Zenf

Load rJava in RStudio using macOS 10.12

Sun, 12 Feb 2017 16:55:11 +0100

This post describes a fix of the .onLoad failed in loadNamespace() for 'rJava' error, which occurs when loading rJava in RStudio using macOS. Credit for this fix goes to Toontje.

The ‘2015’ fix stopped working

In beginning of 2015 I wrote a blog post that shows you how to run rJava in RStudio using OSX 10.10. The fix stopped working. Toontje, a reader of my blog, came for the rescue.

The workaround was tested using the following software:

R 3.3.2
RStudio 1.0.136
macOS 10.12.3 (16D32)
jdk 1.8.0_112
rJava 0.9-8

Loading rJava yields “.onload failed … for rJava” error

I installed rJava using install.packages('rJava') without errors. Loading the library in an R session launched in the terminal works. Loading rJava in RStudio yields the following error:

library('rJava')
Error : .onLoad failed in loadNamespace() for 'rJava', details:
  call: dyn.load(file, DLLpath = DLLpath, ...)
  error: unable to load shared object '/Users/paul/Desktop/rJava_test/packrat/lib/x86_64-apple-darwin15.6.0/3.3.2/rJava/libs/rJava.so':
  dlopen(/Users/paul/Desktop/rJava_test/packrat/lib/x86_64-apple-darwin15.6.0/3.3.2/rJava/libs/rJava.so, 6): Library not loaded: @rpath/libjvm.dylib
  Referenced from: /Users/paul/Desktop/rJava_test/packrat/lib/x86_64-apple-darwin15.6.0/3.3.2/rJava/libs/rJava.so
  Reason: image not found
Error: package or namespace load failed for ‘rJava’

Use “dyn.load(<your_jre_path>)” to fix the problem

Launching RStudio using a custom call specifying the jre library doesn’t work anymore. In the new approach we will load it directly to the running R session.

Check if everything works as expected

At first we check if everything else works as expected.

# `which java` should yield something like this: `/usr/bin/java`
$ which java
/usr/bin/java
# `/usr/libexec/java_home` should yield the current jdk home folder
$ /usr/libexec/java_home
/Library/Java/JavaVirtualMachines/jdk1.8.0_31.jdk/Contents/Home

You can also use /usr/libexec/java_home to switch your java version if you wish. You can download the newest jre from the Oracle here.

If the output is similar to the one above you should be fine!

Run “dyn.load(<your_jre_path>)” in RStudio

Open a new RStudio session. Now, load the jre library using the command:

dyn.load(paste0(system2('/usr/libexec/java_home', stdout = TRUE), '/jre/lib/server/libjvm.dylib'))

This command executes /usr/libexec/java_home and concatinates the stdout to the library path /jre/lib/server/libjvm.dylib.

That’s it. Now you should be able to load library(rJava) without errors.

How to Puzzle Biologists in Your Talk

Tue, 03 Jan 2017 15:00:00 +0100

An incomplete how-to on casting confusion and frustration on your audience.

Sympathy and empathy between computer scientists and biologists cracks down the wall between the disciplines. Still, many talks celebrate the differences and are another brick in said wall.

Here I summarize observations on how to efficiently disconnect biologists from your computer science talks:

Speak a technical computer science language
Abstract problems as much as possible
Explain algorithms instead of problems they solve
Explain as many algorithms as possible
Same for file formats: refer to them instead of the information they contain
Go into details of said algorithms, especially on how they perform
Do not spare on memory usage information
Tell people in which language tools are written in (Biologists are very religious with programming languages)
Everybody should use <insert random programming language here>
Jump a lot between analysis steps and make them as abstract as possible
Avoid examples, they benefit only beginners and the stupid
Scare people from programming on their own because of all the things they can do wrong
Jump between problems quickly
Try not to make yourself obsolete

When you craft the talk, always ask yourself: How can I widen the gap between the sciences and how can I disconnect myself from the audiences reality?

Feel free to add suggestions in the comments.

about

Sun, 06 Mar 2016 20:16:04 +0100

Hi! I’m Paul.

I’m a Postdoc at the Max-Planck-Institute for Metabolism Research in Cologne. I develop visual analytics methods, which combine statistics and visual analysis. I want to derive insight into large scale data sets.

I write about day to day tech-problems as well good practices. Hopefully it will benefit you.

Watch LaTeX Documents using Latexmk

Sun, 06 Mar 2016 16:54:47 +0000

The Watch Document functionality of Textmate automatically compiles Tex documents when saved. This post describes how to achieve this behavior with any editor. It works for Unix-based systems, such as OSX or Linux. An example repository can be found here.

Introduction: Switch from Textmate

I’ve been using Textmate to write and compile my LaTeX Documents since 2007. This Blog post by Jannis Hermanns basically defined my workflow since then. It’s core is the neat Watch Document feature of Textmate, which compiles the document and looks for changes. It recompiles the document as soon as it is saved.

The watch document command of Textmate previews changes when saving the document

This also works with LaTeX projects consisting of multiple files. Simply put the watch on the project master file and it will also look for changes in the associated files. This even extends to included images.

