-#! /usr/bin/env python
# -*- coding: utf-8 -*-
# file legacy_lyxpreview2ppm.py
# Paul A. Rubin, rubin@msu.edu.
# This script takes a LaTeX file and generates a collection of
-# ppm image files, one per previewed snippet.
+# png or ppm image files, one per previewed snippet.
# Example usage:
# legacy_lyxpreview2bitmap.py 0lyxpreview.tex 128 ppm 000000 faf0e6
# Decomposing TEXFILE's name as DIR/BASE.tex, this script will,
# if executed successfully, leave in DIR:
# * a (possibly large) number of image files with names
-# like BASE[0-9]+.ppm
+# like BASE[0-9]+.(ppm|png)
# * a file BASE.metrics, containing info needed by LyX to position
# the images correctly on the screen.
# The script uses several external programs and files:
+# * python 2.4 or later (subprocess module);
# * A latex executable;
# * preview.sty;
# * dvips;
# * gs;
+# * pdflatex (optional);
# * pnmcrop (optional).
+# * pdftocairo (optional).
+# * epstopdf (optional).
# preview.sty is part of the preview-latex project
# http://preview-latex.sourceforge.net/
# Alternatively, it can be obtained from
# CTAN/support/preview-latex/
-# The script uses the deprecated dvi->ps->ppm conversion route.
-# If possible, please grab 'dvipng'; it's faster and more robust.
-# If you have it then this script will not be invoked by
-# lyxpreview2bitmap.py.
-# Warning: this legacy support will be removed one day...
+# What does this script do?
+# [legacy_conversion]
+# 0) Process command-line arguments
+# [legacy_conversion_step1]
+# 1) Call latex to create a DVI file from LaTeX
+# [legacy_conversion_step2]
+# 2) Call dvips to create one PS file for each DVI page
+# [legacy_conversion_step3]
+# 3) If dvips fails look for PDF and call pdftocairo or gs to produce bitmaps
+# 4) Otherwise call pdftocairo or gs on each PostScript file to produce bitmaps
+# [legacy_conversion_pdflatex]
+# 5) Keep track of pages on which gs failed and pass them to pdflatex
+# 6) Call pdftocairo or gs on the PDF output from pdflatex to produce bitmaps
+# 7) Extract and write to file (or return to lyxpreview2bitmap)
+# metrics from both methods (standard and pdflatex)
+
+# The script uses the old dvi->ps->png conversion route,
+# which is good when using PSTricks, TikZ or other packages involving
+# PostScript literals (steps 1, 2, 4).
+# This script also generates bitmaps from PDF created by a call to
+# lyxpreview2bitmap.py passing "pdflatex" to the CONVERTER parameter
+# (step 3).
+# Finally, there's also has a fallback method based on pdflatex, which
+# is required in certain cases, if hyperref is active for instance,
+# (step 5, 6).
+# If possible, dvipng should be used, as it's much faster.
+# If possible, the script will use pdftocairo instead of gs,
+# as it's much faster and gives better results.
import glob, os, pipes, re, string, sys
-from lyxpreview_tools import copyfileobj, error, find_exe, \
- find_exe_or_terminate, make_texcolor, mkstemp, run_command, warning
-
-# Pre-compiled regular expression.
-latex_file_re = re.compile("\.tex$")
+from lyxpreview_tools import check_latex_log, copyfileobj, error, filter_pages,\
+ find_exe, find_exe_or_terminate, join_metrics_and_rename, latex_commands, \
+ latex_file_re, make_texcolor, mkstemp, pdflatex_commands, progress, \
+ run_command, run_latex, warning, write_metrics_info
def usage(prog_name):
- return "Usage: %s <latex file> <dpi> ppm <fg color> <bg color>\n"\
- "\twhere the colors are hexadecimal strings, eg 'faf0e6'"\
+ return "Usage: %s <latex file> <dpi> ppm <fg color> <bg color>\n" \
+ "\twhere the colors are hexadecimal strings, eg 'faf0e6'" \
% prog_name
-
-def extract_metrics_info(log_file, metrics_file):
- metrics = open(metrics_file, 'w')
+# Returns a list of tuples containing page number and ascent fraction
+# extracted from dvipng output.
