| 1 | """Tokenization help for Python programs. |
| 2 | |
| 3 | generate_tokens(readline) is a generator 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). It generates |
| 6 | 5-tuples with these members: |
| 7 | |
| 8 | the token type (see token.py) |
| 9 | the token (a string) |
| 10 | the starting (row, column) indices of the token (a 2-tuple of ints) |
| 11 | the ending (row, column) indices of the token (a 2-tuple of ints) |
| 12 | the original line (string) |
| 13 | |
| 14 | It is designed to match the working of the Python tokenizer exactly, except |
| 15 | that it produces COMMENT tokens for comments and gives type OP for all |
| 16 | operators |
| 17 | |
| 18 | Older entry points |
| 19 | tokenize_loop(readline, tokeneater) |
| 20 | tokenize(readline, tokeneater=printtoken) |
| 21 | are the same, except instead of generating tokens, tokeneater is a callback |
| 22 | function to which the 5 fields described above are passed as 5 arguments, |
| 23 | each time a new token is found.""" |
| 24 | |
| 25 | __author__ = 'Ka-Ping Yee <ping@lfw.org>' |
| 26 | __credits__ = \ |
| 27 | 'GvR, ESR, Tim Peters, Thomas Wouters, Fred Drake, Skip Montanaro' |
| 28 | |
| 29 | import string, re |
| 30 | from token import * |
| 31 | |
| 32 | import token |
| 33 | __all__ = [x for x in dir(token) if x[0] != '_'] + ["COMMENT", "tokenize", |
| 34 | "generate_tokens", "NL"] |
| 35 | del x |
| 36 | del token |
| 37 | |
| 38 | COMMENT = N_TOKENS |
| 39 | tok_name[COMMENT] = 'COMMENT' |
| 40 | NL = N_TOKENS + 1 |
| 41 | tok_name[NL] = 'NL' |
| 42 | N_TOKENS += 2 |
| 43 | |
| 44 | def group(*choices): return '(' + '|'.join(choices) + ')' |
| 45 | def any(*choices): return group(*choices) + '*' |
| 46 | def maybe(*choices): return group(*choices) + '?' |
| 47 | |
| 48 | Whitespace = r'[ \f\t]*' |
| 49 | Comment = r'#[^\r\n]*' |
| 50 | Ignore = Whitespace + any(r'\\\r?\n' + Whitespace) + maybe(Comment) |
| 51 | Name = r'[a-zA-Z_]\w*' |
| 52 | |
| 53 | Hexnumber = r'0[xX][\da-fA-F]*[lL]?' |
| 54 | Octnumber = r'0[0-7]*[lL]?' |
| 55 | Decnumber = r'[1-9]\d*[lL]?' |
| 56 | Intnumber = group(Hexnumber, Octnumber, Decnumber) |
| 57 | Exponent = r'[eE][-+]?\d+' |
| 58 | Pointfloat = group(r'\d+\.\d*', r'\.\d+') + maybe(Exponent) |
| 59 | Expfloat = r'\d+' + Exponent |
| 60 | Floatnumber = group(Pointfloat, Expfloat) |
| 61 | Imagnumber = group(r'\d+[jJ]', Floatnumber + r'[jJ]') |
| 62 | Number = group(Imagnumber, Floatnumber, Intnumber) |
| 63 | |
| 64 | # Tail end of ' string. |
| 65 | Single = r"[^'\\]*(?:\\.[^'\\]*)*'" |
| 66 | # Tail end of " string. |
| 67 | Double = r'[^"\\]*(?:\\.[^"\\]*)*"' |
| 68 | # Tail end of ''' string. |
| 69 | Single3 = r"[^'\\]*(?:(?:\\.|'(?!''))[^'\\]*)*'''" |
| 70 | # Tail end of """ string. |
| 71 | Double3 = r'[^"\\]*(?:(?:\\.|"(?!""))[^"\\]*)*"""' |
| 72 | Triple = group("[uU]?[rR]?'''", '[uU]?[rR]?"""') |
| 73 | # Single-line ' or " string. |
| 74 | String = group(r"[uU]?[rR]?'[^\n'\\]*(?:\\.[^\n'\\]*)*'", |
| 75 | r'[uU]?[rR]?"[^\n"\\]*(?:\\.[^\n"\\]*)*"') |
| 76 | |
