Fixed bad bug with noise

README.md

@@ -2,11 +2,79 @@
 [![codecov.io](https://codecov.io/github/cdiener/corda/coverage.svg?branch=master)](https://codecov.io/github/cdiener/corda?branch=master)
 [![Code Health](https://landscape.io/github/cdiener/corda/master/landscape.svg?style=flat)](https://landscape.io/github/cdiener/corda/master)
 
-# corda
+# CORDA for Python
 
 This is a Python implementation based on the paper of Schultz et. al.
 
 [Reconstruction of Tissue-Specific Metabolic Networks Using CORDA](http://journals.plos.org/ploscompbiol/article/authors?id=10.1371%2Fjournal.pcbi.1004808)
 
-We will try to employ a method similar to FastFVA to recycle the solvers which
-should speed up the calculation a bit.
+## What does it do?
+
+CORDA, short for Cost Optimization Reaction Dependency Assessment is a method
+for the reconstruction of metabolic networks from a given reference model
+(a database of all known reactions) and a confidence mapping for reactions.
+It allows you to reconstruct metabolic models for tissues, patients or specific
+experimental conditions from a set of transcription or proteome measurements.
+
+## How do I install it
+
+CORDA for Python works only for Python 3.4+ and requires
+[cobrapy](http://github.com/opencobra/cobrapy) to work. We recommend
+installation via the [anaconda or miniconda](https://www.continuum.io/downloads)
+distribution. After installing anaconda or miniconda you can install cobrapy
+from the bioconda repository
+
+```bash
+conda install -c bioconda cobra
+```
+After that you can install CORDA using the pip from conda
+
+```bash
+pip install corda
+```
+
+To install the latest development version use
+
+```bash
+pip install https://github.com/cdiener/corda/archive/devel.zip
+```
+
+After CORDA for Python comes out of its infancy I will prepare a conda package
+as well. For now the master branch is usually working and tested whereas all
+new stuff is kept in the devel branch.
+
+## What do I need to run it?
+
+CORDA requires a base model including all reactions that could possibly included
+such as Recon 1/2 or HMR. You will also need gene expression or proteome data
+for our tissue/patient/experimental setting. This data has to be translated into
+5 distinct classes: unknown (0), not expressed/present (-1), low confidence (1),
+medium confidence (2) and high confidence (3). CORDA will then ensure to include
+as many high confidence reactions as possible while minimizing the inclusion
+of absent (-1) reactions while maintaining a set of metabolic requirements.
+
+## How do I use it?
+
+A small tutorial is found at https://cdiener.com/corda.
+
+## What's the advantage over other reconstruction algorithms
+
+I would say there are two major advantages:
+
+1. It does not require any commercial solvers, in fact it works fastest with
+   the free glpk solver that already comes together with cobrapy. For instance
+   for the small central metabolism model (101 irreversible reactions) included
+   together with CORDA the reconstruction uses the following times:
+
+   - cglpk: 0.02 s
+   - cplex: 0.30 s
+   - gurobi: 0.12 s
+   - mosek: 0.23 s
+
+
+2. It's fast. CORDA for Python uses a strategy similar to FastFVA, where a previous
+   solution basis is recycled repeatedly (speed-up of ~4-10 times). A normal
+   reconstruction for Recon 1 with mCADRE can take several hours. With the original
+   Matlab implementation of CORDA this takes about 40 minutes and with CORDA
+   for Python it takes less than 5 minutes. A Recon 2 reconstruction can be
+   achieved in less than 30 minutes.

corda/__init__.py

@@ -1,8 +1,8 @@
 #  __init__.py
-#  
+#
 #  Copyright 2016 Christian Diener <mail[at]cdiener.com>
-#  
+#
 #  MIT license. See LICENSE for more information.
 
-from .util import reaction_confidence, safe_revert_reversible
+from .util import reaction_confidence, test_model
 from .corda import CORDA