from spacy import displacy

def firstLineQuote(text):

pattern_qp = r'\"(.*?)\"[@\-\w\s]+'

pattern_q = r'\"(.*?)\"'

r_qp = re.fullmatch(pattern_qp,text)

r_q = re.fullmatch(pattern_q, text)

if r_qp != None:

return 'personal_quote'

if r_q != None:

return 'quote'

return ''

def generateNER(text):

doc = NLP(text)

ners = []

for word in doc.ents:

if word.label_.lower() in ['event', 'fac', 'gpe', 'law', 'loc', 'money', 'norp', 'org', 'person', 'product', 'work_of_art']:

ners.append(word.label_.lower())

return ','.join(ners)

def analyzeText(text):

print("----------- Actual Text ------------")

print(text)

doc = NLP(text)

print("----------- Spacy render -----------")

displacy.render(doc,style="ent",jupyter=True)

print("----------- Spacy NERs -------------")

for word in doc.ents:

print(word.text,word.label_)

print("----------- POS tag ----------------")

# Token and Tag

for token in doc:

print(token, token.pos_)

nike_data['caption_cleaned'] = nike_data['caption'].apply(lambda x : wrangle(x))

nike_data['NER'] = nike_data['caption_cleaned'].apply(lambda x : generateNER(x))

nike_data['firstLineQuote'] = nike_data['caption_cleaned'].apply(lambda x : firstLineQuote(x.strip().split('\n')[0].strip()))