Project

General

Profile

« Previous | Next » 

Revision 5154

Added by Aaron Marcuse-Kubitza about 12 years ago

tnrs.py: encode(): Also prepend special padding string to empty and whitespace-only strings because these names are otherwise ignored by TNRS (no response row)

View differences:

lib/tnrs.py
64 64 
    return [(r'(?: |(?<=\t)|^)'+re.escape(from_.strip())+r'(?: |(?=\t)|$)',
65 65 
        strings.regexp_repl_esc(to)) for from_, to in decode_map]
66 66 

  		




  
  67
  
    
padding = ' !pad ' # prepend to empty and whitespace-only strings


  		




  67
  68
  
    
encode_map = [


  		




  68
  69
  
    
    ('!', ' !exc '), # escape char


  		




  69
  70
  
    
    ('\t', ' !tab '),


  		




  ......




  75
  76
  
    
    ('\\', ' !bsl '),


  		




  76
  77
  
    
]


  		




  77
  78
  
    
decode_map = strings.flip_map(encode_map)


  		




  
  79
  
    
decode_map.append((padding, ''))


  		




  78
  80
  
    
spliced_decode_map = make_spliced_decode_map(decode_map)


  		




  79
  81
  
    

  		




  80
  
  
    
def encode(str_): return strings.replace_all(encode_map, str_)


  		




  
  82
  
    
def encode(str_):


  		




  
  83
  
    
    str_ = strings.replace_all(encode_map, str_)


  		




  
  84
  
    
    # Empty and whitespace-only strings are ignored by TNRS (no response row)


  		




  
  85
  
    
    if str_.strip() == '': str_ = padding+str_


  		




  
  86
  
    
    return str_


  		




  81
  87
  
    

  		




  82
  88
  
    
def decode(str_): return strings.replace_all_re(spliced_decode_map, str_)


  		




  83
  89
  
    

  		












Also available in:  Unified diff