Version 1.0

pawel-czyz · Aug 16, 2018 · 1d87b5b · 1d87b5b
1 parent 17887a7
commit 1d87b5b
Show file tree

Hide file tree

Showing 10 changed files with 64 additions and 51 deletions.
diff --git a/Makefile b/Makefile
@@ -1,20 +1,21 @@
-# Run unit tests
 test:
-	python3 -m unittest discover -s tests > /dev/null
+	@echo "Running unit tests..."
+	@python3 -m unittest discover -s tests > /dev/null
 
-# Check test coverage using unittest module
 coverage:
-	coverage run --source=cce -m unittest discover -s tests > /dev/null; coverage report
+	@echo "Checking test coverage using unittest module..."
+	@coverage run --source=cce -m unittest discover -s tests > /dev/null; coverage report
 
-# Check test coverage and show the results in browser
 html:
-	coverage run --source=cce -m unittest discover -s tests > /dev/null; coverage html; python -m webbrowser "./htmlcov/index.html" &
+	@echo "Checking test coverage and showing the results in browser..."
+	@coverage run --source=cce -m unittest discover -s tests > /dev/null; coverage html; python -m webbrowser "./htmlcov/index.html" &
 
-# Check compatibility with Python 2.7
 comp:
-	python2 -m unittest discover -s tests > /dev/null
+	@echo "Checking compatibility with Python 2.7..."
+	@python2 -m unittest discover -s tests > /dev/null
 
 install:
-	pip install .
+	@echo "Installing cce module via pip..."
+	@pip install .
 
-.PHONY: test
+.PHONY: test coverage html comp install
diff --git a/Readme.rst b/Readme.rst
@@ -4,8 +4,8 @@ Channel Capacity Estimator
 
 Channel Capacity Estimator (**cce**) is a python module to estimate 
 `information capacity`_ of a communication channel. Mutual information,
-computed as proposed by `Kraskov et al.` (*Physical Review E*, 2004)
-Eq. (8), is maximized over input probabilities by means of a constrained 
+computed as proposed by `Kraskov et al.`_ (*Physical Review E*, 2004,
+Eq. (8)), is maximized over input probabilities by means of a constrained 
 gradient-based stochastic optimization. The only parameter of the Kraskov 
 algorithm is the number of neighbors, *k*, used in the nearest neighbor 
 search. In **cce**, channel input is expected to be of categorical type 
@@ -19,8 +19,9 @@ requirements.txt for a complete list of dependencies.
 
 Module **cce** features the research article "Limits to the rate of 
 information transmission through MAPK pathway" by Grabowski *et al.*, 
-submitted to *PLOS Computational Biology* in 2018. Release 0.4 of the 
-code has been included as supplementary data of this article. 
+submitted to *PLOS Computational Biology* (2018). Version 1.0 of **cce**
+(with pre-built documentation) has been included as supplementary code 
+of this article.
 
 For any updates and fixes to **cce**, please visit project homepage:
 http://pmbm.ippt.pan.pl/software/cce 
@@ -38,7 +39,7 @@ There are three major use cases of **cce**:
 
 In the example below, mutual information is calculated between three sets 
 of points drawn at random from two-dimensional Gaussian distributions,
-located at (0,0), (1,1), and at (3,3) (in SciPy, covariance matrices of 
+located at (0,0), (0,1), and at (3,3) (in SciPy, covariance matrices of 
 all three distributions  by default are identity matrices). Auxiliary 
 function `label_all_with` helps to prepare the list of all points, in 
 which each point is labeled according to its distribution of origin.
@@ -51,15 +52,16 @@ which each point is labeled according to its distribution of origin.
     >>> def label_all_with(label, values): return [(label, v) for v in values]
     >>>
     >>> data = label_all_with('A', mvn(mean=(0,0)).rvs(10000)) \
-             + label_all_with('B', mvn(mean=(1,1)).rvs(10000)) \
+             + label_all_with('B', mvn(mean=(0,1)).rvs(10000)) \
              + label_all_with('C', mvn(mean=(3,3)).rvs(10000))
     >>>
-    >>> wke(data).calculate_mi(k=50)
-    0.9386627422798913
+    >>> wke(data).calculate_mi(k=10)
+    0.9552107248613955
 
