1 | """RFC 3676 format=flowed text processing.
|
---|
2 |
|
---|
3 | This module provides an API to create and display text/plain; format=flowed
|
---|
4 | mimetype text.
|
---|
5 |
|
---|
6 | """
|
---|
7 |
|
---|
8 | # Copyright (C) 2005 Martijn Pieters
|
---|
9 | # Written by Martijn Pieters <mj@zopatista.com>
|
---|
10 | # Development was sponsored by Logicalware (http://www.logicalware.org/)
|
---|
11 | # Licensed as Open Source under the same terms as the Python 2.4.1 license,
|
---|
12 | # as available at http://www.python.org/2.4.1/license.html
|
---|
13 |
|
---|
14 | __revision_id__ = '$Id: formatflowed.py 27 2005-09-17 19:13:48Z mj $'
|
---|
15 |
|
---|
16 | import re
|
---|
17 | import textwrap
|
---|
18 |
|
---|
19 | __all__ = [
|
---|
20 | 'PARAGRAPH',
|
---|
21 | 'FIXED',
|
---|
22 | 'SIGNATURE_SEPARATOR',
|
---|
23 | 'FormatFlowedDecoder',
|
---|
24 | 'FormatFlowedEncoder',
|
---|
25 | 'decode',
|
---|
26 | 'encode',
|
---|
27 | 'convertToWrapped',
|
---|
28 | 'convertToFlowed'
|
---|
29 | ]
|
---|
30 |
|
---|
31 | # Constants denoting the various text chunk types recognized by format=flowed
|
---|
32 | PARAGRAPH, FIXED, SIGNATURE_SEPARATOR = range(3)
|
---|
33 |
|
---|
34 |
|
---|
35 | # -- Public classes ----------------------------------------------------
|
---|
36 |
|
---|
37 |
|
---|
38 | class FormatFlowedDecoder:
|
---|
39 | """Object for converting format=flowed text to other formats
|
---|
40 |
|
---|
41 | The following instance attributes influence the interpretation of
|
---|
42 | format=flowed text:
|
---|
43 | delete_space (default: False)
|
---|
44 | Delete the trailing space before the CRLF on flowed lines before
|
---|
45 | interpreting the line on flowed input, corresponds to the DelSp mime
|
---|
46 | parameter
|
---|
47 | character_set (default: us-ascii)
|
---|
48 | The encoding of text passed in. Text is decoded to unicode using this
|
---|
49 | encoding, using the default error handing scheme.
|
---|
50 |
|
---|
51 | """
|
---|
52 | def __init__(self, delete_space=False, character_set='us-ascii'):
|
---|
53 | self.delete_space = delete_space
|
---|
54 | self.character_set = character_set
|
---|
55 |
|
---|
56 | # -- Private methods -----------------------------------------------
|
---|
57 |
|
---|
58 | def _stripquotes(self, line):
|
---|
59 | """Remove quotemarks from the start of the line
|
---|
60 |
|
---|
61 | Returns the number of quotemarks stripped and the stripped line:
|
---|
62 |
|
---|
63 | >>> decoder = FormatFlowedDecoder()
|
---|
64 | >>> decoder._stripquotes(u'>>> quoted line')
|
---|
65 | (3, u' quoted line')
|
---|
66 |
|
---|
67 | Non-quoted lines are returned unchanged:
|
---|
68 |
|
---|
69 | >>> decoder._stripquotes(u'non-quoted line')
|
---|
70 | (0, u'non-quoted line')
|
---|
71 |
|
---|
72 | """
|
---|
73 | stripped = line.lstrip('>')
|
---|
74 | return len(line) - len(stripped), stripped
|
---|
75 |
|
---|
76 | def _stripstuffing(self, line):
|
---|
77 | """Remove the optional leading space
|
---|
78 |
|
---|
79 | Returns the stripped line:
|
---|
80 |
|
---|
81 | >>> decoder = FormatFlowedDecoder()
|
---|
82 | >>> decoder._stripstuffing(u' stuffed line')
|
---|
83 | u'stuffed line'
|
---|
84 |
|
---|
85 | Non-stuffed lines are returned unchanged:
|
---|
86 |
|
---|
87 | >>> decoder._stripstuffing(u'non-stuffed line')
|
---|
88 | u'non-stuffed line'
|
---|
89 |
|
---|
90 | Additional spacing is preserved:
|
---|
91 |
|
---|
92 | >>> decoder._stripstuffing(u' extra leading space')
|
---|
93 | u' extra leading space'
|
---|
94 |
|
---|
95 | """
|
---|
96 | if line.startswith(u' '):
|
---|
97 | return line[1:]
|
---|
98 | return line
|
---|
99 |
|
---|
100 | def _stripflow(self, line):
|
---|
101 | """Remove the trailing flow space is delete_space is set
|
---|
102 |
|
---|
103 | The instance attribute delete_space is False by default thus this
|
---|
104 | method returns the line unchanged:
|
---|
105 |
|
---|
106 | >>> decoder = FormatFlowedDecoder()
|
---|
107 | >>> decoder._stripflow(u'flowed line ')
|
---|
108 | u'flowed line '
|
---|
109 |
|
---|
110 | But if the delete_space attribute has been set to True the flow space
|
---|
111 | is removed:
|
---|
112 |
|
---|
113 | >>> decoder = FormatFlowedDecoder(delete_space=True)
|
---|
114 | >>> decoder._stripflow(u'flowed line ')
|
---|
115 | u'flowed line'
|
---|
116 |
|
---|
117 | Only one flow space is removed:
|
---|
118 | >>> decoder._stripflow(u'extra whitespace ')
|
---|
119 | u'extra whitespace '
|
---|
120 |
|
---|
121 | """
|
---|
122 | if self.delete_space and line.endswith(u' '):
|
---|
123 | return line[:-1]
|
---|
124 | return line
|
---|
125 |
|
---|
126 | # -- Public API ----------------------------------------------------
|
---|
127 |
|
---|
128 | def decode(self, flowed):
|
---|
129 | """Decode flowed text
|
---|
130 |
|
---|
131 | Returns an iterable serving a sequence of (information, chunk)
|
---|
132 | tuples. information is a dictionary with the following fields:
|
---|
133 | type
|
---|
134 | One of PARAGRAPH, FIXED, SIGNATURE_SEPARATOR
|
---|
135 | quotedepth
|
---|
136 | Number of quotemarks found on the text chunk
|
---|
137 |
|
---|
138 | chunk is a unicode string. All text is unwrapped and without any
|
---|
139 | quotemarks; when displaying these chunks, the appropriate quotemarks
|
---|
140 | should be added again, and chunks of type PARAGRAPH should be
|
---|
141 | displayed wrapped. Chunks of type FIXED should be displayed
|
---|
142 | unwrapped.
