1 # file lyxpreview2bitmap.py
2 # This file is part of LyX, the document processor.
3 # Licence details can be found in the file COPYING.
6 # with much advice from members of the preview-latex project:
7 # David Kastrup, dak@gnu.org and
8 # Jan-Åke Larsson, jalar@mai.liu.se.
10 # Full author contact details are available in file CREDITS
12 # This script takes a LaTeX file and generates a collection of
13 # png or ppm image files, one per previewed snippet.
16 # * python 2.4 or later (subprocess module);
17 # * A latex executable;
21 # * pngtoppm (if outputing ppm format images).
23 # preview.sty and dvipng are part of the preview-latex project
24 # http://preview-latex.sourceforge.net/
26 # preview.sty can alternatively be obtained from
27 # CTAN/support/preview-latex/
30 # lyxpreview2bitmap.py --bg=faf0e6 0lyxpreview.tex
32 # This script takes one obligatory argument:
34 # <input file>: The name of the .tex file to be converted.
36 # and these optional arguments:
38 # --png, --ppm: The desired output format. Either 'png' or 'ppm'.
39 # --dpi=<res>: A scale factor, used to ascertain the resolution of the
40 # generated image which is then passed to gs.
41 # --fg=<color>: The foreground color as a hexadecimal string, eg '000000'.
42 # --bg=<color>: The background color as a hexadecimal string, eg 'faf0e6'.
43 # --latex=<exe>: The converter for latex files. Default is latex.
44 # --bibtex=<exe>: The converter for bibtex files. Default is bibtex.
45 # --lilypond: Preprocess through lilypond-book. Default is false.
46 # --lilypond-book=<exe>:
47 # The converter for lytex files. Default is lilypond-book.
49 # -d, --debug Show the output from external commands.
50 # -h, --help Show an help screen and exit.
51 # -v, --verbose Show progress messages.
53 # Decomposing TEXFILE's name as DIR/BASE.tex, this script will,
54 # if executed successfully, leave in DIR:
55 # * a (possibly large) number of image files with names
57 # * a file BASE.metrics, containing info needed by LyX to position
58 # the images correctly on the screen.
60 # What does this script do?
61 # 1) Call latex/pdflatex/xelatex/whatever (CONVERTER parameter)
62 # 2) If the output is a PDF fallback to legacy
63 # 3) Otherwise check each page of the DVI (with dv2dt) looking for
64 # PostScript literals, not well supported by dvipng. Pages
65 # containing them are passed to the legacy method in a new LaTeX file.
66 # 4) Call dvipng on the pages without PS literals
67 # 5) Join metrics info coming from both methods (legacy and dvipng)
68 # and write them to file
70 # dvipng is fast but gives problem in several cases, like with
71 # PSTricks, TikZ and other packages using PostScript literals
72 # for all these cases the legacy route is taken (step 3).
73 # Moreover dvipng can't work with PDF files, so, if the CONVERTER
74 # paramter is pdflatex we have to fallback to legacy route (step 2).
77 import getopt, glob, os, re, shutil, sys, tempfile
79 import lyxpreview_tools
81 from legacy_lyxpreview2ppm import extract_resolution, legacy_conversion_step1
83 from lyxpreview_tools import bibtex_commands, check_latex_log, copyfileobj, \
84 error, filter_pages, find_exe, find_exe_or_terminate, \
85 join_metrics_and_rename, latex_commands, latex_file_re, make_texcolor, \
86 pdflatex_commands, progress, run_command, run_latex, run_tex, \
87 warning, write_metrics_info
91 Usage: %s <options> <input file>
94 --dpi=<res>: Resolution per inch (default: 128)
95 --png, --ppm: Select the output format (default: png)
96 --fg=<color>: Foreground color (default: black, ie '000000')
97 --bg=<color>: Background color (default: white, ie 'ffffff')
98 --latex=<exe>: Specify the executable for latex (default: latex)
99 --bibtex=<exe>: Specify the executable for bibtex (default: bibtex)
100 --lilypond: Preprocess through lilypond-book (default: false)
101 --lilypond-book=<exe>:
102 The executable for lilypond-book (default: lilypond-book)
104 -d, --debug: Show the output from external commands
105 -h, --help: Show this help screen and exit
106 -v, --verbose: Show progress messages
108 The colors are hexadecimal strings, eg 'faf0e6'."""
109 return msg % prog_name
111 # Returns a list of tuples containing page number and ascent fraction
112 # extracted from dvipng output.
