From 75a9146c37bf3a32c3c421efd252aa5099bfba86 Mon Sep 17 00:00:00 2001
From: L3viathan <git@l3vi.de>
Date: Fri, 6 Feb 2015 14:40:33 +0100
Subject: [PATCH] implement PMI

---
 src/composes/matrix/dense_matrix.py           | 15 +++++++++++++
 src/composes/matrix/sparse_matrix.py          | 14 ++++++++++++
 .../transformation/scaling/pmi_weighting.py   | 22 +++++++++++++++++++
 3 files changed, 51 insertions(+)
 create mode 100644 src/composes/transformation/scaling/pmi_weighting.py

diff --git a/src/composes/matrix/dense_matrix.py b/src/composes/matrix/dense_matrix.py
index ae55185..d642b6a 100644
--- a/src/composes/matrix/dense_matrix.py
+++ b/src/composes/matrix/dense_matrix.py
@@ -245,6 +245,21 @@ def plog(self):
         self.mat[self.mat < 1.0] = 1
         self.mat = np.log(self.mat)
 
+    def log(self):
+        """
+        Applies log to the matrix elements.
+
+        Elements smaller than or equal to 0 (leading to not-defined log)
+        are set to 0. Log is applied on all other elements.
+
+        Modifies the current matrix.
+        """
+
+        #this line uses 3 x size(mat) to run in the worst case
+        #(if we select the entire matrix - depends on the size of the selection)
+        self.mat[self.mat <= 0] = 1
+        self.mat = np.log(self.mat)
+
 
     def assert_positive(self):
         """
diff --git a/src/composes/matrix/sparse_matrix.py b/src/composes/matrix/sparse_matrix.py
index 563188b..461dc22 100644
--- a/src/composes/matrix/sparse_matrix.py
+++ b/src/composes/matrix/sparse_matrix.py
@@ -291,6 +291,20 @@ def plog(self):
         self.mat.data = np.log(self.mat.data)
         self.mat.eliminate_zeros()
 
+    def log(self):
+        """
+        Applies log to the matrix elements.
+
+        Elements smaller than or equal to 0 (leading to not-defined log)
+        are set to 0. Log is applied on all other elements.
+
+        Modifies the current matrix.
+        """
+
+        self.mat.data[self.mat.data <= 0] = 1
+        self.mat.data = np.log(self.mat.data)
+        self.mat.eliminate_zeros()
+
     def get_non_negative(self):
         """
         Turns negative entries to 0.
diff --git a/src/composes/transformation/scaling/pmi_weighting.py b/src/composes/transformation/scaling/pmi_weighting.py
new file mode 100644
index 0000000..f1b1564
--- /dev/null
+++ b/src/composes/transformation/scaling/pmi_weighting.py
@@ -0,0 +1,22 @@
+
+from scaling import Scaling
+from epmi_weighting import EpmiWeighting
+
+class PmiWeighting(Scaling):
+    """
+    Point-wise Mutual Information.
+
+    :math:`pmi(r,c) = log\\frac{P(r,c)}{P(r)P(c)}`
+    """
+
+    _name = "pmi"
+    _uses_column_stats = True
+
+    def apply(self, matrix_, column_marginal=None):
+
+        matrix_ = EpmiWeighting().apply(matrix_, column_marginal)
+        matrix_.log()
+        return matrix_
+
+    def get_column_stats(self, matrix_):
+        return matrix_.sum(0)