#!/usr/bin/env python
# -*- coding: utf-8  -*-
import re, time
import wikipedia, pagegenerators

# Text for translation 
#
# Magic words: %(title)s %(url)s %(parseddate)s
lead = {
	'en': u"""
<p><a href="http://en.wikipedia.org/wiki/Wikipedia:Alternative_text_for_images" title="Wikipedia:Alternative text for images" class="extiw">Alternative text</a> (alt text) is meant for readers who cannot see an image, such as blind readers and readers who use a text or mobile browser. It should summarize an image's appearance, and should not duplicate its caption. Every image should have alt text, except for <a href="http://en.wikipedia.org/wiki/Wikipedia:Alternative_text_for_images#Purely_decorative_images" title="Wikipedia:Alternative text for images" class="extiw">purely decorative images</a>, which should instead have "<code>|link=</code>".</p>
<p>The following table shows images and captions on the left, and alt text and captions on the right: the right column is what a visually impaired reader will hear. This table was computed from the copy of <a href="%(url)s" class="extiw">%(title)s</a> cached on %(parseddate)s.</p>""",

	'simple':u"""
<p>Alternative text (alt text) is suppose to tell the important features to a person who cannot see the image due to non-graphical or mobile browser or blindness.</p>
<p>The following table shows pictures on the left and the alternative text for the picture on the right which a blind person might hear.  This table was computed from the copy of <a href="%(url)s" class="extiw">%(title)s</a> cached on %(parseddate)s.</p>""",
}
ImageDesc = {
                'en': "Image and thumbnail",
                'de': "Bild und miniaturbild",
                'simple': "Pictures",
            }
AltDesc =   {
                'en': "Text description",
                'de': "alttext",
                'simple': "Text for the blind",
            }

# Regular expression for getting images
ImageR = re.compile(r'''(?six)
# Video posters
<div\ class="thumb[^"]*">\s*
<div[^<>]*>\s*
<div[^<>]*>\s*
<div>(?P<poster><a\ href="/w[^"]*"\ [^<>]*><img\ [^<>]*alt="(?P<videoalt>[^"]*)"\ /></a>)</div>\s*
<div.*?</div>\s*
</div>\s*
(?P<videocaption><div[^<>]*>(?P<videomagnify>\s*<div\ class="magnify">.*?</div>|).*?)</div>\s*
</div>\s*
</div>\s*
|
# Regex for border, frame, thumbnail/thumb, ogg video
<div\ class="thumb[^"]*">\s*
<div[^<>]*>(?P<image><a\ href="/w[^"]*"[^<>]*><img\ alt="(?P<thumbalt>[^"]*)"\ [^<>]*class="thumbimage"[^<>]*/></a>)\s*
(?P<thumbcaption><div[^<>]*>(?P<thumbmagnify>\s*<div\ class="magnify">.*?</div>|).*?)</div>\s*
</div>\s*
</div>
|
# Inline images
<a\ href="[^"]*"[^<>]*?><img\b[^<>]*?\ alt="(?P<imagealt>[^"]*)"[^<>]*?\ width="(?P<imagewidth>[^"]*)"\ [^<>]*?/></a>(?!\s*<div\ class="thumbcaption">)
|
# decorative images
<img\b[^<>]*?\ alt="(?P<decalt>[^"]*)"\ [^<>]*?/>
''')
# Image notes
# 	thumb	 class="thumbimage" with class="magnify"
# 	frame	 class="thumbimage" without class="magnify"
# 	border	 class="thumbborder"	
# 	link=	Removes class="image" from <a>

def main():
	genFactory = pagegenerators.GeneratorFactory()
	site = wikipedia.getSite()
	page = wikipedia.Page(site, '')
	for arg in wikipedia.handleArgs():
		if arg.startswith('-page:'):
			page = wikipedia.Page(site, arg[6:])
			site = page.site()
		elif arg.startswith('-interface:'):
			import cgitb; cgitb.enable()
		else:
			if not genFactory.handleArg(arg):
				wikipedia.output('Parameter "%s" not understood' % arg)
	generator = genFactory.getCombinedGenerator() or iter([page])
	
	for page in generator:
		if not page.title():continue
		try:
			# We get address from the 
			html = site.getUrl(site.nice_get_address(page.urlname())+"?"+time.strftime('%Y%m%d%H%M%S'))

			# Check to see if there's a history tab, if not then the page does not exist
			# NOTE only needed if nogzip is set
			if not re.search(r'<li id="ca-history"[^<>]*><a.*?</a></li>', html):
				wikipedia.output("%s does not exist on %s"%(page.aslink(), page.site().sitename()))
				continue	

			# Strip to body content
			html = html[html.index("<!-- start content -->"):html.index("<!-- end content -->")]
		except Exception, e:
			wikipedia.output('Error: %r'%e)
			continue

		# For debugging purposes print all comments
		for cmt in re.findall('<!--.*?-->', html, re.DOTALL):
			print( cmt )

		# Translation substitutions
		wgServer = '%s://%s'%(site.protocol(), site.hostname(),)
		m = re.search(r'<!-- Saved in parser cache [^<>]* timestamp ([0-9]{14}) -->', html)
		metadata = {
			'title':		page.title(),
			'server':		wgServer,
			'url':			wgServer+site.nice_get_address(page.urlname()),
			'parseddate':	m and time.strftime("%d %B %Y at %H:%M", time.strptime(m.group(1), "%Y%m%d%H%M%S")),
		}

		#
		print '<h3>%s</h3>'%page.title().encode('utf-8')
		if site.language() not in ('en', 'simple', ):
			print '<div style="background:#abc; text-align:center;"><h4>Help translate this tool</h4>Leave translations on <a href="http://en.wikipedia.org/wiki/User_talk:Dispenser">my talk page</a>, see the <a href="../sources/altviewer.py">source code</a> for details.</div>'
		print (wikipedia.translate(site, lead)%metadata).encode('utf-8')
		print '<table class="wikitable alt-compare">'
		print '<tr><th>%s</th><th>%s</th></tr>'%(wikipedia.translate(site, ImageDesc), wikipedia.translate(site, AltDesc))
		
		#
		images = ImageR.finditer(html)
		for image in images:
			# TODO: better support for galleries, <del>math equation</del>
			if image.group('decalt') is not None:
				if image.group('decalt') == '' and re.search(r' width="(\d\d|\d)"', image.group()):
					# Exclude images less than 100 pixels
					continue
				elif "/skins-1.5/" in image.group():
					continue
				elif "/w/extensions/" in image.group():
					continue
				elif ' class="tex"' in image.group():
					# Exclude tex equation since they will at some point will have machine generated alt text
					continue

			print '<tr><td align="center"><div>%s</div></td><td align="center"><div>'%(image.group().replace('="/', '="%s/'%wgServer),)
			if image.group('videoalt') is not None:
				print image.group().replace(image.group('poster'), "<div class='alttext %s'>%s</div>"%(image.group('videoalt')=='' and 'defaultalt' or '', image.group('videoalt'))).replace(image.group('videomagnify'), '').replace('="/', '="%s/'%wgServer)
			elif image.group('thumbalt') is not None:
				print image.group().replace(image.group('image'), "<div class='alttext %s'>%s</div>"%(image.group('thumbalt')=='' and 'defaultalt' or '', image.group('thumbalt'))).replace(image.group('thumbmagnify'), '').replace('="/', '="%s/'%wgServer)
			elif image.group('imagealt') is not None:
				print '<div class="alttext %s" style="width:%spx;">%s</div>' % (image.group('imagealt')=='' and 'defaultalt' or '', image.group('imagewidth'), image.group('imagealt') and '<span class="wrapper">%s</span>' % image.group('imagealt') or '')
			elif image.group('decalt') is not None:
				print '<div class="alttext nolink">%s<div class="notice">(Unlinked image)</div></div>' % (image.group('decalt'), )
			else:
				raise
			print '</div></td></tr>'
		print '</table>'

if __name__ == "__main__" and wikipedia.handleUrlAndHeader():
    try:
        wikipedia.startContent(form=True, head="""<style type="text/css">
/* This wasn't included in the global style? */
.magnify {
	float:right;
}

table.alt-compare {
	margin-left:auto;
	margin-right:auto;
	table-layout:fixed;
	width:80%;
}
table.alt-compare div.alttext {
	background:#444;
	color:#fff;
	min-height:2em;
	min-width:90px; /* reasonable width */ 
	padding:0.1em;
}
div.alttext wrapper {
	background-color:#444;
}
table.alt-compare div.alttext:empty {
	border:none;
	background:#f67;
}
table.alt-compare div.alttext.nolink {
	background:sandybrown;
	color:black;
}
table.alt-compare div.alttext.nolink .notice {
	font-size:x-small;
}
div.alttext:empty {
	border:2px solid #dc143c;
	background:#F2F9FF;
	color:#272DA5;
}
table.alt-compare td>div {
	overflow:auto;
}
table div.tright, table div.tleft {
	float: none; 
}
</style>""")
        main()
    finally:
        wikipedia.endContent()
        wikipedia.stopme()