113 # Use write_metrics_info to create the .metrics file with this info
114 def extract_metrics_info(dvipng_stdout):
115 # "\[[0-9]+" can match two kinds of numbers: page numbers from dvipng
116 # and glyph numbers from mktexpk. The glyph numbers always match
117 # "\[[0-9]+\]" while the page number never is followed by "\]". Thus:
118 page_re = re.compile(r"\[([0-9]+)[^]]");
119 metrics_re = re.compile("depth=(-?[0-9]+) height=(-?[0-9]+)")
126 match = page_re.search(dvipng_stdout, pos)
129 page = match.group(1)
131 match = metrics_re.search(dvipng_stdout, pos)
136 # Calculate the 'ascent fraction'.
137 descent = float(match.group(1))
138 ascent = float(match.group(2))
142 # This is an empty image, forbid its display
144 elif ascent >= 0 or descent >= 0:
145 if abs(ascent + descent) > 0.1:
146 frac = ascent / (ascent + descent)
152 results.append((int(page), frac))
153 pos = match.end() + 2
156 error("Failed to extract metrics info from dvipng")
161 def fix_latex_file(latex_file, pdf_output):
162 # python 2 does not allow to declare a string as raw byte so we double
163 # the backslashes and remove the r preffix
164 def_re = re.compile(b"(\\\\newcommandx|\\\\global\\\\long\\\\def)"
167 tmp = tempfile.TemporaryFile()
171 for line in open(latex_file, 'rb').readlines():
172 if not pdf_output and line.startswith(b"\\documentclass"):
174 line += b"\\PassOptionsToPackage{draft}{microtype}\n"
176 match = def_re.match(line)
178 macroname = match.group(2)
179 if macroname in macros:
180 definecmd = match.group(1)
181 if definecmd == b"\\newcommandx":
183 line = line.replace(definecmd, b"\\renewcommandx")
185 macros.append(macroname)
189 copyfileobj(tmp, open(latex_file,"wb"), 1)
194 def convert_to_ppm_format(pngtopnm, basename):
195 png_file_re = re.compile(r"\.png$")
197 for png_file in glob.glob("%s*.png" % basename):
198 ppm_file = png_file_re.sub(".ppm", png_file)
200 p2p_cmd = f'{pngtopnm} "{png_file}"'
201 p2p_status, p2p_stdout = run_command(p2p_cmd)
203 error("Unable to convert %s to ppm format" % png_file)
205 ppm = open(ppm_file, 'w')
206 ppm.write(p2p_stdout)
209 # Returns a tuple of:
210 # ps_pages: list of page indexes of pages containing PS literals
211 # pdf_pages: list of page indexes of pages requiring running pdflatex
212 # page_count: total number of pages
213 # pages_parameter: parameter for dvipng to exclude pages with PostScript/PDF
214 def find_ps_pages(dvi_file):
216 # FIXME: try with pdflatex
217 if not os.path.isfile(dvi_file):
218 error("No DVI output.")
220 # Check for PostScript specials in the dvi, badly supported by dvipng,
221 # and inclusion of PDF/PNG/JPG files.
222 # This is required for correct rendering of PSTricks and TikZ
223 dv2dt = find_exe_or_terminate(["dv2dt"])
224 dv2dt_call = f'{dv2dt} "{dvi_file}"'
226 # The output from dv2dt goes to stdout
227 dv2dt_status, dv2dt_output = run_command(dv2dt_call)
228 psliteral_re = re.compile("^special[1-4] [0-9]+ '(\"|ps:)")
229 hyperref_re = re.compile("^special[1-4] [0-9]+ 'ps:.*/DEST pdfmark")
230 pdffile_re = re.compile("^special[1-4] [0-9]+ 'PSfile=.*\\.(pdf|png|jpg|jpeg|PDF|PNG|JPG|JPEG)")
232 # Parse the dtl file looking for PostScript specials and pdflatex files.
233 # Pages using PostScript specials or pdflatex files are recorded in
234 # ps_pages or pdf_pages, respectively, and then used to create a
235 # different LaTeX file for processing in legacy mode.
