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 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 | import sys |
---|
25 | import os |
---|
26 | import mmap |
---|
27 | |
---|
28 | import pdlparser |
---|
29 | |
---|
30 | class Parser(pdlparser.PDLParser) : |
---|
31 | """A parser for PCL3, PCL4, PCL5 documents.""" |
---|
32 | mediasizes = { # ESC&l####A |
---|
33 | 0 : "Default", |
---|
34 | 1 : "Executive", |
---|
35 | 2 : "Letter", |
---|
36 | 3 : "Legal", |
---|
37 | 6 : "Ledger", |
---|
38 | 25 : "A5", |
---|
39 | 26 : "A4", |
---|
40 | 27 : "A3", |
---|
41 | 45 : "JB5", |
---|
42 | 46 : "JB4", |
---|
43 | 71 : "HagakiPostcard", |
---|
44 | 72 : "OufukuHagakiPostcard", |
---|
45 | 80 : "MonarchEnvelope", |
---|
46 | 81 : "COM10Envelope", |
---|
47 | 90 : "DLEnvelope", |
---|
48 | 91 : "C5Envelope", |
---|
49 | 100 : "B5Envelope", |
---|
50 | 101 : "Custom", |
---|
51 | } |
---|
52 | |
---|
53 | mediasources = { # ESC&l####H |
---|
54 | 0 : "Default", |
---|
55 | 1 : "Main", |
---|
56 | 2 : "Manual", |
---|
57 | 3 : "ManualEnvelope", |
---|
58 | 4 : "Alternate", |
---|
59 | 5 : "OptionalLarge", |
---|
60 | 6 : "EnvelopeFeeder", |
---|
61 | 7 : "Auto", |
---|
62 | 8 : "Tray1", |
---|
63 | } |
---|
64 | |
---|
65 | orientations = { # ESC&l####O |
---|
66 | 0 : "Portrait", |
---|
67 | 1 : "Landscape", |
---|
68 | 2 : "ReversePortrait", |
---|
69 | 3 : "ReverseLandscape", |
---|
70 | } |
---|
71 | |
---|
72 | mediatypes = { # ESC&l####M |
---|
73 | 0 : "Plain", |
---|
74 | 1 : "Bond", |
---|
75 | 2 : "Special", |
---|
76 | 3 : "Glossy", |
---|
77 | 4 : "Transparent", |
---|
78 | } |
---|
79 | |
---|
80 | def isValid(self) : |
---|
81 | """Returns 1 if data is PCL, else 0.""" |
---|
82 | if self.firstblock.startswith("\033E\033") or \ |
---|
83 | (self.firstblock.startswith("\033*rbC") and (not self.lastblock[-3:] == "\f\033@")) or \ |
---|
84 | self.firstblock.startswith("\033%8\033") or \ |
---|
85 | (self.firstblock.find("\033%-12345X") != -1) : |
---|
86 | if self.debug : |
---|
87 | sys.stderr.write("DEBUG: Input file is in the PCL3/4/5 format.\n") |
---|
88 | return 1 |
---|
89 | else : |
---|
90 | return 0 |
---|
91 | |
---|
92 | def setPageDict(self, pages, number, attribute, value) : |
---|
93 | """Initializes a page dictionnary.""" |
---|
94 | dict = pages.setdefault(number, { "copies" : 1, "mediasource" : "Main", "mediasize" : "Default", "mediatype" : "Plain", "orientation" : "Portrait"}) |
---|
95 | dict[attribute] = value |
---|
96 | |
---|
97 | def getJobSize(self) : |
---|
98 | """Count pages in a PCL5 document. |
---|
99 | |
---|
100 | Should also work for PCL3 and PCL4 documents. |
---|
101 | |
---|
102 | Algorithm from pclcount |
---|
103 | (c) 2003, by Eduardo Gielamo Oliveira & Rodolfo Broco Manin |
---|
104 | published under the terms of the GNU General Public Licence v2. |
---|
105 | |
---|
106 | Backported from C to Python by Jerome Alet, then enhanced |
---|
107 | with more PCL tags detected. I think all the necessary PCL tags |
---|
108 | are recognized to correctly handle PCL5 files wrt their number |
---|
109 | of pages. The documentation used for this was : |
---|
110 | |
---|
111 | HP PCL/PJL Reference Set |
---|
112 | PCL5 Printer Language Technical Quick Reference Guide |
---|
113 | http://h20000.www2.hp.com/bc/docs/support/SupportManual/bpl13205/bpl13205.pdf |
---|
114 | """ |
---|
115 | infileno = self.infile.fileno() |
---|
116 | minfile = mmap.mmap(infileno, os.fstat(infileno)[6], prot=mmap.PROT_READ, flags=mmap.MAP_SHARED) |
---|
117 | tagsends = { "&n" : "W", |
---|
118 | "&b" : "W", |
---|
119 | "*i" : "W", |
---|
120 | "*l" : "W", |
---|
121 | "*m" : "W", |
---|
122 | "*v" : "W", |
---|
123 | "*c" : "W", |
---|
124 | "(f" : "W", |