| 77 | # Because of leftmost-then-longest match semantics, be sure to put the |
| 78 | # longest operators first (e.g., if = came before ==, == would get |
| 79 | # recognized as two instances of =). |
| 80 | Operator = group(r"\*\*=?", r">>=?", r"<<=?", r"<>", r"!=", |
| 81 | r"//=?", |
| 82 | r"[+\-*/%&|^=<>]=?", |
| 83 | r"~") |
| 84 | |
| 85 | Bracket = '[][(){}]' |
| 86 | Special = group(r'\r?\n', r'[:;.,`@]') |
| 87 | Funny = group(Operator, Bracket, Special) |
| 88 | |
| 89 | PlainToken = group(Number, Funny, String, Name) |
| 90 | Token = Ignore + PlainToken |
| 91 | |
| 92 | # First (or only) line of ' or " string. |
| 93 | ContStr = group(r"[uU]?[rR]?'[^\n'\\]*(?:\\.[^\n'\\]*)*" + |
| 94 | group("'", r'\\\r?\n'), |
| 95 | r'[uU]?[rR]?"[^\n"\\]*(?:\\.[^\n"\\]*)*' + |
| 96 | group('"', r'\\\r?\n')) |
| 97 | PseudoExtras = group(r'\\\r?\n', Comment, Triple) |
| 98 | PseudoToken = Whitespace + group(PseudoExtras, Number, Funny, ContStr, Name) |
| 99 | |
| 100 | tokenprog, pseudoprog, single3prog, double3prog = map( |
| 101 | re.compile, (Token, PseudoToken, Single3, Double3)) |
| 102 | endprogs = {"'": re.compile(Single), '"': re.compile(Double), |
| 103 | "'''": single3prog, '"""': double3prog, |
| 104 | "r'''": single3prog, 'r"""': double3prog, |
| 105 | "u'''": single3prog, 'u"""': double3prog, |
| 106 | "ur'''": single3prog, 'ur"""': double3prog, |
| 107 | "R'''": single3prog, 'R"""': double3prog, |
| 108 | "U'''": single3prog, 'U"""': double3prog, |
| 109 | "uR'''": single3prog, 'uR"""': double3prog, |
| 110 | "Ur'''": single3prog, 'Ur"""': double3prog, |
| 111 | "UR'''": single3prog, 'UR"""': double3prog, |
| 112 | 'r': None, 'R': None, 'u': None, 'U': None} |
| 113 | |
| 114 | triple_quoted = {} |
| 115 | for t in ("'''", '"""', |
| 116 | "r'''", 'r"""', "R'''", 'R"""', |
| 117 | "u'''", 'u"""', "U'''", 'U"""', |
| 118 | "ur'''", 'ur"""', "Ur'''", 'Ur"""', |
| 119 | "uR'''", 'uR"""', "UR'''", 'UR"""'): |
| 120 | triple_quoted[t] = t |
| 121 | single_quoted = {} |
| 122 | for t in ("'", '"', |
| 123 | "r'", 'r"', "R'", 'R"', |
| 124 | "u'", 'u"', "U'", 'U"', |
| 125 | "ur'", 'ur"', "Ur'", 'Ur"', |
| 126 | "uR'", 'uR"', "UR'", 'UR"' ): |
| 127 | single_quoted[t] = t |
| 128 | |
| 129 | tabsize = 8 |
| 130 | |
| 131 | class TokenError(Exception): pass |
| 132 | |
| 133 | class StopTokenizing(Exception): pass |
| 134 | |
| 135 | def printtoken(type, token, (srow, scol), (erow, ecol), line): # for testing |
| 136 | print "%d,%d-%d,%d:\t%s\t%s" % \ |
| 137 | (srow, scol, erow, ecol, tok_name[type], repr(token)) |
| 138 | |
| 139 | def tokenize(readline, tokeneater=printtoken): |
| 140 | """ |
| 141 | The tokenize() function accepts two parameters: one representing the |
| 142 | input stream, and one providing an output mechanism for tokenize(). |
| 143 | |
| 144 | The first parameter, readline, must be a callable object which provides |
| 145 | the same interface as the readline() method of built-in file objects. |
| 146 | Each call to the function should return one line of input as a string. |
| 147 | |
