[_]]
    (implicit ev: Aux[L, M, R]) :
    Aux[M[H] :: L, M, H :: R] =
      new UnliftType[M[H] :: L, M] { type Out = H :: R }

  implicit def unliftNoType[H, L <: HList, R <: HList, M[_]]

pyetras

ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQDlx7plcFbtfnGBLr8oKm02hBANckhOwSp0wXfVznaJ3Azlr4a5gduT2Vpb/F5Ja9WYKnQe5QrOusE2ZPvn/oxAtKRlum1aHjySb8sTLw/Zfd2HPW1FdCCY1DsQE0A2ltuXZYDxc/oAzlHQcZWFcUKy6QyWQ4Z8bi4MVCukJPcxjFebh1hg5a0gNgQ8DzvkE9RE8mf/vfb2ONfPes6Zzz/CVNqO4pVuqTJKv6993gh7e8bklYBxuGeOiJNBHJ4yFr/cAdo2+B1LlLYmXBWoNV5eehLdY4K9yuwI5/YdLMYIby9M7cx5EWLMT7ORFpRqGW8u36fq8xtuVl5AmYgPIROT Pietras@your-ass-is-mine.local

pyetras

#AppStore: Install XCode

#install homebrew
ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"

brew doctor
brew update

curl https://gist.githubusercontent.com/pyetras/4b59e1902cfaf9af3c98/raw/Caskfile > Caskfile
curl https://gist.githubusercontent.com/pyetras/4b59e1902cfaf9af3c98/raw/Brewfile > Brewfile

pyetras

why is this interesing?

[RPlot14.pdf]

This diagram visualizes top 100 contributors to rails/rails collaborating on other open-source projects on github. Each cell represents two projects that shared some contributors; darker cells indicate that more rails/rails people collaborated in both projects.

This data was pulled from GitHub Archive stored in Google BigQuery. It contains a total number of over 227 * 10^6 events from 2011 till the present day. BigQuery enables querying this huge dataset with a SQL-like language and reasonable performance. We would start by finding all GitHub users (or actors) that were involved with the rails/rails repository and counting their activities.

pyetras

select 
  all.actor, 
  datediff(all.last_at, all.first_at) as length, 
  all.cnt as all, all.cnt_meaningful as meaningful, 
  datediff(all.last_at, all.first_at)/all.cnt as freq,
  all.cnt_contributions > 0 as is_contributor,
  all.cnt_contributions as contributions
from 
  (SELECT 
    actor_attributes_login as actor,