Setup

Creating a virtual-env

virtualenv cloudcv_env
source cloudcv_env/bin/activate

Cloning

git clone https://github.com/batra-mlp-lab/pcloudcv.git pcloudcv
cd pcloudcv/pcloudcv

Pre-requisites

1. Please ensure that redis-server is installed and running in background. Follow the instructions on this page - till the end of the page. Redis Quickstart

   On OSX: http://www.richardsumilang.com/blog/2014/04/04/how-to-install-redis-on-os-x/ ( I hope this is easier. Didn't try it myself)

2. Port 8000 should be free

3. The h5py in requirements.txt needs to have HDF5 installed before. On Ubuntu 12.04 it's called "libhdf5-serial-dev", otherwise it cannot find the path to "hdf5.h". Using Synaptic seems best to have all the paths set properly. On Ubuntu: sudo apt-get install libhdf5-dev On OSX : brew tap homebrew/science , brew install hdf5

4. In addition, a C compiler and the necesarry Python development environment are needed:

   Ubuntu/Debian: sudo apt-get install python-dev build-essential (untested, please report back if more packages are needed here) Mac OS X: Install XCode (usually bundled with the Mac OS X installation disks)

   On ubuntu 12:

   sudo apt-get install libtiff4-dev libjpeg8-dev zlib1g-dev libfreetype6-dev liblcms2-dev libwebp-dev    tcl8.5-dev tk8.5-dev python-tk
   
   

   On Ubuntu 14:

   sudo apt-get install libtiff5-dev libjpeg8-dev zlib1g-dev libfreetype6-dev liblcms2-dev libwebp-dev tcl8.6-dev tk8.6-dev python-tk
   

   On Mac:

   brew install libtiff libjpeg webp little-cms2
   
   

Installing Dependencies

for req in $(cat requirements.txt); do pip install $req; done

The dependencies are not installed in order when using requirements.txt and therefore this is a workaround. This will ensure proper installation of all the packages.

Config File
It is important to ensure that config file is well setup. You can look at the sample config.json and modify it to suit your needs

1. path:
   This is the input path. It should be the full path to where the files are located.
   It can either take this form if you want to upload images from local directory:

   "path": "local: /home/dexter/Pictures/test_upload/kitchen",
   

   or this form if you have files in your dropbox. Please note that dropbox is relative to /Apps/CloudCV folder.

   "path": "dropbox: /1",
   

2. exec: This executable should contain the name of the executable you want to run

3. output: This should contain the full path to the output directory. Please note that there shouldn't be a local: prefix before the output directory path. Also in-case the input is an input directory, the output path is not considered and the results are uploaded in their corresponding jobid folder.

Getting Started There are four Ipython Notebooks that are working correctly - Classification Visualization, Image Stitching, Feature Extraction, DB_Integration. Make sure to copy config.json.example to config.json. Then change the path to images are in your local or dropbox directory.

Try these Ipython Notebooks to get an understanding of the functionalities that we are trying to provide and report bugs/ submit patches.

I am working on writing the documentation for this API and will update the repository very soon.

Running Test Code To run the test code, do the following from the terminal

python run.py config.json

you can additionally give following optional arguments

1. -I "local: /full/path/to/input" or -I "dropbox:/relative/path"
2. -O "/full/path/to/output/folder"
3. -E "executable name"
4. --nologin This will not try registering/authenticating the user with CloudCV. This will however disable dropbox functinality.