Lib/tokenize.py

   1 """Tokenization help for Python programs.
   2
   3 This module exports a function called 'tokenize()' that breaks a stream of
   4 text into Python tokens.  It accepts a readline-like method which is called
   5 repeatedly to get the next line of input (or "" for EOF) and a "token-eater"
   6 function which is called once for each token found.  The latter function is
   7 passed the token type, a string containing the token, the starting and
   8 ending (row, column) coordinates of the token, and the original line.  It is
   9 designed to match the working of the Python tokenizer exactly, except that
  10 it produces COMMENT tokens for comments and gives type OP for all operators."""
  11
  12 __version__ = "Ka-Ping Yee, 26 October 1997; patched, GvR 3/30/98"
  13
  14 import string, re
  15 from token import *
  16
  17 COMMENT = N_TOKENS
  18 tok_name[COMMENT] = 'COMMENT'
  19 NL = N_TOKENS + 1
  20 tok_name[NL] = 'NL'
  21
  22
  23 # Changes from 1.3:
  24 #     Ignore now accepts \f as whitespace.  Operator now includes '**'.
  25 #     Ignore and Special now accept \n or \r\n at the end of a line.
  26 #     Imagnumber is new.  Expfloat is corrected to reject '0e4'.
  27 # Note: to quote a backslash in a regex, it must be doubled in a r'aw' string.
  28
  29 def group(*choices): return '(' + '|'.join(choices) + ')'
  30 def any(*choices): return apply(group, choices) + '*'
  31 def maybe(*choices): return apply(group, choices) + '?'
  32
  33 Whitespace = r'[ \f\t]*'
  34 Comment = r'#[^\r\n]*'
  35 Ignore = Whitespace + any(r'\\\r?\n' + Whitespace) + maybe(Comment)
  36 Name = r'[a-zA-Z_]\w*'
  37
  38 Hexnumber = r'0[xX][\da-fA-F]*[lL]?'
  39 Octnumber = r'0[0-7]*[lL]?'
  40 Decnumber = r'[1-9]\d*[lL]?'
  41 Intnumber = group(Hexnumber, Octnumber, Decnumber)
  42 Exponent = r'[eE][-+]?\d+'
  43 Pointfloat = group(r'\d+\.\d*', r'\.\d+') + maybe(Exponent)
  44 Expfloat = r'[1-9]\d*' + Exponent
  45 Floatnumber = group(Pointfloat, Expfloat)
  46 Imagnumber = group(r'0[jJ]', r'[1-9]\d*[jJ]', Floatnumber + r'[jJ]')
  47 Number = group(Imagnumber, Floatnumber, Intnumber)
  48
  49 # Tail end of ' string.
  50 Single = r"[^'\\]*(?:\\.[^'\\]*)*'"
  51 # Tail end of " string.
  52 Double = r'[^"\\]*(?:\\.[^"\\]*)*"'
  53 # Tail end of ''' string.
  54 Single3 = r"[^'\\]*(?:(?:\\.|'(?!''))[^'\\]*)*'''"
  55 # Tail end of """ string.
  56 Double3 = r'[^"\\]*(?:(?:\\.|"(?!""))[^"\\]*)*"""'
  57 Triple = group("[uU]?[rR]?'''", '[uU]?[rR]?"""')
  58 # Single-line ' or " string.
  59 String = group(r"[uU]?[rR]?'[^\n'\\]*(?:\\.[^\n'\\]*)*'",
  60                r'[uU]?[rR]?"[^\n"\\]*(?:\\.[^\n"\\]*)*"')
  61
  62 # Because of leftmost-then-longest match semantics, be sure to put the
  63 # longest operators first (e.g., if = came before ==, == would get
  64 # recognized as two instances of =).
