Merge branch 'master' of git.xpub.nl:/var/www/git.xpub.nl/repos/OuNoPo-make

Makefile

@@ -61,7 +61,7 @@ output/tagged-words.txt: ocr/output.txt ## Analyzes OCR'ed text using a Part of
 #  >>> import nltk
 #  >>> nltk.download('averaged_perceptron_tagger')
 
-output/chatbot.txt: ocr/output.txt ## DESCRIBE WHAT IT DOES. Dependencies: python3's chatterbot
+output/chatbot.txt: ocr/output.txt ## Comments a text with a simple chatbot. Dependencies: python3's chatterbot
 	cat $< | python3 src/textbotconversation.py $(@)
 
 

src/wordtagger.py

@@ -1,26 +1,28 @@
 import nltk
 from sys import stdin, stdout
 
-# Step 1: define input and set up a list
+# Define input
 input = stdin.read()
-taggedwordlist = []
 
-string = input
-words = nltk.word_tokenize(string)
-taggedwordlist = nltk.pos_tag(words)
-
-for word, pos in nltk.pos_tag(words):
+# FILTER FUNCTIONS
+# This function cuts a string into words. Then runs a POS tagger for each word. Returns a list with tags
+def postagger(string):
+    words = nltk.word_tokenize(string)
     taggedwordlist = nltk.pos_tag(words)
-    # print('{0} is a {1}'.format(word,pos)) # Command out to print the analysis step
-
 
-taglist = [ pos for word,pos in taggedwordlist ]
+    for word, pos in nltk.pos_tag(words):
+        taggedwordlist = nltk.pos_tag(words)
+        #print('{0} is a {1}'.format(word,pos)) # Comment out to print the analysis step
 
-#print(taglist)
+    taglist = [ pos for word,pos in taggedwordlist ]
+    #print(taglist)
+    return taglist;
 
-readabletaglist = []
+# This function changes the tags to readable equivalents (NNP to noun for example)
+def postagger_readable(list):
+    readabletaglist = []
 
-for tag in taglist:
+    for tag in list:
         if tag in {"NNP","NNS","NN","NNPS"}:
             readabletag = 'noun'
         elif tag in {'VB','VBD','VBG','VBN','VBP','VBZ'}:
@@ -67,5 +69,14 @@ for tag in taglist:
             readabletag = tag
 
         readabletaglist.append(readabletag)
+    return readabletaglist;
+
+
+# This function creates the output
+def main():
+    taglist = postagger(input)
+    readabletaglist = postagger_readable(taglist)
+    stdout.write(' '.join(readabletaglist))
+    stdout.write('\n')
 
-stdout.write(' '.join(readabletaglist))
+main()