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 class="footer_version">
  38 %(footer_name_version)s
  39 </p>
  40 <p class="footer_report">
  41 %(footer_report_links)s
  42 </p>
  43 </div>
  44 '''
  45 footer_name_version = _doc ('This page is for %(package_name)s-%(package_version)s (%(branch_str)s).')
  46 # ugh, must not have "_doc" in strings because it is naively replaced with "_" in hacked gettext process
  47 footer_report_links = _doc ('Your <a href="%(suggest_Docs_url)s">suggestions for the documentation</a> are welcome, please report errors to our <a href="%(mail_address_url)s">bug list</a>.')
  48
  49
  50 mail_address = 'http://post.gmane.org/post.php?group=gmane.comp.gnu.lilypond.bugs'
  51 suggest_Docs_url = 'http://lilypond.org/web/devel/participating/documentation-adding'
  52
  53 header_tag = '<!-- header_tag -->'
  54 header_tag_re = re.compile (header_tag)
  55
  56 footer_tag = '<!-- footer_tag -->'
  57 footer_tag_re = re.compile (footer_tag)
  58
  59 lang_available = _doc ("Other languages: %s.")
  60 browser_lang = _doc ('About <A HREF="%s">automatic language selection</A>.')
  61 browser_language_url = "/web/about/browser-language"
  62
  63 LANGUAGES_TEMPLATE = '''
  64 <p id="languages">
  65  %(language_available)s
  66  <br/>
  67  %(browser_language)s
  68 </p>
  69 '''
  70
  71
  72 html_re = re.compile ('(.*?)(?:[.]([^/.]*))?[.]html$')
  73 pages_dict = {}
  74
  75 def build_pages_dict (filelist):
  76     """Build dictionary of available translations of each page"""
  77     global pages_dict
  78     for f in filelist:
  79         m = html_re.match (f)
  80         if m:
  81             g = m.groups()
  82             if len (g) <= 1 or g[1] == None:
  83                 e = ''
  84             else:
  85                 e = g[1]
  86             if not g[0] in pages_dict:
  87                 pages_dict[g[0]] = [e]
  88             else:
  89                 pages_dict[g[0]].append (e)
  90
  91 def source_links_replace (m, source_val):
  92     return 'href="' + os.path.join (source_val, m.group (1)) + '"'
  93
  94 splitted_docs_re = re.compile ('(input/lsr/out-www/lilypond-snippets|\
  95 Documentation/user/out-www/(lilypond|music-glossary|lilypond-program|\
  96 lilypond-learning))/')
  97
  98 snippets_ref_re = re.compile (r'href="(\.\./)?lilypond-snippets')
  99 user_ref_re = re.compile ('href="(?:\.\./)?lilypond\
 100 (-internals|-learning|-program|(?!-snippets))')
 101
 102 docindex_link_re = re.compile (r'href="index.html"')
 103
 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     # we also need to replace in the lsr, which is already processed above!
 120     if 'input/' in prefix or 'Documentation/topdocs' in prefix:
 121         # fix the link from the regtest, lsr and topdoc pages to the doc index
 122         # (rewrite prefix to obtain the relative path of the doc index page)
 123         rel_link = re.sub (r'out-www/.*$', '', prefix)
 124         rel_link = re.sub (r'[^/]*/', '../', rel_link)
 125         if 'input/regression' in prefix:
 126             indexfile = "Documentation/devel"
 127         else:
 128             indexfile = "index"
 129         s = docindex_link_re.sub ('href="' + rel_link + indexfile + '.html\"', s)
 130
 131     source_path = os.path.join (os.path.dirname (prefix), 'source')
 132     if not os.path.islink (source_path):
 133         return s
 134     source_val = os.readlink (source_path)
 135     return re.sub ('href="source/(.*?)"', lambda m: source_links_replace (m, source_val), s)
 136
 137 body_tag_re = re.compile ('(?i)<body([^>]*)>')
 138 html_tag_re = re.compile ('(?i)<html>')
 139 doctype_re = re.compile ('(?i)<!DOCTYPE')
 140 doctype = '<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">\n'
 141 css_re = re.compile ('(?i)<link rel="stylesheet" type="text/css" ([^>]*)href="[^">]*?lilypond.*\.css"([^>]*)>')
 142 end_head_tag_re = re.compile ('(?i)</head>')
 143 css_link = """    <link rel="stylesheet" type="text/css" title="Patrick McCarty's design" href="%(rel)sDocumentation/lilypond-mccarty.css">
 144     <link rel="alternate stylesheet" type="text/css" href="%(rel)sDocumentation/lilypond.css" title="Andrew Hawryluk's design">
 145     <link rel="alternate stylesheet" type="text/css" href="%(rel)sDocumentation/lilypond-blue.css" title="Kurt Kroon's blue design">
 146     <!--[if lte IE 7]>
 147     <link href="%(rel)sDocumentation/lilypond-ie-fixes.css" rel="stylesheet" type="text/css">
 148     <![endif]-->
 149 """
 150
 151
 152 def add_header (s, prefix):
 153     """Add header (<body>, doctype and CSS)"""
 154     if header_tag_re.search (s) == None:
 155         body = '<body\\1>'
 156         (s, n) = body_tag_re.subn (body + header, s, 1)
 157         if not n:
 158             (s, n) = html_tag_re.subn ('<html>' + header, s, 1)
 159             if not n:
 160                 s = header + s
 161
 162         s = header_tag + '\n' + s
 163
 164         if doctype_re.search (s) == None:
 165             s = doctype + s
 166
 167         if css_re.search (s) == None:
 168             depth = (prefix.count ('/') - 1) * '../'
 169             s = end_head_tag_re.sub ((css_link % {'rel': depth}) + '</head>', s)
 170     return s
 171
 172 title_tag_re = re.compile ('.*?<title>(.*?)</title>', re.DOTALL)
 173 AT_web_title_re = re.compile ('@WEB-TITLE@')
 174
 175 def add_title (s):
 176     # urg
 177     # maybe find first node?
