Home	Trees	Indices	Help	PyXB hosted on

Source Code for Module pyxb.utils.activestate

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

Home	Trees	Indices	Help	PyXB hosted on