If you use a PDF viewer that supports SyncTeX, such as Skim, you can even jump to the current position of the cursor from the TeX document to the PDF and vice versa.

Once I got used to this convenience, compiling LaTeX documents by hand felt like a chore. Unfortunately, using Textmate began to feel likewise. Since I do all my coding in Sublime Text, I looked for a similar workflow for that editor. Turns out, it doesn’t matter which editor you use, you can simply incorporate latexmk to achieve this functionality. Even better, latexmk is shipped per default with the most popular LaTeX distributions, so you should have it installed already.

Compile LaTeX Projects using Latexmk

Consider the following LaTeX document article.tex:

\documentclass[12pt]{article}
\begin{document}
Hi!
\end{document}

Using latexmk, you can compile this into a PDF file by entering the following command:

latexmk -pdf article.tex

In order to watch for changes, you can simply add the -pvc option, which stands for “preview continuously”:

latexmk -pdf -pvc article.tex

As you can see, the command will not terminate back into the prompt, as latexmk watches for updated files (targets) associated with the previewed file. You have to shutdown it manually using CTRL-C. This is already close to the desired behavior. I suggest adding the following options:

-bibtex to compile associated BibTeX files,
-f to continue compiling albeit found errors,
-quiet to make latexmk less verbose and show only important messages,
-use-make to resolve missing files after the pdflatex call using custom dependencies, and
-pdflatex="pdflatex -synctex=1 -interaction=nonstopmode" to add options to the pdflatex command, most important the synctex command to allow jumping between the source PDF and the text editor.

For a precise description of each command, refer to the latexmk readme. All options together yield the following command:

latexmk -quiet -bibtex -pvc -f -pdf -pdflatex="pdflatex -synctex=1 -interaction=nonstopmode" article.tex

This is exactly the behavior I was looking for. And as a big plus, it is independent of your editor choice. Using it with Sublime Text and the LaTeXTools Plugin along with a Skim even allows to jump between source and compiled PDF using the enabled synctex feature.

Jump between the IDE (Sublime Text 3 with LatexTools) and the PDF Viewer (Skim) using the Synctex feature

Use the Power of `make` with `latexmk`

Utilizing this knowledge with makefiles is easy. I know that makefiles often look scary due to their seemingly obscure syntax, but actually they are easy to set up and will spare you a lot of work. If you want to have a quick introduction to make, I strongly recommend Mike Bostocks beautiful blog post on it. The commands applied here, however, are basic and should be self explanatory.

The makefile I propose is an adapted version from this stackoverflow post. I built the makefile to work as follows:

make compiles the document, but does not watch for changes,
make watch compiles the document and watches for changes using the latexmk -pvc command, and
make clean removes all files produced using the LaTeX compilation process, including the resulting PDF.

# makefile
# File adapted from this stackoverflow question: https://tex.stackexchange.com/questions/40738/how-to-properly-make-a-latex-project

# The first rule in a Makefile is the one executed by default ("make"). It
# should always be the "all" rule, so that "make" and "make all" are identical.
all: article.pdf

# MAIN LATEXMK RULE

# -pdf tells latexmk to generate a PDF instead of DVI.
# -pdflatex="" tells latexmk to call a specific backend with specific options.
# -use-make tells latexmk to call make for generating missing files.
# -interaction=nonstopmode keeps the pdflatex backend from stopping at a
# missing file reference and interactively asking you for an alternative.
# -synctex=1 is required to jump between the source PDF and the text editor.
# -pvc (preview continuously) watches the directory for changes.
# -quiet suppresses most status messages (https://tex.stackexchange.com/questions/40783/can-i-make-latexmk-quieter).
article.pdf: article.tex
    latexmk -quiet -bibtex $(PREVIEW_CONTINUOUSLY) -f -pdf -pdflatex="pdflatex -synctex=1 -interaction=nonstopmode" -use-make article.tex

# The .PHONY rule keeps make from processing a file named "watch" or "clean".
.PHONY: watch
# Set the PREVIEW_CONTINUOUSLY variable to -pvc to switch latexmk into the preview continuously mode
watch: PREVIEW_CONTINUOUSLY=-pvc
watch: article.pdf

.PHONY: clean
# -bibtex also removes the .bbl files (http://tex.stackexchange.com/a/83384/79184).
clean:
    latexmk -CA -bibtex

The all task creates the target file “article.pdf” using latexmk. You’ll notice the $(PREVIEW_CONTINUOUSLY) where the -pvc option should be. $(PREVIEW_CONTINUOUSLY) refers to the variable PREVIEW_CONTINUOUSLY, which is empty for all calls except for make watch, where PREVIEW_CONTINUOUSLY is set to -pvc. The make clean command removes all LaTeX output files, including the BibTeX files as well as the PDF.

If you ran make and now want to run make preview, it will get the message “Nothing to be done for preview”, since make sees that the target PDF is up to date. The -B option forces make to create the targets, so make -B preview will work.

Closing Remarks

You can find a minimal example of a LaTeX project using this setup in this Github repository: github.com/paulklemm/latexmk-synctex-example. Feel free to play around with your own makefile rules and explore the possibilities.