+# Use write_metrics_info to create the .metrics file with this info
+def legacy_extract_metrics_info(log_file):
log_re = re.compile("Preview: ([ST])")
data_re = re.compile("(-?[0-9]+) (-?[0-9]+) (-?[0-9]+) (-?[0-9]+)")
tp_descent = 0.0
success = 0
+ results = []
try:
for line in open(log_file, 'r').readlines():
match = log_re.match(line)
error("Unexpected data in %s\n%s" % (log_file, line))
if snippet:
- ascent = string.atoi(match.group(2))
- descent = string.atoi(match.group(3))
+ ascent = string.atof(match.group(2))
+ descent = string.atof(match.group(3))
frac = 0.5
- if ascent >= 0 and descent >= 0:
- ascent = float(ascent) + tp_ascent
- descent = float(descent) - tp_descent
+ if ascent == 0 and descent == 0:
+ # This is an empty image, forbid its display
+ frac = -1.0
+ elif ascent >= 0 or descent >= 0:
+ ascent = ascent + tp_ascent
+ descent = descent - tp_descent
if abs(ascent + descent) > 0.1:
frac = ascent / (ascent + descent)
if frac < 0 or frac > 1:
frac = 0.5
- metrics.write("Snippet %s %f\n" % (match.group(1), frac))
+ results.append((int(match.group(1)), frac))
else:
tp_descent = string.atof(match.group(2))
except:
# Unable to open the file, but do nothing here because
# the calling function will act on the value of 'success'.
- warning('Warning in extract_metrics_info! Unable to open "%s"' % log_file)
+ warning('Warning in legacy_extract_metrics_info! Unable to open "%s"' % log_file)
warning(`sys.exc_type` + ',' + `sys.exc_value`)
- return success
+ if success == 0:
+ error("Failed to extract metrics info from %s" % log_file)
+ return results
def extract_resolution(log_file, dpi):
fontsize_re = re.compile("Preview: Fontsize")
return dpi * (10.0 / fontsize) * (1000.0 / magnification)
-def legacy_latex_file(latex_file, fg_color, bg_color, bg_color_gr):
- use_preview_dvi_re = re.compile("(\s*\\\\usepackage\[[^]]+)(dvips\]{preview})")
- use_preview_pdf_re = re.compile("(\s*\\\\usepackage\[[^]]+)(pdftex\]{preview})")
+def legacy_latex_file(latex_file, fg_color, bg_color):
+ use_preview_re = re.compile(r"\s*\\usepackage\[([^]]+)\]{preview}")
+ fg_color_gr = make_texcolor(fg_color, True)
+ bg_color_gr = make_texcolor(bg_color, True)
tmp = mkstemp()
success = 0
try:
- for line in open(latex_file, 'r').readlines():
- match = use_preview_dvi_re.match(line)
- if match == None:
- match = use_preview_pdf_re.match(line)
- if match == None:
- tmp.write(line)
- continue
- success = 1
- tmp.write(" \\usepackage{color}\n" \
- " \\pagecolor[rgb]{%s}\n" \
- "%s\n" \
- % (bg_color_gr, match.group()))
- continue
-
- success = 1
- tmp.write("%stightpage,%s\n" \
- " \\AtBeginDocument{\\AtBeginDvi{%%\n" \
- " \\special{!userdict begin/bop-hook{//bop-hook exec\n" \
- " <%s%s>{255 div}forall setrgbcolor\n" \
- " clippath fill setrgbcolor}bind def end}}}\n" \
- % (match.group(1), match.group(2), fg_color, bg_color))
-
+ f = open(latex_file, 'r')
except:
