zacc806
/
pdfminer.six
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

add html exect layout mode; default changed. 

git-svn-id: https://pdfminerr.googlecode.com/svn/trunk/pdfminer@272 1aa58f4a-7d42-0410-adbc-911cccaed67c
pull/1/head
yusuke.shinyama.dummy 2010-11-14 10:07:41 +00:00
parent 08c5c66917
commit 476ecf7e32
2 changed files with 146 additions and 74 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

178
pdfminer/converter.py
View File

@ -91,10 +91,22 @@ class PDFLayoutAnalyzer(PDFTextDevice):
return return
def render_char(self, matrix, font, fontsize, scaling, rise, cid): def render_char(self, matrix, font, fontsize, scaling, rise, cid):
item = LTChar(matrix, font, fontsize, scaling, rise, cid) try:
text = font.to_unichr(cid)
assert isinstance(text, unicode), text
except PDFUnicodeNotDefined:
text = self.handle_undefined_char(font, cid)
textwidth = font.char_width(cid)
textdisp = font.char_disp(cid)
item = LTChar(matrix, font, fontsize, scaling, rise, text, textwidth, textdisp)
self.cur_item.add(item) self.cur_item.add(item)
return item.adv return item.adv
def handle_undefined_char(self, font, cid):
if self.debug:
print >>sys.stderr, 'undefined: %r, %r' % (font, cid)
return '?'
def receive_layout(self, ltpage): def receive_layout(self, ltpage):
return return
@ -126,10 +138,6 @@ class PDFConverter(PDFLayoutAnalyzer):
self.codec = codec self.codec = codec
return return
def write(self, text):
self.outfp.write(enc(text, self.codec))
return
def write_image(self, image): def write_image(self, image):
stream = image.stream stream = image.stream
filters = stream.get_filters() filters = stream.get_filters()
@ -163,7 +171,7 @@ class TextConverter(PDFConverter):
self.showpageno = showpageno self.showpageno = showpageno
return return
def write(self, text): def write_text(self, text):
self.outfp.write(text.encode(self.codec, 'ignore')) self.outfp.write(text.encode(self.codec, 'ignore'))
return return
@ -173,13 +181,13 @@ class TextConverter(PDFConverter):
for child in item: for child in item:
render(child) render(child)
elif isinstance(item, LTText): elif isinstance(item, LTText):
self.write(item.text) self.write_text(item.text)
if isinstance(item, LTTextBox): if isinstance(item, LTTextBox):
self.write('\n') self.write_text('\n')
if self.showpageno: if self.showpageno:
self.write('Page %s\n' % ltpage.pageid) self.write_text('Page %s\n' % ltpage.pageid)
render(ltpage) render(ltpage)
self.write('\f') self.write_text('\f')
return return
@ -191,92 +199,152 @@ class HTMLConverter(PDFConverter):
#'char': 'green', #'char': 'green',
#'figure': 'yellow', #'figure': 'yellow',
#'textline': 'magenta', #'textline': 'magenta',
'textbox': 'cyan', #'textbox': 'cyan',
'textgroup': 'red', #'textgroup': 'red',
'polygon': 'black', 'polygon': 'black',
'page': 'gray', 'page': 'gray',
} }
TEXT_COLORS = { TEXT_COLORS = {
'textbox': 'blue', #'textbox': 'blue',
'char': 'black', 'char': 'black',
} }
def __init__(self, rsrcmgr, outfp, codec='utf-8', pageno=1, laparams=None, def __init__(self, rsrcmgr, outfp, codec='utf-8', pageno=1, laparams=None,
scale=1, showpageno=True, pagepad=50, outdir=None): scale=1, exact=False, showpageno=True, pagepad=50, outdir=None):
PDFConverter.__init__(self, rsrcmgr, outfp, codec=codec, pageno=pageno, laparams=laparams) PDFConverter.__init__(self, rsrcmgr, outfp, codec=codec, pageno=pageno, laparams=laparams)
self.exact = exact
self.showpageno = showpageno self.showpageno = showpageno
self.pagepad = pagepad self.pagepad = pagepad
self.outdir = outdir self.outdir = outdir
