califyn/title-capitalization

title-capitalization's Issues

Breaks with IndexError on "(m/w/d)" string

Amazing Project! However when using the text enhancer I'll have to do extensive text cleanup. When running

input_text = "aushilfe (m/w/d) als reinigungskraft"
output_text = apply_te(input_text, lan='de')

python errors with

IndexError                                Traceback (most recent call last)

[<ipython-input-9-11b317305bd2>](https://localhost:8080/#) in <module>()
      1 # input_text = model.examples[0]
      2 input_text = "aushilfe (m/w/d) als reinigungskraft"
----> 3 output_text = apply_te(input_text, lan='de')
      4 print(f"Input: \n{input_text}\nOutput:\n{output_text}")

[<ipython-input-2-196122d4e396>](https://localhost:8080/#) in apply_te(text, lan)
     15 
     16 def apply_te(text, lan='en'):
---> 17     return model.enhance_text(text, lan)

<torch_package_0>.release_module.py in enhance_text(self, text, lan, len_limit)
    118 
    119         if len(text.split()) < len_limit:
--> 120             enhanced = self.enhance_textblock(text, lan_id)
    121         else:
    122             enhanced = self.enhance_long_textblock(text, lan_id, len_limit)

<torch_package_0>.release_module.py in enhance_textblock(self, text, lan_id)
     73                                                                       punct[0][1:-1],
     74                                                                       capital[0][1:-1],
---> 75                                                                       self.index2punct))
     76 
     77     def count_occurrences(self, text, char):

<torch_package_0>.release_module.py in enhance_tokens(tokens, punct, capital, index2punct)
    163                 if len(token) < 2:
    164                     print(token)
--> 165                 token = '##' + token[2].upper() + token[3:]
    166         if c == 2:
    167             token = token.upper()

IndexError: string index out of range

The same text without the (m/w/d) works as expected