buildscripts/postprocess_html.py

   1 #!@PYTHON@
   2
   3 """
   4 Postprocess HTML files:
   5 add footer, tweak links, add language selection menu.
   6 """
   7 import re
   8 import os
   9 import time
  10 import operator
  11
  12 import langdefs
  13
  14 # This is to try to make the docball not too big with almost duplicate files
  15 # see process_links()
  16 non_copied_pages = ['Documentation/user/out-www/lilypond-big-page',
  17                     'Documentation/user/out-www/lilypond-internals-big-page',
  18                     'Documentation/user/out-www/lilypond-learning-big-page',
  19                     'Documentation/user/out-www/lilypond-program-big-page',
  20                     'Documentation/user/out-www/music-glossary-big-page',
  21                     'out-www/examples',
  22                     'Documentation/topdocs',
  23                     'Documentation/bibliography',
  24                     'Documentation/out-www/THANKS',
  25                     'Documentation/out-www/DEDICATION',
  26                     'Documentation/out-www/devel',
  27                     'input/']
  28
  29 def _doc (s):
  30     return s
  31
  32 header = r"""
  33 """
  34
  35 footer = '''
  36 <div class="footer">
  37 <p>
  38 %(footer_name_version)s
  39 <br>
  40 <address>
  41 %(footer_report_errors)s </address>
  42 <br>
  43 %(footer_suggest_docs)s
  44 <br>
  45 </p>
  46 </div>
  47 '''
  48 footer_name_version = _doc ('This page is for %(package_name)s-%(package_version)s (%(branch_str)s).')
  49 footer_report_errors = _doc ('Report errors to <a href="%(mail_address_url)s">%(mail_address)s</a>.')
  50 # ugh, must not have "_doc" in strings because it is naively replaced with "_" in hacked gettext process
  51 footer_suggest_docs = _doc ('Your <a href="%(suggest_Docs_url)s">suggestions for the documentation</a> are welcome.')
  52
  53 mail_address = 'http://post.gmane.org/post.php?group=gmane.comp.gnu.lilypond.bugs'
  54 suggest_Docs_url = 'http://lilypond.org/web/devel/participating/documentation-adding'
  55
  56 header_tag = '<!-- header_tag -->'
  57 header_tag_re = re.compile (header_tag)
  58
  59 footer_tag = '<!-- footer_tag -->'
  60 footer_tag_re = re.compile (footer_tag)
  61
  62 lang_available = _doc ("Other languages: %s.")
  63 browser_lang = _doc ('About <A HREF="%s">automatic language selection</A>.')
  64 browser_language_url = "/web/about/browser-language"
  65
  66 LANGUAGES_TEMPLATE = '''
  67 <p id="languages">
  68  %(language_available)s
  69  <br/>
  70  %(browser_language)s
  71 </p>
  72 '''
  73
  74
  75 html_re = re.compile ('(.*?)(?:[.]([^/.]*))?[.]html$')
  76 pages_dict = {}
  77
  78 def build_pages_dict (filelist):
  79     """Build dictionary of available translations of each page"""
  80     global pages_dict
  81     for f in filelist:
  82         m = html_re.match (f)
  83         if m:
  84             g = m.groups()
  85             if len (g) <= 1 or g[1] == None:
  86                 e = ''
  87             else:
  88                 e = g[1]
  89             if not g[0] in pages_dict:
  90                 pages_dict[g[0]] = [e]
  91             else:
  92                 pages_dict[g[0]].append (e)
  93
  94 def source_links_replace (m, source_val):
  95     return 'href="' + os.path.join (source_val, m.group (1)) + '"'
  96
  97 splitted_docs_re = re.compile ('(input/lsr/out-www/lilypond-snippets|\
  98 Documentation/user/out-www/(lilypond|music-glossary|lilypond-program|\
  99 lilypond-learning))/')
 100
 101 snippets_ref_re = re.compile (r'href="(\.\./)?lilypond-snippets')
 102 user_ref_re = re.compile ('href="(?:\.\./)?lilypond\
 103 (-internals|-learning|-program|(?!-snippets))')
 104
 105 ## Windows does not support symlinks.
