1 | #! /usr/bin/env python |
---|
2 | # -*- coding: ISO-8859-15 -*- |
---|
3 | # |
---|
4 | # pkpgcounter : a generic Page Description Language parser |
---|
5 | # |
---|
6 | # (c) 2003, 2004, 2005, 2006 Jerome Alet <alet@librelogiciel.com> |
---|
7 | # This program is free software; you can redistribute it and/or modify |
---|
8 | # it under the terms of the GNU General Public License as published by |
---|
9 | # the Free Software Foundation; either version 2 of the License, or |
---|
10 | # (at your option) any later version. |
---|
11 | # |
---|
12 | # This program is distributed in the hope that it will be useful, |
---|
13 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
---|
14 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
---|
15 | # GNU General Public License for more details. |
---|
16 | # |
---|
17 | # You should have received a copy of the GNU General Public License |
---|
18 | # along with this program; if not, write to the Free Software |
---|
19 | # Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. |
---|
20 | # |
---|
21 | # $Id$ |
---|
22 | # |
---|
23 | |
---|
24 | """This modules implements a page counter for PCL3/4/5 documents.""" |
---|
25 | |
---|
26 | import sys |
---|
27 | import os |
---|
28 | import mmap |
---|
29 | from struct import unpack |
---|
30 | |
---|
31 | import pdlparser |
---|
32 | import pjl |
---|
33 | |
---|
34 | FORMFEED = chr(12) |
---|
35 | ESCAPE = chr(27) |
---|
36 | |
---|
37 | class Parser(pdlparser.PDLParser) : |
---|
38 | """A parser for PCL3, PCL4, PCL5 documents.""" |
---|
39 | totiffcommand = 'pcl6 -sDEVICE=pdfwrite -dPARANOIDSAFER -dNOPAUSE -dBATCH -dQUIET -sOutputFile=- - | gs -sDEVICE=tiff24nc -dPARANOIDSAFER -dNOPAUSE -dBATCH -dQUIET -r%(dpi)i -sOutputFile="%(fname)s" -' |
---|
40 | mediasizes = { # ESC&l####A |
---|
41 | 0 : "Default", |
---|
42 | 1 : "Executive", |
---|
43 | 2 : "Letter", |
---|
44 | 3 : "Legal", |
---|
45 | 6 : "Ledger", |
---|
46 | 25 : "A5", |
---|
47 | 26 : "A4", |
---|
48 | 27 : "A3", |
---|
49 | 45 : "JB5", |
---|
50 | 46 : "JB4", |
---|
51 | 71 : "HagakiPostcard", |
---|
52 | 72 : "OufukuHagakiPostcard", |
---|
53 | 80 : "MonarchEnvelope", |
---|
54 | 81 : "COM10Envelope", |
---|
55 | 90 : "DLEnvelope", |
---|
56 | 91 : "C5Envelope", |
---|
57 | 100 : "B5Envelope", |
---|
58 | 101 : "Custom", |
---|
59 | } |
---|
60 | |
---|
61 | mediasources = { # ESC&l####H |
---|
62 | 0 : "Default", |
---|
63 | 1 : "Main", |
---|
64 | 2 : "Manual", |
---|
65 | 3 : "ManualEnvelope", |
---|
66 | 4 : "Alternate", |
---|
67 | 5 : "OptionalLarge", |
---|
68 | 6 : "EnvelopeFeeder", |
---|
69 | 7 : "Auto", |
---|
70 | 8 : "Tray1", |
---|
71 | } |
---|
72 | |
---|
73 | orientations = { # ESC&l####O |
---|
74 | 0 : "Portrait", |
---|
75 | 1 : "Landscape", |
---|
76 | 2 : "ReversePortrait", |
---|
77 | 3 : "ReverseLandscape", |
---|
78 | } |
---|
79 | |
---|
80 | mediatypes = { # ESC&l####M |
---|
81 | 0 : "Plain", |
---|
82 | 1 : "Bond", |
---|
83 | 2 : "Special", |
---|
84 | 3 : "Glossy", |
---|
85 | 4 : "Transparent", |
---|
86 | } |
---|
87 | |
---|
88 | def isValid(self) : |
---|
89 | """Returns True if data is PCL3/4/5, else False.""" |
---|
90 | if self.firstblock.startswith("\033E\033") or \ |
---|
91 | (self.firstblock.startswith("\033*rbC") and (not self.lastblock[-3:] == "\f\033@")) or \ |
---|
92 | self.firstblock.startswith("\033%8\033") or \ |
---|
93 | (self.firstblock.find("\033%-12345X") != -1) or \ |
---|
94 | (self.firstblock.find("@PJL ENTER LANGUAGE=PCL\012\015\033") != -1) or \ |
---|
