Everyone who has tried to mess around with their Ubuntu distro knows the pain of having to reinstall Ubuntu and set it up to their liking again

These are the scripts that I use to set my Ubuntu up as quick as possible. Feel free to fork it and create your own version, and any contributions are more than welcome :)

First download/clone this repository

Run chmod u+x *.sh to make all scripts executable Then execute them in the terminal in the sequence of filenames.

• 1-BasicSetUp.sh - Sets up terminal configuration, a download accelerator, anaconda python, and shell aliases.
• 2-GenSoftware.sh - Sets up tools for programming(editor, etc), and other general purpose software I use
• 3-ML-Gpu.sh - Compiles commonly used ML/DL/RL libraries from source, so that it is optimized to run on your computer
• opencvDirectInstall.sh - Compiles the latest tag of OpenCV+Contrib from source on your machine with focus on optimization of execution of OpenCV code.
  
  

• Default python will be changed to Anaconda, with the latest Python 3, and conda environments called py27 and py36 running Python2.7 and Python3.6 respectively will be your alternate Python2 and Python 3.6 environments. Anaconda Python will be installed in /opt/anaconda3/ so that it is accessible by multiple users
• Default shell is changed to Zim, a zsh plugin, instead of bash. Why zsh? Because it simply has a much better autocomplete. And why zim? Because it's much faster than Oh My Zsh and Prezto

• maxvol : Will set your volume to 150%
• download <webpage-name>: Download the webpage and all sub-directories linked to it
• server : Sets up a server for file sharing in your local network. Whatever is in your current directory will be visible on the ip. It will also print the possible set of IP addresses. To access from another computer, shoot up a browser and simply hit ip_add:port
• gpom : Alias for git push origin master. Will push your current directory
• jn : Starts a jupyter notebook in that directory
• jl : Starts a jupyter lab in that directory
• ydl "URL": Downloads the song at URL at 128kbps, 44.1kHz in m4a format with the title and song name automatically set in the metadata
• update: Runs sudo apt-get update && sudo apt-get dist-upgrade && sudo apt-get autoremove -y
• tsux: Create a tmux session with -u (so that the icons(battery, etc) are properly displayed at the bottom), and with a window with htop, nvidia-smi -l 1 and lm-sensors automatically activated.
  • Reason for not making this the default tmux: You cannot attach tmux sessions if you alias the tmux command itself

Tmux, Tilda, Ubuntu-Restricted-Extras, Lyx, VLC, Chromium and Firefox, Dropbox, Gparted, Boot-Repair, Shutter,Grub Customizer, Ffmpeg, Qt5, gimp, meld(To be used with git mergetool), axel, tor & i2p, redshift, lm-sensors (might've missed some)

• Machine Learning Libraries: Tensorflow built from source, optimized for user hardware. Theano, Keras, OpenAI Gym and Pytorch installed from pip
• OpenCV: Compiled from source, multithreaded and optimized to use your hardware
• Autopep8
• scdl - a soundcloud downloader
• org-e - An app to sort and declutter folders (like ~/Downloads/)
• youtube-dl: A youtube downloader

• If you have Anaconda Python, OpenCV will be linked to Anaconda Python by default, not the Linux default python. If you would like to compile for the Linux default Python, remove Anaconda from your path before running the opencvDirectInstall.sh script

• If you would like to install with OpenCV for CUDA, change the flag in the file opencvDirectInstall.sh for WITH_CUDA to ON

• After OpenCV installation, if you get an error of the sort illegal hardware instructions when you try to run a python or c++ program, that is because your CPU is an older one (Pentium/Celeron/...). You can overcome this by adding the following to the end of the cmake (just before the ..)

   -D ENABLE_SSE=OFF \
   -D ENABLE_SSE2=OFF \
   -D ENABLE_SSE3=OFF ..

  If you still want to be able to receive the benefits of CPU optimization to whatever extent you can, then hit cat /proc/cpuinfo and see what sses are available under flags

• These scripts are written and tested on the following configurations -

  • Ubuntu 14.04 & 16.04
  • 32-bit and 64-bit Intel Processors
  • ML-GPU.sh - NVIDIA GPUs including but not limited to GeForce GTX 1080, 1070, 940MX, 850M, and Titan X

  Although it should work on other configurations out of the box, I have not tested them

• A docker image for this set-up (last updated Jan 30th, 2017)
• A Ubuntu customization dedicated to robotics

• Fail Proof checks
• configuring default wifi settings