|
---|
143 |
|
---|
144 |
|
---|
145 | Examples
|
---|
146 | --------
|
---|
147 |
|
---|
148 | Here is a simple example:
|
---|
149 |
|
---|
150 | >>> CRLF = '\\r\\n'
|
---|
151 | >>> decoder = FormatFlowedDecoder()
|
---|
152 | >>> result = decoder.decode(CRLF.join((
|
---|
153 | ... ">> `Take some more tea,' the March Hare said to Alice, very ",
|
---|
154 | ... ">> earnestly.",
|
---|
155 | ... ">",
|
---|
156 | ... "> `I've had nothing yet,' Alice replied in an offended ",
|
---|
157 | ... "> tone, `so I can't take more.'",
|
---|
158 | ... "",
|
---|
159 | ... "`You mean you can't take less,' said the Hatter: `it's very ",
|
---|
160 | ... "easy to take more than nothing.'",
|
---|
161 | ... "",
|
---|
162 | ... "-- ",
|
---|
163 | ... "Lewis Carroll")))
|
---|
164 | >>> list(result) == [
|
---|
165 | ... ({'quotedepth': 2, 'type': PARAGRAPH},
|
---|
166 | ... u"`Take some more tea,' the March Hare said to Alice, "
|
---|
167 | ... u"very earnestly."),
|
---|
168 | ... ({'quotedepth': 1, 'type': FIXED}, u""),
|
---|
169 | ... ({'quotedepth': 1, 'type': PARAGRAPH},
|
---|
170 | ... u"`I've had nothing yet,' Alice replied in an offended "
|
---|
171 | ... u"tone, `so I can't take more.'"),
|
---|
172 | ... ({'quotedepth': 0, 'type': FIXED}, u""),
|
---|
173 | ... ({'quotedepth': 0, 'type': PARAGRAPH},
|
---|
174 | ... u"`You mean you can't take less,' said the Hatter: `it's "
|
---|
175 | ... u"very easy to take more than nothing.'"),
|
---|
176 | ... ({'quotedepth': 0, 'type': FIXED}, u""),
|
---|
177 | ... ({'quotedepth': 0, 'type': SIGNATURE_SEPARATOR}, u"-- "),
|
---|
178 | ... ({'quotedepth': 0, 'type': FIXED}, u"Lewis Carroll")
|
---|
179 | ... ]
|
---|
180 | True
|
---|
181 |
|
---|
182 |
|
---|
183 | Improperly closed paragraphs
|
---|
184 | ----------------------------
|
---|
185 |
|
---|
186 | The decoder can deal with various cases of improperly format=flowed
|
---|
187 | messages. Paragraphs normally end with a fixed line, but the following
|
---|
188 | cases are also considered paragraph-closing cases:
|
---|
189 |
|
---|
190 | - A change in quotedepth:
|
---|
191 |
|
---|
192 | >>> result = decoder.decode(CRLF.join((
|
---|
193 | ... "> Depth one paragraph with flow space. ",
|
---|
194 | ... ">> Depth two paragraph with flow space. ",
|
---|
195 | ... "Depth zero paragraph with fixed line.")))
|
---|
196 | >>> list(result) == [
|
---|
197 | ... ({'quotedepth': 1, 'type': PARAGRAPH},
|
---|
198 | ... u"Depth one paragraph with flow space. "),
|
---|
199 | ... ({'quotedepth': 2, 'type': PARAGRAPH},
|
---|
200 | ... u"Depth two paragraph with flow space. "),
|
---|
201 | ... ({'quotedepth': 0, 'type': FIXED},
|
---|
202 | ... u"Depth zero paragraph with fixed line.")]
|
---|
203 | True
|
---|
204 |
|
---|
205 | - A signature separator:
|
---|
206 |
|
---|
207 | >>> result = decoder.decode(CRLF.join((
|
---|
208 | ... "A paragraph with flow space. ",
|
---|
209 | ... "-- ")))
|
---|
210 | >>> list(result) == [
|
---|
211 | ... ({'quotedepth': 0, 'type': PARAGRAPH},
|
---|
212 | ... u"A paragraph with flow space. "),
|
---|
213 | ... ({'quotedepth': 0, 'type': SIGNATURE_SEPARATOR}, u"-- ")]
|
---|
214 | True
|
---|
215 |
|
---|
216 | - The end of the message:
|
---|
217 |
|
---|
218 | >>> result = decoder.decode(CRLF.join((
|
---|
219 | ... "A paragraph with flow space. ",)))
|
---|
220 | >>> list(result) == [
|
---|
221 | ... ({'quotedepth': 0, 'type': PARAGRAPH},
|
---|
222 | ... u"A paragraph with flow space. ")]
|
---|
223 | True
|
---|
224 |
|
---|
225 |
|
---|
226 | Decoder options
|
---|
227 | ---------------
|
---|
228 |
|
---|
229 | The delete_space attribute of the FormatFlowedDecoder class can be used
|
---|
230 | to control wether or not the trailing space on flowed lines should be
|
---|
231 | retained; this is used to encode flowed text where spaces are rare:
|
---|
232 |
|
---|
233 | >>> decoder = FormatFlowedDecoder(delete_space=True)
|
---|
234 | >>> result = decoder.decode(CRLF.join((
|
---|
235 | ... "Contrived example with a word- ",
|
---|
236 | ... "break across the paragraph.")))
|
---|
237 | >>> list(result) == [
|
---|
238 | ... ({'quotedepth': 0, 'type': PARAGRAPH},
|
---|
239 | ... u'Contrived example with a word-break across the '
|
---|
240 | ... u'paragraph.')]
|
---|
241 | True
|
---|
242 |
|
---|
243 | Note that the characterset determines what how to interpret a space
|
---|
244 | and a quote marker. The cp037 characterset does not encode these
|
---|
245 | characters the same way, for example:
|
---|
246 |
|
---|
247 | >>> decoder = FormatFlowedDecoder(character_set='cp037')
|
---|
248 | >>> result = decoder.decode(CRLF.join((
|
---|
249 | ... "n@\\xe3\\x88\\x89\\xa2@\\x89\\xa2@\\x81@\\x98\\xa4\\x96\\xa3"
|
---|
250 | ... "\\x85\\x84@\\x97\\x81\\x99\\x81\\x87\\x99\\x81\\x97\\x88@",
|
---|
251 | ... "n@\\x85\\x95\\x83\\x96\\x84\\x85\\x84@\\x89\\x95@\\x83\\x97"
|
---|
252 | ... "\\xf0\\xf3\\xf7K")))
|
---|
253 | >>> list(result) == [
|
---|
254 | ... ({'quotedepth': 1, 'type': PARAGRAPH},
|
---|
255 | ... u'This is a quoted paragraph encoded in cp037.')]
|
---|
256 | True
|
---|
257 |
|
---|
258 | """
|
---|
259 | para = u''
|
---|
260 | pinfo = {'type': PARAGRAPH}
|
---|
261 | for line in flowed.split('\r\n'):
|
---|
262 | line = line.decode(self.character_set)
|
---|
263 | quotedepth, line = self._stripquotes(line)
|
---|
264 | line = self._stripstuffing(line)
|
---|
265 | if line == '-- ':
|
---|
266 | # signature separator
|
---|
267 | if para:
|
---|
268 | # exception case: flowed line followed by sig-sep
|
---|
269 | yield (pinfo, para)
|
---|
270 | pinfo = {'type': PARAGRAPH}
|
---|
271 | para = u''
|
---|
272 | yield ({'type': SIGNATURE_SEPARATOR,
|
---|
273 | 'quotedepth': quotedepth}, line)
|
---|
274 | continue
|
---|
275 | if line.endswith(u' '):
|
---|
276 | # flowed line; collect into a paragraph
|
---|
277 | if quotedepth != pinfo.get('quotedepth', quotedepth):
|
---|
278 | # exception case: flowed line followed by quotedepth change
|
---|
279 | yield (pinfo, para)
|
---|
280 | pinfo = {'type': PARAGRAPH}
|
---|
281 | para = u''
|
---|
282 | para += self._stripflow(line)
|
---|
283 | pinfo['quotedepth'] = quotedepth
|
---|
284 | continue
|
---|
285 | # fixed line
|
---|
286 | if para:
|
---|
287 | # completed paragraph
|
---|
288 | if quotedepth != pinfo.get('quotedepth', quotedepth):
|
---|
289 | # exception case: flowed line followed by quotedepth change
|
---|
290 | yield (pinfo, para)
|
---|
291 | pinfo = {'type': PARAGRAPH}
|
---|
292 | para = u''
|
---|
293 | else:
|
---|
294 | yield (pinfo, para + line)
|
---|
295 | pinfo = {'type': PARAGRAPH}
|
---|
296 | para = u''
|
---|
297 | continue
|
---|
298 | yield ({'type': FIXED, 'quotedepth': quotedepth}, line)
|
---|
299 |
|
---|
300 | if para:
|
---|
301 | # exception case: last line was a flowed line
|
---|
302 | yield (pinfo, para)
|
---|
303 |
|
---|
304 |
|
---|
305 | class FormatFlowedEncoder:
|
---|
306 | """Object to generate format=flowed text
|
---|
307 |
|
---|
308 | The following attributes influence the flowed formatting of text:
|
---|
309 | extra_space (default: False)
|
---|
310 | Use an extra space to create flowed lines; this requires that the
|
---|
311 | DelSpace flag will be set true on the Content-Type mime header. Use
|
---|
312 | this flag on texts that have little or no spaces to break on.
|
---|
313 | character_set (default: us-ascii)
|
---|
314 | Encode the output to this character set.
|
---|
315 | spacestuff_quoted (default: True)
|
---|
316 | Always spacestuff quoted chunks, i.e. place a space between the quote
|
---|
317 | markers and the text.
|
---|
318 | width (default: 78)
|
---|
319 | The maximum line width generated for flowed paragraphs; fixed lines
|
---|
320 | can still exceed this width. This value does not include the CRLF
|
---|
321 | line endings.
|
---|
322 |
|
---|
323 | """
|
---|
324 | def __init__(self, extra_space=False, character_set='us-ascii',
|
---|
325 | spacestuff_quoted=True, width=78):
|
---|
326 | self.extra_space = extra_space
|
---|
327 | self.character_set = character_set
|
---|
328 | self.spacestuff_quoted = spacestuff_quoted
|
---|
329 | self.width = width
|
---|
330 |
|
---|
331 | def _spacestuff(self, line, force=False):
|
---|
332 | """Prepend a space to lines starting with ' ', '>' or 'From'
|
---|
333 |
|
---|
334 | Returns the altered line. Set 'force' to True to skip the tests and
|
---|
335 | always prepend the space regardless:
|
---|
336 |
|
---|
337 | >>> encoder = FormatFlowedEncoder()
|
---|
338 | >>> encoder._spacestuff(u' leading space needs to be preserved')
|
---|
339 | u' leading space needs to be preserved'
|
---|
340 | >>> encoder._spacestuff(u'> can be confused for a quotemark')
|
---|
341 | u' > can be confused for a quotemark'
|
---|
342 | >>> encoder._spacestuff(u'From is often escaped by MTAs')
|
---|
343 | u' From is often escaped by MTAs'
|
---|
344 | >>> encoder._spacestuff(u'Padding is considered harmless')
|
---|
345 | u'Padding is considered harmless'
|
---|
346 | >>> encoder._spacestuff(u'So forcing it is fine', True)
|
---|
347 | u' So forcing it is fine'
|
---|
348 |
|
---|
349 | Note that empty lines can never be spacestuffed:
|
---|
350 |
|
---|
351 | >>> encoder._spacestuff(u'')
|
---|
352 | u''
|
---|
353 |
|
---|
354 | """
|
---|
355 | if not line:
|
---|
356 | return line
|
---|
357 | # Although the RFC doesn't say so explicitly, in practice 'From' only
|
---|
358 | # needs escaping when (1) not quoted and (2) actually encoded as
|
---|
359 | # 'From' (so independent of the unicode sequence u'From').
|
---|
360 | # For simplicity's sake, we spacestuff it any time a line starts with
|
---|
361 | # it before adding quotemarks and encoding the line.
|
---|
362 | if force or line[0] in (' ', '>') or line.startswith('From'):
|
---|
363 | return u' ' + line
|
---|
364 | return line
|
---|
365 |
|
---|
366 | # -- Public API ----------------------------------------------------
|
---|
367 |
|
---|
368 | def encode(self, chunks):
|
---|
369 | """Encode chunks of text to format=flowed
|
---|
370 |
|
---|
371 | chunks
|
---|
372 | An iterable sequence of (information, text) tuples, where information
|
---|
373 | is a dictionary with 'type' and 'quotedepth' keys. The 'type' value
|
---|
374 | is one of PARAGRAPH, FIXED or SIGNATURE-SEPARATOR, and the
|
---|
375 | 'quotedepth' value a positive integer indicating the quoting depth.
|
---|
376 | text should be the unicode text to be encoded.
|
---|
377 |
|
---|
378 | Example
|
---|
379 | -------
|
---|
380 |
|
---|
381 | To illustrate, an example:
|
---|
382 |
|
---|
383 | >>> chunks = (
|
---|
384 | ... ({'quotedepth': 2, 'type': PARAGRAPH},
|
---|
385 | ... u"`Take some more tea,' the March Hare said to Alice, "
|
---|
386 | ... u"very earnestly."),
|
---|
387 | ... ({'quotedepth': 1, 'type': FIXED}, u""),
|
---|
388 | ... ({'quotedepth': 1, 'type': PARAGRAPH},
|
---|
389 | ... u"`I've had nothing yet,' Alice replied in an offended "
|
---|
390 | ... u"tone, `so I can't take more.'"),
|
---|
391 | ... ({'quotedepth': 0, 'type': FIXED}, u""),
|
---|
392 | ... ({'quotedepth': 0, 'type': PARAGRAPH},
|
---|
393 | ... u"`You mean you can't take less,' said the Hatter: `it's "
|
---|
394 | ... u"very easy to take more than nothing.'"),
|
---|
395 | ... ({'quotedepth': 0, 'type': FIXED}, u""),
|
---|
396 | ... ({'quotedepth': 0, 'type': SIGNATURE_SEPARATOR}, u"-- "),
|
---|
397 | ... ({'quotedepth': 0, 'type': PARAGRAPH}, u"Carol Lewis"),
|
---|
398 | ... )
|
---|
399 | >>> result = FormatFlowedEncoder(width=45).encode(chunks)
|
---|
400 | >>> result.split('\\r\\n') == [
|
---|
401 | ... ">> `Take some more tea,' the March Hare said ",
|
---|
402 | ... ">> to Alice, very earnestly.",
|
---|
403 | ... ">",
|
---|
404 | ... "> `I've had nothing yet,' Alice replied in ",
|
---|
405 | ... "> an offended tone, `so I can't take more.'",
|
---|
406 | ... "",
|
---|
407 | ... "`You mean you can't take less,' said the ",
|
---|
408 | ... "Hatter: `it's very easy to take more than ",
|
---|
409 | ... "nothing.'",
|
---|
410 | ... "",
|
---|
411 | ... "-- ",
|
---|
412 | ... "Carol Lewis",
|
---|
413 | ... ""]
|
---|
414 | True
|
---|
415 |
|
---|
416 | """
|
---|
417 | encoded = []
|
---|
418 | for info, text in chunks:
|
---|
419 | encoded.append(self.encodeChunk(text, **info))
|
---|
420 | return ''.join(encoded)
|
---|
421 |
|
---|
422 | def encodeChunk(self, chunk, type=PARAGRAPH, quotedepth=0):
|
---|
423 | """Encode a chunk of text to format=flowed
|
---|
424 |
|
---|
425 | The chunk is encoded to format=flowed text, controlled by the
|
---|
426 | following arguments.
|
---|
427 | chunk
|
---|
428 | The unicode text to be encoded. Newlines are considered to be
|
---|
429 | whitespace and will be converted to spaces.
|
---|
430 | type (default: PARAGRAPH)
|
---|
431 | Chunk type; one of PARAGRAPH, FIXED or SIGNATURE_SEPARATOR. When
|
---|
432 | called with type SIGNATURE_SEPARATOR the chunk is ignored and '-- '
|
---|
433 | is written out.
|
---|
434 | quotedepth (default: 0)
|
---|
435 | The quote depth of the chunk.
|
---|
436 |
|
---|
437 |
|
---|
438 | Examples
|
---|
439 | --------
|
---|
440 |
|
---|
441 | The encoder has to deal with three types of text chunks. To illustrate,
|
---|
442 | we create a encoder instance geared:
|
---|
443 |
|
---|
444 | >>> encoder = FormatFlowedEncoder(width=45)
|
---|
445 |
|
---|
446 | We can then use this encoder to encode some examples of these different
|
---|
447 | types:
|
---|
448 |
|
---|
449 | - fixed lines:
|
---|
450 |
|
---|
451 | >>> encoder.encodeChunk(u'A fixed line remains unaltered', FIXED)
|
---|
452 | 'A fixed line remains unaltered\\r\\n'
|
---|
453 | >>> encoder.encodeChunk(u'Although quoting is prepended', FIXED, 2)
|
---|
454 | '>> Although quoting is prepended\\r\\n'
|
---|
455 | >>> encoder.encodeChunk(u'Trailing spaces are removed ', FIXED)
|
---|
456 | 'Trailing spaces are removed\\r\\n'
|
---|
457 | >>> encoder.encodeChunk(u'> and special first chars are fluffed',
|
---|
458 | ... FIXED)
|
---|
459 | ' > and special first chars are fluffed\\r\\n'
|
---|
460 |
|
---|
461 | - a paragraph (the default type):
|
---|
462 |
|
---|
463 | >>> result = encoder.encodeChunk(
|
---|
464 | ... u"`Take some more tea,' the March Hare said to Alice, "
|
---|
465 | ... u"very earnestly.")
|
---|
466 | >>> result == ("`Take some more tea,' the March Hare said \\r\\n"
|
---|
467 | ... "to Alice, very earnestly.\\r\\n")
|
---|
468 | True
|
---|
469 | >>> result = encoder.encodeChunk(
|
---|
470 | ... u"`I've had nothing yet,' Alice replied in an offended "
|
---|
471 | ... u"tone, `so I can't take more.'", PARAGRAPH, 1)
|
---|
472 | >>> result == ("> `I've had nothing yet,' Alice replied in \\r\\n"
|
---|
473 | ... "> an offended tone, `so I can't take more.'\\r\\n")
|
---|
474 | True
|
---|
475 | >>> result = encoder.encodeChunk(
|
---|
476 | ... u'The wrapping deals quite well with > eratic '
|
---|
477 | ... u'spacing and space fluffs characters where needed.')
|
---|
478 | >>> result == ("The wrapping deals quite well with \\r\\n"
|
---|
479 | ... " > eratic spacing and space fluffs \\r\\n"
|
---|
480 | ... "characters where needed.\\r\\n")
|
---|
481 | True
|
---|
482 |
|
---|
483 | - signature separators:
|
---|
484 |
|
---|
485 | >>> encoder.encodeChunk(u'-- ', SIGNATURE_SEPARATOR)
|
---|
486 | '-- \\r\\n'
|
---|
487 | >>> encoder.encodeChunk(u'-- ', SIGNATURE_SEPARATOR, 3)
|
---|
488 | '>>> -- \\r\\n'
|
---|
489 |
|
---|
490 | Note that the actual chunk value is ignored for this type:
|
---|
491 |
|
---|
492 | >>> encoder.encodeChunk(u'foobar', SIGNATURE_SEPARATOR)
|
---|
493 | '-- \\r\\n'
|
---|
494 |
|
---|
495 |
|
---|
496 | Encoder options
|
---|
497 | ---------------
|
---|
498 |
|
---|
499 | The encoding can be influenced by several instance attributes; the
|
---|
500 | width attribute was used for the paragraph demonstrations. Others
|
---|
501 | include 'extra_space', 'character_set' and 'spacestuff_quoted':
|
---|
502 |
|
---|
503 | - extra_space generates extra spaces on flowed lines so flowed lines
|
---|
504 | can be broken on something other than whitespace:
|
---|
505 |
|
---|
506 | >>> encoder = FormatFlowedEncoder(extra_space=True, width=45)
|
---|
507 | >>> result = encoder.encodeChunk(
|
---|
508 | ... u'This is useful for texts with many word-breaks or few '
|
---|
509 | ... u'spaces')
|
---|
510 | >>> result == ("This is useful for texts with many word- \\r\\n"
|
---|
511 | ... "breaks or few spaces\\r\\n")
|
---|
512 | True
|
---|
513 |
|
---|
514 | - character_set controls the output encoding:
|
---|
515 |
|
---|
516 | >>> encoder = FormatFlowedEncoder(character_set='cp037')
|
---|
517 | >>> result = encoder.encodeChunk(u'Can you read me now?',
|
---|
518 | ... quotedepth=1)
|
---|
519 | >>> result == ('n@\\xc3\\x81\\x95@\\xa8\\x96\\xa4@\\x99\\x85\\x81'
|
---|
520 | ... '\\x84@\\x94\\x85@\\x95\x96\\xa6o\\r\\n')
|
---|
521 | True
|
---|
522 |
|
---|
523 | - spacestuff_quoted causes quoted lines to be spacestuffed by default;
|
---|
524 | this makes for slightly more readable quoted text output. It is on
|
---|
525 | by default, but can be switched off:
|
---|
526 |
|
---|
527 | >>> encoder = FormatFlowedEncoder(spacestuff_quoted=False)
|
---|
528 | >>> encoder.encodeChunk(u'Look Ma! No space!', quotedepth=1)
|
---|
529 | '>Look Ma! No space!\\r\\n'
|
---|
530 |
|
---|
531 |
|
---|
532 | RFC 2822 compliance
|
---|
533 | -------------------
|
---|
534 |
|
---|
535 | Note that RFC 2822 requires that generated lines never exceed the
|
---|
536 | hard limit of 998 characters without the CRLF at the end. The encoder
|
---|
537 | has to enforce this by chopping the lines up into pieces not exceeding
|
---|
538 | that length:
|
---|
539 |
|
---|
540 | >>> encoder = FormatFlowedEncoder()
|
---|
541 | >>> result = encoder.encodeChunk(u'-' * 1500, FIXED)
|
---|
542 | >>> result = result.split('\\r\\n')
|
---|
543 | >>> len(result)
|
---|
544 | 3
|
---|
545 | >>> len(result[0])
|
---|
546 | 998
|
---|
547 | >>> result == ['-' * 998, '-' * 502, '']
|
---|
548 | True
|
---|
549 |
|
---|
550 | """
|
---|
551 | # cleanup: replace newlines with spaces and remove trailing spaces
|
---|
552 | chunk = ' '.join(chunk.rstrip().splitlines())
|
---|
553 |
|
---|
554 | # Pre-encode quoting
|
---|
555 | quotemarker = u'>' * quotedepth
|
---|
556 | quotemarker = quotemarker.encode(self.character_set)
|
---|
557 | forcestuff = self.spacestuff_quoted and quotedepth > 0
|
---|
558 |
|
---|
559 | if type == SIGNATURE_SEPARATOR:
|
---|
560 | chunk = u'-- '
|
---|
561 |
|
---|
562 | if type == PARAGRAPH:
|
---|
563 | # Maximum width is reduced by stuffing and quotemarkers
|
---|
564 | width = self.width - len(quotemarker) - 2
|
---|
565 | if width <= 0:
|
---|
566 | raise ValueError('Not enough width for both quoting and text')
|
---|
567 | wrapper = _FlowedTextWrapper(width, self.extra_space)
|
---|
568 | chunk = wrapper.wrap(chunk)
|
---|
569 | else:
|
---|
570 | chunk = [chunk]
|
---|
571 |
|
---|
572 | lines = []
|
---|
573 | for line in chunk:
|
---|
574 | # add space to flowed lines (all but last); this is an extra space
|
---|
575 | # if the wrapping of paragraphs included spaces at the end of the
|
---|
576 | # lines.
