3 # Part of Objavi2, which turns html manuals into books
5 # Copyright (C) 2009 Douglas Bagnall
7 # This program is free software; you can redistribute it and/or modify
8 # it under the terms of the GNU General Public License as published by
9 # the Free Software Foundation; either version 2 of the License, or
10 # (at your option) any later version.
12 # This program is distributed in the hope that it will be useful,
13 # but WITHOUT ANY WARRANTY; without even the implied warranty of
14 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 # GNU General Public License for more details.
17 # You should have received a copy of the GNU General Public License along
18 # with this program; if not, write to the Free Software Foundation, Inc.,
19 # 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
21 """Make a pdf from the specified book."""
22 from __future__
import with_statement
27 from urllib2
import urlopen
28 from getopt
import gnu_getopt
30 from fmbook
import log
, Book
33 from config
import SERVER_DEFAULTS
, DEFAULT_SERVER
35 FORM_TEMPLATE
= os
.path
.abspath('templates/form.html')
36 PROGRESS_TEMPLATE
= os
.path
.abspath('templates/progress.html')
39 #spaces?, digits!, dot?, digits?, spaces?
40 #return re.compile(r'^\s*[+-]?\d+\.?\d*\s*$').match
47 def isfloat_or_auto(s
):
48 return isfloat(s
) or s
.lower() in ('', 'auto')
51 # 10 or 13 digits with any number of hyphens, perhaps with check-digit missing
53 return (re
.match(r
'^\d+[\dXx*]$', s
) and len(s
) in (9, 10, 12, 13))
56 # ARG_VALIDATORS is a mapping between the expected cgi arguments and
57 # functions to validate their values. (None means no validation).
59 "book": re
.compile(r
'^(\w+/?)*\w+$').match
, # can be: BlahBlah/Blah_Blah
60 "css": None, # an url, empty (for default), or css content
61 "title": lambda x
: len(x
) < 999,
62 #"header": None, # header text, UNUSED
64 "license": config
.LICENSES
.__contains
__,
65 "server": SERVER_DEFAULTS
.__contains
__,
66 "engine": config
.ENGINES
.__contains
__,
67 "booksize": config
.PAGE_SIZE_DATA
.__contains
__,
68 "page_width": isfloat
,
69 "page_height": isfloat
,
70 "gutter": isfloat_or_auto
,
71 "top_margin": isfloat_or_auto
,
72 "side_margin": isfloat_or_auto
,
73 "bottom_margin": isfloat_or_auto
,
74 "columns": isfloat_or_auto
,
75 "column_margin": isfloat_or_auto
,
76 "cgi-context": lambda x
: x
.lower() in '1true0false',
77 "mode": config
.CGI_MODES
.__contains
__,
78 "pdftype": lambda x
: config
.CGI_MODES
.get(x
, [False])[0],
79 "rotate": u
"yes".__eq
__,
80 "grey_scale": u
"yes".__eq
__,
83 __doc__
+= '\nValid arguments are: %s.\n' % ', '.join(ARG_VALIDATORS
.keys())
86 """Read and validate CGI or commandline arguments, putting the
87 good ones into the returned dictionary. Command line arguments
88 should be in the form --title='A Book'.
90 query
= cgi
.FieldStorage()
91 options
, args
= gnu_getopt(sys
.argv
[1:], '', [x
+ '=' for x
in ARG_VALIDATORS
])
92 options
= dict(options
)
94 for key
, validator
in ARG_VALIDATORS
.items():
95 value
= query
.getfirst(key
, options
.get('--' + key
, None))
96 log('%s: %s' % (key
, value
), debug
='STARTUP')
98 if validator
is not None and not validator(value
):
99 log("argument '%s' is not valid ('%s')" % (key
, value
))
103 log(data
, debug
='STARTUP')
106 def get_server_list():
107 return sorted(SERVER_DEFAULTS
.keys())
109 def get_book_list(server
):
110 """Ask the server for a list of books. Floss Manual TWikis keep such a list at
111 /bin/view/TWiki/WebLeftBarWebsList?skin=text but it needs a bit of processing
113 If BOOK_LIST_CACHE is non-zero, the book list won't be re-fetched
114 in that many seconds, rather it will be read from disk.
116 if config
.BOOK_LIST_CACHE
:
117 cache_name
= os
.path
.join(config
.BOOK_LIST_CACHE_DIR
, '%s.booklist' % server
)
118 if (os
.path
.exists(cache_name
) and
119 os
.stat(cache_name
).st_mtime
+ config
.BOOK_LIST_CACHE
> time
.time()):
125 url
= 'http://%s/bin/view/TWiki/WebLeftBarWebsList?skin=text' % server
131 items
= sorted(re
.findall(r
'/bin/view/([\w/]+)/WebHome', s
))
132 if config
.BOOK_LIST_CACHE
:
133 f
= open(cache_name
, 'w')
134 f
.write('\n'.join(items
))
139 #order by increasing areal size.
