4 Postprocess HTML files.
13 # This is to try to make the docball not too big with almost duplicate files
15 non_copied_pages = ['Documentation/user/out-www/lilypond-big-page',
16 'Documentation/user/out-www/lilypond-internals-big-page',
17 'Documentation/user/out-www/lilypond-learning-big-page',
18 'Documentation/user/out-www/lilypond-program-big-page',
19 'Documentation/user/out-www/music-glossary-big-page',
21 'Documentation/topdocs',
22 'Documentation/bibliography',
23 'Documentation/out-www/THANKS',
24 'Documentation/out-www/DEDICATION',
25 'Documentation/out-www/devel',
35 <div style="background-color: #e8ffe8; padding: 2; border: #c0ffc0 1px solid;">
38 %(footer_name_version)s
41 %(footer_report_errors)s </address>
43 %(footer_suggest_docs)s
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.')
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'
56 header_tag = '<!-- header_tag -->'
57 header_tag_re = re.compile (header_tag)
59 footer_tag = '<!-- footer_tag -->'
60 footer_tag_re = re.compile (footer_tag)
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"
66 LANGUAGES_TEMPLATE = '''
68 %(language_available)s
75 html_re = re.compile ('(.*?)(?:[.]([^/.]*))?[.]html$')
78 def build_pages_dict (filelist):
79 """Build dictionary of available translations of each page"""
85 if len (g) <= 1 or g[1] == None:
89 if not g[0] in pages_dict:
90 pages_dict[g[0]] = [e]
92 pages_dict[g[0]].append (e)
94 def source_links_replace (m, source_val):
95 return 'href="' + os.path.join (source_val, m.group (1)) + '"'
97 splitted_docs_re = re.compile ('(input/lsr/out-www/lilypond-snippets|Documentation/user/out-www/(lilypond|music-glossary|lilypond-program|lilypond-learning))/')
99 snippets_ref_re = re.compile (r'href="(\.\./)?lilypond-snippets')
100 user_ref_re = re.compile (r'href="(?:\.\./)?lilypond(-internals|-learning|-program|(?!-snippets))')
102 ## Windows does not support symlinks.
103 # This function avoids creating symlinks for splitted HTML manuals
104 # Get rid of symlinks in GNUmakefile.in (local-WWW-post)
105 # this also fixes missing PNGs only present in translated docs
106 def hack_urls (s, prefix):
107 if splitted_docs_re.match (prefix):
108 s = re.sub ('(href|src)="(../lily-.*?|.*?[.]png)"', '\\1="../\\2"', s)
110 # fix xrefs between documents in different directories ad hoc
111 if 'user/out-www/lilypond' in prefix:
112 s = snippets_ref_re.sub ('href="source/input/lsr/lilypond-snippets', s)
113 elif 'input/lsr' in prefix:
114 s = user_ref_re.sub ('href="source/Documentation/user/lilypond\\1', s)
116 source_path = os.path.join (os.path.dirname (prefix), 'source')
117 if not os.path.islink (source_path):
119 source_val = os.readlink (source_path)
120 return re.sub ('href="source/(.*?)"', lambda m: source_links_replace (m, source_val), s)
122 body_tag_re = re.compile ('(?i)<body([^>]*)>')
123 html_tag_re = re.compile ('(?i)<html>')
124 doctype_re = re.compile ('(?i)<!DOCTYPE')
125 doctype = '<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">\n'
128 """Add header (<body> and doctype)"""
129 if header_tag_re.search (s) == None:
130 body = '<body bgcolor="white" text="black" \\1>'
131 (s, n) = body_tag_re.subn (body + header, s, 1)
133 (s, n) = html_tag_re.subn ('<html>' + header, s, 1)
137 s = header_tag + '\n' + s
139 if doctype_re.search (s) == None:
143 title_tag_re = re.compile ('.*?<title>(.*?)</title>', re.DOTALL)
144 AT_web_title_re = re.compile ('@WEB-TITLE@')
148 # maybe find first node?
149 fallback_web_title = '-- --'
150 m = title_tag_re.match (s)
152 fallback_web_title = m.group (1)
153 s = AT_web_title_re.sub (fallback_web_title, s)
156 footer_insert_re = re.compile ('<!--\s*FOOTER\s*-->')
157 end_body_re = re.compile ('(?i)</body>')
158 end_html_re = re.compile ('(?i)</html>')
160 def add_footer (s, footer_text):
162 (s, n) = footer_insert_re.subn (footer_text + '\n' + '<!-- FOOTER -->', s, 1)
164 (s, n) = end_body_re.subn (footer_text + '\n' + '</body>', s, 1)
166 (s, n) = end_html_re.subn (footer_text + '\n' + '</html>', s, 1)
168 s += footer_text + '\n'
171 def find_translations (prefix, lang_ext):
172 """find available translations of a page"""
175 for l in langdefs.LANGUAGES:
178 if e in pages_dict[prefix]:
180 elif lang_ext == '' and l.enabled and reduce (operator.and_, [not prefix.startswith (s) for s in non_copied_pages]):
181 # English version of missing translated pages will be written
183 return available, missing
185 online_links_re = re.compile ('''(href|src)=['"]([^/][.]*[^.:'"]*)([.]html|[.]png)(#[^"']*|)['"]''')
186 offline_links_re = re.compile ('''href=['"]([^/][.]*[^.:'"]*)([.]html)(#[^"']*|)['"]''')
187 big_page_name_re = re.compile ('''(.+?)-big-page''')
189 def process_i18n_big_page_links (match, prefix, lang_ext):
190 big_page_name = big_page_name_re.match (match.group (1))
192 destination_path = os.path.normpath (os.path.join (os.path.dirname (prefix),
193 big_page_name.group (0)))
194 if not lang_ext in pages_dict[destination_path]:
195 return match.group (0)
196 return 'href="' + match.group (1) + '.' + lang_ext \
197 + match.group (2) + match.group (3) + '"'
199 def process_links (s, prefix, lang_ext, file_name, missing, target):
201 if target == 'online':
202 # Strip .html, .png suffix for auto language selection (content
203 # negotiation). The menu must keep the full extension, so do
204 # this before adding the menu.