 178     fallback_web_title = '-- --'
 179     m = title_tag_re.match (s)
 180     if m:
 181         fallback_web_title = m.group (1)
 182     s = AT_web_title_re.sub (fallback_web_title, s)
 183     return s
 184
 185 footer_insert_re = re.compile ('<!--\s*FOOTER\s*-->')
 186 end_body_re = re.compile ('(?i)</body>')
 187 end_html_re = re.compile ('(?i)</html>')
 188
 189 def add_footer (s, footer_text):
 190     """add footer"""
 191     (s, n) = footer_insert_re.subn (footer_text + '\n' + '<!-- FOOTER -->', s, 1)
 192     if not n:
 193         (s, n) = end_body_re.subn (footer_text + '\n' + '</body>', s, 1)
 194     if not n:
 195         (s, n) = end_html_re.subn (footer_text + '\n' + '</html>', s, 1)
 196     if not n:
 197         s += footer_text + '\n'
 198     return s
 199
 200 def find_translations (prefix, lang_ext):
 201     """find available translations of a page"""
 202     available = []
 203     missing = []
 204     for l in langdefs.LANGUAGES:
 205         e = l.webext
 206         if lang_ext != e:
 207             if e in pages_dict[prefix]:
 208                 available.append (l)
 209             elif lang_ext == '' and l.enabled and reduce (operator.and_,
 210                                                           [not prefix.startswith (s)
 211                                                            for s in non_copied_pages]):
 212                 # English version of missing translated pages will be written
 213                 missing.append (e)
 214     return available, missing
 215
 216 online_links_re = re.compile ('''(href|src)=['"]\
 217 ((?!Compiling-from-source.html")[^/][.]*[^.:'"]*)\
 218 ([.]html)(#[^"']*|)['"]''')
 219 offline_links_re = re.compile ('href=[\'"]\
 220 ((?!Compiling-from-source.html")[^/][.]*[^.:\'"]*)([.]html)(#[^"\']*|)[\'"]')
 221 big_page_name_re = re.compile ('''(.+?)-big-page''')
 222
 223 def process_i18n_big_page_links (match, prefix, lang_ext):
 224     big_page_name = big_page_name_re.match (match.group (1))
 225     if big_page_name:
 226         destination_path = os.path.normpath (os.path.join (os.path.dirname (prefix),
 227                                                            big_page_name.group (0)))
 228         if not lang_ext in pages_dict[destination_path]:
 229             return match.group (0)
 230     return 'href="' + match.group (1) + '.' + lang_ext \
 231         + match.group (2) + match.group (3) + '"'
 232
 233 def process_links (s, prefix, lang_ext, file_name, missing, target):
 234     page_flavors = {}
 235     if target == 'online':
 236         # Strip .html, suffix for auto language selection (content
 237         # negotiation).  The menu must keep the full extension, so do
 238         # this before adding the menu.
 239         page_flavors[file_name] = \
 240             [lang_ext, online_links_re.sub ('\\1="\\2\\4"', s)]
 241     elif target == 'offline':
 242         # in LANG doc index: don't rewrite .html suffixes
 243         # as not all .LANG.html pages exist;
 244         # the doc index should be translated and contain links with the right suffixes
 245         if prefix == 'Documentation/out-www/index':
 246             page_flavors[file_name] = [lang_ext, s]
 247         elif lang_ext == '':
 248             page_flavors[file_name] = [lang_ext, s]
 249             for e in missing:
 250                 page_flavors[langdefs.lang_file_name (prefix, e, '.html')] = \
 251                     [e, offline_links_re.sub ('href="\\1.' + e + '\\2\\3"', s)]
 252         else:
 253             # For saving bandwidth and disk space, we don't duplicate big pages
 254             # in English, so we must process translated big pages links differently.