Run rJava with RStudio using OSX 10.10

Fri, 20 Feb 2015 14:30:10 +0000

Update 2017-02-12: Newer fix available

The workaround described in this post seized working for newer version of rJava and RStudio. I wrote a new blog post using Toontje’s fix here: Load rJava in RStudio using macOS 10.12.

The post below refers to older version of rJava, RStudio and macOS.

The Problem

I’ve had some trouble getting rJava to run with RStudio. My current solution is a small workaround based on this post on the RStudio support pages.

The first thing I did was installing and running rJava through RStudio and attempt to load it.

install.packages('rJava')
# a bunch of output follows
# ...
# Now we want to load rJava, which yields an error
library(rJava)
Error : .onLoad failed in loadNamespace() for 'rJava', details:
  call: dyn.load(file, DLLpath = DLLpath, ...)
  error: unable to load shared object '/usr/local/Cellar/r/3.1.2_1/R.framework/Versions/3.1/Resources/library/rJava/libs/rJava.so':
  dlopen(/usr/local/Cellar/r/3.1.2_1/R.framework/Versions/3.1/Resources/library/rJava/libs/rJava.so, 6): Library not loaded: @rpath/libjvm.dylib
  Referenced from: /usr/local/Cellar/r/3.1.2_1/R.framework/Versions/3.1/Resources/library/rJava/libs/rJava.so
  Reason: image not found
Error: package or namespace load failed for ‘rJava’

When I’ve ran R through the Terminal and ran up library(rJava), everything worked just fine. Somehow, RStudio is not able to properly load the java path. A small workaround helps here.

The Workaround

We will launch RStudio through the terminal with a custom call, which gives it the proper Java path!

Check if everything works as expected

At first we check a couple of things.

# `which java` should yield something like this: `/usr/bin/java`
$ which java
/usr/bin/java
# `/usr/libexec/java_home` should yield the current jdk home folder
$ /usr/libexec/java_home
/Library/Java/JavaVirtualMachines/jdk1.8.0_31.jdk/Contents/Home

By the way, you can also use /usr/libexec/java_home to switch your java version if you wish. You can also download the newest jre from the Oracle here.

If the output is similar to the one above you should be fine!

Run `RStudio` through the Terminal using a custom command

Now you open RStudio through the terminal using the following command, which sets the LD_LIBRARY_PATH by hard. Make sure you quit all running instances of RStudio before you do that!

$ LD_LIBRARY_PATH=$(/usr/libexec/java_home)/jre/lib/server: open -a RStudio

RStudio launches. Loading library(rJava) should work just fine! If you want to use rJava with RStudio, you have to run it using this command (until they hopefully fix it!)

Make an Alias for the new Command with your Shell

You can register an alias with your shell. For bash you open the bash profile under ~/.bash_profile and add the following alias:

alias rstudio='LD_LIBRARY_PATH=$(/usr/libexec/java_home)/jre/lib/server: open -a RStudio'

Thats it

I hope this gets fixed soon. Until then, hopefully my workaround works for you guys! See you :).

Update 2015-12-09

In my current setting of OSX 10.11.2 (15C50), R version 3.2.2 (2015-08-14) and rJava 0.9-8, this solution does not work. Unfortunately, I have no idea how to fix it. Here is my current situation.

Installing the rJava from CRAN fails

Since El Capitan seems to have problems with passing evironment variables, the current CRAN version of rJava will crash during the installation process. It yields the following error:

install.packages("rJava")
# Output until error skipped
configure: error: One or more JNI types differ from the corresponding native type. You may need to use non-standard compiler flags or a different compiler in order to fix this.
ERROR: configuration failed for package ‘rJava’
* removing ‘/usr/local/lib/R/3.2/site-library/rJava’

This can, however, be easily fixed by installing rJava from RForge.net as stated here. Running install.packages('<PATH TO DOWNLOAD>/rJava_0.9-8.tar.gz', repos = NULL, type="source") installed the current rJava version properly.

Loading rJava in RStudio fails

I’ve set up my homebrew R version properly using R CMD javareconf JAVA_CPPFLAGS="-I/System/Library/Frameworks/JavaVM.framework/Headers -I/Library/Java/JavaVirtualMachines/jdk1.8.0_31.jdk/" as statet in the caveats of the install. Also, /usr/libexec/java_home yields the proper path.

Loading rJava in R through the shell works fine. Loading RStudio with LD_LIBRARY_PATH=$(/usr/libexec/java_home)/jre/lib/server: open -a RStudio and loading rJava does not. It yields the same error as before:

library(rJava)
Error : .onLoad failed in loadNamespace() for 'rJava', details:
  call: dyn.load(file, DLLpath = DLLpath, ...)
  error: unable to load shared object '/usr/local/lib/R/3.2/site-library/rJava/libs/rJava.so':
  dlopen(/usr/local/lib/R/3.2/site-library/rJava/libs/rJava.so, 6): Library not loaded: @rpath/libjvm.dylib
  Referenced from: /usr/local/lib/R/3.2/site-library/rJava/libs/rJava.so
  Reason: image not found
Error: package or namespace load failed for ‘rJava’

The Solution?

If you have any ideas on how to solve this, please let me know.

