#!/usr/bin/python3 import os, sys, re def filterSVG(t,verbose=False): """ remove .svg in image names, replace by .pdf """ if verbose: sys.stderr.write("replacing .svg extensions by .pdf for included graphics\n") return re.sub(r"\.svg", ".pdf", t) def filterPNG(t,verbose=False): """ remove .png in image names, replace by .pdf """ if verbose: sys.stderr.write("replacing .png extensions by .pdf for included graphics\n") return re.sub(r"\.png", ".pdf", t) def filterJPG(t,verbose=False): """ remove .jpg in image names, replace by .pdf """ if verbose: sys.stderr.write("replacing .jpg extensions by .pdf for included graphics\n") return re.sub(r"\.jpg", ".pdf", t) def filterNastyUnicode(t,verbose=False): """ remove problematic Unicode characters, like dots, thin spaces, special minus sign """ if verbose: sys.stderr.write("removing nasty unicode chars\n") toReplace={ "\u2005": " ", "\u2003": " ", "\u200a": " ", "\u22ef": r"\\dots", "\u2212": "-", "↑": "", "↓": "", " ": "~", "’": "'", "«": "\\\\guillemotleft{}", "»": "\\\\guillemotright{}", "×": r"\\times{}" } for code, repl in toReplace.items(): t=re.sub(code, repl, t) return t def filterSphinxIncludeGraphics(t, verbose=False): if verbose: sys.stderr.write("remove empty lines between SphinxIncludeGraphics\n") pattern=re.compile(r"\\noindent\\sphinxincludegraphics.*") lines=t.split("\n") new=[lines[0], lines[1]] # always keep the two first lines for i in range(2, len(lines)): if pattern.match(new[-2]) and new[-1]=="" and pattern.match(lines[i]): new[-1]=lines[i] # this drops the empty line else: new.append(lines[i]) return "\n".join(new) def filterCurlyBraces(t,verbose=False): """ remove .png in image names, replace by .pdf """ if verbose: sys.stderr.write("removing extra curly braces around image filenames\n") return re.sub(r"(\\noindent\\sphinxincludegraphics.*){{(.*)}\.pdf}", r"\1{\2.pdf}", t) def filterGreek(t,verbose=False): """ remove .png in image names, replace by .pdf """ if verbose: sys.stderr.write("replacing UTF-8 greek characters\n") t = re.sub(r"π", r"$\\pi$", t) t = re.sub(r"θ", r"$\\theta$", t) t = re.sub(r"(\S+)Ω", r"$\1\\Omega$", t) return re.sub(r"Ω", r"$\\Omega$", t) def filterSphinx(t,verbose=False): """ remove .png in image names, replace by .pdf """ if verbose: sys.stderr.write("customizing some directives\n") t = re.sub(r"\\usepackage{sphinx}", r"\\usepackage{sphinx}"+"\n"+r"\\usepackage{pdfpages}", t) return re.sub(r"\\sphinxmaketitle", r"\\includepdf{coverpage.pdf}"+"\n"+r"\\includepdf{preface.pdf}", t) def l10n(t, lang): """ make localization changes in the source text t """ babel = { "en": "english", "fr": "french", "es": "spanish", } if lang not in babel: return t return re.sub(r"\\usepackage({babel})", r"\\usepackage[{name}]\1".format(name=babel[lang]), t) filters=( filterSVG, filterPNG, filterJPG, filterNastyUnicode, filterSphinxIncludeGraphics, filterCurlyBraces, filterGreek, filterSphinx, ) if __name__=="__main__": buildDir=sys.argv[1] texFile=sys.argv[2] t="" with open(buildDir+"/"+texFile, encoding="utf-8") as infile: t=infile.read() for f in filters: t=f(t, verbose=True) lang = os.path.abspath(buildDir).split("/")[-3] t = l10n(t,lang) with open(buildDir+"/"+texFile+".tmp","w", encoding="utf-8") as outfile: outfile.write(t) os.rename(buildDir+"/"+texFile+".tmp", buildDir+"/"+texFile)