From aa55578ebdce64acfe4284caf57b4450e8c91659 Mon Sep 17 00:00:00 2001
From: Hui Lan <lanhui@zjnu.edu.cn>
Date: Tue, 9 Feb 2021 12:30:00 +0800
Subject: overlap.py: compare predicted edges with gold standard (osu
 AtRegNet.csv) edges.

---
 Code/overlap.py | 76 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 76 insertions(+)
 create mode 100644 Code/overlap.py

diff --git a/Code/overlap.py b/Code/overlap.py
new file mode 100644
index 0000000..2630b23
--- /dev/null
+++ b/Code/overlap.py
@@ -0,0 +1,76 @@
+# Author: Hui Lan <lanhui@zjnu.edu.cn>
+class Overlap:
+
+    ''' Compare two networks and compute hit rate '''
+    
+    def __init__(self, predicted_edges={}, tau1=None, gold_standard_edges={}, tau2=None):
+        ''' predicted_edges: dictionary {'TF * Target': score}
+            gold_standard_edges: dictionary {'TF * Target': score}
+        '''
+        def get_edges_above_threshold(d, tau):
+            d2 = {}
+            for k in d:
+                if d[k] > tau:
+                    d2[k] = score
+            return d2
+        
+        self.pred = predicted_edges
+        self.true = gold_standard_edges
+        if tau1 != None:
+            self.pred = get_edges_above_threshold(self.pred, tau1)
+        if tau2 != None:
+            self.true = get_edges_above_threshold(self.true, tau2)
+
+
+    def getTP(self):
+        tp = 0
+        for k in self.pred:
+            if k in self.true:
+                tp += 1
+                
+        return tp
+
+    def getFP(self):
+        fp = 0
+        for k in self.pred:
+            if not k in self.true:
+                fp += 1
+                
+        return fp
+
+    def getNumberOfPositivesInPred(self):
+        return len(self.pred)
+
+    def getNumberOfPositivesInTrue(self):
+        return len(self.true)
+
+
+
+if __name__ == '__main__':
+    f = open('../Data/temp/AtRegNet.20210208.csv')
+    AtRegNet_dict = {}
+    for line in f:
+        line = line.strip()
+        lst = line.split(',')
+        if lst[0] != 'TFName' and len(lst) > 4:
+            tf = lst[1].upper().strip()
+            target = lst[4].upper().strip()
+            AtRegNet_dict[tf+target] = 100
+    f.close()
+
+    f = open('../Data/temp/edges.txt')
+    BrainEdges_dict = {}
+    for line in f:
+        line = line.strip()
+        lst = line.split('\t')
+        tf = lst[1].split()[0]
+        target = lst[0].split()[0]
+        score = float(lst[8])
+        BrainEdges_dict[tf+target] = score
+    f.close()
+
+    overlap = Overlap(BrainEdges_dict, 3, AtRegNet_dict, 0)
+    print('TP:%d, PP:%d, Hit rate: %4.7f whiling comparing with AtRegNet.20210208.csv.' % (overlap.getTP(), overlap.getNumberOfPositivesInPred(), overlap.getTP()/overlap.getNumberOfPositivesInPred()))
+
+
+    
-- 
cgit v1.2.1