Python/MongoDB Data Warehouse and Data Glue

*Metrique can help bring data into an intuitive, indexable data object collection that supports transparent historical version snapshotting, advanced ad-hoc server-side querying, including (mongodb) aggregations and (mongodb) mapreduce, along with python, ipython, pandas, numpy, matplotlib, and so on, is well integrated with the scientific python computing stack.

Author: "Chris Ward" <cward@redhat.com>

Sources: https://github.com/drpoovilleorg/metrique

metrique

metriqued

metriquec

Make sure you have the following OS stuff installed before doing anything else. The examples given below use yum and assume fedora rpm package names:

yum install python python-devel python-setuptools
yum install git gcc gcc-c++ gcc-gfortran

Also, make sure python pip, distribute and setuptools are installed up2date:

easy_install -U pip
pip install -U distribute
pip install -U setuptools

We strongly suggest install metrique* into a virtual environment. If you don't understand what this means or are only interested to install the metrique client to interact with an existing metrique host, skip this. These steps are entirely optional.

To install virtualenv, run:

pip install virtualenv

To create a new virtual environment to install metrique into:

mkdir ~/virtenvs
virtualenv --no-site-packages ~/virtenvs/metriqueenv

# activate the virtual environment
source ~/virtenvs/metriqueenv/bin/activate

# add the following line to .bashrc to quickly enable the virtenv
# echo alias met='source ~/virtenvs/metriqueenv/bin/activate' >> ~/.bashrc

# then prepare the environment; pip install metrique, ...

We also strongly suggest installing and using IPython notebook instead of standard python shell for interactive data exploration.

To install ipython notebook install the following OS packages:

yum install libpng-devel freetype-devel 

Then install ipython with pip:

pip install ipython

If you see any error, not otherwise mentioned here, Google.