95 | (self.firstblock.startswith(chr(0xcd)+chr(0xca)) and self.firstblock.find("\033E\033")) : |
---|
96 | self.logdebug("DEBUG: Input file is in the PCL3/4/5 format.") |
---|
97 | return True |
---|
98 | else : |
---|
99 | return False |
---|
100 | |
---|
101 | def setPageDict(self, pages, number, attribute, value) : |
---|
102 | """Initializes a page dictionnary.""" |
---|
103 | dic = pages.setdefault(number, { "copies" : 1, "mediasource" : "Main", "mediasize" : "Default", "mediatype" : "Plain", "orientation" : "Portrait", "escaped" : "", "duplex": 0}) |
---|
104 | dic[attribute] = value |
---|
105 | |
---|
106 | def readByte(self) : |
---|
107 | """Reads a byte from the input stream.""" |
---|
108 | tag = ord(self.minfile[self.pos]) |
---|
109 | #self.logdebug("%08x ===> %02x" % (self.pos, tag)) |
---|
110 | self.pos += 1 |
---|
111 | return tag |
---|
112 | |
---|
113 | def endPage(self) : |
---|
114 | """Handle the FF marker.""" |
---|
115 | self.logdebug("FORMFEED %i" % self.pagecount) |
---|
116 | self.pagecount += 1 |
---|
117 | |
---|
118 | def escPercent(self) : |
---|
119 | """Handles the ESC% sequence.""" |
---|
120 | if self.minfile[self.pos : self.pos+7] == r"-12345X" : |
---|
121 | self.logdebug("Generic ESCAPE sequence at %08x" % self.pos) |
---|
122 | self.pos += 7 |
---|
123 | |
---|
124 | def handleTag(self, tagtable) : |
---|
125 | """Handles tags.""" |
---|
126 | tagtable[self.readByte()]() |
---|
127 | |
---|
128 | def escape(self) : |
---|
129 | """Handles the ESC character.""" |
---|
130 | #self.logdebug("ESCAPE") |
---|
131 | self.handleTag(self.esctags) |
---|
132 | |
---|
133 | def escAmp(self) : |
---|
134 | """Handles the ESC& sequence.""" |
---|
135 | #self.logdebug("AMP") |
---|
136 | self.handleTag(self.escamptags) |
---|
137 | |
---|
138 | def escStar(self) : |
---|
139 | """Handles the ESC* sequence.""" |
---|
140 | #self.logdebug("STAR") |
---|
141 | self.handleTag(self.escstartags) |
---|
142 | |
---|
143 | def escLeftPar(self) : |
---|
144 | """Handles the ESC( sequence.""" |
---|
145 | #self.logdebug("LEFTPAR") |
---|
146 | self.handleTag(self.escleftpartags) |
---|
147 | |
---|
148 | def escRightPar(self) : |
---|
149 | """Handles the ESC( sequence.""" |
---|
150 | #self.logdebug("RIGHTPAR") |
---|
151 | self.handleTag(self.escrightpartags) |
---|
152 | |
---|
153 | def escE(self) : |
---|
154 | """Handles the ESCE sequence.""" |
---|
155 | #self.logdebug("RESET") |
---|
156 | self.resets += 1 |
---|
157 | |
---|
158 | def escAmpl(self) : |
---|
159 | """Handles the ESC&l sequence.""" |
---|
160 | while 1 : |
---|
161 | (value, end) = self.getInteger() |
---|
162 | if value is None : |
---|
163 | return |
---|
164 | if end in ('h', 'H') : |
---|
165 | mediasource = self.mediasources.get(value, str(value)) |
---|
166 | self.mediasourcesvalues.append(mediasource) |
---|
167 | self.logdebug("MEDIASOURCE %s" % mediasource) |
---|
168 | elif end in ('a', 'A') : |
---|
169 | mediasize = self.mediasizes.get(value, str(value)) |
---|
170 | self.mediasizesvalues.append(mediasize) |
---|
171 | self.logdebug("MEDIASIZE %s" % mediasize) |
---|
172 | elif end in ('o', 'O') : |
---|
173 | orientation = self.orientations.get(value, str(value)) |
---|
174 | self.orientationsvalues.append(orientation) |
---|
175 | self.logdebug("ORIENTATION %s" % orientation) |
---|
176 | elif end in ('m', 'M') : |
---|
177 | mediatype = self.mediatypes.get(value, str(value)) |
---|
178 | self.mediatypesvalues.append(mediatype) |
---|
179 | self.logdebug("MEDIATYPE %s" % mediatype) |
---|
