summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
-rw-r--r--Lib/tokenize.py60
1 files changed, 60 insertions, 0 deletions
diff --git a/Lib/tokenize.py b/Lib/tokenize.py
new file mode 100644
index 0000000..6b3d991
--- /dev/null
+++ b/Lib/tokenize.py
@@ -0,0 +1,60 @@
+# This module compiles a regular expression that recognizes Python tokens.
+# It is designed to match the working of the Python tokenizer exactly.
+# It takes care of everything except indentation;
+# note that un-escaped newlines are tokens, too.
+# tokenprog.regs[3] gives the location of the token without whitespace
+# It also defines various subexpressions, but doesn't compile them.
+# See the function test() below for an example of how to use.
+
+import regex
+
+# Note: to get a quoted backslash in a regexp, it must be quadrupled.
+
+Ignore = '[ \t]*\(\\\\\n[ \t]*\)*\(#.*\)?'
+
+Name = '[a-zA-Z_][a-zA-Z0-9_]*'
+
+Hexnumber = '0[xX][0-9a-fA-F]*[lL]?'
+Octnumber = '0[0-7]*[lL]?'
+Decnumber = '[1-9][0-9]*[lL]?'
+Intnumber = Hexnumber + '\|' + Octnumber + '\|' + Decnumber
+Exponent = '[eE][-+]?[0-9]+'
+Pointfloat = '\([0-9]+\.[0-9]*\|\.[0-9]+\)\(' + Exponent + '\)?'
+Expfloat = '[0-9]+' + Exponent
+Floatnumber = Pointfloat + '\|' + Expfloat
+Number = Intnumber + '\|' + Floatnumber
+
+String = '\'\(\\\\.\|[^\\\n\']\)*\''
+
+Operator = '~\|\+\|-\|\*\|/\|%\|\^\|&\||\|<<\|>>\|==\|<=\|<>\|!=\|>=\|=\|<\|>'
+Bracket = '[][(){}]'
+Special = '[:;.,`\n]'
+Funny = Operator + '\|' + Bracket + '\|' + Special
+
+PlainToken = Name + '\|' + Number + '\|' + String + '\|' + Funny
+
+Token = Ignore + '\(' + PlainToken + '\)'
+
+try:
+ save_syntax = regex.set_syntax(0) # Use default syntax
+ tokenprog = regex.compile(Token)
+finally:
+ dummy = regex.set_syntax(save_syntax) # Restore original syntax
+
+
+def test(file):
+ f = open(file, 'r')
+ while 1:
+ line = f.readline()
+ if not line: break
+ i, n = 0, len(line)
+ while i < n:
+ j = tokenprog.match(line, i)
+ if j < 0:
+ print 'No token at', `line[i:i+20]` + '...'
+ i = i+1
+ else:
+ i = i+j
+ a, b = tokenprog.regs[3]
+ if a < b:
+ print 'Token:', `line[a:b]`