 In this example, probabilities of input distributions, henceforth referred
 to as *weights*, are assumed to be equal for all input distributions. Format
-of data is akin to [('A', array([-0.4, 2.8])), ('A', array([-0.9, -0.1])), ..., ('B', array([1.7, 0.9])), ..., ('C', array([3.2, 3.3])), ...).
+of data is akin to [('A', array([-0.4, 2.8])), ('A', array([-0.9, -0.1])), 
+..., ('B', array([1.7, 0.9])), ..., ('C', array([3.2, 3.3])), ...).
 Entries of data are not required to be grouped according to the label.
 Distribution labels can be given as strings, not just single characters. 
 Instead of NumPy arrays, ordinary lists with coordinates will be also 
@@ -81,12 +83,12 @@ input distributions:
     >>> def label_all_with(label, values): return [(label, v) for v in values]
     >>>
     >>> data = label_all_with('A', mvn(mean=(0,0)).rvs(10000)) \
-             + label_all_with('B', mvn(mean=(1,1)).rvs(10000)) \
+             + label_all_with('B', mvn(mean=(0,1)).rvs(10000)) \
              + label_all_with('C', mvn(mean=(3,3)).rvs(10000))
     >>>
-    >>> weights = {'A': 3/6, 'B': 1/6, 'C': 2/6}
-    >>> wke(data).calculate_weighted_mi(weights=weights, k=50)
-    0.9420502318804324  
+    >>> weights = {'A': 2/6, 'B': 1/6, 'C': 3/6}
+    >>> wke(data).calculate_weighted_mi(weights=weights, k=10)
+    1.0065891280377155
 
 (This example involves random numbers, so your result may vary slightly.)
 
@@ -102,26 +104,38 @@ input distributions:
     >>> def label_all_with(label, values): return [(label, v) for v in values]
     >>>
     >>> data = label_all_with('A', mvn(mean=(0,0)).rvs(10000)) \
-             + label_all_with('B', mvn(mean=(1,1)).rvs(10000)) \
+             + label_all_with('B', mvn(mean=(0,1)).rvs(10000)) \
              + label_all_with('C', mvn(mean=(3,3)).rvs(10000))
     >>>
-    >>> wke(data).calculate_maximized_mi(k=50)
-    (0.98616722147976, {'A': 0.38123083, 'B': 0.16443817, 'C': 0.45433092})
+    >>> wke(data).calculate_maximized_mi(k=10)
+    (1.0154510500713743, {'A': 0.33343804, 'B': 0.19158363, 'C': 0.4749783})
 
 The output tuple contains the maximized mutual information (channel capacity) 
 and probabilities of input distributions that maximize mutual information (argmax). 
 Optimization is performed within TensorFlow with multiple threads and takes 
 less than a minute on a quad-core processor.
 (This example involves random numbers, so your result may vary slightly.)
 
+
 Testing
 -------
-To launch a suite of unit tests run:
+To launch a suite of unit tests, run:
 
 .. code:: bash
 
     $ make test
 
+
+Documentation
+-------------
+Developer's code documentation may be generated with
+
+.. code:: bash
+
+   $ cd docs
+   $ make html
+
+
 Installation
 ------------
 To install **cce** locally via pip, run:
@@ -139,16 +153,14 @@ Then, you can directly start using the package:
     >>> ...
 
 
-
-
 Authors
 -------
 
 The code was developed by `Frederic Grabowski`_ and `Paweł Czyż`_,
 with some guidance from `Marek Kochańczyk`_ and under supervision of 
 `Tomasz Lipniacki`_ from the `Laboratory of Modeling in Biology and Medicine`_,
 `Institute of Fundamental Technological Reasearch, Polish Academy of Sciences`_
-in Warsaw.
+(IPPT PAN) in Warsaw.
 