self.scale = scale self.scale = scale
self.outfp.write('<html><head>\n') self.fontscale = 0.7
self.outfp.write('<meta http-equiv="Content-Type" content="text/html; charset=%s">\n' % self.write('<html><head>\n')
self.codec) self.write('<meta http-equiv="Content-Type" content="text/html; charset=%s">\n' % self.codec)
self.outfp.write('</head><body>\n') self.write('</head><body>\n')
self.yoffset = self.pagepad self.yoffset = self.pagepad
self._font = None
self._fontstack = []
return return
def write_rect(self, color, width, x, y, w, h): def write(self, text):
self.outfp.write(text)
return
def write_text(self, text):
self.write(enc(text, self.codec))
return
def place_rect(self, color, borderwidth, x, y, w, h):
color = self.RECT_COLORS.get(color) color = self.RECT_COLORS.get(color)
if color is not None: if color is not None:
self.outfp.write('<span style="position:absolute; border: %s %dpx solid; ' self.write('<span style="position:absolute; border: %s %dpx solid; '
'left:%dpx; top:%dpx; width:%dpx; height:%dpx;"></span>\n' % 'left:%dpx; top:%dpx; width:%dpx; height:%dpx;"></span>\n' %
(color, width, (color, borderwidth,
x*self.scale, (self.yoffset-y)*self.scale, x*self.scale, (self.yoffset-y)*self.scale,
w*self.scale, h*self.scale)) w*self.scale, h*self.scale))
return return
def write_text(self, color, text, x, y, size): def place_image(self, item, borderwidth, x, y, w, h):
if self.outdir is not None:
name = self.write_image(item)
self.write('<img src="%s" border="%d" style="position:absolute; left:%dpx; top:%dpx;" '
'width="%d" height="%d" />\n' %
(enc(name), borderwidth,
x*self.scale, (self.yoffset-y)*self.scale,
w*self.scale, h*self.scale))
return
def place_text(self, color, text, x, y, size):
color = self.TEXT_COLORS.get(color) color = self.TEXT_COLORS.get(color)
if color is not None: if color is not None:
self.outfp.write('<span style="position:absolute; color:%s; left:%dpx; top:%dpx; font-size:%dpx;">' % self.write('<span style="position:absolute; color:%s; left:%dpx; top:%dpx; font-size:%dpx;">' %
(color, x*self.scale, (self.yoffset-y)*self.scale, size*self.scale)) (color, x*self.scale, (self.yoffset-y)*self.scale, size*self.scale))
self.write(text) self.write_text(text)
self.outfp.write('</span>\n') self.write('</span>\n')
return
def begin_textbox(self, color, borderwidth, x, y, w, h, writing_mode):
self._fontstack.append(self._font)
self._font = None
self.write('<div style="position:absolute; border: %s %dpx solid; writing-mode:%s; '
'left:%dpx; top:%dpx; width:%dpx; height:%dpx;">' %
(color, borderwidth, writing_mode,
x*self.scale, (self.yoffset-y)*self.scale,
w*self.scale, h*self.scale))
return
def put_text(self, text, fontname, fontsize):
font = (fontname, fontsize)
if font != self._font:
if self._font is not None:
self.write('</span>')
self.write('<span style="font-size:%dpx">' %
(fontsize * self.scale * self.fontscale))
self._font = font
self.write_text(text)
return
def end_textbox(self, color):
if self._font is not None:
self.write('</span>')
self._font = self._fontstack.pop()
self.write('</div>')
return return
def receive_layout(self, ltpage): def receive_layout(self, ltpage):
def render(item): def render(item):
if isinstance(item, LTPage): if isinstance(item, LTPage):
self.yoffset += item.y1 self.yoffset += item.y1
self.write_rect('page', 1, item.x0, item.y1, item.width, item.height) self.place_rect('page', 1, item.x0, item.y1, item.width, item.height)
if self.showpageno: if self.showpageno:
self.outfp.write('<div style="position:absolute; top:%dpx;">' % self.write('<div style="position:absolute; top:%dpx;">' %