236 # If hyperref is detected, the corresponding page is recorded in pdf_pages.
244 for line in dv2dt_output.split("\n"):
246 if line.startswith("bop"):
252 if line.startswith("eop") and (page_has_ps or page_has_pdf):
253 # We save in a list all the PostScript/PDF pages
255 ps_pages.append(page_index)
257 pdf_pages.append(page_index)
258 ps_or_pdf_pages.append(page_index)
260 if psliteral_re.match(line) != None:
261 # Literal PostScript special detected!
262 # If hyperref is detected, put this page on the pdf pages list
263 if hyperref_re.match(line) != None:
268 elif pdffile_re.match(line) != None:
269 # Inclusion of pdflatex image file detected!
272 # Create the -pp parameter for dvipng
274 if len(ps_or_pdf_pages) > 0 and len(ps_or_pdf_pages) < page_index:
275 # Don't process Postscript/PDF pages with dvipng by selecting the
276 # wanted pages through the -pp parameter. E.g., dvipng -pp 4-12,14,64
277 pages_parameter = " -pp "
281 # Use page ranges, as a list of pages could exceed command line
282 # maximum length (especially under Win32)
283 for index in range(1, page_index + 1):
284 if (not index in ps_or_pdf_pages) and skip:
285 # We were skipping pages but current page shouldn't be skipped.
286 # Add this page to -pp, it could stay alone or become the
288 pages_parameter += str(index)
289 # Save the starting index to avoid things such as "11-11"
291 # We're not skipping anymore
293 elif (index in ps_or_pdf_pages) and (not skip):
294 # We weren't skipping but current page should be skipped
295 if last != index - 1:
296 # If the start index of the range is the previous page
297 # then it's not a range
298 pages_parameter += "-" + str(index - 1)
301 pages_parameter += ","
305 # Remove the trailing separator
306 pages_parameter = pages_parameter.rstrip(",")
307 # We've to manage the case in which the last page is closing a range
308 if (not index in ps_or_pdf_pages) and (not skip) and (last != index):
309 pages_parameter += "-" + str(index)
311 return (ps_pages, pdf_pages, page_index, pages_parameter)
322 output_format = "png"
323 script_name = argv[0]
325 # Parse and manipulate the command line arguments.
327 (opts, args) = getopt.gnu_getopt(argv[1:], "dhv", ["bibtex=", "bg=",
328 "debug", "dpi=", "fg=", "help", "latex=", "lilypond",
329 "lilypond-book=", "png", "ppm", "verbose"])
330 except getopt.GetoptError as err:
331 error(f"{err}\n{usage(script_name)}")
334 for opt, val in opts:
335 if opt in ("-h", "--help"):
336 print(usage(script_name))
338 elif opt == "--bibtex":
342 elif opt in ("-d", "--debug"):
343 lyxpreview_tools.debug = True
348 error("Cannot convert %s to an integer value" % val)
351 elif opt == "--latex":
353 elif opt == "--lilypond":
355 elif opt == "--lilypond-book":
356 lilypond_book = [val]
357 elif opt in ("--png", "--ppm"):
358 output_format = opt[2:]
359 elif opt in ("-v", "--verbose"):
360 lyxpreview_tools.verbose = True
362 # Determine input file
364 err = "A single input file is required, %s given" % (len(args) or "none")
365 error(f"{err}\n{usage(script_name)}")
368 dir, latex_file = os.path.split(input_path)
370 # Check for the input file
371 if not os.path.exists(input_path):
372 error('File "%s" not found.' % input_path)
376 if lyxpreview_tools.verbose:
377 f_out = open('verbose.txt', 'a')
382 progress("Running Python %s" % str(sys.version_info[:3]))
383 progress("Starting %s..." % script_name)
385 progress("Use win32_modules: %d" % lyxpreview_tools.use_win32_modules)
386 progress("Output format: %s" % output_format)
387 progress("Foreground color: %s" % fg_color)
388 progress("Background color: %s" % bg_color)
389 progress("Resolution (dpi): %s" % dpi)
390 progress("File to process: %s" % input_path)
392 fg_color = bytes(fg_color, 'ascii')
393 bg_color = bytes(bg_color, 'ascii')
395 fg_color_dvipng = make_texcolor(fg_color, False).decode('ascii')
396 bg_color_dvipng = make_texcolor(bg_color, False).decode('ascii')
398 # External programs used by the script.