 
 License

diff --git a/cce/estimator.py b/cce/estimator.py
@@ -8,9 +8,9 @@
 from scipy.spatial import cKDTree
 from scipy.special import digamma
 import numpy as np
-from cce.preprocess import normalize, add_noise_if_duplicates
-from cce.optimize import weight_optimizer
-from cce.score import weight_loss
+from cce.preprocessing import normalize, add_noise_if_duplicates
+from cce.optimization import weight_optimizer
+from cce.scoring import weight_loss
 
 
 class WeightedKraskovEstimator:
@@ -39,7 +39,7 @@ def __init__(self, data: list = None, leaf_size: int = 16):
         self._number_of_points_for_label = defaultdict(lambda: 0)
         self._number_of_labels = None
 
-        # Immersed data -- X is mapped from cateogorical data into reals
+        # Immersed data -- X is mapped from categorical data into reals
         # using _huge_dist, and we store spaces X x Y and just Y.
         self._immersed_data_full = None
         self._immersed_data_coordinates = None
@@ -226,7 +226,7 @@ def calculate_weighted_mi(self, weights: dict, k: int) -> float:
 
 
     def optimize_weights(self) -> tuple:
-        """Function optimizing weights using weight_optimizer.
+        """Optimizes probabilities of input distributions (weights).
 
         Returns
         -------
@@ -281,7 +281,7 @@ def _turn_into_neigh_list(self, indices, special_point_label):
 
 
     def calculate_neighborhoods(self, k: int):
-        """Function that prepares neighborhood_array.
+        """Prepares neighborhood_array.
 
         Parameters
         ----------

diff --git a/cce/optimize.py → cce/optimization.py b/cce/optimize.py → cce/optimization.py
diff --git a/cce/preprocess.py → cce/preprocessing.py b/cce/preprocess.py → cce/preprocessing.py
@@ -16,7 +16,7 @@ def _project_coords(data: list) -> list:
 
 
 def normalize(data: list) -> list:
-    """Perform input data normalization
+    """Performs input data normalization.
 
     Parameters
     ----------
@@ -44,7 +44,7 @@ def normalize(data: list) -> list:
 
 
 def unique(arr) -> bool:
-    """Check if all points in the array of coordinates are unique.
+    """Checks if all points in the array of coordinates are unique.
 
     Parameters
     ----------
@@ -60,7 +60,7 @@ def unique(arr) -> bool:
 
 
 def add_noise_if_duplicates(data: list) -> list:
-    """Add noise to input data
+    """Adds noise to input data.
 
     Parameters
     ----------

diff --git a/cce/score.py → cce/scoring.py b/cce/score.py → cce/scoring.py
diff --git a/docs/Makefile b/docs/Makefile
@@ -3,7 +3,7 @@
 
 # You can set these variables from the command line.
 SPHINXOPTS    =
-SPHINXBUILD   = python -msphinx
+SPHINXBUILD   = python3 -msphinx
 SPHINXPROJ    = ChannelCapacityEstimator
 SOURCEDIR     = .
 BUILDDIR      = _build

diff --git a/docs/conf.py b/docs/conf.py
@@ -66,7 +66,7 @@
 # The short X.Y version.
 version = '1.0'
 # The full version, including alpha/beta/rc tags.
-release = '1.0'
+release = '1.0.0'
 
 # The language for content autogenerated by Sphinx. Refer to documentation
 # for a list of supported languages.

diff --git a/docs/source/cce.rst b/docs/source/cce.rst
@@ -4,18 +4,18 @@ cce package
 Submodules
 ----------
 
-cce\.score module
------------------
+cce\.scoring module
+-------------------
 
-.. automodule:: cce.score
+.. automodule:: cce.scoring
     :members:
     :undoc-members:
     :show-inheritance:
 
-cce\.preprocess module
-----------------------
+cce\.preprocessing module
+-------------------------
 
-.. automodule:: cce.preprocess
+.. automodule:: cce.preprocessing
     :members:
     :undoc-members:
     :show-inheritance:
@@ -28,10 +28,10 @@ cce\.estimator module
     :undoc-members:
     :show-inheritance:
 
-cce\.optimize module
---------------------
+cce\.optimization module
+------------------------
 
-.. automodule:: cce.optimize
+.. automodule:: cce.optimization
     :members:
     :undoc-members:
     :show-inheritance:

diff --git a/tests/test_preprocess.py b/tests/test_preprocess.py
@@ -1,5 +1,5 @@
 import unittest
-from cce.preprocess import normalize
+from cce.preprocessing import normalize
 
 LARGE_VALUES_SMALL_SPREAD = [('1', [1e9, 1e9]),
                              ('1', [1e9+1, 1e9+1]),