def render(item):
if isinstance(item, LTPage):
self.outfp.write('<page id="%s" bbox="%s" rotate="%d">\n' %
(item.id, strbbox(item.bbox), item.rotate))
for child in item:
render(child)
self.outfp.write('</page>\n')
elif isinstance(item, LTLine) and item.direction:
self.outfp.write('<line linewidth="%d" direction="%s" bbox="%s" />\n' % (item.linewidth, item.direction, strbbox(item.bbox)))
elif isinstance(item, LTRect):
self.outfp.write('<rect linewidth="%d" bbox="%s" />\n' % (item.linewidth, strbbox(item.bbox)))
elif isinstance(item, LTPolygon):
self.outfp.write('<polygon linewidth="%d" bbox="%s" pts="%s"/>\n' % (item.linewidth, strbbox(item.bbox), item.get_pts()))
elif isinstance(item, LTFigure):
self.outfp.write('<figure id="%s" bbox="%s">\n' % (item.id, strbbox(item.bbox)))
for child in item:
render(child)
self.outfp.write('</figure>\n')
elif isinstance(item, LTTextLine):
self.outfp.write('<textline bbox="%s">\n' % strbbox(item.bbox))
for child in item:
render(child)
self.outfp.write('</textline>\n')
elif isinstance(item, LTTextBox):
self.outfp.write('<textbox id="%s" bbox="%s">\n' % (item.id, strbbox(item.bbox)))
for child in item:
render(child)
self.outfp.write('</textbox>\n')
elif isinstance(item, LTTextItem):
self.outfp.write('<text font="%s" vertical="%s" bbox="%s" fontsize="%.3f">' %
(enc(item.font.fontname), item.is_vertical(),
strbbox(item.bbox), item.fontsize))
self.write(item.text)
self.outfp.write('</text>\n')
elif isinstance(item, LTText):
self.outfp.write('<text>%s</text>\n' % item.text)
elif isinstance(item, LTImage):
x = ''
if self.outdir:
name = self.write_image(item)
if name:
x = 'name="%s" ' % enc(name)
self.outfp.write('<image %stype="%s" width="%d" height="%d" />\n' % (x, item.type, item.width, item.height))
else:
assert 0, item
return
def __repr__(self):
if self.debug:
return ('<text matrix=%s font=%r fontsize=%.1f bbox=%s adv=%s text=%r>' %
('[%.1f, %.1f, %.1f, %.1f, (%.1f, %.1f)]' % self.matrix,
self.font, self.fontsize, strbbox(self.bbox),
'(%.1f, %.1f)' % self.adv,
self.text))
else:
return '<text %r>' % self.text
def begin_page(self, page, ctm):
self.outfp.write('<page id="%s" bbox="%s" rotate="%d">' %
(self.pageno, strbbox(page.mediabox), page.rotate))
return
def __repr__(self):
return ('<page id=%r bbox=%s rotate=%r>' % (self.id, strbbox(self.bbox), self.rotate))
def __repr__(self):
return ('<textbox %s(%s) %r...>' % (strbbox(self.bbox), self.direction, self.get_text()[:20]))
def __repr__(self):
return ('<textline %s(%s)>' % (strbbox(self.bbox), self.direction))
def __repr__(self):
return ('<figure id=%r bbox=%s matrix=%r>' % (self.id, strbbox(self.bbox), self.matrix))
def __repr__(self):
return ('<group %s>' % strbbox(self.bbox))
def __repr__(self):
return ('<item bbox=%s>' % strbbox(self.bbox))