Use Github for Scientific Writing

Wed, 16 Jul 2014 16:36:35 +0000

GIT is a good tool for managing scientific writing, whether you write on your own or with collaborators. A rich set of tools for supporting collaborative work makes Github a good place for storing LaTeX source there.

There are many reasons why I adore github! For me, it stands for all the good things associated with open content – the beautiful mindset behind it where many people participate to create amazing things! Openness, which is still largely missing in scientific research.

Collaborative work on a scientific paper is carried out in the worst case by emailing LaTeX code. I don’t think I have to talk about the disadvantages of this approach. Tools like Dropbox allow to share folders with different users, but also have major disadvantages. They share everything in a folder, regardless of source files or auxiliary files created during the typesetting process. Many LaTeX editors store configuration files in the project folders and are therefore synced as well. And the constant dropbox pop-up informing me that someone currently works on a text is maddening.

GIT was designed for code-projects, but also works very well for the scientific writing process. If you are new to GIT, check out this great interactive tutorial, teaching you everything you need to know in about 15 minutes at try.github.io.

To clarify: GIT is the name of the versioning system, github is a web site for storing GIT-repositories and it comes with additional features. The major advantages of GIT/github include:

ignoring auxiliary files created by TeX during the typesetting process or configuration files from the LaTeX editor,
track changes using commit messages, creating DIFFs and using branches,
discuss ideas using the build in Wiki and bug tracking/milestone system,
modern Git-Clients, which make it easy to use for people too shy to use the command line (which is by far not as scary as you might assume).

Github repositories are by default open for everyone to see. If you don’t want that, you have to create a private repository. Github offers five free private repositories for students or University associates (https://education.github.com/).

In this blog post I show how to set up a git repository for a scientific paper, recommend a structure and give examples on how you can use features of GIT and github to boost your productivity! Note that I will demonstrate the use of GIT through the console. You can also use the github app (Mac, Windows) or any other GIT client for this.

Setup

Creating the New Repository

Creating a new repository at Github is easy as pie. When you are logged in, you can use the toolbar at the top right corner to create a new repository.

In the next window you can create your new repository. Give it a meaningful title and make sure to select the TeX .gitignore template at the bottom. If you want to create a private repository, you can request it using this link: https://education.github.com/. The people at Github are pretty fast, it took them only a few hours to verify my account.

After you’ve created the repository it should look similar to this:

Add your Files to the Repository

Chances are that you already wrote some stuff or have unpacked the authors packages of the conference or journal. Git makes it easy to add these files to the existing repository. At first, lets clone the newly created repository. You find the http address to your repository on the right bottom of its github page under HTTPS clone URL.

{% codeblock lang:bash %} $ cd ~/Desktop $ git clone https://github.com/Powernap/My-Awesome-Paper.git Cloning into ‘My-Awesome-Paper’… remote: Counting objects: 4, done. remote: Compressing objects: 100% (³⁄₃), done. remote: Total 4 (delta 0), reused 0 (delta 0) Unpacking objects: 100% (⁴⁄₄), done. Checking connectivity… done. {% endcodeblock %}

Now you should have the repository on your desktop. You can add all existing files to the folder via drag and drop. Here are the commands you potentially need to add the new files to the repository:

git status is a useful command for checking the current status of a repository is . It also should ignore the files specified in the .gitignore file.
git add <someFile> adds files to a repository
git add -A adds all files and folders listed under untracked files
git commit makes a commit to a repository. Note that these commits are still only part of your local copy of the repository – unlike other versioning systems, such as SVN, you can create commits while being offline.
git push uploads all changes to the repository stored at github.

Here is an example for doing this:

# You can use any folder you want!
$ cd ~/Desktop/My-Awesome-Paper
$ git status
    On branch master
    Your branch is up-to-date with 'origin/master'.

    Untracked files:
      (use "git add <file>..." to include in what will be committed)

        Example.tex
        SCITEPRESS.eps
        SCITEPRESS.sty
        apalike.bst
        apalike.sty
        article.cls
        example.bib

    nothing added to commit but untracked files present (use "git add" to track)
# Now add all files under `Untracked files`
$ git add -A
# '-m' directly adds the commit message
$ git commit -m "Added conference style files"
    [master f01710d] Added conference style files
     7 files changed, 2318 insertions(+)
    create mode 100755 Example.tex
    create mode 100755 SCITEPRESS.eps
    create mode 100755 SCITEPRESS.sty
    create mode 100755 apalike.bst
    create mode 100755 apalike.sty
    create mode 100755 article.cls
    create mode 100755 example.bib
# Now everything is at the 'head' of your local copy of the repository.
# Time to push it to the github server!
$ git push
    Counting objects: 10, done.
    Delta compression using up to 8 threads.
    Compressing objects: 100% (9/9), done.
    Writing objects: 100% (9/9), 382.59 KiB | 0 bytes/s, done.
    Total 9 (delta 0), reused 0 (delta 0)
    To https://github.com/Powernap/My-Awesome-Paper.git
       0e07357..f01710d  master -> master

If you find it hard to follow, you should look into this quick tutorial.