140 def calc_size(name
, pointsize
, klass
):
142 mmx
= pointsize
[0] * config
.POINT_2_MM
143 mmy
= pointsize
[1] * config
.POINT_2_MM
144 return (mmx
* mmy
, name
, klass
,
145 '%s (%dmm x %dmm)' % (name
, mmx
, mmy
))
147 return (0, name
, klass
, name
) # presumably 'custom'
149 return [x
[1:] for x
in sorted(calc_size(k
, v
.get('pointsize'), v
.get('class', ''))
150 for k
, v
in config
.PAGE_SIZE_DATA
.iteritems())
154 def optionise(items
, default
=None):
155 """Make a list of strings into an html option string, as would fit
156 inside <select> tags."""
159 if isinstance(x
, str):
162 # couple: value, name
164 options
.append('<option selected="selected" value="%s">%s</option>' % x
)
166 options
.append('<option value="%s">%s</option>' % x
)
168 # triple: value, class, name
170 options
.append('<option selected="selected" value="%s" class="%s">%s</option>' % x
)
172 options
.append('<option value="%s" class="%s">%s</option>' % x
)
174 return '\n'.join(options
)
177 """Make a list of strings into html <li> items, to fit in a <ul>
179 return '\n'.join('<li>%s</li>' % x
for x
in items
)
182 def get_default_css(server
=DEFAULT_SERVER
, mode
='book'):
183 """Get the default CSS text for the selected server"""
185 cssfile
= SERVER_DEFAULTS
[server
]['css-%s' % mode
]
193 """Links to various example pdfs."""
195 for script
in os
.listdir(config
.FONT_EXAMPLE_SCRIPT_DIR
):
196 if not script
.isalnum():
197 log("warning: font-sample %s won't work; skipping" % script
)
199 links
.append('<a href="%s?script=%s">%s</a>' % (config
.FONT_LIST_URL
, script
, script
))
203 def make_progress_page(book
, bookname
, mode
):
204 """Return a function that will notify the user of progress. In
205 CGI context this means making an html page to display the
206 messages, which are then sent as javascript snippets on the same
209 return lambda message
: '******* got message "%s"' %message
211 f
= open(PROGRESS_TEMPLATE
)
214 progress_list
= ''.join('<li id="%s">%s</li>\n' % x
[:2] for x
in config
.PROGRESS_POINTS
219 'bookname': bookname
,
220 'progress_list': progress_list
,
223 def progress_notifier(message
):
224 print ('<script type="text/javascript">\n'
225 'objavi_show_progress("%s");\n'
226 '</script>' % message
228 if message
== 'finished':
229 print '</body></html>'
231 return progress_notifier
233 def make_book_name(book
, server
, suffix
='.pdf'):
234 lang
= SERVER_DEFAULTS
.get(server
, SERVER_DEFAULTS
[DEFAULT_SERVER
])['lang']
235 book
= ''.join(x
for x
in book
if x
.isalnum())
236 return '%s-%s-%s%s' % (book
, lang
,
237 time
.strftime('%Y.%m.%d-%H.%M.%S'),
241 def get_page_settings(args
):
242 """Find the size and any optional layout settings.
244 args['booksize'] is either a keyword describing a size or
245 'custom'. If it is custom, the form is inspected for specific
246 dimensions -- otherwise these are ignored.
248 The margins, gutter, number of columns, and column
249 margins all set themselves automatically based on the page
250 dimensions, but they can be overridden. Any that are are
252 # get all the values including sizes first
253 # the sizes are found as 'page_width' and 'page_height',
254 # but the Book class expects them as a 'pointsize' tuple, so
255 # they are easily ignored.
257 for k
, extrema
in config
.PAGE_EXTREMA
.iteritems():
259 v
= float(args
.get(k
))
260 except (ValueError, TypeError):
261 #log("don't like %r as a float value for %s!" % (args.get(k), k))
263 min_val
, max_val
, multiplier
= extrema
264 if v
< min_val
or v
> max_val
:
265 log('rejecting %s: outside %s' % (v
,) + extrema
)
267 log('found %s=%s' % (k
, v
))
268 settings
[k
] = v
* multiplier
#convert to points in many cases
270 # now if args['size'] is not 'custom', the width and height found
272 size
= args
.get('booksize', config
.DEFAULT_SIZE
)
273 settings
.update(config
.PAGE_SIZE_DATA
[size
])
275 #if args['mode'] is 'newspaper', then the number of columns is
276 #automatically determined unless set -- otherwise default is 1.
277 if args
['mode'] == 'newspaper' and settings
.get('columns') is None:
278 settings
['columns'] = 'auto'
280 if args
.get('grey_scale'):
281 settings
['grey_scale'] = True
284 #will raise KeyError if width, height aren't set
285 settings
['pointsize'] = (settings
['page_width'], settings
['page_height'])
286 del settings
['page_width']
287 del settings
['page_height']
289 settings
['engine'] = args
.get('engine', config
.DEFAULT_ENGINE
)
293 def output_and_exit(f
):
294 """Decorator: prefix function output with http headers and exit
295 immediately after."""