|
---|
577 | if line != chunk[-1]:
|
---|
578 | line += ' '
|
---|
579 | line = self._spacestuff(line, forcestuff)
|
---|
580 | line = quotemarker + line.encode(self.character_set)
|
---|
581 |
|
---|
582 | # Enforce a hard limit of 998 characters per line (excluding CRLF)
|
---|
583 | # Unfortunately we can only enforce this *after* encoding,
|
---|
584 | # otherwise we could flow lines that are too long.
|
---|
585 | while len(line) > 998:
|
---|
586 | lines.append(line[:998])
|
---|
587 | line = line[998:]
|
---|
588 |
|
---|
589 | lines.append(line)
|
---|
590 |
|
---|
591 | lines.append('') # ensure last ending CRLF
|
---|
592 | return '\r\n'.join(lines)
|
---|
593 |
|
---|
594 |
|
---|
595 | # -- Convenience functions ---------------------------------------------
|
---|
596 |
|
---|
597 |
|
---|
598 | def decode(flowed, **kwargs):
|
---|
599 | """Convert format=flowed text
|
---|
600 |
|
---|
601 | See the FormatFlowedDecoder.decode docstring for more information. All
|
---|
602 | keyword arguments are passed to the FormatFlowedDecoder instance.
|
---|
603 |
|
---|
604 | """
|
---|
605 | decoder = FormatFlowedDecoder(**kwargs)
|
---|
606 | return decoder.decode(flowed)
|
---|
607 |
|
---|
608 | def encode(chunks, **kwargs):
|
---|
609 | """Convert chunks of text to format=flowed
|
---|
610 |
|
---|
611 | See the FormatFlowedEncoder.encode docstring for more information. All
|
---|
612 | keyword arguments are passed to the FormatFlowedEncoder instance.
|
---|
613 |
|
---|
614 | """
|
---|
615 | encoder = FormatFlowedEncoder(**kwargs)
|
---|
616 | return encoder.encode(chunks)
|
---|
617 |
|
---|
618 | def convertToWrapped(flowed, width=78, quote=u'>', wrap_fixed=True, **kwargs):
|
---|
619 | """Covert flowed text to encoded and wrapped text
|
---|
620 |
|
---|
621 | Create text suitable for a proportional font, fixed with, plain text
|
---|
622 | display. The argements are interpreted as follows:
|
---|
623 | flowed
|
---|
624 | The format=flowed formatted text to convert
|
---|
625 | width (default: 78)
|
---|
626 | The maximum line length at which to wrap paragraphs.
|
---|
627 | quote (default: u'>')
|
---|
628 | Character sequence to use to mark quote depths; it is multiplied with
|
---|
629 | the quotedepth to quote a line. If this sequence does not end in a
|
---|
630 | space a space is added between the quotemars and the line.
|
---|
631 | wrap_fixed (default: True)
|
---|
632 | If true, fixed text chunks are wrapped to the given width as well,
|
---|
633 | including hard word breaks if a word exceeds the line width
|
---|
634 |
|
---|
635 | The remaining arguments are used as arguments to FormatFlowedDecoder.
