Home	Trees	Indices	Help	PyXB hosted on

Source Code for Module pyxb.utils.activestate

 1  # -*- coding: utf-8 -*- 
 2  # This file contains code adopted from http://code.activestate.com/recipes/. 
 3  # Copyright and licensing information is associated with each incorporated 
 4  # recipe.  At this time, all code is licensed under PSF. 
 5   
 6  # http://code.activestate.com/recipes/363841/ 
 7  # by Lars Tiede, derivative from: 
 8  # http://code.activestate.com/recipes/52257/ 
 9  # by Paul Prescod 
10  # Licensed under PSF 
11  # No changes 
12 -def detectXMLEncoding(fp): 
13      """ Attempts to detect the character encoding of the xml file 
14      given by a file object fp. fp must not be a codec wrapped file 
15      object! 
16   
17      The return value can be: 
18          - if detection of the BOM succeeds, the codec name of the 
19          corresponding unicode charset is returned 
20   
21          - if BOM detection fails, the xml declaration is searched for 
22          the encoding attribute and its value returned. the "<" 
23          character has to be the very first in the file then (it's xml 
24          standard after all). 
25   
26          - if BOM and xml declaration fail, None is returned. According 
27          to xml 1.0 it should be utf_8 then, but it wasn't detected by 
28          the means offered here. at least one can be pretty sure that a 
29          character coding including most of ASCII is used :-/ 
30      """ 
31      ### detection using BOM 
32       
33      ## the BOMs we know, by their pattern 
34      bomDict={ # bytepattern : name               
35               (0x00, 0x00, 0xFE, 0xFF) : "utf_32_be",         
36               (0xFF, 0xFE, 0x00, 0x00) : "utf_32_le", 
37               (0xFE, 0xFF, None, None) : "utf_16_be",  
38               (0xFF, 0xFE, None, None) : "utf_16_le",  
39               (0xEF, 0xBB, 0xBF, None) : "utf_8", 
40              } 
41   
42      ## go to beginning of file and get the first 4 bytes 
43      oldFP = fp.tell() 
44      fp.seek(0) 
45      (byte1, byte2, byte3, byte4) = tuple(map(ord, fp.read(4))) 
46   
47      ## try bom detection using 4 bytes, 3 bytes, or 2 bytes 
48      bomDetection = bomDict.get((byte1, byte2, byte3, byte4)) 
49      if not bomDetection : 
50          bomDetection = bomDict.get((byte1, byte2, byte3, None)) 
51          if not bomDetection : 
52              bomDetection = bomDict.get((byte1, byte2, None, None)) 
53   
54      ## if BOM detected, we're done :-) 
55      if bomDetection : 
56          fp.seek(oldFP) 
57          return bomDetection 
58   
59   
60      ## still here? BOM detection failed. 
61      ##  now that BOM detection has failed we assume one byte character 
62      ##  encoding behaving ASCII - of course one could think of nice 
63      ##  algorithms further investigating on that matter, but I won't for now. 
64       
65   
66      ### search xml declaration for encoding attribute 
67      import re 
68   
69      ## assume xml declaration fits into the first 2 KB (*cough*) 
70      fp.seek(0) 
71      buffer = fp.read(2048) 
72   
73      ## set up regular expression 
74      xmlDeclPattern = r""" 
75      ^<\?xml             # w/o BOM, xmldecl starts with <?xml at the first byte 
76      .+?                 # some chars (version info), matched minimal 
77      encoding=           # encoding attribute begins 
78      ["']                # attribute start delimiter 
79      (?P<encstr>         # what's matched in the brackets will be named encstr 
80       [^"']+              # every character not delimiter (not overly exact!) 
81      )                   # closes the brackets pair for the named group 
82      ["']                # attribute end delimiter 
83      .*?                 # some chars optionally (standalone decl or whitespace) 
84      \?>                 # xmldecl end 
85      """ 
86   
87      xmlDeclRE = re.compile(xmlDeclPattern, re.VERBOSE) 
88   
89      ## search and extract encoding string 
90      match = xmlDeclRE.search(buffer) 
91      fp.seek(oldFP) 
92      if match : 
93          return match.group("encstr") 
94      else : 
95          return None 
96   

Home	Trees	Indices	Help	PyXB hosted on