399 latex = find_exe_or_terminate(latex or latex_commands)
400 bibtex = find_exe(bibtex or bibtex_commands)
402 lilypond_book = find_exe_or_terminate(lilypond_book or
403 ["lilypond-book --safe"])
405 # These flavors of latex are known to produce pdf output
406 pdf_output = latex in pdflatex_commands
408 progress("Latex command: %s" % latex)
409 progress("Latex produces pdf output: %s" % pdf_output)
410 progress("Bibtex command: %s" % bibtex)
411 progress("Lilypond-book command: %s" % lilypond_book)
412 progress("Preprocess through lilypond-book: %s" % lilypond)
413 progress("Altering the latex file for font size and colors")
415 # Make sure that multiple defined macros and the microtype package
416 # don't cause issues in the latex file.
417 fix_latex_file(latex_file, pdf_output)
420 progress("Preprocess the latex file through %s" % lilypond_book)
422 lilypond_book += " --pdf"
423 lilypond_book += " --latex-program=%s" % latex.split()[0]
425 # Make a copy of the latex file
426 lytex_file = latex_file_re.sub(".lytex", latex_file)
427 shutil.copyfile(latex_file, lytex_file)
429 # Preprocess the latex file through lilypond-book.
430 lytex_status, lytex_stdout = run_tex(lilypond_book, lytex_file)
433 progress("Using the legacy conversion method (PDF support)")
434 return legacy_conversion_step1(latex_file, dpi, output_format, fg_color,
435 bg_color, latex, pdf_output)
437 # This can go once dvipng becomes widespread.
438 dvipng = find_exe(["dvipng"])
440 progress("Using the legacy conversion method (dvipng not found)")
441 return legacy_conversion_step1(latex_file, dpi, output_format, fg_color,
442 bg_color, latex, pdf_output)
444 dv2dt = find_exe(["dv2dt"])
446 progress("Using the legacy conversion method (dv2dt not found)")
447 return legacy_conversion_step1(latex_file, dpi, output_format, fg_color,
448 bg_color, latex, pdf_output)
451 if output_format == "ppm":
452 pngtopnm = find_exe(["pngtopnm"])
454 progress("Using the legacy conversion method (pngtopnm not found)")
455 return legacy_conversion_step1(latex_file, dpi, output_format,
456 fg_color, bg_color, latex, pdf_output)
458 # Compile the latex file.
460 latex_status, latex_stdout = run_latex(latex, latex_file, bibtex)
461 latex_log = latex_file_re.sub(".log", latex_file)
463 progress("Will try to recover from %s failure" % latex)
464 error_pages = check_latex_log(latex_log)
466 # The dvi output file name
467 dvi_file = latex_file_re.sub(".dvi", latex_file)
469 # If there's no DVI output, look for PDF and go to legacy or fail
470 if not os.path.isfile(dvi_file):
471 # No DVI, is there a PDF?
472 pdf_file = latex_file_re.sub(".pdf", latex_file)
473 if os.path.isfile(pdf_file):
474 progress("%s produced a PDF output, fallback to legacy." \
475 % (os.path.basename(latex)))
476 progress("Using the legacy conversion method (PDF support)")
477 return legacy_conversion_step1(latex_file, dpi, output_format,
478 fg_color, bg_color, latex, True)
480 error("No DVI or PDF output. %s failed." \
481 % (os.path.basename(latex)))
483 # Look for PS literals or inclusion of pdflatex files in DVI pages
484 # ps_pages: list of indexes of pages containing PS literals
485 # pdf_pages: list of indexes of pages requiring running pdflatex
486 # page_count: total number of pages
487 # pages_parameter: parameter for dvipng to exclude pages with PostScript
488 (ps_pages, pdf_pages, page_count, pages_parameter) = find_ps_pages(dvi_file)
490 # If all pages need PostScript or pdflatex, directly use the legacy method.