|
---|
636 |
|
---|
637 | Here is a simple example:
|
---|
638 |
|
---|
639 | >>> CRLF = '\\r\\n'
|
---|
640 | >>> result = convertToWrapped(CRLF.join((
|
---|
641 | ... ">> `Take some more tea,' the March Hare said to Alice, very ",
|
---|
642 | ... ">> earnestly.",
|
---|
643 | ... ">",
|
---|
644 | ... "> `I've had nothing yet,' Alice replied in an offended ",
|
---|
645 | ... "> tone, `so I can't take more.'",
|
---|
646 | ... "",
|
---|
647 | ... "`You mean you can't take less,' said the Hatter: `it's very ",
|
---|
648 | ... "easy to take more than nothing.'",
|
---|
649 | ... "",
|
---|
650 | ... "-- ",
|
---|
651 | ... "Lewis Caroll")), width=60)
|
---|
652 | >>> result.split('\\n') == [
|
---|
653 | ... ">> `Take some more tea,' the March Hare said to Alice, very",
|
---|
654 | ... ">> earnestly.",
|
---|
655 | ... "> ",
|
---|
656 | ... "> `I've had nothing yet,' Alice replied in an offended tone,",
|
---|
657 | ... "> `so I can't take more.'",
|
---|
658 | ... "",
|
---|
659 | ... "`You mean you can't take less,' said the Hatter: `it's very",
|
---|
660 | ... "easy to take more than nothing.'",
|
---|
661 | ... "",
|
---|
662 | ... "-- ",
|
---|
663 | ... "Lewis Caroll"]
|
---|
664 | True
|
---|
665 |
|
---|
666 | """
|
---|
667 | result = []
|
---|
668 | for info, chunk in decode(flowed, **kwargs):
|
---|
669 | type = info['type']
|
---|
670 | quotedepth = info['quotedepth']
|
---|
671 | quotemarker = quotedepth and quote * quotedepth or u''
|
---|
672 | if quotemarker and quote[-1] != u' ':
|
---|
673 | quotemarker += u' '
|
---|
674 | if type == FIXED and not wrap_fixed:
|
---|
675 | result.append(quotemarker + chunk)
|
---|
676 | elif not chunk or type == SIGNATURE_SEPARATOR:
|
---|
677 | result.append(quotemarker + chunk)
|
---|
678 | else:
|
---|
679 | result.extend(textwrap.wrap(chunk, width,
|
---|
680 | replace_whitespace=False,
|
---|
681 | initial_indent=quotemarker,
|
---|
682 | subsequent_indent=quotemarker))
|
---|
683 | return u'\n'.join(result)
|
---|
684 |
|
---|
685 | def convertToFlowed(text, quotechars=u'>|%', **kwargs):
|
---|
686 | """Convert plain text to format=flowed
|
---|
687 |
|
---|
688 | Attempt to interpret the plain text as paragraphs and fixed lines,
|
---|
689 | creating a format=flowed encoded text. The paragraph detection is fairly
|
---|
690 | simple and probably not suitable for real-world email.
|
---|
691 |
|
---|
692 | text
|
---|
693 | Unicode text to be converted. Paragraphs are detected based on
|
---|
694 | whitelines between them, making all lines with extra linespace at the
|
---|
695 | start fixed to preserve that whitespace.
|
---|
696 | quotechars (default: u'>|%')
|
---|
697 | A set of characters recognized as quote markers; used to detect quote
|
---|
698 | depth.
|
---|
699 |
|
---|
700 | Additional kwargs are passed on to FormatFlowedEncoder.
|
---|
701 |
|
---|
702 | """
|
---|
703 | encoder = FormatFlowedEncoder(**kwargs)
|
---|
704 | return encoder.encode(_parseFlowableChunks(text, quotechars))
|
---|
705 |
|
---|
706 |
|
---|
707 | # -- Private classes and methods ---------------------------------------
|
---|
708 |
|
---|
709 |
|
---|
710 | class _FlowedTextWrapper(textwrap.TextWrapper):
|
---|
711 | """Custom text wrapper for flowed text
|
---|
712 |
|
---|
713 | When not using extra spaces, only break on spaces; when we are using
|
---|
714 | extra spaces, don't swallow whitespace at the start and end of lines, but
|
---|
715 | do break long words (as they can be reconstructed with DelSpace on).
|
---|
716 |
|
---|
717 | """
|
---|
718 | def __init__(self, width=78, extra_space=False):
|
---|
719 | textwrap.TextWrapper.__init__(self, width,
|
---|
720 | break_long_words=extra_space)
|
---|
721 | self.extra_space = extra_space
|
---|
722 | if not extra_space:
|
---|
723 | self.wordsep_re = re.compile(r'(\s+)')
|
---|
724 |
|
---|
725 | def _handle_long_word(self, reversed_chunks, cur_line, cur_len, width):
|
---|
726 | # _handle_long_word taken from python 2.5 CVS speed optimisation
|
---|
727 | # Can be removed if this is used with python 2.5
|
---|
728 | space_left = max(width - cur_len, 1)
|
---|
729 | if self.break_long_words:
|
---|
730 | cur_line.append(reversed_chunks[-1][:space_left])
|
---|
731 | reversed_chunks[-1] = reversed_chunks[-1][space_left:]
|
---|
732 | elif not cur_line:
|
---|
733 | cur_line.append(reversed_chunks.pop())
|
---|
734 |
|
---|
735 | def _wrap(self, chunks):
|
---|
736 | # Simplified and customized version of textwrap.TextWrapper
|
---|
737 | # Based on textwrapper rev. 1.37 in python CVS, with speed optimisation
|
---|
738 | lines = []
|
---|
739 | chunks.reverse()
|
---|
740 | while chunks:
|
---|
741 | cur_line = []
|
---|
742 | cur_len = 0
|
---|
743 | width = self.width
|
---|
744 |
|
---|
745 | # Don't strip space at the start of a line when using extra_space
|
---|
746 | # because spaces are significant there.
|
---|
747 | if chunks[-1].strip() == '' and lines and not self.extra_space:
|
---|
748 | del chunks[-1]
|
---|
749 |
|
---|
750 | while chunks:
|
---|
751 | l = len(chunks[-1])
|
---|
752 | if cur_len + l <= width:
|
---|
753 | cur_line.append(chunks.pop())
|
---|
754 | cur_len += l
|
---|
755 | else:
|
---|
756 | break
|
---|
757 |
|
---|
758 | if chunks and len(chunks[-1]) > width:
|
---|
759 | self._handle_long_word(chunks, cur_line, cur_len, width)
|
---|
760 |
|
---|
761 | # Don't drop space at end of line if using extra_space for
|
---|
762 | # marking flowed lines because otherwise there is no space between
|
---|
763 | # this line and the next when decoding the flowed text
|
---|
764 | if cur_line and cur_line[-1].strip() == '' and not self.extra_space:
|
---|
765 | del cur_line[-1]
|
---|
766 |
|
---|
767 | if cur_line:
|
---|
768 | lines.append(''.join(cur_line))
|
---|
769 | return lines
|
---|
770 |
|
---|
771 |
|
---|
772 | def _parseFlowableChunks(text, quotechars='>%|'):
|
---|
773 | """Parse out encodeble chunks, determining chunk type
|
---|
774 |
|
---|
775 | First step is to remove and count quoting marks, determining the quotedepth
|
---|
776 | of the text. Then the type of the lines is detected.