180 | elif end == 'X' : |
---|
181 | self.copies.append(value) |
---|
182 | self.logdebug("COPIES %i" % value) |
---|
183 | |
---|
184 | def escAmpa(self) : |
---|
185 | """Handles the ESC&a sequence.""" |
---|
186 | while 1 : |
---|
187 | (value, end) = self.getInteger() |
---|
188 | if value is None : |
---|
189 | return |
---|
190 | if end == 'G' : |
---|
191 | self.logdebug("BACKSIDES %i" % value) |
---|
192 | self.backsides.append(value) |
---|
193 | |
---|
194 | def escAmpb(self) : |
---|
195 | """Handles the ESC&b sequence.""" |
---|
196 | while 1 : |
---|
197 | (value, end) = self.getInteger() |
---|
198 | if value is None : |
---|
199 | return |
---|
200 | if end == 'W' : |
---|
201 | self.pos += value |
---|
202 | #self.logdebug("SKIPTO %08x" % self.pos) |
---|
203 | |
---|
204 | def escAmpn(self) : |
---|
205 | """Handles the ESC&n sequence.""" |
---|
206 | while 1 : |
---|
207 | (value, end) = self.getInteger() |
---|
208 | if value is None : |
---|
209 | return |
---|
210 | if end == 'W' : |
---|
211 | self.pos += value |
---|
212 | #self.logdebug("SKIPTO %08x" % self.pos) |
---|
213 | |
---|
214 | def escAmpp(self) : |
---|
215 | """Handles the ESC&p sequence.""" |
---|
216 | while 1 : |
---|
217 | (value, end) = self.getInteger() |
---|
218 | if value is None : |
---|
219 | return |
---|
220 | if end == 'X' : |
---|
221 | self.pos += value |
---|
222 | #self.logdebug("SKIPTO %08x" % self.pos) |
---|
223 | |
---|
224 | def escAmpu(self) : |
---|
225 | """Handles the ESC&u sequence.""" |
---|
226 | while 1 : |
---|
227 | (value, end) = self.getInteger() |
---|
228 | if value is None : |
---|
229 | return |
---|
230 | |
---|
231 | def escStarb(self) : |
---|
232 | """Handles the ESC*b sequence.""" |
---|
233 | while 1 : |
---|
234 | (value, end) = self.getInteger() |
---|
235 | if (end is None) and (value is None) : |
---|
236 | return |
---|
237 | if end in ('V', 'W', 'v', 'w') : |
---|
238 | self.pos += (value or 0) |
---|
239 | #self.logdebug("SKIPTO %08x" % self.pos) |
---|
240 | |
---|
241 | def escStarcgilmv(self) : |
---|
242 | """Handles the ESC*c, ESC*g, ESC*i, ESC*l, ESC*m, ESC*v sequences.""" |
---|
243 | while 1 : |
---|
244 | (value, end) = self.getInteger() |
---|
245 | if value is None : |
---|
246 | return |
---|
247 | if end == 'W' : |
---|
248 | self.pos += value |
---|
249 | #self.logdebug("SKIPTO %08x" % self.pos) |
---|
250 | |
---|
251 | def escStaro(self) : |
---|
252 | """Handles the ESC*o sequence.""" |
---|
253 | while 1 : |
---|
254 | (value, end) = self.getInteger() |
---|
255 | if value is None : |
---|
256 | return |
---|
257 | |
---|
258 | def escStarp(self) : |
---|
259 | """Handles the ESC*p sequence.""" |
---|
260 | while 1 : |
---|
261 | (value, end) = self.getInteger() |
---|
262 | if value is None : |
---|
263 | return |
---|
264 | |
---|
265 | def escStarr(self) : |
---|
266 | """Handles the ESC*r sequence.""" |
---|
267 | while 1 : |
---|
268 | (value, end) = self.getInteger() |
---|
269 | if value is None : |
---|
270 | if end is None : |
---|
271 | return |
---|
272 | elif end in ('B', 'C') : |
---|
273 | #self.logdebug("EndGFX") |
---|
274 | if not self.startgfx : |
---|
275 | self.logdebug("EndGFX found before StartGFX, ignored.") |
---|
276 | else : |
---|
277 | self.endgfx.append(1) |
---|
278 | if end == 'A' and (0 <= value <= 3) : |
---|
279 | #self.logdebug("StartGFX %i" % value) |
---|
280 | self.startgfx.append(value) |
---|
281 | |
---|
282 | def escStart(self) : |
---|
283 | """Handles the ESC*t sequence.""" |
---|
284 | while 1 : |