491 if len(ps_pages) == page_count:
492 progress("Using the legacy conversion method (PostScript support)")
493 return legacy_conversion_step1(latex_file, dpi, output_format, fg_color,
494 bg_color, latex, pdf_output)
495 elif len(pdf_pages) == page_count:
496 progress("Using the legacy conversion method (PDF support)")
497 return legacy_conversion_step1(latex_file, dpi, output_format, fg_color,
498 bg_color, "pdflatex", True)
500 # Retrieve resolution
501 resolution = extract_resolution(latex_log, dpi)
503 # Run the dvi file through dvipng.
504 dvipng_call = '%s -Ttight -depth -height -D %d -fg "%s" -bg "%s" %s "%s"' \
505 % (dvipng, resolution, fg_color_dvipng, bg_color_dvipng, pages_parameter, dvi_file)
506 dvipng_status, dvipng_stdout = run_command(dvipng_call)
509 warning("%s failed to generate images from %s... fallback to legacy method" \
510 % (os.path.basename(dvipng), dvi_file))
511 progress("Using the legacy conversion method (dvipng failed)")
512 return legacy_conversion_step1(latex_file, dpi, output_format, fg_color,
513 bg_color, latex, pdf_output)
515 # Extract metrics info from dvipng_stdout.
516 metrics_file = latex_file_re.sub(".metrics", latex_file)
517 dvipng_metrics = extract_metrics_info(dvipng_stdout)
519 # If some pages require PostScript pass them to legacy method
520 if len(ps_pages) > 0:
521 # Create a new LaTeX file just for the snippets needing
523 legacy_latex_file = latex_file_re.sub("_legacy.tex", latex_file)
524 filter_pages(latex_file, legacy_latex_file, ps_pages)
526 # Pass the new LaTeX file to the legacy method
527 progress("Pages %s include postscript specials" % ps_pages)
528 progress("Using the legacy conversion method (PostScript support)")
529 legacy_status, legacy_metrics = legacy_conversion_step1(legacy_latex_file,
530 dpi, output_format, fg_color, bg_color, latex, pdf_output, True)
532 # Now we need to mix metrics data from dvipng and the legacy method
533 original_bitmap = latex_file_re.sub("%d." + output_format, legacy_latex_file)
534 destination_bitmap = latex_file_re.sub("%d." + output_format, latex_file)
536 # Join metrics from dvipng and legacy, and rename legacy bitmaps
537 join_metrics_and_rename(dvipng_metrics, legacy_metrics, ps_pages,
538 original_bitmap, destination_bitmap)
540 # If some pages require running pdflatex pass them to legacy method
541 if len(pdf_pages) > 0:
542 # Create a new LaTeX file just for the snippets needing
544 legacy_latex_file = latex_file_re.sub("_legacy.tex", latex_file)
545 filter_pages(latex_file, legacy_latex_file, pdf_pages)
547 # Pass the new LaTeX file to the legacy method
548 progress("Pages %s require processing with pdflatex" % pdf_pages)
549 progress("Using the legacy conversion method (PDF support)")
550 legacy_status, legacy_metrics = legacy_conversion_step1(legacy_latex_file,
551 dpi, output_format, fg_color, bg_color, "pdflatex", True, True)
553 # Now we need to mix metrics data from dvipng and the legacy method
554 original_bitmap = latex_file_re.sub("%d." + output_format, legacy_latex_file)
555 destination_bitmap = latex_file_re.sub("%d." + output_format, latex_file)
557 # Join metrics from dvipng and legacy, and rename legacy bitmaps
558 join_metrics_and_rename(dvipng_metrics, legacy_metrics, pdf_pages,
559 original_bitmap, destination_bitmap)
561 # Invalidate metrics for pages that produced errors
562 if len(error_pages) > 0:
564 for index in error_pages:
565 if index not in ps_pages and index not in pdf_pages:
566 dvipng_metrics.pop(index - 1)
567 dvipng_metrics.insert(index - 1, (index, -1.0))
570 warning("Failed to produce %d preview snippet(s)" % error_count)
572 # Convert images to ppm format if necessary.
573 if output_format == "ppm":
574 convert_to_ppm_format(pngtopnm, latex_file_re.sub("", latex_file))
576 # Actually create the .metrics file
577 write_metrics_info(dvipng_metrics, metrics_file)
579 return (0, dvipng_metrics)
581 if __name__ == "__main__":
582 sys.exit(main(sys.argv)[0])