Added todays python exersize, and it works

receng.py

@@ -0,0 +1,79 @@
+import numpy
+import rawdata
+data = rawdata.raw_scores
+
+def diff(v1,v2) :
+    v = numpy.zeros(len(v1))
+    for i in range(len(v1)) :
+       if v1[i]<0 or v2[i]<0:  
+           v[i] = 0
+       else :
+           v[i] = v1[i] - v2[i]
+    return v
+
+class Engine(object):
+
+   def set_data(self,data):
+       self.d = data
+
+   def create_matrix(self):
+       self.matrix = numpy.zeros( (self.NP, self.NR) )
+       for i in range(self.NR) :
+           referee = self.referees[i]
+           for j in range(self.NP) :
+               paper = self.papers[j]
+               if self.papers[j] not in self.d[referee] :
+                   self.matrix[j,i] = -1
+               else :
+                   self.matrix[j,i] = self.d[referee][paper]
+
+   def create_papers(self):
+       p = []
+       for referee in self.d.values(): 
+          for paper in referee.keys() :
+              p.append(paper)
+       sp = set(p)
+       self.papers = list(sp)
+       self.NP = len(self.papers)
+
+   def create_referees(self):
+       self.referees = self.d.keys()       
+       self.NR = len(self.referees)
+
+   def similarity(self,r1,r2):      
+       v1 = numpy.array(self.matrix[:,r1])
+       v2 = numpy.array(self.matrix[:,r2])
+#       print v1
+#       print v2
+       v = diff(v1,v2)   
+#       print v  
+       return numpy.linalg.norm( v )
+
+   def pearson(self,r1,r2):      
+       v1 = numpy.array(self.matrix[:,r1])
+       v2 = numpy.array(self.matrix[:,r2])
+       v1l = []
+       v2l = []
+       for i in range(self.NP) :
+         if v1[i]>=0 and v2[i] >= 0 :
+            v1l.append(v1[i])
+            v2l.append(v2[i])
+       if len(v1l) > 0 :
+         return numpy.cov( v1l, v2l )[0,1]
+       else :
+         return 0
+
+#   def tanimoto(self,r1,r2):
+
+
+
+engine = Engine()
+engine.set_data(data)
+engine.create_referees()
+engine.create_papers()
+engine.create_matrix()
+print engine.referees
+print engine.papers
+print engine.matrix
+print engine.similarity(2,4)
+print engine.pearson(2,3)