---|
285 | (value, end) = self.getInteger() |
---|
286 | if value is None : |
---|
287 | return |
---|
288 | |
---|
289 | def escRightorLeftParsf(self) : |
---|
290 | """Handles the ESC(s, ESC)s, ESC(f sequences.""" |
---|
291 | while 1 : |
---|
292 | (value, end) = self.getInteger() |
---|
293 | if value is None : |
---|
294 | return |
---|
295 | if end == 'W' : |
---|
296 | self.pos += value |
---|
297 | #self.logdebug("SKIPTO %08x" % self.pos) |
---|
298 | |
---|
299 | def getInteger(self) : |
---|
300 | """Returns an integer value and the end character.""" |
---|
301 | sign = 1 |
---|
302 | value = None |
---|
303 | while 1 : |
---|
304 | char = chr(self.readByte()) |
---|
305 | if char in (ESCAPE, FORMFEED) : |
---|
306 | self.pos -= 1 # Adjust position |
---|
307 | return (None, None) |
---|
308 | if char == '-' : |
---|
309 | sign = -1 |
---|
310 | elif not char.isdigit() : |
---|
311 | if value is not None : |
---|
312 | return (sign*value, char) |
---|
313 | else : |
---|
314 | return (value, char) |
---|
315 | else : |
---|
316 | value = ((value or 0) * 10) + int(char) |
---|
317 | |
---|
318 | def getJobSize(self) : |
---|
319 | """Count pages in a PCL5 document. |
---|
320 | |
---|
321 | Should also work for PCL3 and PCL4 documents. |
---|
322 | |
---|
323 | Algorithm from pclcount |
---|
324 | (c) 2003, by Eduardo Gielamo Oliveira & Rodolfo Broco Manin |
---|
325 | published under the terms of the GNU General Public Licence v2. |
---|
326 | |
---|
327 | Backported from C to Python by Jerome Alet, then enhanced |
---|
328 | with more PCL tags detected. I think all the necessary PCL tags |
---|
329 | are recognized to correctly handle PCL5 files wrt their number |
---|
330 | of pages. The documentation used for this was : |
---|
331 | |
---|
332 | HP PCL/PJL Reference Set |
---|
333 | PCL5 Printer Language Technical Quick Reference Guide |
---|
334 | http://h20000.www2.hp.com/bc/docs/support/SupportManual/bpl13205/bpl13205.pdf |
---|
335 | """ |
---|
336 | infileno = self.infile.fileno() |
---|
337 | self.minfile = minfile = mmap.mmap(infileno, os.fstat(infileno)[6], prot=mmap.PROT_READ, flags=mmap.MAP_SHARED) |
---|
338 | self.pagecount = 0 |
---|
339 | self.resets = 0 |
---|
340 | self.backsides = [] |
---|
341 | self.copies = [] |
---|
342 | self.mediasourcesvalues = [] |
---|
343 | self.mediasizesvalues = [] |
---|
344 | self.orientationsvalues = [] |
---|
345 | self.mediatypesvalues = [] |
---|
346 | self.startgfx = [] |
---|
347 | self.endgfx = [] |
---|
348 | |
---|
349 | tags = [ lambda : None] * 256 |
---|
350 | tags[ord(FORMFEED)] = self.endPage |
---|
351 | tags[ord(ESCAPE)] = self.escape |
---|
352 | |
---|
353 | self.esctags = [ lambda : None ] * 256 |
---|
354 | self.esctags[ord('%')] = self.escPercent |
---|
355 | self.esctags[ord('*')] = self.escStar |
---|
356 | self.esctags[ord('&')] = self.escAmp |
---|
357 | self.esctags[ord('(')] = self.escLeftPar |
---|
358 | self.esctags[ord(')')] = self.escRightPar |
---|
359 | self.esctags[ord('E')] = self.escE |
---|
360 | |
---|
361 | self.escamptags = [lambda : None ] * 256 |
---|
362 | self.escamptags[ord('a')] = self.escAmpa |
---|
363 | self.escamptags[ord('b')] = self.escAmpb |
---|
364 | self.escamptags[ord('l')] = self.escAmpl |
---|
365 | self.escamptags[ord('n')] = self.escAmpn |
---|
366 | self.escamptags[ord('p')] = self.escAmpp |
---|
367 | self.escamptags[ord('u')] = self.escAmpu |
---|
368 | |
---|
369 | self.escstartags = [ lambda : None ] * 256 |
---|
370 | self.escstartags[ord('b')] = self.escStarb |
