Update scripts

hp/hp.py

@@ -45,6 +45,10 @@ while i < len(lines):
 
 print("Parsing lines")
 corpus = " ".join(truelines)
+
+#with open("hp_parsed.txt", "w") as f:
+#    f.writelines(corpus)
+
 tokens = corpus.split(" ")
 words = set(tokens)
 id_to_word = dict(enumerate(words))
@@ -55,8 +59,10 @@ from numpy import matrix as M, array
 #m = M( [ [ 0 for _ in range(len(words)) ] for _ in range(len(words)) ] )
 
 print("allocating array")
+# ich zähle alle folgeworte i→j
 m = array([0])
 m.resize(N, N)
+# die wahrscheinlichkeiten
 M = array([0.0])
 M.resize(N, N)