 106 # This function avoids creating symlinks for splitted HTML manuals
 107 # Get rid of symlinks in GNUmakefile.in (local-WWW-post)
 108 # this also fixes missing PNGs only present in translated docs
 109 def hack_urls (s, prefix):
 110     if splitted_docs_re.match (prefix):
 111         s = re.sub ('(href|src)="(../lily-.*?|.*?[.]png)"', '\\1="../\\2"', s)
 112
 113     # fix xrefs between documents in different directories ad hoc
 114     if 'user/out-www/lilypond' in prefix:
 115         s = snippets_ref_re.sub ('href="source/input/lsr/lilypond-snippets', s)
 116     elif 'input/lsr' in prefix:
 117         s = user_ref_re.sub ('href="source/Documentation/user/lilypond\\1', s)
 118
 119     source_path = os.path.join (os.path.dirname (prefix), 'source')
 120     if not os.path.islink (source_path):
 121         return s
 122     source_val = os.readlink (source_path)
 123     return re.sub ('href="source/(.*?)"', lambda m: source_links_replace (m, source_val), s)
 124
 125 body_tag_re = re.compile ('(?i)<body([^>]*)>')
 126 html_tag_re = re.compile ('(?i)<html>')
 127 doctype_re = re.compile ('(?i)<!DOCTYPE')
 128 doctype = '<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">\n'
 129 css_re = re.compile ('(?i)<link rel="stylesheet" type="text/css" href="[^">]*?lilypond.css">')
 130 end_head_tag_re = re.compile ('(?i)</head>')
 131 css_link = '<link rel="stylesheet" type="text/css" href="%sDocumentation/lilypond.css">\n'
 132
 133 def add_header (s, prefix):
 134     """Add header (<body>, doctype and CSS)"""
 135     if header_tag_re.search (s) == None:
 136         body = '<body\\1>'
 137         (s, n) = body_tag_re.subn (body + header, s, 1)
 138         if not n:
 139             (s, n) = html_tag_re.subn ('<html>' + header, s, 1)
 140             if not n:
 141                 s = header + s
 142
 143         s = header_tag + '\n' + s
 144
 145         if doctype_re.search (s) == None:
 146             s = doctype + s
 147
 148         if css_re.search (s) == None:
 149             depth = (prefix.count ('/') - 1) * '../'
 150             s = end_head_tag_re.sub ((css_link % depth) + '</head>', s)
 151     return s
 152
 153 title_tag_re = re.compile ('.*?<title>(.*?)</title>', re.DOTALL)
 154 AT_web_title_re = re.compile ('@WEB-TITLE@')
 155
 156 def add_title (s):
 157     # urg
 158     # maybe find first node?
 159     fallback_web_title = '-- --'
 160     m = title_tag_re.match (s)
 161     if m:
 162         fallback_web_title = m.group (1)
 163     s = AT_web_title_re.sub (fallback_web_title, s)
 164     return s
 165
 166 footer_insert_re = re.compile ('<!--\s*FOOTER\s*-->')
 167 end_body_re = re.compile ('(?i)</body>')
 168 end_html_re = re.compile ('(?i)</html>')
 169
 170 def add_footer (s, footer_text):
 171     """add footer"""
 172     (s, n) = footer_insert_re.subn (footer_text + '\n' + '<!-- FOOTER -->', s, 1)
 173     if not n:
 174         (s, n) = end_body_re.subn (footer_text + '\n' + '</body>', s, 1)
 175     if not n:
 176         (s, n) = end_html_re.subn (footer_text + '\n' + '</html>', s, 1)
 177     if not n:
 178         s += footer_text + '\n'
 179     return s
 180
 181 def find_translations (prefix, lang_ext):
 182     """find available translations of a page"""
 183     available = []
 184     missing = []
 185     for l in langdefs.LANGUAGES:
 186         e = l.webext
 187         if lang_ext != e:
 188             if e in pages_dict[prefix]:
 189                 available.append (l)
 190             elif lang_ext == '' and l.enabled and reduce (operator.and_,
 191                                                           [not prefix.startswith (s)
 192                                                            for s in non_copied_pages]):
 193                 # English version of missing translated pages will be written
 194                 missing.append (e)
 195     return available, missing
 196
 197 online_links_re = re.compile ('''(href|src)=['"]\
 198 ((?!Compiling-from-source.html")[^/][.]*[^.:'"]*)\
 199 ([.]html|[.]png)(#[^"']*|)['"]''')
 200 offline_links_re = re.compile ('href=[\'"]\
 201 ((?!Compiling-from-source.html")[^/][.]*[^.:\'"]*)([.]html)(#[^"\']*|)[\'"]')
 202 big_page_name_re = re.compile ('''(.+?)-big-page''')
 203
 204 def process_i18n_big_page_links (match, prefix, lang_ext):
 205     big_page_name = big_page_name_re.match (match.group (1))
 206     if big_page_name:
 207         destination_path = os.path.normpath (os.path.join (os.path.dirname (prefix),
 208                                                            big_page_name.group (0)))
 209         if not lang_ext in pages_dict[destination_path]:
 210             return match.group (0)
 211     return 'href="' + match.group (1) + '.' + lang_ext \
 212         + match.group (2) + match.group (3) + '"'
 213
 214 def process_links (s, prefix, lang_ext, file_name, missing, target):
 215     page_flavors = {}
 216     if target == 'online':
 217         # Strip .html, .png suffix for auto language selection (content
 218         # negotiation).  The menu must keep the full extension, so do
 219         # this before adding the menu.
 220         page_flavors[file_name] = \
 221             [lang_ext, online_links_re.sub ('\\1="\\2\\4"', s)]
 222     elif target == 'offline':
 223         # in LANG doc index: don't rewrite .html suffixes
 224         # as not all .LANG.html pages exist;
 225         # the doc index should be translated and contain links with the right suffixes
 226         if prefix == 'Documentation/out-www/index':
 227             page_flavors[file_name] = [lang_ext, s]
 228         elif lang_ext == '':
 229             page_flavors[file_name] = [lang_ext, s]
 230             for e in missing:
 231                 page_flavors[langdefs.lang_file_name (prefix, e, '.html')] = \
 232                     [e, offline_links_re.sub ('href="\\1.' + e + '\\2\\3"', s)]
 233         else:
 234             # For saving bandwidth and disk space, we don't duplicate big pages
 235             # in English, so we must process translated big pages links differently.