| 148 | The second parameter, tokeneater, must also be a callable object. It is |
| 149 | called once for each token, with five arguments, corresponding to the |
| 150 | tuples generated by generate_tokens(). |
| 151 | """ |
| 152 | try: |
| 153 | tokenize_loop(readline, tokeneater) |
| 154 | except StopTokenizing: |
| 155 | pass |
| 156 | |
| 157 | # backwards compatible interface |
| 158 | def tokenize_loop(readline, tokeneater): |
| 159 | for token_info in generate_tokens(readline): |
| 160 | tokeneater(*token_info) |
| 161 | |
| 162 | def generate_tokens(readline): |
| 163 | """ |
| 164 | The generate_tokens() generator requires one argment, readline, which |
| 165 | must be a callable object which provides the same interface as the |
| 166 | readline() method of built-in file objects. Each call to the function |
| 167 | should return one line of input as a string. |
| 168 | |
| 169 | The generator produces 5-tuples with these members: the token type; the |
| 170 | token string; a 2-tuple (srow, scol) of ints specifying the row and |
| 171 | column where the token begins in the source; a 2-tuple (erow, ecol) of |
| 172 | ints specifying the row and column where the token ends in the source; |
| 173 | and the line on which the token was found. The line passed is the |
| 174 | logical line; continuation lines are included. |
| 175 | """ |
| 176 | lnum = parenlev = continued = 0 |
| 177 | namechars, numchars = string.ascii_letters + '_', '0123456789' |
| 178 | contstr, needcont = '', 0 |
| 179 | contline = None |
| 180 | indents = [0] |
| 181 | |
| 182 | while 1: # loop over lines in stream |
| 183 | line = readline() |
| 184 | lnum = lnum + 1 |
| 185 | pos, max = 0, len(line) |
| 186 | |
| 187 | if contstr: # continued string |
| 188 | if not line: |
| 189 | raise TokenError, ("EOF in multi-line string", strstart) |
| 190 | endmatch = endprog.match(line) |
| 191 | if endmatch: |
| 192 | pos = end = endmatch.end(0) |
| 193 | yield (STRING, contstr + line[:end], |
| 194 | strstart, (lnum, end), contline + line) |
| 195 | contstr, needcont = '', 0 |
| 196 | contline = None |
| 197 | elif needcont and line[-2:] != '\\\n' and line[-3:] != '\\\r\n': |
| 198 | yield (ERRORTOKEN, contstr + line, |
| 199 | strstart, (lnum, len(line)), contline) |
| 200 | contstr = '' |
| 201 | contline = None |
| 202 | continue |
| 203 | else: |
| 204 | contstr = contstr + line |
| 205 | contline = contline + line |
| 206 | continue |
| 207 | |
| 208 | elif parenlev == 0 and not continued: # new statement |
| 209 | if not line: break |
| 210 | column = 0 |
| 211 | while pos < max: # measure leading whitespace |
| 212 | if line[pos] == ' ': column = column + 1 |
| 213 | elif line[pos] == '\t': column = (column/tabsize + 1)*tabsize |
| 214 | elif line[pos] == '\f': column = 0 |
| 215 | else: break |
| 216 | pos = pos + 1 |
| 217 | if pos == max: break |
| 218 | |
| 219 | if line[pos] in '#\r\n': # skip comments or blank lines |
| 220 | yield ((NL, COMMENT)[line[pos] == '#'], line[pos:], |
| 221 | (lnum, pos), (lnum, len(line)), line) |