((self.yoffset-item.y1)*self.scale)) ((self.yoffset-item.y1)*self.scale))
self.outfp.write('<a name="%s">Page %s</a></div>\n' % (item.pageid, item.pageid)) self.write('<a name="%s">Page %s</a></div>\n' % (item.pageid, item.pageid))
for child in item: for child in item:
render(child) render(child)
elif isinstance(item, LTChar):
self.write_rect('char', 1, item.x0, item.y1, item.width, item.height)
self.write_text('char', item.text, item.x0, item.y1, item.size)
elif isinstance(item, LTPolygon): elif isinstance(item, LTPolygon):
self.write_rect('polygon', 1, item.x0, item.y1, item.width, item.height) self.place_rect('polygon', 1, item.x0, item.y1, item.width, item.height)
elif isinstance(item, LTTextLine):
self.write_rect('textline', 1, item.x0, item.y1, item.width, item.height)
for child in item:
render(child)
elif isinstance(item, LTTextBox):
self.write_rect('textbox', 1, item.x0, item.y1, item.width, item.height)
for child in item:
render(child)
self.write_text('textbox', str(item.index+1), item.x0, item.y1, 20)
elif isinstance(item, LTFigure): elif isinstance(item, LTFigure):
self.write_rect('figure', 1, item.x0, item.y1, item.width, item.height) self.place_rect('figure', 1, item.x0, item.y1, item.width, item.height)
for child in item: for child in item:
render(child) render(child)
elif isinstance(item, LTImage): elif isinstance(item, LTImage):
if self.outdir: self.place_image(item, 1, item.x0, item.y1, item.width, item.height)
name = self.write_image(item) else:
self.outfp.write('<img src="%s" style="position:absolute; left:%dpx; top:%dpx;" ' if self.exact:
'width="%d" height="%d" />\n' % if isinstance(item, LTTextLine):
(enc(name), self.place_rect('textline', 1, item.x0, item.y1, item.width, item.height)
item.x0*self.scale, (self.yoffset-item.y1)*self.scale, for child in item:
item.width*self.scale, item.height*self.scale)) render(child)
elif isinstance(item, LTTextBox):
self.place_rect('textbox', 1, item.x0, item.y1, item.width, item.height)
self.place_text('textbox', str(item.index+1), item.x0, item.y1, 20)
for child in item:
render(child)
elif isinstance(item, LTChar):
self.place_rect('char', 1, item.x0, item.y1, item.width, item.height)
self.place_text('char', item.text, item.x0, item.y1, item.size)
else:
if isinstance(item, LTTextLine):
for child in item:
render(child)
self.write('<br>')
elif isinstance(item, LTTextBox):
self.begin_textbox('textbox', 1, item.x0, item.y1, item.width, item.height,
item.get_writing_mode())
for child in item:
render(child)
self.end_textbox('textbox')
elif isinstance(item, LTChar):
self.put_text(item.text, item.fontname, item.size)
elif isinstance(item, LTText):
self.write_text(item.text)
return return
render(ltpage) render(ltpage)
if ltpage.layout: if ltpage.layout:
def show_layout(item): def show_layout(item):
if isinstance(item, LTTextGroup): if isinstance(item, LTTextGroup):
self.write_rect('textgroup', 1, item.x0, item.y1, item.width, item.height) self.place_rect('textgroup', 1, item.x0, item.y1, item.width, item.height)
for child in item: for child in item:
show_layout(child) show_layout(child)
return return
@ -285,9 +353,9 @@ class HTMLConverter(PDFConverter):
return return
def close(self): def close(self):
self.outfp.write('<div style="position:absolute; top:0px;">Page: %s</div>\n' % self.write('<div style="position:absolute; top:0px;">Page: %s</div>\n' %
', '.join('<a href="#%s">%s</a>' % (i,i) for i in xrange(1,self.pageno))) ', '.join('<a href="#%s">%s</a>' % (i,i) for i in xrange(1,self.pageno)))