# Unable to open the file, but do nothing here because
# the calling function will act on the value of 'success'.
warning('Warning in legacy_latex_file! Unable to open "%s"' % latex_file)
warning(`sys.exc_type` + ',' + `sys.exc_value`)
+ for line in f.readlines():
+ if success:
+ tmp.write(line)
+ continue
+ match = use_preview_re.match(line)
+ if match == None:
+ tmp.write(line)
+ continue
+ success = 1
+ # Package order: color should be loaded before preview
+ # Preview options: add the options lyx and tightpage
+ tmp.write(r"""
+\usepackage{color}
+\definecolor{fg}{rgb}{%s}
+\definecolor{bg}{rgb}{%s}
+\pagecolor{bg}
+\usepackage[%s,tightpage]{preview}
+\makeatletter
+\def\t@a{cmr}
+\if\f@family\t@a
+\IfFileExists{lmodern.sty}{\usepackage{lmodern}}{\usepackage{ae,aecompl}}
+\fi
+\g@addto@macro\preview{\begingroup\color{bg}\special{ps::clippath fill}\color{fg}}
+\g@addto@macro\endpreview{\endgroup}
+\makeatother
+""" % (fg_color_gr, bg_color_gr, match.group(1)))
+
if success:
copyfileobj(tmp, open(latex_file,"wb"), 1)
copyfileobj(tmp, open(file,"wb"), 1)
-def legacy_conversion(argv):
+def legacy_conversion(argv, skipMetrics = False):
# Parse and manipulate the command line arguments.
- if len(argv) != 6:
+ if len(argv) == 7:
+ latex = [argv[6]]
+ elif len(argv) != 6:
error(usage(argv[0]))
+ else:
+ latex = None
dir, latex_file = os.path.split(argv[1])
if len(dir) != 0:
fg_color = argv[4]
bg_color = argv[5]
- bg_color_gr = make_texcolor(argv[5], True)
# External programs used by the script.
- path = string.split(os.environ["PATH"], os.pathsep)
- latex = find_exe_or_terminate(["latex", "pplatex", "platex", "latex2e"], path)
+ latex = find_exe_or_terminate(latex or latex_commands)
- # Move color information into the latex file.
- if not legacy_latex_file(latex_file, fg_color, bg_color, bg_color_gr):
- error("Unable to move color info into the latex file")
+ pdf_output = latex in pdflatex_commands
- # Compile the latex file.
- latex_call = '%s "%s"' % (latex, latex_file)
+ return legacy_conversion_step1(latex_file, dpi, output_format, fg_color,
+ bg_color, latex, pdf_output, skipMetrics)
- latex_status, latex_stdout = run_command(latex_call)
- if latex_status != None:
- error("%s failed to compile %s" \
- % (os.path.basename(latex), latex_file))
- return legacy_conversion_step2(latex_file, dpi, output_format)
+# Add color info to the latex file, since ghostscript doesn't
+# have the option to set foreground and background colors on
+# the command line. Run the resulting file through latex.
+def legacy_conversion_step1(latex_file, dpi, output_format, fg_color, bg_color,
+ latex, pdf_output = False, skipMetrics = False):
+ # Move color information, lyx and tightpage options into the latex file.
+ if not legacy_latex_file(latex_file, fg_color, bg_color):
+ error("""Unable to move the color information, and the lyx and tightpage
+ options of preview-latex, into the latex file""")
-def legacy_conversion_step2(latex_file, dpi, output_format):
+ # Compile the latex file.
+ latex_status, latex_stdout = run_latex(latex, latex_file)
+ if latex_status:
+ progress("Will try to recover from %s failure" % latex)
+
+ if pdf_output:
+ return legacy_conversion_step3(latex_file, dpi, output_format, True, skipMetrics)
+ else:
+ return legacy_conversion_step2(latex_file, dpi, output_format, skipMetrics)
+
+# Creates a new LaTeX file from the original with pages specified in
+# failed_pages, pass it through pdflatex and updates the metrics
+# from the standard legacy route
+def legacy_conversion_pdflatex(latex_file, failed_pages, legacy_metrics,
+ use_pdftocairo, conv, gs_device, gs_ext, alpha, resolution, output_format):
+
+ error_count = 0
+
+ # Search for pdflatex executable
+ pdflatex = find_exe(["pdflatex"])
+ if pdflatex == None:
+ warning("Can't find pdflatex. Some pages failed with all the possible routes.")