 255             if 'big-page' in prefix:
 256                 page_flavors[file_name] = \
 257                     [lang_ext,
 258                      offline_links_re.sub \
 259                          (lambda match: process_i18n_big_page_links (match, prefix, lang_ext),
 260                           s)]
 261             else:
 262                 page_flavors[file_name] = \
 263                     [lang_ext,
 264                      offline_links_re.sub ('href="\\1.' + lang_ext + '\\2\\3"', s)]
 265     return page_flavors
 266
 267 def add_menu (page_flavors, prefix, available, target, translation):
 268     for k in page_flavors:
 269         language_menu = ''
 270         languages = ''
 271         if page_flavors[k][0] != '':
 272             t = translation[page_flavors[k][0]]
 273         else:
 274             t = _doc
 275         for lang in available:
 276             lang_file = lang.file_name (os.path.basename (prefix), '.html')
 277             if language_menu != '':
 278                 language_menu += ', '
 279             language_menu += '<a href="%s">%s</a>' % (lang_file, t (lang.name))
 280         if target == 'offline':
 281             browser_language = ''
 282         elif target == 'online':
 283             browser_language = t (browser_lang) % browser_language_url
 284         if language_menu:
 285             language_available = t (lang_available) % language_menu
 286             languages = LANGUAGES_TEMPLATE % vars ()
 287         page_flavors[k][1] = add_footer (page_flavors[k][1], languages)
 288     return page_flavors
 289
 290
 291 def process_html_files (package_name = '',
 292                         package_version = '',
 293                         target = 'offline',
 294                         name_filter = lambda s: s):
 295     """Add header, footer and tweak links to a number of HTML files
 296
 297     Arguments:
 298      package_name=NAME         set package_name to NAME
 299      package_version=VERSION   set package version to VERSION
 300      targets=offline|online    set page processing depending on the target
 301           offline is for reading HTML pages locally
 302           online is for hosting the HTML pages on a website with content
 303             negotiation
 304      name_filter               a HTML file name filter
 305     """
 306     translation = langdefs.translation
 307     localtime = time.strftime ('%c %Z', time.localtime (time.time ()))
 308
 309     if "http://" in mail_address:
 310         mail_address_url = mail_address
 311     else:
 312         mail_address_url= 'mailto:' + mail_address
 313
 314     versiontup = package_version.split ('.')
 315     branch_str = _doc ('stable-branch')
 316     if int (versiontup[1]) %  2:
 317         branch_str = _doc ('development-branch')
 318
 319     # Initialize dictionaries for string formatting
 320     subst = {}
 321     subst[''] = dict ([i for i in globals ().items() if type (i[1]) is str])
 322     subst[''].update (dict ([i for i in locals ().items() if type (i[1]) is str]))
 323     for l in translation:
 324         e = langdefs.LANGDICT[l].webext
 325         if e:
 326             subst[e] = {}
 327             for name in subst['']:
 328                 subst[e][name] = translation[l] (subst[''][name])
 329     # Do deeper string formatting as early as possible,
 330     # so only one '%' formatting pass is needed later
 331     for e in subst:
 332         subst[e]['footer_name_version'] = subst[e]['footer_name_version'] % subst[e]
 333         subst[e]['footer_report_links'] = subst[e]['footer_report_links'] % subst[e]
 334
 335     for prefix, ext_list in pages_dict.items ():
 336         for lang_ext in ext_list:
 337             file_name = langdefs.lang_file_name (prefix, lang_ext, '.html')
 338             in_f = open (file_name)
 339             s = in_f.read()
 340             in_f.close()
 341
 342             s = s.replace ('%', '%%')
 343             s = hack_urls (s, prefix)
 344             s = add_header (s, prefix)
 345
 346             ### add footer
 347             if footer_tag_re.search (s) == None:
 348                 s = add_footer (s, footer_tag + footer)
 349
 350                 available, missing = find_translations (prefix, lang_ext)
 351                 page_flavors = process_links (s, prefix, lang_ext, file_name, missing, target)
 352                 # Add menu after stripping: must not have autoselection for language menu.
 353                 page_flavors = add_menu (page_flavors, prefix, available, target, translation)
 354             for k in page_flavors:
 355                 page_flavors[k][1] = page_flavors[k][1] % subst[page_flavors[k][0]]
 356                 out_f = open (name_filter (k), 'w')
 357                 out_f.write (page_flavors[k][1])
 358                 out_f.close()
 359         # if the page is translated, a .en.html symlink is necessary for content negotiation
 360         if target == 'online' and ext_list != ['']:
 361             os.symlink (os.path.basename (prefix) + '.html', name_filter (prefix + '.en.html'))