  65 Operator = group(r"\*\*=?", r">>=?", r"<<=?", r"<>", r"!=",
  66                  r"[+\-*/%&|^=<>]=?",
  67                  r"~")
  68
  69 Bracket = '[][(){}]'
  70 Special = group(r'\r?\n', r'[:;.,`]')
  71 Funny = group(Operator, Bracket, Special)
  72
  73 PlainToken = group(Number, Funny, String, Name)
  74 Token = Ignore + PlainToken
  75
  76 # First (or only) line of ' or " string.
  77 ContStr = group(r"[uU]?[rR]?'[^\n'\\]*(?:\\.[^\n'\\]*)*" +
  78                 group("'", r'\\\r?\n'),
  79                 r'[uU]?[rR]?"[^\n"\\]*(?:\\.[^\n"\\]*)*' +
  80                 group('"', r'\\\r?\n'))
  81 PseudoExtras = group(r'\\\r?\n', Comment, Triple)
  82 PseudoToken = Whitespace + group(PseudoExtras, Number, Funny, ContStr, Name)
  83
  84 tokenprog, pseudoprog, single3prog, double3prog = map(
  85     re.compile, (Token, PseudoToken, Single3, Double3))
  86 endprogs = {"'": re.compile(Single), '"': re.compile(Double),
  87             "'''": single3prog, '"""': double3prog,
  88             "r'''": single3prog, 'r"""': double3prog,
  89             "u'''": single3prog, 'u"""': double3prog,
  90             "ur'''": single3prog, 'ur"""': double3prog,
  91             "R'''": single3prog, 'R"""': double3prog,
  92             "U'''": single3prog, 'U"""': double3prog,
  93             "uR'''": single3prog, 'uR"""': double3prog,
  94             "Ur'''": single3prog, 'Ur"""': double3prog,
  95             "UR'''": single3prog, 'UR"""': double3prog,
  96             'r': None, 'R': None, 'u': None, 'U': None}
  97
  98 tabsize = 8
  99
 100 class TokenError(Exception):
 101     pass
 102
 103 def printtoken(type, token, (srow, scol), (erow, ecol), line): # for testing
 104     print "%d,%d-%d,%d:\t%s\t%s" % \
 105         (srow, scol, erow, ecol, tok_name[type], repr(token))
 106
 107 def tokenize(readline, tokeneater=printtoken):
 108     lnum = parenlev = continued = 0
 109     namechars, numchars = string.letters + '_', string.digits
 110     contstr, needcont = '', 0
 111     contline = None
 112     indents = [0]
 113
 114     while 1:                                   # loop over lines in stream
 115         line = readline()
 116         lnum = lnum + 1
 117         pos, max = 0, len(line)
 118
 119         if contstr:                            # continued string
 120             if not line:
 121                 raise TokenError, ("EOF in multi-line string", strstart)
 122             endmatch = endprog.match(line)
 123             if endmatch:
 124                 pos = end = endmatch.end(0)
 125                 tokeneater(STRING, contstr + line[:end],
 126                            strstart, (lnum, end), contline + line)
 127                 contstr, needcont = '', 0
 128                 contline = None
 129             elif needcont and line[-2:] != '\\\n' and line[-3:] != '\\\r\n':
 130                 tokeneater(ERRORTOKEN, contstr + line,
 131                            strstart, (lnum, len(line)), contline)
 132                 contstr = ''
 133                 contline = None
 134                 continue
 135             else:
 136                 contstr = contstr + line
 137                 contline = contline + line
 138                 continue
 139
 140         elif parenlev == 0 and not continued:  # new statement
 141             if not line: break
 142             column = 0
 143             while pos < max:                   # measure leading whitespace
 144                 if line[pos] == ' ': column = column + 1
 145                 elif line[pos] == '\t': column = (column/tabsize + 1)*tabsize
 146                 elif line[pos] == '\f': column = 0
 147                 else: break
 148                 pos = pos + 1
 149             if pos == max: break
 150
 151             if line[pos] in '#\r\n':           # skip comments or blank lines
 152                 tokeneater((NL, COMMENT)[line[pos] == '#'], line[pos:],
 153                            (lnum, pos), (lnum, len(line)), line)
 154                 continue
 155
 156             if column > indents[-1]:           # count indents or dedents
 157                 indents.append(column)
 158                 tokeneater(INDENT, line[:pos], (lnum, 0), (lnum, pos), line)