| 222 | continue |
| 223 | |
| 224 | if column > indents[-1]: # count indents or dedents |
| 225 | indents.append(column) |
| 226 | yield (INDENT, line[:pos], (lnum, 0), (lnum, pos), line) |
| 227 | while column < indents[-1]: |
| 228 | if column not in indents: |
| 229 | raise IndentationError( |
| 230 | "unindent does not match any outer indentation level") |
| 231 | indents = indents[:-1] |
| 232 | yield (DEDENT, '', (lnum, pos), (lnum, pos), line) |
| 233 | |
| 234 | else: # continued statement |
| 235 | if not line: |
| 236 | raise TokenError, ("EOF in multi-line statement", (lnum, 0)) |
| 237 | continued = 0 |
| 238 | |
| 239 | while pos < max: |
| 240 | pseudomatch = pseudoprog.match(line, pos) |
| 241 | if pseudomatch: # scan for tokens |
| 242 | start, end = pseudomatch.span(1) |
| 243 | spos, epos, pos = (lnum, start), (lnum, end), end |
| 244 | token, initial = line[start:end], line[start] |
| 245 | |
| 246 | if initial in numchars or \ |
| 247 | (initial == '.' and token != '.'): # ordinary number |
| 248 | yield (NUMBER, token, spos, epos, line) |
| 249 | elif initial in '\r\n': |
| 250 | yield (parenlev > 0 and NL or NEWLINE, |
| 251 | token, spos, epos, line) |
| 252 | elif initial == '#': |
| 253 | yield (COMMENT, token, spos, epos, line) |
| 254 | elif token in triple_quoted: |
| 255 | endprog = endprogs[token] |
| 256 | endmatch = endprog.match(line, pos) |
| 257 | if endmatch: # all on one line |
| 258 | pos = endmatch.end(0) |
| 259 | token = line[start:pos] |
| 260 | yield (STRING, token, spos, (lnum, pos), line) |
| 261 | else: |
| 262 | strstart = (lnum, start) # multiple lines |
| 263 | contstr = line[start:] |
| 264 | contline = line |
| 265 | break |
| 266 | elif initial in single_quoted or \ |
| 267 | token[:2] in single_quoted or \ |
| 268 | token[:3] in single_quoted: |
| 269 | if token[-1] == '\n': # continued string |
| 270 | strstart = (lnum, start) |
| 271 | endprog = (endprogs[initial] or endprogs[token[1]] or |
| 272 | endprogs[token[2]]) |
| 273 | contstr, needcont = line[start:], 1 |
| 274 | contline = line |
| 275 | break |
| 276 | else: # ordinary string |
| 277 | yield (STRING, token, spos, epos, line) |
| 278 | elif initial in namechars: # ordinary name |
| 279 | yield (NAME, token, spos, epos, line) |
| 280 | elif initial == '\\': # continued stmt |
| 281 | continued = 1 |
| 282 | else: |
| 283 | if initial in '([{': parenlev = parenlev + 1 |
| 284 | elif initial in ')]}': parenlev = parenlev - 1 |
| 285 | yield (OP, token, spos, epos, line) |
| 286 | else: |
| 287 | yield (ERRORTOKEN, line[pos], |
| 288 | (lnum, pos), (lnum, pos+1), line) |
| 289 | pos = pos + 1 |
| 290 | |
| 291 | for indent in indents[1:]: # pop remaining indent levels |
| 292 | yield (DEDENT, '', (lnum, 0), (lnum, 0), '') |
| 293 | yield (ENDMARKER, '', (lnum, 0), (lnum, 0), '') |
| 294 | |
| 295 | if __name__ == '__main__': # testing |
| 296 | import sys |
| 297 | if len(sys.argv) > 1: tokenize(open(sys.argv[1]).readline) |
| 298 | else: tokenize(sys.stdin.readline) |