---|
371 | self.escstartags[ord('o')] = self.escStaro |
---|
372 | self.escstartags[ord('p')] = self.escStarp |
---|
373 | self.escstartags[ord('r')] = self.escStarr |
---|
374 | self.escstartags[ord('t')] = self.escStart |
---|
375 | self.escstartags[ord('c')] = self.escStarcgilmv |
---|
376 | self.escstartags[ord('g')] = self.escStarcgilmv |
---|
377 | self.escstartags[ord('i')] = self.escStarcgilmv |
---|
378 | self.escstartags[ord('l')] = self.escStarcgilmv |
---|
379 | self.escstartags[ord('m')] = self.escStarcgilmv |
---|
380 | self.escstartags[ord('v')] = self.escStarcgilmv |
---|
381 | |
---|
382 | self.escleftpartags = [ lambda : None ] * 256 |
---|
383 | self.escleftpartags[ord('s')] = self.escRightorLeftParsf |
---|
384 | self.escleftpartags[ord('f')] = self.escRightorLeftParsf |
---|
385 | |
---|
386 | self.escrightpartags = [ lambda : None ] * 256 |
---|
387 | self.escrightpartags[ord('s')] = self.escRightorLeftParsf |
---|
388 | |
---|
389 | self.pos = 0 |
---|
390 | try : |
---|
391 | try : |
---|
392 | while 1 : |
---|
393 | tags[self.readByte()]() |
---|
394 | except IndexError : # EOF ? |
---|
395 | pass |
---|
396 | finally : |
---|
397 | self.minfile.close() |
---|
398 | |
---|
399 | self.logdebug("Pagecount : \t\t\t%i" % self.pagecount) |
---|
400 | self.logdebug("Resets : \t\t\t%i" % self.resets) |
---|
401 | self.logdebug("Copies : \t\t\t%s" % self.copies) |
---|
402 | self.logdebug("NbCopiesMarks : \t\t%i" % len(self.copies)) |
---|
403 | self.logdebug("MediaTypes : \t\t\t%s" % self.mediatypesvalues) |
---|
404 | self.logdebug("NbMediaTypes : \t\t\t%i" % len(self.mediatypesvalues)) |
---|
405 | self.logdebug("MediaSizes : \t\t\t%s" % self.mediasizesvalues) |
---|
406 | self.logdebug("NbMediaSizes : \t\t\t%i" % len(self.mediasizesvalues)) |
---|
407 | self.logdebug("MediaSources : \t\t\t%s" % self.mediasourcesvalues) |
---|
408 | nbmediasourcesdefault = len([m for m in self.mediasourcesvalues if m == 'Default']) |
---|
409 | self.logdebug("MediaSourcesDefault : \t\t%i" % nbmediasourcesdefault) |
---|
410 | self.logdebug("MediaSourcesNOTDefault : \t%i" % (len(self.mediasourcesvalues) - nbmediasourcesdefault)) |
---|
411 | self.logdebug("Orientations : \t\t\t%s" % self.orientationsvalues) |
---|
412 | self.logdebug("NbOrientations : \t\t\t%i" % len(self.orientationsvalues)) |
---|
413 | self.logdebug("StartGfx : \t\t\t%s" % len(self.startgfx)) |
---|
414 | self.logdebug("EndGfx : \t\t\t%s" % len(self.endgfx)) |
---|
415 | self.logdebug("BackSides : \t\t\t%s" % self.backsides) |
---|
416 | self.logdebug("NbBackSides : \t\t\t%i" % len(self.backsides)) |
---|
417 | |
---|
418 | return self.pagecount or nbmediasourcesdefault |
---|
419 | |
---|
420 | def test() : |
---|
421 | """Test function.""" |
---|
422 | if (len(sys.argv) < 2) or ((not sys.stdin.isatty()) and ("-" not in sys.argv[1:])) : |
---|
423 | sys.argv.append("-") |
---|
424 | totalsize = 0 |
---|
425 | for arg in sys.argv[1:] : |
---|
426 | if arg == "-" : |
---|
427 | infile = sys.stdin |
---|
428 | mustclose = 0 |
---|
429 | else : |
---|
430 | infile = open(arg, "rb") |
---|
431 | mustclose = 1 |
---|
432 | try : |
---|
433 | parser = Parser(infile, debug=1) |
---|
434 | totalsize += parser.getJobSize() |
---|
435 | except pdlparser.PDLParserError, msg : |
---|
436 | sys.stderr.write("ERROR: %s\n" % msg) |
---|
437 | sys.stderr.flush() |
---|
438 | if mustclose : |
---|
439 | infile.close() |
---|
440 | print "%s" % totalsize |
---|
441 | |
---|
442 | if __name__ == "__main__" : |
---|
443 | test() |
---|