self.outfp.write('</body></html>\n') self.write('</body></html>\n')
return return
@ -302,6 +370,10 @@ class XMLConverter(PDFConverter):
self.outfp.write('<pages>\n') self.outfp.write('<pages>\n')
return return
def write_text(self, text):
self.outfp.write(enc(text, self.codec))
return
def receive_layout(self, ltpage): def receive_layout(self, ltpage):
def render(item): def render(item):
if isinstance(item, LTPage): if isinstance(item, LTPage):
@ -341,7 +413,7 @@ class XMLConverter(PDFConverter):
elif isinstance(item, LTChar): elif isinstance(item, LTChar):
self.outfp.write('<text font="%s" bbox="%s" size="%.3f">' % self.outfp.write('<text font="%s" bbox="%s" size="%.3f">' %
(enc(item.font.fontname), bbox2str(item.bbox), item.size)) (enc(item.font.fontname), bbox2str(item.bbox), item.size))
self.write(item.text) self.write_text(item.text)
self.outfp.write('</text>\n') self.outfp.write('</text>\n')
elif isinstance(item, LTText): elif isinstance(item, LTText):
self.outfp.write('<text>%s</text>\n' % item.text) self.outfp.write('<text>%s</text>\n' % item.text)

30
pdfminer/layout.py
View File

@ -183,22 +183,16 @@ class LTChar(LTItem, LTText):
debug = 0 debug = 0
def __init__(self, matrix, font, fontsize, scaling, rise, cid): def __init__(self, matrix, font, fontsize, scaling, rise, text, textwidth, textdisp):
self.matrix = matrix
self.font = font
self.fontsize = fontsize
self.adv = font.char_width(cid) * fontsize * scaling
try:
text = font.to_unichr(cid)
assert isinstance(text, unicode), text
except PDFUnicodeNotDefined:
text = '?'
LTText.__init__(self, text) LTText.__init__(self, text)
self.matrix = matrix
self.fontname = font.fontname
self.adv = textwidth * fontsize * scaling
# compute the boundary rectangle. # compute the boundary rectangle.
if self.font.is_vertical(): if font.is_vertical():
# vertical # vertical
width = font.get_width() * fontsize width = font.get_width() * fontsize
(vx,vy) = font.char_disp(cid) (vx,vy) = textdisp
if vx is None: if vx is None:
vx = width/2 vx = width/2
else: else:
@ -224,7 +218,7 @@ class LTChar(LTItem, LTText):
if y1 < y0: if y1 < y0:
(y0,y1) = (y1,y0) (y0,y1) = (y1,y0)
LTItem.__init__(self, (x0,y0,x1,y1)) LTItem.__init__(self, (x0,y0,x1,y1))
if self.font.is_vertical(): if font.is_vertical():
self.size = self.width self.size = self.width
else: else:
self.size = self.height self.size = self.height
@ -232,9 +226,9 @@ class LTChar(LTItem, LTText):
def __repr__(self): def __repr__(self):
if self.debug: if self.debug:
return ('<%s %s matrix=%s font=%r fontsize=%.1f adv=%s text=%r>' % return ('<%s %s matrix=%s font=%r adv=%s text=%r>' %
(self.__class__.__name__, bbox2str(self.bbox), (self.__class__.__name__, bbox2str(self.bbox),
matrix2str(self.matrix), self.font, self.fontsize, matrix2str(self.matrix), self.fontname,
self.adv, self.text)) self.adv, self.text))
else: else:
return '<char %r>' % self.text return '<char %r>' % self.text
@ -378,12 +372,18 @@ class LTTextBoxHorizontal(LTTextBox):
self._objs = csort(self._objs, key=lambda obj: -obj.y1) self._objs = csort(self._objs, key=lambda obj: -obj.y1)
return LTTextBox.finish(self) return LTTextBox.finish(self)
def get_writing_mode(self):
return 'lr-tb'
class LTTextBoxVertical(LTTextBox): class LTTextBoxVertical(LTTextBox):
def finish(self): def finish(self):
self._objs = csort(self._objs, key=lambda obj: -obj.x1) self._objs = csort(self._objs, key=lambda obj: -obj.x1)
return LTTextBox.finish(self) return LTTextBox.finish(self)
def get_writing_mode(self):
return 'tb-rl'
## LTTextGroup ## LTTextGroup
## ##