---|
125 | "(s" : "W", |
---|
126 | ")s" : "W", |
---|
127 | "&p" : "X", |
---|
128 | # "&l" : "XHAOM", # treated specially |
---|
129 | "&a" : "G", # TODO : 0 means next side, 1 front side, 2 back side |
---|
130 | "*g" : "W", |
---|
131 | "*r" : "sbABC", |
---|
132 | "*t" : "R", |
---|
133 | # "*b" : "VW", # treated specially because it occurs very often |
---|
134 | } |
---|
135 | pagecount = resets = ejects = backsides = startgfx = endgfx = 0 |
---|
136 | starb = ampl = ispcl3 = escstart = 0 |
---|
137 | mediasourcecount = mediasizecount = orientationcount = mediatypecount = 0 |
---|
138 | tag = None |
---|
139 | pages = {} |
---|
140 | pos = 0 |
---|
141 | try : |
---|
142 | while 1 : |
---|
143 | char = minfile[pos] ; pos += 1 |
---|
144 | if char == "\014" : |
---|
145 | pagecount += 1 |
---|
146 | elif char == "\033" : |
---|
147 | starb = ampl = 0 |
---|
148 | # |
---|
149 | # <ESC>*b###y#m###v###w... -> PCL3 raster graphics |
---|
150 | # <ESC>*b###W -> Start of a raster data row/block |
---|
151 | # <ESC>*b###V -> Start of a raster data plane |
---|
152 | # <ESC>*c###W -> Start of a user defined pattern |
---|
153 | # <ESC>*i###W -> Start of a viewing illuminant block |
---|
154 | # <ESC>*l###W -> Start of a color lookup table |
---|
155 | # <ESC>*m###W -> Start of a download dither matrix block |
---|
156 | # <ESC>*v###W -> Start of a configure image data block |
---|
157 | # <ESC>*r1A -> Start Gfx |
---|
158 | # <ESC>(s###W -> Start of a characters description block |
---|
159 | # <ESC>)s###W -> Start of a fonts description block |
---|
160 | # <ESC>(f###W -> Start of a symbol set block |
---|
161 | # <ESC>&b###W -> Start of configuration data block |
---|
162 | # <ESC>&l###X -> Number of copies for current page |
---|
163 | # <ESC>&n###W -> Starts an alphanumeric string ID block |
---|
164 | # <ESC>&p###X -> Start of a non printable characters block |
---|
165 | # <ESC>&a2G -> Back side when duplex mode as generated by rastertohp |
---|
166 | # <ESC>*g###W -> Needed for planes in PCL3 output |
---|
167 | # <ESC>&l###H (or only 0 ?) -> Eject if NumPlanes > 1, as generated by rastertohp. Also defines mediasource |
---|
168 | # <ESC>&l###A -> mediasize |
---|
169 | # <ESC>&l###O -> orientation |
---|
170 | # <ESC>&l###M -> mediatype |
---|
171 | # <ESC>*t###R -> gfx resolution |
---|
172 | # |
---|
173 | tagstart = minfile[pos] ; pos += 1 |
---|
174 | if tagstart in "E9=YZ" : # one byte PCL tag |
---|
175 | if tagstart == "E" : |
---|
176 | resets += 1 |
---|
177 | continue # skip to next tag |
---|
178 | tag = tagstart + minfile[pos] ; pos += 1 |
---|
179 | if tag == "*b" : |
---|
180 | starb = 1 |
---|
181 | tagend = "VW" |
---|
182 | elif tag == "&l" : |
---|
183 | ampl = 1 |
---|
184 | tagend = "XHAOM" |
---|
185 | else : |
---|
186 | try : |
---|
187 | tagend = tagsends[tag] |
---|
188 | except KeyError : |
---|
189 | continue # Unsupported PCL tag |
---|
190 | # Now read the numeric argument |
---|
191 | size = 0 |
---|
192 | while 1 : |
---|
193 | char = minfile[pos] ; pos += 1 |
---|
194 | if not char.isdigit() : |
---|
195 | break |
---|
196 | size = (size * 10) + int(char) |
---|
197 | if char in tagend : |
---|
198 | if tag == "&l" : |
---|
199 | if char == "X" : |
---|
200 | self.setPageDict(pages, pagecount, "copies", size) |
---|
201 | elif char == "H" : |
---|
202 | self.setPageDict(pages, pagecount, "mediasource", self.mediasources.get(size, str(size))) |
---|
203 | mediasourcecount += 1 |
---|
204 | ejects += 1 |
---|
205 | elif char == "A" : |
---|
206 | self.setPageDict(pages, pagecount, "mediasize", self.mediasizes.get(size, str(size))) |
---|
207 | mediasizecount += 1 |
---|
208 | elif char == "O" : |
---|