+ failed_pages = []
+ else:
+ # Create a new LaTeX file from the original but only with failed pages
+ pdf_latex_file = latex_file_re.sub("_pdflatex.tex", latex_file)
+ filter_pages(latex_file, pdf_latex_file, failed_pages)
+
+ # pdflatex call
+ error_pages = []
+ pdflatex_status, pdflatex_stdout = run_latex(pdflatex, pdf_latex_file)
+ if pdflatex_status:
+ error_pages = check_latex_log(latex_file_re.sub(".log", pdf_latex_file))
+
+ pdf_file = latex_file_re.sub(".pdf", pdf_latex_file)
+ latex_file_root = latex_file_re.sub("", pdf_latex_file)
+
+ # Converter call to produce bitmaps
+ if use_pdftocairo:
+ conv_call = '%s -png -transp -r %d "%s" "%s"' \
+ % (conv, resolution, pdf_file, latex_file_root)
+ conv_status, conv_stdout = run_command(conv_call)
+ if not conv_status:
+ seqnum_re = re.compile("-([0-9]+)")
+ for name in glob.glob("%s-*.png" % latex_file_root):
+ match = seqnum_re.search(name)
+ if match != None:
+ new_name = seqnum_re.sub(str(int(match.group(1))), name)
+ os.rename(name, new_name)
+ else:
+ conv_call = '%s -dNOPAUSE -dBATCH -dSAFER -sDEVICE=%s ' \
+ '-sOutputFile="%s%%d.%s" ' \
+ '-dGraphicsAlphaBit=%d -dTextAlphaBits=%d ' \
+ '-r%f "%s"' \
+ % (conv, gs_device, latex_file_root, \
+ gs_ext, alpha, alpha, resolution, pdf_file)
+ conv_status, conv_stdout = run_command(conv_call)
+
+ if conv_status:
+ # Give up!
+ warning("Some pages failed with all the possible routes")
+ failed_pages = []
+ else:
+ # We've done it!
+ pdf_log_file = latex_file_re.sub(".log", pdf_latex_file)
+ pdf_metrics = legacy_extract_metrics_info(pdf_log_file)
+
+ # Invalidate metrics for pages that produced errors
+ if len(error_pages) > 0:
+ for index in error_pages:
+ pdf_metrics.pop(index - 1)
+ pdf_metrics.insert(index - 1, (index, -1.0))
+ error_count += 1
+
+ original_bitmap = latex_file_re.sub("%d." + output_format, pdf_latex_file)
+ destination_bitmap = latex_file_re.sub("%d." + output_format, latex_file)
+
+ # Join the metrics with the those from dvips and rename the bitmap images
+ join_metrics_and_rename(legacy_metrics, pdf_metrics, failed_pages,
+ original_bitmap, destination_bitmap)
+
+ return error_count
+
+
+# The file has been processed through latex and we expect dvi output.
+# Run dvips, taking note whether it was successful.
+def legacy_conversion_step2(latex_file, dpi, output_format, skipMetrics = False):
# External programs used by the script.
- path = string.split(os.environ["PATH"], os.pathsep)
- dvips = find_exe_or_terminate(["dvips"], path)
- gs = find_exe_or_terminate(["gswin32c", "gs"], path)
- pnmcrop = find_exe(["pnmcrop"], path)
+ dvips = find_exe_or_terminate(["dvips"])
# Run the dvi file through dvips.
dvi_file = latex_file_re.sub(".dvi", latex_file)
ps_file = latex_file_re.sub(".ps", latex_file)
- pdf_file = latex_file_re.sub(".pdf", latex_file)
- dvips_call = '%s -o "%s" "%s"' % (dvips, ps_file, dvi_file)
+ dvips_call = '%s -i -o "%s" "%s"' % (dvips, ps_file, dvi_file)
dvips_failed = False
dvips_status, dvips_stdout = run_command(dvips_call)
- if dvips_status != None:
+ if dvips_status:
warning('Failed: %s %s ... looking for PDF' \
% (os.path.basename(dvips), dvi_file))
dvips_failed = True
- # Extract resolution data for gs from the log file.
