changed graph centrality prior, added sentence position prior

ex2/2.py

@@ -105,7 +105,7 @@ def build_summary(sentences,prior_func,weight_func,t):
         with open(source_path + text_file,'r',encoding='Latin-1') as file: #source_path + text_file
             text = file.read()
         sentences = text_to_sentences(text)
-        summary = build_summary(sentences,uniform_prior,cos_sim_weight,thresh)
+        summary = build_summary(sentences,uniform_prior,uniform_weight,thresh)
         with open(source_path + text_file,'r',encoding='Latin-1') as summary_file: #sums_path+ 'Ext-' + text_file    '../ex1/textsum.txt'
             MAP += AP(summary,summary_file.read())
             #print(MAP)

ex2/priors.py

@@ -3,6 +3,20 @@ def uniform_prior(sent_index,graph,sentences):
 
 #receives graph matrix for convenience...
 def degree_centrality_prior(sent_index,graph,sentences):
-	links = graph[sent_index]
-	nonzero = np.nonzero(links)[0]
-	return len(nonzero)/len(links)
+	total_links = 0
+	degree = 0 
+	for i in range(len(graph)):
+		nonzeros = len(np.nonzero(graph[i])[0])
+		if i == sent_index:
+			degree = nonzeros
+		total_links += nonzeros
+	if total_links == 0:
+		return 0
+	return degree/total_links
+
+
+def sentence_position_prior(sent_index,graph,sentences):
+	total = 0
+	for i in range(len(sentences)):
+		total += 1
+	return (sent_index + 1) / total

functions.py

@@ -35,7 +35,7 @@ def AP(systemSummaries, targetSummaries):
 
 #sentences is a list, returns cossim matrix
 def get_cosine_similarities_matrix(sentences):	
-	vec = TfidfVectorizer()
+	vec = TfidfVectorizer(stop_words='english')
 
 	X = vec.fit_transform(sentences)
 	return cosine_similarity(X)