209 | self.setPageDict(pages, pagecount, "orientation", self.orientations.get(size, str(size))) |
---|
210 | orientationcount += 1 |
---|
211 | elif char == "M" : |
---|
212 | self.setPageDict(pages, pagecount, "mediatype", self.mediatypes.get(size, str(size))) |
---|
213 | mediatypecount += 1 |
---|
214 | elif tag == "*r" : |
---|
215 | # Special tests for PCL3 |
---|
216 | if (char == "s") and size : |
---|
217 | while 1 : |
---|
218 | char = minfile[pos] ; pos += 1 |
---|
219 | if char == "A" : |
---|
220 | break |
---|
221 | elif (char == "b") and (minfile[pos] == "C") and not size : |
---|
222 | ispcl3 = 1 # Certainely a PCL3 file |
---|
223 | startgfx += (char == "A") and (minfile[pos - 2] in ("0", "1", "2", "3")) # Start Gfx |
---|
224 | endgfx += (not size) and (char in ("C", "B")) # End Gfx |
---|
225 | elif tag == "*t" : |
---|
226 | escstart += 1 |
---|
227 | elif (tag == "&a") and (size == 2) : |
---|
228 | backsides += 1 # Back side in duplex mode |
---|
229 | else : |
---|
230 | # we just ignore the block. |
---|
231 | if tag == "&n" : |
---|
232 | # we have to take care of the operation id byte |
---|
233 | # which is before the string itself |
---|
234 | size += 1 |
---|
235 | pos += size |
---|
236 | else : |
---|
237 | if starb : |
---|
238 | # special handling of PCL3 in which |
---|
239 | # *b introduces combined ESCape sequences |
---|
240 | size = 0 |
---|
241 | while 1 : |
---|
242 | char = minfile[pos] ; pos += 1 |
---|
243 | if not char.isdigit() : |
---|
244 | break |
---|
245 | size = (size * 10) + int(char) |
---|
246 | if char in ("w", "v") : |
---|
247 | ispcl3 = 1 # certainely a PCL3 document |
---|
248 | pos += size - 1 |
---|
249 | elif char in ("y", "m") : |
---|
250 | ispcl3 = 1 # certainely a PCL3 document |
---|
251 | pos -= 1 # fix position : we were ahead |
---|
252 | elif ampl : |
---|
253 | # special handling of PCL3 in which |
---|
254 | # &l introduces combined ESCape sequences |
---|
255 | size = 0 |
---|
256 | while 1 : |
---|
257 | char = minfile[pos] ; pos += 1 |
---|
258 | if not char.isdigit() : |
---|
259 | break |
---|
260 | size = (size * 10) + int(char) |
---|
261 | if char in ("a", "o", "h", "m") : |
---|
262 | ispcl3 = 1 # certainely a PCL3 document |
---|
263 | pos -= 1 # fix position : we were ahead |
---|
264 | if char == "h" : |
---|
265 | self.setPageDict(pages, pagecount, "mediasource", self.mediasources.get(size, str(size))) |
---|
266 | mediasourcecount += 1 |
---|
267 | elif char == "a" : |
---|
268 | self.setPageDict(pages, pagecount, "mediasize", self.mediasizes.get(size, str(size))) |
---|
269 | mediasizecount += 1 |
---|
270 | elif char == "o" : |
---|
271 | self.setPageDict(pages, pagecount, "orientation", self.orientations.get(size, str(size))) |
---|
272 | orientationcount += 1 |
---|
273 | elif char == "m" : |
---|
274 | self.setPageDict(pages, pagecount, "mediatype", self.mediatypes.get(size, str(size))) |
---|
275 | mediatypecount += 1 |
---|
276 | except IndexError : # EOF ? |
---|
277 | minfile.close() # reached EOF |
---|
278 | |
---|
279 | # if pagecount is still 0, we will use the number |
---|
280 | # of resets instead of the number of form feed characters. |
---|
281 | # but the number of resets is always at least 2 with a valid |
---|
282 | # pcl file : one at the very start and one at the very end |
---|
283 | # of the job's data. So we substract 2 from the number of |
---|
284 | # resets. And since on our test data we needed to substract |
---|
285 | # 1 more, we finally substract 3, and will test several |
---|
286 | # PCL files with this. If resets < 2, then the file is |
---|
287 | # probably not a valid PCL file, so we use 0 |
---|
288 | |
---|
289 | if self.debug : |
---|
290 | sys.stderr.write("pagecount : %s\n" % pagecount) |
