0

So I copied a code from another StackOverflow post (How do I make a better visualization of word associations for a given text?). However, even without changing a single line I keep getting: free variable 'is_noun' referenced before assignment in enclosing scope

...why?

For reference, this is the code:

def word_association_graph(text):
    nouns_in_text = []

    for sent in text.split('.')[:-1]:   
        tokenized = nltk.word_tokenize(sent)
        nouns=[word for (word, pos) in nltk.pos_tag(tokenized) if is_noun(pos)]
        nouns_in_text.append(' '.join([word for word in nouns if not (word=='' or len(word)==1)]))

    nouns_list = []
    is_noun = lambda pos: pos[:2] == 'NN'

    for sent in nouns_in_text:
        temp = sent.split(' ')
        for word in temp:
            if word not in nouns_list:
                nouns_list.append(word)

    df = pd.DataFrame(np.zeros(shape=(len(nouns_list),2)), columns=['Nouns', 'Verbs & Adjectives'])
    df['Nouns'] = nouns_list

    is_adjective_or_verb = lambda pos: pos[:2]=='JJ' or pos[:2]=='VB'
    for sent in text.split('.'):
        for noun in nouns_list:
            if noun in sent:
                tokenized = nltk.word_tokenize(sent)
                adjectives_or_verbs = [word for (word, pos) in nltk.pos_tag(tokenized) if is_adjective_or_verb(pos)]
                ind = df[df['Nouns']==noun].index[0]
                df['Verbs & Adjectives'][ind]=adjectives_or_verbs

    fig = plt.figure(figsize=(30,20))
    G = nx.Graph()

    for i in range(len(df)):
        G.add_node(df['Nouns'][i])
        for word in df['Verbs & Adjectives'][i]:
            G.add_edges_from([(df['Nouns'][i], word)])

    pos = nx.spring_layout(G)
    nx.draw(G, with_labels=True, font_size=20) #font_weight='bold', ```

1 Answers1

1

You are using is_noun in the first for loop before it is even defined. This definition happens only after. Move your is_noun = lambda pos: pos[:2] == 'NN' line at the beginning of the function code.

mozway
  • 194,879
  • 13
  • 39
  • 75