205 page_flavors[file_name] = \
206 [lang_ext, online_links_re.sub ('\\1="\\2\\4"', s)]
207 elif target == 'offline':
208 # in LANG doc index: don't rewrite .html suffixes
209 # as not all .LANG.html pages exist;
210 # the doc index should be translated and contain links with the right suffixes
211 if prefix == 'Documentation/out-www/index':
212 page_flavors[file_name] = [lang_ext, s]
214 page_flavors[file_name] = [lang_ext, s]
216 page_flavors[langdefs.lang_file_name (prefix, e, '.html')] = \
217 [e, offline_links_re.sub ('href="\\1.' + e + '\\2\\3"', s)]
219 # For saving bandwidth and disk space, we don't duplicate big pages
220 # in English, so we must process translated big pages links differently.
221 if 'big-page' in prefix:
222 page_flavors[file_name] = \
224 offline_links_re.sub \
225 (lambda match: process_i18n_big_page_links (match, prefix, lang_ext),
228 page_flavors[file_name] = \
230 offline_links_re.sub ('href="\\1.' + lang_ext + '\\2\\3"', s)]
233 def add_menu (page_flavors, prefix, available, target, translation):
234 for k in page_flavors:
237 if page_flavors[k][0] != '':
238 t = translation[page_flavors[k][0]]
241 for lang in available:
242 lang_file = lang.file_name (os.path.basename (prefix), '.html')
243 if language_menu != '':
244 language_menu += ', '
245 language_menu += '<a href="%s">%s</a>' % (lang_file, t (lang.name))
246 if target == 'offline':
247 browser_language = ''
248 elif target == 'online':
249 browser_language = t (browser_lang) % browser_language_url
251 language_available = t (lang_available) % language_menu
252 languages = LANGUAGES_TEMPLATE % vars ()
253 page_flavors[k][1] = add_footer (page_flavors[k][1], languages)
257 def process_html_files (package_name = '',
258 package_version = '',
260 name_filter = lambda s: s):
261 """Add header, footer and tweak links to a number of HTML files
264 package_name=NAME set package_name to NAME
265 package_version=VERSION set package version to VERSION
266 targets=offline|online set page processing depending on the target
267 offline is for reading HTML pages locally
268 online is for hosting the HTML pages on a website with content
270 name_filter a HTML file name filter
272 translation = langdefs.translation
273 localtime = time.strftime ('%c %Z', time.localtime (time.time ()))
275 if "http://" in mail_address:
276 mail_address_url = mail_address
278 mail_address_url= 'mailto:' + mail_address
280 versiontup = package_version.split ('.')
281 branch_str = _doc ('stable-branch')
282 if int (versiontup[1]) % 2:
283 branch_str = _doc ('development-branch')
285 # Initialize dictionaries for string formatting
287 subst[''] = dict ([i for i in globals ().items() if type (i[1]) is str])
288 subst[''].update (dict ([i for i in locals ().items() if type (i[1]) is str]))
289 for l in translation:
290 e = langdefs.LANGDICT[l].webext
293 for name in subst['']:
294 subst[e][name] = translation[l] (subst[''][name])
295 # Do deeper string formatting as early as possible,
296 # so only one '%' formatting pass is needed later
298 subst[e]['footer_name_version'] = subst[e]['footer_name_version'] % subst[e]
299 subst[e]['footer_report_errors'] = subst[e]['footer_report_errors'] % subst[e]
300 subst[e]['footer_suggest_docs'] = subst[e]['footer_suggest_docs'] % subst[e]
302 for prefix, ext_list in pages_dict.items ():
303 for lang_ext in ext_list:
304 file_name = langdefs.lang_file_name (prefix, lang_ext, '.html')
305 in_f = open (file_name)
309 s = s.replace ('%', '%%')
310 s = hack_urls (s, prefix)
314 if footer_tag_re.search (s) == None:
315 s = add_footer (s, footer_tag + footer)
317 available, missing = find_translations (prefix, lang_ext)
318 page_flavors = process_links (s, prefix, lang_ext, file_name, missing, target)
319 # Add menu after stripping: must not have autoselection for language menu.
320 page_flavors = add_menu (page_flavors, prefix, available, target, translation)
321 for k in page_flavors:
322 page_flavors[k][1] = page_flavors[k][1] % subst[page_flavors[k][0]]
323 out_f = open (name_filter (k), 'w')
324 out_f.write (page_flavors[k][1])
326 # if the page is translated, a .en.html symlink is necessary for content negotiation
327 if target == 'online' and ext_list != ['']:
328 os.symlink (os.path.basename (prefix) + '.html', name_filter (prefix + '.en.html'))