# -*- coding: iso-8859-1 -*- r""" = pdf action = [[TableOfContents]] == Purpose == This action is used to get a pdf copy of the page. It does an on-the-fly, server-side, html to pdf conversion using ''htmldoc'', and sends an `application/pdf` content to the web browser. == Calling sequence == {{{ http://my.server/MyWiki/MyPage?action=pdf http://my.server/MyWiki/MyPage?action=pdf&style=NAME http://my.server/MyWiki/MyPage?action=pdf&rev=NUM http://my.server/MyWiki/MyPage?action=pdf&pageaction=ACTION http://my.server/MyWiki/MyPage?action=pdf&help=1}}} style::: `page` or `book`. These styles can be redefined or extended in the wiki config (see below). Asking for an unknown style will display an error and list available styles. rev::: Specifies a page version number. pageaction::: Specifies an action for rendering the page. By default and if omitted, the `print` action will be used. help::: Displays this help. == Procedure == Usage: * Append `&action=pdf` to any URL * or insert the macro `[[pdf]]` (coming soon!) in a page, which will insert a small control panel Hints on the `book` style: * Have at least one heading, else ''htmldoc'' will return an error. * ''htmldoc'' will ignore any content before the first heading. * ''htmldoc'' automatically generates a table of contents which is looking nicer (has page numbers). To avoid double TOCs, place the `[[TableOfContents]]` macro at start of your page, before any heading. This way, the output of the macro will be ignored by ''htmldoc''. Suggestions: * Modify your themes to include a pdf link/icon besides print. * Modify the `AttachFile` standard action to add a pdf link. == Installation == 1. install ''htmldoc'' on the wiki server (on a Debian box: `apt-get install htmldoc`) 1. copy the [http://moinmoin.wikiwikiweb.de/ActionMarket?action=AttachFile&do=get&target=pdf.py pdf.py] action to the `data/plugin/action` directory of your wiki == Optional: customizing styles == To install new styles, or modify the default ones, edit `wikiconfig.py` and add: {{{ htmldoc_styles = { "page": "--verbose --no-localfiles --no-compression --jpeg " \ "--header t.D --footer ./. --size a4 --left 0.5in " \ "--webpage", "book": "--verbose --no-localfiles --no-compression --jpeg " \ "--header t.D --footer ./. --size a4 --left 0.5in", } }}} By this means, you can change the page format from `a4` to `letter`, or add new styles to the dictionary. For a complete list of available ''htmldoc'' options, see http://www.easysw.com/htmldoc/docfiles/8-cmdref.html == Modification History == {{{ @copyright: 2006 by Pascal Bauermeister @license: GNU GPL, see COPYING for details. 2006-05-24 v1.0.0 PascalBauermeister Initial revision 2006-05-26 v1.0.1 PascalBauermeister * Set env var HTMLDOC_NOCGI to solve CGI issue 2006-05-26 v1.0.2 PascalBauermeister * Relative image URLs turned absolute was bogus. It is less bogus now. }}} """ import os, mimetypes, time, zipfile from MoinMoin import config, user, util, wikiutil, packages from MoinMoin.Page import Page from MoinMoin.util import MoinMoinNoFooter, filesys action_name = __name__.split('.')[-1] def_style = 'page' def error_msg (pagename, request, msg): Page (request, pagename).send_page (request, msg=msg) def format (src_text, request, formatter): # parse the text (in wiki source format) and make HTML, # after diverting sys.stdout to a string import StringIO from MoinMoin.parser import wiki str_out = StringIO.StringIO () # create str to collect output request.redirect (str_out) # divert output to that string p = Page (request, "$$$") formatter.setPage (p) try: # parse text wiki.Parser (src_text, request).format (formatter) finally: request.redirect () # restore output formatter.setPage (request.page) return str_out.getvalue () # return what was generated def get_style (request, name): """Defines built-in, default styles. These styles can be exended or replaced by declaring a dictionary similar to def_styles below, but named htmldoc_styles, in the wikiconfig.py file. """ _ = request.getText def_styles = { "page": "--verbose --no-localfiles --no-compression --jpeg " \ "--header t.D --footer ./. --size a4 --left 0.5in " \ "--webpage", "book": "--verbose --no-localfiles --no-compression --jpeg " \ "--header t.D --footer ./. --size a4 --left 0.5in", } try: # get styles from config, to update the default ones cfg_styles = request.cfg.htmldoc_styles styles = def_styles styles.update (cfg_styles) except AttributeError: styles = def_styles try: # get wanted style style = styles [name] except KeyError: msg = _("Unknown style: ") + name + "
" + \ _("Possible styles: ") + ', '.join (styles.keys ()) + "
" + \ _("Default style: ") + def_style error_msg (request.page.page_name, request, msg) return None return style def escape (str): return str.replace ('&','&').replace ('<', '<').replace ('>', '>') def execute (pagename, request): """ Main dispatcher for the action. """ _ = request.getText debug = 0 revision = None pageaction = None style = def_style msg = None if request.form.has_key ('rev'): revision = request.form ['rev'] [0] if request.form.has_key ('pageaction'): pageaction = request.form ['pageaction'] [0] if request.form.has_key ('style'): style = request.form ['style'] [0] if request.form.has_key ('help'): help = format (__doc__, request, request.formatter) error_msg (pagename, request, help) return if request.form.has_key ('debug'): debug = 1 # make pdf html = get_html (request, revision, pageaction, debug) flags = get_style (request, style) if not flags: return data = html2pdf (request, html, flags, debug) if not data: return # send it filename = pagename.replace ('/', '-') + ".pdf" send_pdf (filename, request, data, debug) return def send_pdf (filename, request, data, debug): import shutil, StringIO # get mimetype type = "application/pdf" if debug: type = "text/plain" # send header request.http_headers([ "Content-Type: %s" % type, "Content-Length: %d" % len (data), # TODO: fix the encoding here, plain 8 bit is not allowed # according to the RFCs There is no solution that is # compatible to IE except stripping non-ascii chars "Content-Disposition: inline; filename=\"%s\"" % filename.encode (config.charset), ]) # send data sio = StringIO.StringIO (data) shutil.copyfileobj (sio, request, 8192) raise MoinMoinNoFooter # regex to match the source url of tags import re RX_IMG = re.compile (r'()', re.I) def get_html (request, revision, pageaction, debug): """ get the html body of this page @param request: the request for the page @param revision: wanted revision (None for head) @param pageaction: action for page (None for print) """ import copy, sys, StringIO # Get HTML for the page: # make a copy of the request, run it and intercept the output orig_form = {} orig_form.update (request.form) # set form: if pageaction == action_name : pageaction = None if pageaction: request.form ["action"] = [pageaction] else: request.form ["action"] = [u"print"] if revision: request.form ["rev"] = [revision] # divert stdout, re-run the request old_stdout = sys.stdout out = StringIO.StringIO () sys.stdout = out copy.copy (request).run () html = out.getvalue () out.close () # restore things sys.stdout = old_stdout request.form = orig_form # Do some post-processing on th ehtml # lf -> cr html = html.replace ('\r\n', '\n') html = html.replace ('\r', '\n') # remove headers lines = html.split ('\n') while lines [0].strip (): del lines [0] html = '\n'.join (lines) # look for non-html content (TODO: better done by examining # headers) start = html.strip () [0:20] .lower () if not (start.startswith ("%s" % html # make images URLs absolute base = request.getQualifiedURL() # this is bogus because other attributes may preceed src... ## html = html.replace (', and replace URL by the # same prefixed by the URL base: matches = [] for match in RX_IMG.finditer (html): m = match.groups () [1] if m not in matches: html = html.replace (m, base+m) matches.append (m) # this is not really proof... but I'm too lazy to do an HTML # parser for this ! return html def html2pdf (request, html, flags, debug): import os cmd = "HTMLDOC_NOCGI=1 htmldoc -t pdf %s -" % flags inp, out, err = os.popen3 (cmd, 'b') inp.write (html) inp.close () pdf = out.read () out.close () msg = "Command: 
" + escape (cmd) + "
returned:
"+ \
          escape (err.read ()).replace ('\n','
') + "
" err.close () # as it is difficult to get the htmldoc return code, we check for # error by checking the produced pdf length if not len (pdf): error_msg (request.page.page_name, request, msg) if debug: return html return pdf