+ return legacy_conversion_step3(latex_file, dpi, output_format, dvips_failed, skipMetrics)
+
+
+# Either latex and dvips have been run and we have a ps file, or
+# pdflatex has been run and we have a pdf file. Proceed with pdftocairo or gs.
+def legacy_conversion_step3(latex_file, dpi, output_format, dvips_failed, skipMetrics = False):
+ # External programs used by the script.
+ gs = find_exe_or_terminate(["gswin32c", "gswin64c", "gs"])
+ pnmcrop = find_exe(["pnmcrop"])
+ pdftocairo = find_exe(["pdftocairo"])
+ epstopdf = find_exe(["epstopdf"])
+ use_pdftocairo = pdftocairo != None and output_format == "png"
+ if use_pdftocairo:
+ conv = pdftocairo
+ else:
+ conv = gs
+
+ # Files to process
+ pdf_file = latex_file_re.sub(".pdf", latex_file)
+ ps_file = latex_file_re.sub(".ps", latex_file)
+
+ # The latex file name without extension
+ latex_file_root = latex_file_re.sub("", latex_file)
+
+ # Extract resolution data for the converter from the log file.
log_file = latex_file_re.sub(".log", latex_file)
resolution = extract_resolution(log_file, dpi)
+ # Check whether some pages produced errors
+ error_pages = check_latex_log(log_file)
+
# Older versions of gs have problems with a large degree of
# anti-aliasing at high resolutions
alpha = 4
gs_device = "pnmraw"
gs_ext = "ppm"
- # Generate the bitmap images
- gs_call = '%s -dNOPAUSE -dBATCH -dSAFER -sDEVICE=%s ' \
- '-sOutputFile="%s%%d.%s" ' \
- '-dGraphicsAlphaBit=%d -dTextAlphaBits=%d ' \
- '-r%f "%s"' \
- % (gs, gs_device, latex_file_re.sub("", latex_file), \
- gs_ext, alpha, alpha, resolution, ps_file)
+ # Extract the metrics from the log file
+ legacy_metrics = legacy_extract_metrics_info(log_file)
- if dvips_failed:
- gs_call = '%s -dNOPAUSE -dBATCH -dSAFER -sDEVICE=%s ' \
- '-sOutputFile="%s%%d.%s" ' \
- '-dGraphicsAlphaBit=%d -dTextAlphaBits=%d ' \
- '-r%f "%s"' \
- % (gs, gs_device, latex_file_re.sub("", latex_file), \
- gs_ext, alpha, alpha, resolution, pdf_file)
-
- gs_status, gs_stdout = run_command(gs_call)
- if gs_status != None:
- error("Failed: %s %s" % (os.path.basename(gs), ps_file))
-
- # Crop the images
- if pnmcrop != None:
- crop_files(pnmcrop, latex_file_re.sub("", latex_file))
-
- # Extract metrics info from the log file.