|
---|
777 |
|
---|
778 | Paragraphs are determined by terminating lines; terminating lines are
|
---|
779 | changes in quoting (depth or quoting used, signatures or fixed lines (see
|
---|
780 | below)
|
---|
781 |
|
---|
782 | Fixed lines are used for lines with nothing but whitespace and for lines
|
---|
783 | with whitespace prepended (indented lines).
|
---|
784 |
|
---|
785 | Any line with only two dashes at the start and whitespace is a signature
|
---|
786 | seperator.
|
---|
787 |
|
---|
788 | Example code:
|
---|
789 |
|
---|
790 | >>> result = _parseFlowableChunks(u'\\n'.join((
|
---|
791 | ... u'Normal text, as long as they are not delimited by empty ',
|
---|
792 | ... u'lines will be considered paragraphs and will be parsed as ',
|
---|
793 | ... u'such.',
|
---|
794 | ... u'',
|
---|
795 | ... u'> > Quoting will be detected as well, and as long as it is ',
|
---|
796 | ... u'> > consistent text will be collected into one paragraph.',
|
---|
797 | ... u'> Changes in depth trigger a new paragraph.',
|
---|
798 | ... u'> Leading whitespace makes for fixed lines.',
|
---|
799 | ... u'Signature separators are dealt with accordingly:',
|
---|
800 | ... u'-- '
|
---|
801 | ... )))
|
---|
802 | >>> result.next() == ({'type': PARAGRAPH, 'quotedepth': 0},
|
---|
803 | ... u'Normal text, as long as they are not delimited by empty '
|
---|
804 | ... u'lines will be considered paragraphs and will be parsed as '
|
---|
805 | ... u'such.')
|
---|
806 | True
|
---|
807 | >>> result.next() == ({'type': FIXED, 'quotedepth': 0}, u'')
|
---|
808 | True
|
---|
809 | >>> result.next() == ({'type': PARAGRAPH, 'quotedepth': 2},
|
---|
810 | ... u'Quoting will be detected as well, and as long as it is '
|
---|
811 | ... u'consistent text will be collected into one paragraph.')
|
---|
812 | True
|
---|
813 | >>> result.next() == ({'type': PARAGRAPH, 'quotedepth': 1},
|
---|
814 | ... u'Changes in depth trigger a new paragraph.')
|
---|
815 | True
|
---|
816 | >>> result.next() == ({'type': FIXED, 'quotedepth': 1},
|
---|
817 | ... u' Leading whitespace makes for fixed lines.')
|
---|
818 | True
|
---|
819 | >>> result.next() == ({'type': PARAGRAPH, 'quotedepth': 0},
|
---|
820 | ... u'Signature separators are dealt with accordingly:')
|
---|
821 | True
|
---|
822 | >>> result.next() == ({'type': SIGNATURE_SEPARATOR, 'quotedepth': 0},
|
---|
823 | ... u'-- ')
|
---|
824 | True
|
---|
825 | >>> result.next()
|
---|
826 | Traceback (most recent call last):
|
---|
827 | ...
|
---|
828 | StopIteration
|
---|
829 |
|
---|
830 | """
|
---|
831 | # Match quotemarks with limited whitespace around them
|
---|
832 | qm_match = re.compile('(^\s{0,2}([%s]\s?)+)' % quotechars).match
|
---|
833 | # Find all quotemarks
|
---|
834 | qm_findall = re.compile('[%s]' % quotechars).findall
|
---|
835 |
|
---|
836 | quotedepth = 0
|
---|
837 | quotemarks = ''
|
---|
838 | para = u''
|
---|
839 |
|
---|
840 | for line in text.splitlines():
|
---|
841 | has_quotes = qm_match(line)
|
---|
842 | same_quotes = quotemarks and line.startswith(quotemarks)
|
---|
843 | if (has_quotes and not same_quotes) or (not has_quotes and quotedepth):
|
---|
844 | # Change in quoting
|
---|
845 | if para:
|
---|
846 | yield {'type': PARAGRAPH, 'quotedepth': quotedepth}, para
|
---|
847 | para = u''
|
---|
848 |
|
---|
849 | quotemarks = has_quotes and has_quotes.group(0) or u''
|
---|
850 | quotedepth = len(qm_findall(quotemarks))
|
---|
851 |
|
---|
852 | line = line[len(quotemarks):]
|
---|
853 |
|
---|
854 | if line.rstrip() == u'--':
|
---|
855 | # signature separator
|
---|
856 | if para:
|
---|
857 | yield {'type': PARAGRAPH, 'quotedepth': quotedepth}, para
|
---|
858 | para = u''
|
---|
859 |
|
---|
860 | yield {'type': SIGNATURE_SEPARATOR, 'quotedepth': quotedepth}, line
|
---|
861 | continue
|
---|
862 |
|
---|
863 | if line.strip() == u'' or line.lstrip() != line:
|
---|
864 | # Fixed line
|
---|
865 | if para:
|
---|
866 | yield {'type': PARAGRAPH, 'quotedepth': quotedepth}, para
|
---|
867 | para = u''
|
---|
868 |
|
---|
869 | yield {'type': FIXED, 'quotedepth': quotedepth}, line
|
---|
870 | continue
|
---|
871 |
|
---|
872 | # Paragraph line; store and loop to next line
|
---|
873 | para += line
|
---|
874 |
|
---|
875 | if para:
|
---|
876 | yield {'type': PARAGRAPH, 'quotedepth': quotedepth}, para
|
---|
877 |
|
---|
878 | def _test(verbose=False):
|
---|
879 | import doctest
|
---|
880 | return doctest.testmod(verbose=verbose)
|
---|
881 |
|
---|
882 | if __name__ == '__main__':
|
---|
883 | import sys
|
---|
884 | _test('-v' in sys.argv)
|
---|