 236             if 'big-page' in prefix:
 237                 page_flavors[file_name] = \
 238                     [lang_ext,
 239                      offline_links_re.sub \
 240                          (lambda match: process_i18n_big_page_links (match, prefix, lang_ext),
 241                           s)]
 242             else:
 243                 page_flavors[file_name] = \
 244                     [lang_ext,
 245                      offline_links_re.sub ('href="\\1.' + lang_ext + '\\2\\3"', s)]
 246     return page_flavors
 247
 248 def add_menu (page_flavors, prefix, available, target, translation):
 249     for k in page_flavors:
 250         language_menu = ''
 251         languages = ''
 252         if page_flavors[k][0] != '':
 253             t = translation[page_flavors[k][0]]
 254         else:
 255             t = _doc
 256         for lang in available:
 257             lang_file = lang.file_name (os.path.basename (prefix), '.html')
 258             if language_menu != '':
 259                 language_menu += ', '
 260             language_menu += '<a href="%s">%s</a>' % (lang_file, t (lang.name))
 261         if target == 'offline':
 262             browser_language = ''
 263         elif target == 'online':
 264             browser_language = t (browser_lang) % browser_language_url
 265         if language_menu:
 266             language_available = t (lang_available) % language_menu
 267             languages = LANGUAGES_TEMPLATE % vars ()
 268         page_flavors[k][1] = add_footer (page_flavors[k][1], languages)
 269     return page_flavors
 270
 271
 272 def process_html_files (package_name = '',
 273                         package_version = '',
 274                         target = 'offline',
 275                         name_filter = lambda s: s):
 276     """Add header, footer and tweak links to a number of HTML files
 277
 278     Arguments:
 279      package_name=NAME         set package_name to NAME
 280      package_version=VERSION   set package version to VERSION
 281      targets=offline|online    set page processing depending on the target
 282           offline is for reading HTML pages locally
 283           online is for hosting the HTML pages on a website with content
 284             negotiation
 285      name_filter               a HTML file name filter
 286     """
 287     translation = langdefs.translation
 288     localtime = time.strftime ('%c %Z', time.localtime (time.time ()))
 289
 290     if "http://" in mail_address:
 291         mail_address_url = mail_address
 292     else:
 293         mail_address_url= 'mailto:' + mail_address
 294
 295     versiontup = package_version.split ('.')
 296     branch_str = _doc ('stable-branch')
 297     if int (versiontup[1]) %  2:
 298         branch_str = _doc ('development-branch')
 299
 300     # Initialize dictionaries for string formatting
 301     subst = {}
 302     subst[''] = dict ([i for i in globals ().items() if type (i[1]) is str])
 303     subst[''].update (dict ([i for i in locals ().items() if type (i[1]) is str]))
 304     for l in translation:
 305         e = langdefs.LANGDICT[l].webext
 306         if e:
 307             subst[e] = {}
 308             for name in subst['']:
 309                 subst[e][name] = translation[l] (subst[''][name])
 310     # Do deeper string formatting as early as possible,
 311     # so only one '%' formatting pass is needed later
 312     for e in subst:
 313         subst[e]['footer_name_version'] = subst[e]['footer_name_version'] % subst[e]
 314         subst[e]['footer_report_errors'] = subst[e]['footer_report_errors'] % subst[e]
 315         subst[e]['footer_suggest_docs'] = subst[e]['footer_suggest_docs'] % subst[e]
 316
 317     for prefix, ext_list in pages_dict.items ():
 318         for lang_ext in ext_list:
 319             file_name = langdefs.lang_file_name (prefix, lang_ext, '.html')
 320             in_f = open (file_name)
 321             s = in_f.read()
 322             in_f.close()
 323
 324             s = s.replace ('%', '%%')
 325             s = hack_urls (s, prefix)
 326             s = add_header (s, prefix)
 327
 328             ### add footer
 329             if footer_tag_re.search (s) == None:
 330                 s = add_footer (s, footer_tag + footer)
 331
 332                 available, missing = find_translations (prefix, lang_ext)
 333                 page_flavors = process_links (s, prefix, lang_ext, file_name, missing, target)
 334                 # Add menu after stripping: must not have autoselection for language menu.
 335                 page_flavors = add_menu (page_flavors, prefix, available, target, translation)
 336             for k in page_flavors:
 337                 page_flavors[k][1] = page_flavors[k][1] % subst[page_flavors[k][0]]
 338                 out_f = open (name_filter (k), 'w')
 339                 out_f.write (page_flavors[k][1])
 340                 out_f.close()
 341         # if the page is translated, a .en.html symlink is necessary for content negotiation
 342         if target == 'online' and ext_list != ['']:
 343             os.symlink (os.path.basename (prefix) + '.html', name_filter (prefix + '.en.html'))