The `.gitignore` File

It is a good habit to add a .gitignore file to your repository, which contains a list of files ignored by git. LaTeX produces many auxiliary files during its typesetting process, which differ depending on the OS you use and the installed TeX version. It is a good idea to exclude them, especially when collaborating with other people on a paper. Adding files is straightforward, add file names and separate them using newlines. You can even use wildcards using the * symbol, or exclude entire folders (e.g. output/). Whether to add PDF-files to the repository is a matter of taste. I always commit the current state of my work as pdf because it makes it easy to look at versions of previous commits.

Use the Power of `GIT` and `github`

Once set up, the full power of git and github lies beneath your fingertips.

Commit often!

When working with versioning systems, it is recommended to commit early and often! Changes to the text are by far more comprehensible and it avoids conflicts when working with multiple people on a text!

Updating the repository is straightforward, just run git add <your File>, create a git commit and git push it to the server. If you just want to update all files in the repository, you can use git commit -a instead of tediously git add‘ing all changed files.

All commits can be viewed at your projects github page.

Here you can see the changes made to the repository in every commit. Each line can also be commented to trigger a discussion or give feedback! These views are called DIFFs.

Structure your `LaTeX` Documents

It is often a good idea to divide your chapters or sections into individual files and include them in a master file via the \input command. A barebone TeX file could look like this:

\documentclass[12pt]{article}
\begin{document}
\input{Chapter-1.tex}
\input{Chapter-2.tex}
\end{document}

Besides keeping the project structured, it reduces the number of conflicts when you also work with co-authors. I also found it useful to separate each sentence in the following way:

\section{Proof of Everything}
%%
\noindent This is my introduction to this chapter.
%%
I divide every sentence with a comment.
%%
This way it is easy to identify huge sentences that are hard to read and to understand and everybody knows how hard these sentences are to find in the text when you have a lot of text and well you know what I mean.
%%
Sentences can also be easily commented out this way.
%%
Changes made to Tex files structured like this are well readable using the github diff view.

Having all authors of a scientific paper in one place is rare. Exchange of ideas and feedback via email is tedious. The build in Wiki for each repository is a great way for collecting and summarizing ideas. Even in public repositories you can restrict access to the wiki to collaborators only if you want to keep it more private.

The issue and milestone system is packed with useful features for discussing the current process of the work. Discussions can even be referenced to commits. All information is kept at one place, communication is synchronized and you don’t need to look up old mails with chunks of information any more!

Use Branches

Say you want to write a new section, but it will take a couple of days. At the same time, you do not want to change the repository all the time, forcing your co-authors to git pull every time they also want to edit the paper. GIT was designed with such cases in mind! You can simply create your own sub-repository, called a branch, edit it and merge it together with the master-branch.

# Create new 'my-section' branch
$ git branch my-section
# List all branches
$ git branch
  * master
    my-section
# Switch to 'my-section' branch
$ git checkout my-section
  Switched to branch 'my-section'
# Make changes and upload them (-a adds all locally changed files to the repo)
$ git add my-section.tex
$ git commit -a -m "added my new section"
  [my-section fce493f] added my new section
   2 files changed, 9 insertions(+)
   create mode 100644 my-section.tex
# Switch back to the Master branch and merge the sections
$ git checkout master
  Switched to branch 'master'
  Your branch is up-to-date with 'origin/master'.
$ git merge my-section
  Updating f01710d..fce493f
  Fast-forward
   Example.tex    | 1 +
   my-section.tex | 8 ++++++++
   2 files changed, 9 insertions(+)
   create mode 100644 my-section.tex
$ git push

Create Releases

Releases are a great feature to freeze commits. I use it to save important milestones, such as the blind submission for reviewing.

Use All the Features!

Just kidding. But github offers a ton of other features I did not cover in this post. To get an impression about githubs power, head over to its feature page.

They even work on their own editor Atom, which supports LaTeX but isn’t as functional as the packages in Textmate or Sublime Text.

Closing Remarks

There is no right way to write. All these examples are based on my own experience, my own way to work and what I found enough to share it with you. Surely there are much more different ways to use GIT and github for writing. Also, there are great cloud-based LaTeX editors, which also allow collaborative editing (e.g. ShareLaTeX). Please feel free to make remarks in the comment section!

Since I feel that openness plays a crucial role in scientific research, we do good in using these systems and break out of our isolated bubbles and have open discussions about our work and how it develops. Tools like github allow us to do so!

Encore

This is what reviewing revisions can look like if your co-authors use this workflow. I find this pretty awesome 💙.

Watch Twitch using VLC in macOS

Sat, 07 Dec 2013 00:20:42 +0000

If you own a Retina-Macbook you problably struggle with fairly high CPU load and bad performance when watching Twitch-channels using the standard flash-based player in the browser.

A tool called livestreamer can be used to bring Twitch streams to the beloved VLC player (which also uses the GPU to process videos). This way you are not just able to reduce the used resources, the streams also feel much smoother, especially for high resolutions.

Twitch will not be able to stream ads if you use this solution. Please be fair and subscribe to channels you like and you support.

[Update 2017-05-09] Streamlink as replacement for abandoned Livestreamer

The authors of Livestreamer abandoned the project. I recommend you use a fork of Livestreamer called Streamlink.

