INSTALL

On Ubuntu

1- install the current Tulip from sources
http://sourceforge.net/projects/auber/files/tulip/tulip-4.7.0/tulip-4.7.0_src.tar.gz/download

a- install synaptic (apt-get install synaptic) so we can install the dev packages (with headers etc.)

- install cmake
- install cmake-gui (cmake-qt-gui)
- install zlib (zlib1g-dev)
- install glut (freeglut3-dev)
- install glew (libglew-dev)
- install ljpeg (ibjpeg-dev)
- install png (libpng-dev)
- install freetype (libfreetype6-dev)
- install qt (libqt4-dev)
- install qmake (qt4-qmake)
- install blas (libopenblas-dev)
- install lapack (liblapack-dev & liblapacke-dev)
- install fortran (gfortran)
- install python (libpython2.7-dev)

b- launch cmake-gui (we are now preparing the compilation of Tulip), a few settings first

- place the source code where you've uncompressed the tar.gz (hereafter "tlproot/")
- place the builds where you want (tlproot/build is advised)
hit generate to have a bunch of variables showing:
- place CMAKE-INSTALL-PREFIX where you want (advised tlproot/install)

now hit the button configure twice, and generate (once or twice I forgot) 
(you may get a few warnings concerning doxygen, sphinx, binutils, quazip, these are optionnal dependencies)


- ## normally no need to set it, but if the build fails, you may need to place CMAKE-BUILD-TYPE as Debug

go to your BUILD folder (tlproot/build/)
>> make && make install

... it should take a while


- install python pip (apt-get install python-pip)
(so we can install the python dependencies)

- install tornado (pip install tornado)
- install scipy (pip install scipy - this one can be a bitch as well, but if you have all the packages right installed, you're safe, it's just long :))
- install numpy (pip install numpy)
- install scikit-learn (pip install scikit-learn)

####################
# here starts detangler
####################

- import from git the branch demo (the branch "ready to use"):
git clone -b demo --single-branch https://github.com/renoust/Detangler.git

- edit the config file in Detangler/server, to put the location of libtulip(-python ...).so
in out example it's: 
      tulip_lib_path="tlproot/install/lib/"
      tulip_python_path="tlproot/install/lib/python/"

execute the server in
Detangler/server/
tulip-server.sh

now the prototype only works well with "Chrome"
to launch the prototype, in chrome:

http://localhost:31497
# this weird port was set to bypass internal administrative policies, but you can change it in the code

Then you can load some datasets...