- metrics_file = latex_file_re.sub(".metrics", latex_file)
- if not extract_metrics_info(log_file, metrics_file):
- error("Failed to extract metrics info from %s" % log_file)
+ # List of pages which failed to produce a correct output
+ failed_pages = []
- return 0
+ # Generate the bitmap images
+ if dvips_failed:
+ # dvips failed, maybe there's a PDF, try to produce bitmaps
+ if use_pdftocairo:
+ conv_call = '%s -png -transp -r %d "%s" "%s"' \
+ % (pdftocairo, resolution, pdf_file, latex_file_root)
+
+ conv_status, conv_stdout = run_command(conv_call)
+ if not conv_status:
+ seqnum_re = re.compile("-([0-9]+)")
+ for name in glob.glob("%s-*.png" % latex_file_root):
+ match = seqnum_re.search(name)
+ if match != None:
+ new_name = seqnum_re.sub(str(int(match.group(1))), name)
+ os.rename(name, new_name)
+ else:
+ conv_call = '%s -dNOPAUSE -dBATCH -dSAFER -sDEVICE=%s ' \
+ '-sOutputFile="%s%%d.%s" ' \
+ '-dGraphicsAlphaBit=%d -dTextAlphaBits=%d ' \
+ '-r%f "%s"' \
+ % (gs, gs_device, latex_file_root, \
+ gs_ext, alpha, alpha, resolution, pdf_file)
+
+ conv_status, conv_stdout = run_command(conv_call)
+
+ if conv_status:
+ error("Failed: %s %s" % (os.path.basename(conv), pdf_file))
+ else:
+ # Model for calling the converter on each file
+ if use_pdftocairo and epstopdf != None:
+ conv_call = '%s -png -transp -singlefile -r %d "%%s" "%s%%d"' \
+ % (pdftocairo, resolution, latex_file_root)
+ else:
+ conv_call = '%s -dNOPAUSE -dBATCH -dSAFER -sDEVICE=%s ' \
+ '-sOutputFile="%s%%d.%s" ' \
+ '-dGraphicsAlphaBit=%d -dTextAlphaBits=%d ' \
+ '-r%f "%%s"' \
+ % (gs, gs_device, latex_file_root, \
+ gs_ext, alpha, alpha, resolution)
+
+ i = 0
+ # Collect all the PostScript files (like *.001, *.002, ...)
+ ps_files = glob.glob("%s.[0-9][0-9][0-9]" % latex_file_root)
+ ps_files.sort()
+
+ # Call the converter for each file
+ for file in ps_files:
+ i = i + 1
+ progress("Processing page %s, file %s" % (i, file))
+ if use_pdftocairo and epstopdf != None:
+ conv_name = "epstopdf"
+ conv_status, conv_stdout = run_command("%s --outfile=%s.pdf %s"
+ % (epstopdf, file, file))
+ if not conv_status:
+ conv_name = "pdftocairo"
+ file = file + ".pdf"
+ conv_status, conv_stdout = run_command(conv_call % (file, i))
+ else:
+ conv_name = "ghostscript"
+ conv_status, conv_stdout = run_command(conv_call % (i, file))
+
+ if conv_status:
+ # The converter failed, keep track of this
+ warning("%s failed on page %s, file %s" % (conv_name, i, file))
+ failed_pages.append(i)
+
+ # Pass failed pages to pdflatex
+ if len(failed_pages) > 0:
+ warning("Now trying to obtain failed previews through pdflatex")
+ error_count = legacy_conversion_pdflatex(latex_file, failed_pages,
+ legacy_metrics, use_pdftocairo, conv, gs_device, gs_ext, alpha,
+ resolution, output_format)
+ else:
+ error_count = 0
+
+ # Invalidate metrics for pages that produced errors
+ if len(error_pages) > 0:
+ for index in error_pages:
+ if index not in failed_pages:
+ legacy_metrics.pop(index - 1)
+ legacy_metrics.insert(index - 1, (index, -1.0))
+ error_count += 1
+
+ # Crop the ppm images
+ if pnmcrop != None and output_format == "ppm":
+ crop_files(pnmcrop, latex_file_root)
+
+ # Allow to skip .metrics creation for custom management
+ # (see the dvipng method)
+ if not skipMetrics:
+ # Extract metrics info from the log file.
+ metrics_file = latex_file_re.sub(".metrics", latex_file)
+ write_metrics_info(legacy_metrics, metrics_file)
+ if error_count:
+ warning("Failed to produce %d preview snippet(s)" % error_count)
+
+ return (0, legacy_metrics)
if __name__ == "__main__":
- legacy_conversion(sys.argv)
+ sys.exit(legacy_conversion(sys.argv)[0])