You install it via pip:

pip install streamlink

[Update 2014-02-20]

If you are using the glorious Alfred App you might be interested in this workflow, which allows for the same thing: http://www.packal.org/workflow/twitchstreamer.

[Update 2014-03-06]

Added updating instructions.

Install Livestreamer

Download and install rtmpdump from here
Download python-setuptools from here
Open Terminal.app

Navigate to the folder where you downloaded the Egg-File and install it

cd ~/Downloads
sh setuptools-0.6c11-py2.7.egg
# Maybe you need to run it as sudo - `sudo sh setuptools-0.6c11-py2.7.egg`
{% endcodeblock %}
- Clone `Livestreamer` GIT Repository and install `Livestreamer`
{% codeblock lang:bash %}
git clone git://github.com/chrippa/livestreamer.git
cd livestreamer
python setup.py install
# Again, maybe you need to run it as sudo - `sudo python setup.py install`

Now you should be able to view Twitch channels in VLC

Using Livestreamer

Say you want to view this Twitch channel in VLC: http://www.twitch.tv/wcs_europe. All you have to to is go into the Terminal.app and type

livestreamer http://www.twitch.tv/wcs_europe [quality]
# You might as well skip the `http://www.` part

Here, [quality] has to be a quality setting from the stream, usually ranging between low, medium, high and source. If you leave it empty, livestreamer will tell you, which options you can choose from. Setting the parameter to best tells livestreamer to use the highest quality available.

Et voilá. Enjoy your stream.

Update Livestreamer

If you see a message like this when launching livestreamer you might update to the latest version:

[cli][info] A new version of Livestreamer (1.7.4) is available!

To do so, navigate to any folder, for example ~/Downloads, and run the following commands:

cd ~/Downloads
git clone git://github.com/chrippa/livestreamer.git
cd livestreamer
python setup.py install

After that you can remove the livestreamer folder from the folder you cloned the GIT repository to (~/Downloads/livestreamer in our example).

References

Links and instructions how to install everything is taken from different Posts in this [Gamespot Thread](http://forum.gamesports.net/dota/showthread.php?45027-How-to-watch-Twitch-TV-in-VLC-player-(MAC-OSX-HOW-TO)
For further information, visit this site from the livestreamer developer, which offers more detailed instructions

Install Python with VTK and Qt support for use with Eclipse in OS X Mountain Lion

Mon, 20 Aug 2012 00:00:00 +0000

In this post I’d like to show you how to install Python with support for VTK and Qt using Apple OS X 10.8 Mountain Lion. Then I set up Eclipse with the new Python installation in order to use all of it’s auto-completion and debugging capabilities. Please leave a comment with your experiences using this.

The instructions for the installation of Python and Qt consists mostly of information from these two posts [1] [2], so please credit these guys for this, not me.

If you run a Windows or Linux machine and you might want to do yourself a favor and use the DRE which is included in DeVIDE.

Install and configure Homebrew

As a very first step you need to install Homebrew, the friendly packet manager for Mac OS X [3]. This makes it really easy to install all the necessary stuff. While OS X already ships with versions of Python, Ruby and other stuff, the versions they provide are often deprecated and installing own packages can be quite frustrating. Homebrew installs all it’s stuff isolated from the system under /usr/local/cellar. This way you can install your own libraries without touching the components installed from Apple.

In order make sure that Terminal searches the bin folder of Homebrew first before the System bin folder (which is /usr/bin), you need to add it to the PATH-variable. This can be done by editing the .bash_profile file in you user root folder /Users/<your username>/.bash_profile (further referenced as ~/). If this file doesn’t exist, create it! Now you add

# ~/.bash_profile
export PATH=/usr/local/bin:$PATH
export PATH=/usr/local/share/python:$PATH

to the file. The second entry is later used to get the right version of PIP. Note: If you have a Terminal session up and running you need to restart it in order to apply the changes to the PATH. If you have trouble with hidden files on the system, this Tool can come in handy for you!

Before you install anything to Homebrew, be sure to run brew doctor and brew update to check if all dependencies are installed (xCode, Command Line Tools and x11). When brew doctor gives you green light you are good to go!

Install Python

The very first step is that you install your own version of Python by running the following command:

brew install python --framework --universal

In order to make it the default Python installation of the system you need to execute

cd /System/Library/Frameworks/Python.framework/Versions
sudo rm Current
ln -s /usr/local/Cellar/python/2.7.2/Frameworks/Python.framework/Versions/Current

This changes the symlink to our new Python installation. You will be promted for you password since this is an action which requires admin rights.

If everything worked, which python should return /usr/local/bin/python.

Next step is the installation of PIP using the command

easy_install pip

Now which pip should return /usr/local/share/python/pip - otherwise the PATH is not properly set as described above.

Now you can install other Packages like Numpy or Scipy - depending on what you need. Note: If you want to install Scipy you need to install the gfortran compiler.

pip install numpy
brew install gfortran
pip install scipy

Note: To date of this post, you have to use the developer version of Scipy to make it work under Mountain Lion. This can be done using

pip install -e git+https://github.com/scipy/scipy#egg=scipy-dev

This will create a folder ~/src and install Scipy from there.