298 print "Content-type: text/html; charset=utf-8\n"
304 def mode_booklist(args
):
305 print optionise(get_book_list(args
.get('server', config
.DEFAULT_SERVER
)),
306 default
=args
.get('book'))
310 #XX sending as text/html, but it doesn't really matter
311 print get_default_css(args
.get('server', config
.DEFAULT_SERVER
), args
.get('pdftype', 'book'))
316 f
= open(FORM_TEMPLATE
)
319 f
= open(config
.FONT_LIST_INCLUDE
)
320 font_list
= [x
.strip() for x
in f
if x
.strip()]
322 server
= args
.get('server', config
.DEFAULT_SERVER
)
323 book
= args
.get('book')
324 size
= args
.get('booksize', config
.DEFAULT_SIZE
)
325 engine
= args
.get('engine', config
.DEFAULT_ENGINE
)
327 'server_options': optionise(get_server_list(), default
=server
),
328 'book_options': optionise(get_book_list(server
), default
=book
),
329 'size_options': optionise(get_size_list(), default
=size
),
330 'engines': optionise(config
.ENGINES
.keys(), default
=engine
),
331 'pdf_types': optionise(sorted(k
for k
, v
in config
.CGI_MODES
.iteritems() if v
[0])),
332 'css': get_default_css(server
),
333 'font_links': listify(font_links()),
334 'font_list': listify(font_list
),
335 'default_license' : config
.DEFAULT_LICENSE
,
336 'licenses' : optionise(config
.LICENSES
, default
=config
.DEFAULT_LICENSE
),
342 for id, title
, type, source
, classes
, epilogue
in config
.FORM_INPUTS
:
343 val
= d
.get(source
, '')
344 e
= config
.FORM_ELEMENT_TYPES
[type] % locals()
345 form
.append('\n<div id="%(id)s_div" class="form-item %(classes)s">\n'
346 '<div class="input_title">%(title)s</div>\n'
347 '<div class="input_contents"> %(e)s %(epilogue)s\n</div>'
348 '</div>\n' % locals())
351 _valid_inputs
= set(ARG_VALIDATORS
)
352 _form_inputs
= set(x
[0] for x
in config
.FORM_INPUTS
if x
[2] != 'ul')
353 log("valid but not used inputs: %s" % (_valid_inputs
- _form_inputs
))
354 log("invalid form inputs: %s" % (_form_inputs
- _valid_inputs
))
356 print template
% {'form': ''.join(form
)}
361 # so we're making a pdf.
362 mode
= args
.get('mode', 'book')
363 bookid
= args
.get('book')
364 server
= args
.get('server', config
.DEFAULT_SERVER
)
365 engine
= args
.get('engine', config
.DEFAULT_ENGINE
)
366 page_settings
= get_page_settings(args
)
367 bookname
= make_book_name(bookid
, server
)
368 progress_bar
= make_progress_page(bookid
, bookname
, mode
)
370 with
Book(bookid
, server
, bookname
, page_settings
=page_settings
, engine
=engine
,
371 watcher
=progress_bar
, isbn
=args
.get('isbn'),
372 license
=args
.get('license')) as book
:
376 book
.set_title(args
.get('title'))
377 book
.add_css(args
.get('css'), mode
)
378 book
.add_section_titles()
382 elif mode
in ('web', 'newspaper'):
383 book
.make_simple_pdf(mode
)
388 book
.notify_watcher('finished')
390 #These ones are similar enought to be handled by the one function
391 mode_newspaper
= mode_book
396 def mode_openoffice(args
):
397 """Make an openoffice document. A whole lot of the inputs have no
399 bookid
= args
.get('book')
400 server
= args
.get('server', config
.DEFAULT_SERVER
)
401 #page_settings = get_page_settings(args)
402 bookname
= make_book_name(bookid
, server
, '.odt')
403 progress_bar
= make_progress_page(bookid
, bookname
, 'openoffice')
405 with
Book(bookid
, server
, bookname
,
406 watcher
=progress_bar
, isbn
=args
.get('isbn'),
407 license
=args
.get('license')) as book
:
411 book
.set_title(args
.get('title'))
412 book
.add_css(args
.get('css'), 'openoffice')
413 book
.add_section_titles()
415 book
.notify_watcher('finished')
420 mode
= args
.get('mode')
421 if mode
is None and 'book' in args
:
425 CGI_CONTEXT
= 'SERVER_NAME' in os
.environ
or args
.get('cgi-context', 'no').lower() in '1true'
427 if not args
and not CGI_CONTEXT
:
431 output_function
= globals().get('mode_%s' % mode
, mode_form
)
432 output_function(args
)
434 if __name__
== '__main__':
435 if config
.CGITB_DOMAINS
and os
.environ
.get('REMOTE_ADDR') in config
.CGITB_DOMAINS
: