| 1 | #!/usr/bin/env python |
|---|
| 2 | # -*- coding: utf-8 -*- |
|---|
| 3 | |
|---|
| 4 | """ |
|---|
| 5 | Caller will hand this library a buffer and ask it to either convert |
|---|
| 6 | it or auto-detect the type. |
|---|
| 7 | |
|---|
| 8 | Based on http://code.activestate.com/recipes/52257/ |
|---|
| 9 | |
|---|
| 10 | Licensed under the PSF License |
|---|
| 11 | """ |
|---|
| 12 | from gluon._compat import to_unicode |
|---|
| 13 | import codecs |
|---|
| 14 | |
|---|
| 15 | # None represents a potentially variable byte. "##" in the XML spec... |
|---|
| 16 | autodetect_dict = { # bytepattern : ("name", |
|---|
| 17 | (0x00, 0x00, 0xFE, 0xFF): ("ucs4_be"), |
|---|
| 18 | (0xFF, 0xFE, 0x00, 0x00): ("ucs4_le"), |
|---|
| 19 | (0xFE, 0xFF, None, None): ("utf_16_be"), |
|---|
| 20 | (0xFF, 0xFE, None, None): ("utf_16_le"), |
|---|
| 21 | (0x00, 0x3C, 0x00, 0x3F): ("utf_16_be"), |
|---|
| 22 | (0x3C, 0x00, 0x3F, 0x00): ("utf_16_le"), |
|---|
| 23 | (0x3C, 0x3F, 0x78, 0x6D): ("utf_8"), |
|---|
| 24 | (0x4C, 0x6F, 0xA7, 0x94): ("EBCDIC") |
|---|
| 25 | } |
|---|
| 26 | |
|---|
| 27 | |
|---|
| 28 | def autoDetectXMLEncoding(buffer): |
|---|
| 29 | """ buffer -> encoding_name |
|---|
| 30 | The buffer should be at least 4 bytes long. |
|---|
| 31 | Returns None if encoding cannot be detected. |
|---|
| 32 | Note that encoding_name might not have an installed |
|---|
| 33 | decoder (e.g. EBCDIC) |
|---|
| 34 | """ |
|---|
| 35 | # a more efficient implementation would not decode the whole |
|---|
| 36 | # buffer at once but otherwise we'd have to decode a character at |
|---|
| 37 | # a time looking for the quote character...that's a pain |
|---|
| 38 | |
|---|
| 39 | encoding = "utf_8" |
|---|
| 40 | # according to the XML spec, this is the default this code successively tries to refine the default |
|---|
| 41 | # whenever it fails to refine, it falls back to the last place encoding was set. |
|---|
| 42 | |
|---|
| 43 | if len(buffer) >= 4: |
|---|
| 44 | bytes = (byte1, byte2, byte3, byte4) = tuple(map(ord, buffer[0:4])) |
|---|
| 45 | enc_info = autodetect_dict.get(bytes, None) |
|---|
| 46 | if not enc_info: # try autodetection again removing potentially |
|---|
| 47 | # variable bytes |
|---|
| 48 | bytes = (byte1, byte2, None, None) |
|---|
| 49 | enc_info = autodetect_dict.get(bytes) |
|---|
| 50 | else: |
|---|
| 51 | enc_info = None |
|---|
| 52 | |
|---|
| 53 | if enc_info: |
|---|
| 54 | encoding = enc_info # we've got a guess... these are the new defaults |
|---|
| 55 | |
|---|
| 56 | # try to find a more precise encoding using xml declaration |
|---|
| 57 | secret_decoder_ring = codecs.lookup(encoding)[1] |
|---|
| 58 | (decoded, length) = secret_decoder_ring(buffer) |
|---|
| 59 | first_line = decoded.split("\n")[0] |
|---|
| 60 | if first_line and first_line.startswith("<?xml"): |
|---|
| 61 | encoding_pos = first_line.find("encoding") |
|---|
| 62 | if encoding_pos != -1: |
|---|
| 63 | # look for double quote |
|---|
| 64 | quote_pos = first_line.find('"', encoding_pos) |
|---|
| 65 | |
|---|
| 66 | if quote_pos == -1: # look for single quote |
|---|
| 67 | quote_pos = first_line.find("'", encoding_pos) |
|---|
| 68 | |
|---|
| 69 | if quote_pos > -1: |
|---|
| 70 | quote_char, rest = (first_line[quote_pos], |
|---|
| 71 | first_line[quote_pos + 1:]) |
|---|
| 72 | encoding = rest[:rest.find(quote_char)] |
|---|
| 73 | |
|---|
| 74 | return encoding |
|---|
| 75 | |
|---|
| 76 | |
|---|
| 77 | def decoder(buffer): |
|---|
| 78 | encoding = autoDetectXMLEncoding(buffer) |
|---|
| 79 | return to_unicode(buffer, charset=encoding) |
|---|