Install Qt

To install Qt and Pyqt-bindings you execute

brew install qt
brew install pyqt

Install VTK

To install VTK you need to run

brew install vtk --tcl --qt --python --pyqt --examples

This took about 60 minutes on my early 2008 Macbook Pro, so take your time for a well-earned rest.

When this is done you should be able to import vtk into your Python environment without any errors.

Setup Python for use with Eclipse

This part is oriented on the introduction on how to include the DRE into Eclipse, which can be found under [4] and in the following screencast. I just made a few enhancements to make it work for this application. Note that the steps of the screencast don’t work exactly on our installation!

To use Python together with Eclipse you have to install the Classic Version from the official homepage. Next, install PyDev into Eclipse. Now you need to tell Eclipse which Python installation it should use. This can be done in Eclipse → Settings → Pydev → Interpreter-Python → New. In the next Dialog you have to add Path to the Homebrew Python.

Confirm the next Prompt which should look like the following screenshot and you are done.

Now you are good to go. Create a new PyDev Project, select the new created interpreter and play around with VTK - it all should be working with auto completion and the other neat Eclipse features. The Code example Charl used in the Screencast should compile:

# vtkDonutTest.py
import vtk
# create a donut polydata source
s = vtk.vtkSuperquadricSource()
s.ToroidalOn()
# connect it to a polydatamapper
m = vtk.vtkPolyDataMapper()
m.SetInput(s.GetOutput())
# and finally create an actor to represent the donut in the scene
a = vtk.vtkActor()
a.SetMapper(m)

# the renderer is the 3D scene
ren = vtk.vtkRenderer()
ren.AddActor(a)
# renderwindow to contain the 3D scene
rw = vtk.vtkRenderWindow()
rw.AddRenderer(ren)

# we want the user to interact
iren = vtk.vtkRenderWindowInteractor()
iren.SetRenderWindow(rw)

# get the whole business going
iren.Initialize()
rw.Render()
iren.Start()

To test the Qt-Support you can run the following code [5]:

# vtkQtTest.py
import vtk
from PyQt4 import QtGui
from vtk.qt4.QVTKRenderWindowInteractor import QVTKRenderWindowInteractor

"""A simple example that uses the QVTKRenderWindowInteractor class."""
# every QT app needs an app
app = QtGui.QApplication(['QVTKRenderWindowInteractor'])

# create the widget
widget = QVTKRenderWindowInteractor()
widget.Initialize()
widget.Start()
# if you dont want the 'q' key to exit comment this.
widget.AddObserver("ExitEvent", lambda o, e, a=app: a.quit())

ren = vtk.vtkRenderer()
widget.GetRenderWindow().AddRenderer(ren)

cone = vtk.vtkConeSource()
cone.SetResolution(8)

coneMapper = vtk.vtkPolyDataMapper()
coneMapper.SetInput(cone.GetOutput())

coneActor = vtk.vtkActor()
coneActor.SetMapper(coneMapper)

ren.AddActor(coneActor)

# show the widget
widget.show()
# start event processing
app.exec_()

Add VTK-Module to Qt-Designer

If you want to use the Qt-Designer for the use with VTK, things get a little bit tricky. The Qt-Designer comes with our installation of Qt under /usr/local/Cellar/qt/4.8.2/Designer.app and is a WYSISWYG-Editor for Qt-UIs. VTK already has installed a plugin for it, it just has to be sym-linked into the Qt-Designer-Path. You can do this using the following commands, but check your current versions before running these.

cd /usr/local/Cellar/qt/4.8.2/plugins/designer
ln -s /usr/local/Cellar/vtk/5.10.0/plugins/designer/libQVTKWidgetPlugin.dylib libQVTKWidgetPlugin.dylib

Now you should have a QVTKWidget available in the Qt-Designer. The tricky part begins now. Unfortunately when you try to include a Qt-UI which contains a QVTKWidget, it fails to compile, because QVTKWidget belongs to the Class QVTKWidget.QVTKWidget, while the correct class is vtk.QVTKWidget. You can fix this using the following workaround.

When you have created your UI using the Qt-Designer you can convert it to Python-Code using the command

pyuic4 myTestApp.ui > myTestApp_ui.py

Import the new file into your Project and change the line

from QVTKWidget import QVTKWidget

from vtk import QVTKWidget

Now it will compile and run. The tradeoff is that you have to manually change this line every time you want to make changes to the UI using the Qt-Designer. To give an example for a converted UI file which runs, look at the following code.