 159             while column < indents[-1]:
 160                 indents = indents[:-1]
 161                 tokeneater(DEDENT, '', (lnum, pos), (lnum, pos), line)
 162
 163         else:                                  # continued statement
 164             if not line:
 165                 raise TokenError, ("EOF in multi-line statement", (lnum, 0))
 166             continued = 0
 167
 168         while pos < max:
 169             pseudomatch = pseudoprog.match(line, pos)
 170             if pseudomatch:                                # scan for tokens
 171                 start, end = pseudomatch.span(1)
 172                 spos, epos, pos = (lnum, start), (lnum, end), end
 173                 token, initial = line[start:end], line[start]
 174
 175                 if initial in numchars \
 176                     or (initial == '.' and token != '.'):  # ordinary number
 177                     tokeneater(NUMBER, token, spos, epos, line)
 178                 elif initial in '\r\n':
 179                     tokeneater(parenlev > 0 and NL or NEWLINE,
 180                                token, spos, epos, line)
 181                 elif initial == '#':
 182                     tokeneater(COMMENT, token, spos, epos, line)
 183                 elif token in ("'''", '"""',               # triple-quoted
 184                                "r'''", 'r"""', "R'''", 'R"""',
 185                                "u'''", 'u"""', "U'''", 'U"""',
 186                                "ur'''", 'ur"""', "Ur'''", 'Ur"""',
 187                                "uR'''", 'uR"""', "UR'''", 'UR"""'):
 188                     endprog = endprogs[token]
 189                     endmatch = endprog.match(line, pos)
 190                     if endmatch:                           # all on one line
 191                         pos = endmatch.end(0)
 192                         token = line[start:pos]
 193                         tokeneater(STRING, token, spos, (lnum, pos), line)
 194                     else:
 195                         strstart = (lnum, start)           # multiple lines
 196                         contstr = line[start:]
 197                         contline = line
 198                         break
 199                 elif initial in ("'", '"') or \
 200                     token[:2] in ("r'", 'r"', "R'", 'R"',
 201                                   "u'", 'u"', "U'", 'U"') or \
 202                     token[:3] in ("ur'", 'ur"', "Ur'", 'Ur"',
 203                                   "uR'", 'uR"', "UR'", 'UR"' ):
 204                     if token[-1] == '\n':                  # continued string
 205                         strstart = (lnum, start)
 206                         endprog = (endprogs[initial] or endprogs[token[1]] or
 207                                    endprogs[token[2]])
 208                         contstr, needcont = line[start:], 1
 209                         contline = line
 210                         break
 211                     else:                                  # ordinary string
 212                         tokeneater(STRING, token, spos, epos, line)
 213                 elif initial in namechars:                 # ordinary name
 214                     tokeneater(NAME, token, spos, epos, line)
 215                 elif initial == '\\':                      # continued stmt
 216                     continued = 1
 217                 else:
 218                     if initial in '([{': parenlev = parenlev + 1
 219                     elif initial in ')]}': parenlev = parenlev - 1
 220                     tokeneater(OP, token, spos, epos, line)
 221             else:
 222                 tokeneater(ERRORTOKEN, line[pos],
 223                            (lnum, pos), (lnum, pos+1), line)
 224                 pos = pos + 1
 225
 226     for indent in indents[1:]:                 # pop remaining indent levels
 227         tokeneater(DEDENT, '', (lnum, 0), (lnum, 0), '')
 228     tokeneater(ENDMARKER, '', (lnum, 0), (lnum, 0), '')
 229
 230 if __name__ == '__main__':                     # testing
 231     import sys
 232     if len(sys.argv) > 1: tokenize(open(sys.argv[1]).readline)
 233     else: tokenize(sys.stdin.readline)