---|
291 | sys.stderr.write("resets : %s\n" % resets) |
---|
292 | sys.stderr.write("ejects : %s\n" % ejects) |
---|
293 | sys.stderr.write("backsides : %s\n" % backsides) |
---|
294 | sys.stderr.write("startgfx : %s\n" % startgfx) |
---|
295 | sys.stderr.write("endgfx : %s\n" % endgfx) |
---|
296 | sys.stderr.write("mediasourcecount : %s\n" % mediasourcecount) |
---|
297 | sys.stderr.write("mediasizecount : %s\n" % mediasizecount) |
---|
298 | sys.stderr.write("orientationcount : %s\n" % orientationcount) |
---|
299 | sys.stderr.write("mediatypecount : %s\n" % mediatypecount) |
---|
300 | sys.stderr.write("escstart : %s\n" % escstart) |
---|
301 | |
---|
302 | # if not pagecount : |
---|
303 | # pagecount = (pagecount or ((resets - 3) * (resets > 2))) |
---|
304 | # else : |
---|
305 | # # here we add counters for other ways new pages may have |
---|
306 | # # been printed and ejected by the printer |
---|
307 | # pagecount += ejects + backsides |
---|
308 | # |
---|
309 | # # now handle number of copies for each page (may differ). |
---|
310 | # # in duplex mode, number of copies may be sent only once. |
---|
311 | # for pnum in range(pagecount) : |
---|
312 | # # if no number of copies defined, take the preceding one else the one set before any page else 1. |
---|
313 | # page = pages.get(pnum, pages.get(pnum - 1, pages.get(0, { "copies" : 1 }))) |
---|
314 | # pagecount += (page["copies"] - 1) |
---|
315 | # |
---|
316 | # # in PCL3 files, there's one Start Gfx tag per page |
---|
317 | # if ispcl3 : |
---|
318 | # if endgfx == int(startgfx / 2) : # special case for cdj1600 |
---|
319 | # pagecount = endgfx |
---|
320 | # elif startgfx : |
---|
321 | # pagecount = startgfx |
---|
322 | # elif endgfx : |
---|
323 | # pagecount = endgfx |
---|
324 | |
---|
325 | |
---|
326 | if pagecount == mediasourcecount == escstart : |
---|
327 | pass # should be OK. |
---|
328 | elif (not startgfx) and (not endgfx) : |
---|
329 | pagecount = ejects or pagecount |
---|
330 | elif startgfx == endgfx : |
---|
331 | pagecount = startgfx |
---|
332 | elif startgfx == (endgfx - 1) : |
---|
333 | pagecount = startgfx |
---|
334 | elif (startgfx == 1) and not endgfx : |
---|
335 | pass |
---|
336 | else : |
---|
337 | pagecount = abs(startgfx - endgfx) |
---|
338 | |
---|
339 | if self.debug : |
---|
340 | for pnum in range(pagecount) : |
---|
341 | # if no number of copies defined, take the preceding one else the one set before any page else 1. |
---|
342 | page = pages.get(pnum, pages.get(pnum - 1, pages.get(0, { "copies" : 1, "mediasource" : "Main", "mediasize" : "Default", "mediatype" : "Plain", "orientation" : "Portrait"}))) |
---|
343 | sys.stderr.write("%s*%s*%s*%s*%s\n" % (page["copies"], page["mediatype"], page["mediasize"], page["orientation"], page["mediasource"])) |
---|
344 | |
---|
345 | return pagecount |
---|
346 | |
---|
347 | def test() : |
---|
348 | """Test function.""" |
---|
349 | if (len(sys.argv) < 2) or ((not sys.stdin.isatty()) and ("-" not in sys.argv[1:])) : |
---|
350 | sys.argv.append("-") |
---|
351 | totalsize = 0 |
---|
352 | for arg in sys.argv[1:] : |
---|
353 | if arg == "-" : |
---|
354 | infile = sys.stdin |
---|
355 | mustclose = 0 |
---|
356 | else : |
---|
357 | infile = open(arg, "rb") |
---|
358 | mustclose = 1 |
---|
359 | try : |
---|
360 | parser = Parser(infile, debug=1) |
---|
361 | totalsize += parser.getJobSize() |
---|
362 | except pdlparser.PDLParserError, msg : |
---|
363 | sys.stderr.write("ERROR: %s\n" % msg) |
---|
364 | sys.stderr.flush() |
---|
365 | if mustclose : |
---|
366 | infile.close() |
---|
367 | print "%s" % totalsize |
---|
368 | |
---|
369 | if __name__ == "__main__" : |
---|
370 | test() |
---|