# pythonQtVTKTest.py
import sys
import vtk
from PyQt4 import QtGui
from pythonQtVTKTest_ui import Ui_Form_Main

class StartQT4(QtGui.QMainWindow):
    def __init__(self, parent=None):
        QtGui.QWidget.__init__(self, parent)
        self.ui = Ui_Form_Main()
        self.ui.setupUi(self)
        self.draw_stuff()    
    def draw_stuff(self):
        # create a donut polydata source
        s = vtk.vtkSuperquadricSource()
        s.ToroidalOn()
        # connect it to a polydatamapper
        m = vtk.vtkPolyDataMapper()
        m.SetInput(s.GetOutput())
        # and finally create an actor to represent the donut in the scene
        a = vtk.vtkActor()
        a.SetMapper(m)

        # the renderer is the 3D scene
        ren = vtk.vtkRenderer()
        ren.AddActor(a)
        # renderwindow to contain the 3D scene
        rw = vtk.vtkRenderWindow()
        rw.AddRenderer(ren)
        self.ui.qvtkWidget.GetRenderWindow().AddRenderer(ren)

        # we want the user to interact
        iren = vtk.vtkRenderWindowInteractor()
        iren.SetRenderWindow(rw)

        # get the whole business going
        iren.Initialize()
        rw.Render()
        #iren.Start()

if __name__ == "__main__":
    app = QtGui.QApplication(sys.argv)
    myapp = StartQT4()
    myapp.show()
    sys.exit(app.exec_())

# pythonQtVTKTest_UI.py
# -*- coding: utf-8 -*-

# Form implementation generated from reading ui file 'pythonQtVTKTest.ui'
#
# Created: Tue Aug 21 15:40:57 2012
#      by: PyQt4 UI code generator 4.9.4
#
# WARNING! All changes made in this file will be lost!

from PyQt4 import QtCore, QtGui

try:
    _fromUtf8 = QtCore.QString.fromUtf8
except AttributeError:
    _fromUtf8 = lambda s: s

class Ui_Form_Main(object):
    def setupUi(self, Form_Main):
        Form_Main.setObjectName(_fromUtf8("Form_Main"))
        Form_Main.setEnabled(True)
        Form_Main.resize(623, 462)
        self.qvtkWidget = QVTKWidget(Form_Main)
        self.qvtkWidget.setGeometry(QtCore.QRect(10, 10, 601, 441))
        self.qvtkWidget.setObjectName(_fromUtf8("qvtkWidget"))

        self.retranslateUi(Form_Main)
        QtCore.QMetaObject.connectSlotsByName(Form_Main)

    def retranslateUi(self, Form_Main):
        Form_Main.setWindowTitle(QtGui.QApplication.translate("Form_Main", "VTK Test", None, QtGui.QApplication.UnicodeUTF8))

from vtk import QVTKWidget

If anyone knows a simpler solution to this, please post it in the comments!

Using MacTeX and Homebrew

Just a short footnote for those of you who use MacTeX. MacTeX stores it’s files under /usr/local/texlive next to the Homebrew stuff. They coexist peaceful next to each other, you just have to apply some diplomacy using the following command [6].

sudo chown -R $USER:staff /usr/local

You might also consult the brew doctor, he will tell you what to do.

Thats it

Thank you for reading, I hope these instructions help you in some way. Please feel free to leave a comment. Have a nice day!

References

[1] http://www.thisisthegreenroom.com/2011/installing-python-numpy-scipy-matplotlib-and-ipython-on-lion/
[2] http://jbowles.github.com/blog/2012/04/22/python/
[3] http://mxcl.github.com/homebrew/
[4] http://code.google.com/p/devide/wiki/EclipseDRE
[5] https://svn.enthought.com/svn/enthought/Mayavi/trunk/enthought/tvtk/pyface/ui/qt4/QVTKRenderWindowInteractor.py
[6] https://github.com/ThinkRuby/thinkruby/wiki/.

Zenf

Load rJava in RStudio using macOS 10.12

The ‘2015’ fix stopped working

Loading rJava yields “.onload failed … for rJava” error

Use “dyn.load(<your_jre_path>)” to fix the problem

Check if everything works as expected

Run “dyn.load(<your_jre_path>)” in RStudio

How to Puzzle Biologists in Your Talk

about

Hi! I’m Paul.

Watch LaTeX Documents using Latexmk

Introduction: Switch from Textmate

The watch document command of Textmate previews changes when saving the document

Compile LaTeX Projects using Latexmk

Jump between the IDE (Sublime Text 3 with LatexTools) and the PDF Viewer (Skim) using the Synctex feature

Use the Power of make with latexmk

Closing Remarks

Run rJava with RStudio using OSX 10.10

Update 2017-02-12: Newer fix available

The Problem

The Workaround

Check if everything works as expected

Run RStudio through the Terminal using a custom command

Make an Alias for the new Command with your Shell

Thats it

Update 2015-12-09

Installing the rJava from CRAN fails

Loading rJava in RStudio fails

The Solution?

Use Github for Scientific Writing

Setup

Creating the New Repository

Add your Files to the Repository

The .gitignore File

Use the Power of GIT and github

Commit often!

Structure your LaTeX Documents

Use the Social Features!

Use Branches

Create Releases

Use All the Features!

Closing Remarks

Encore

Watch Twitch using VLC in macOS

Install Livestreamer

Using Livestreamer

Update Livestreamer

References

Install Python with VTK and Qt support for use with Eclipse in OS X Mountain Lion

Install and configure Homebrew

Install Python

Install Qt

Install VTK

Setup Python for use with Eclipse

Add VTK-Module to Qt-Designer

Using MacTeX and Homebrew

Thats it

References

This might also be useful for you

Use the Power of `make` with `latexmk`

Run `RStudio` through the Terminal using a custom command

The `.gitignore` File

Use the Power of `GIT` and `github`

Structure your `LaTeX` Documents