share/extensions/text_extract.py

9274e24cf014e4d132518578c221b04c60974d63JazzyNico#!/usr/bin/env python
9274e24cf014e4d132518578c221b04c60974d63JazzyNico"""
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoCopyright (C) 2011 Nicolas Dufour (jazzynico)
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoDirection code from the Restack extension, by Rob Antonishen
9274e24cf014e4d132518578c221b04c60974d63JazzyNico
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoPermission is hereby granted, free of charge, to any person obtaining a copy
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoof this software and associated documentation files (the "Software"), to deal
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoin the Software without restriction, including without limitation the rights
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoto use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9274e24cf014e4d132518578c221b04c60974d63JazzyNicocopies of the Software, and to permit persons to whom the Software is
9274e24cf014e4d132518578c221b04c60974d63JazzyNicofurnished to do so, subject to the following conditions:
9274e24cf014e4d132518578c221b04c60974d63JazzyNico
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoThe above copyright notice and this permission notice shall be included in
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoall copies or substantial portions of the Software.
9274e24cf014e4d132518578c221b04c60974d63JazzyNico
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoTHE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoIMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoFITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoAUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoLIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoOUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoTHE SOFTWARE.
9274e24cf014e4d132518578c221b04c60974d63JazzyNico
9274e24cf014e4d132518578c221b04c60974d63JazzyNico"""
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNico# standard library
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNicoimport chardataeffect
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNicofrom copy import deepcopy
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNicoimport csv
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNicoimport math
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNicoimport os
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNicoimport string
9274e24cf014e4d132518578c221b04c60974d63JazzyNicotry:
9274e24cf014e4d132518578c221b04c60974d63JazzyNico    from subprocess import Popen, PIPE
9274e24cf014e4d132518578c221b04c60974d63JazzyNico    bsubprocess = True
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoexcept:
9274e24cf014e4d132518578c221b04c60974d63JazzyNico    bsubprocess = False
9274e24cf014e4d132518578c221b04c60974d63JazzyNico
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNico# local library
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNicoimport inkex
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNico
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoclass Extract(inkex.Effect):
9274e24cf014e4d132518578c221b04c60974d63JazzyNico    def __init__(self):
9274e24cf014e4d132518578c221b04c60974d63JazzyNico        inkex.Effect.__init__(self)
9274e24cf014e4d132518578c221b04c60974d63JazzyNico        self.OptionParser.add_option("-d", "--direction",
629aea629b26536692b4cb4a2adebecfd59c6193bryce                        action="store", type="string",
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                        dest="direction", default="tb",
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                        help="direction to extract text")
9274e24cf014e4d132518578c221b04c60974d63JazzyNico        self.OptionParser.add_option("-x", "--xanchor",
629aea629b26536692b4cb4a2adebecfd59c6193bryce                        action="store", type="string",
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                        dest="xanchor", default="m",
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                        help="horizontal point to compare")
9274e24cf014e4d132518578c221b04c60974d63JazzyNico        self.OptionParser.add_option("-y", "--yanchor",
629aea629b26536692b4cb4a2adebecfd59c6193bryce                        action="store", type="string",
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                        dest="yanchor", default="m",
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                        help="vertical point to compare")
9274e24cf014e4d132518578c221b04c60974d63JazzyNico
9274e24cf014e4d132518578c221b04c60974d63JazzyNico    def effect(self):
9274e24cf014e4d132518578c221b04c60974d63JazzyNico        if len(self.selected)==0:
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            for node in self.document.xpath('//svg:text | //svg:flowRoot', namespaces=inkex.NSS):
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                self.selected[node.get('id')] = node
629aea629b26536692b4cb4a2adebecfd59c6193bryce
9274e24cf014e4d132518578c221b04c60974d63JazzyNico        if len( self.selected ) > 0:
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            objlist = []
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            svg = self.document.getroot()
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            parentnode = self.current_layer
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            file = self.args[ -1 ]
629aea629b26536692b4cb4a2adebecfd59c6193bryce
629aea629b26536692b4cb4a2adebecfd59c6193bryce            # get all bounding boxes in file by calling inkscape again with the --query-all command line option
629aea629b26536692b4cb4a2adebecfd59c6193bryce            # it returns a comma separated list structured id,x,y,w,h
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            if bsubprocess:
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                p = Popen('inkscape --query-all "%s"' % (file), shell=True, stdout=PIPE, stderr=PIPE)
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                err = p.stderr
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                f = p.communicate()[0]
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                try:
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                    reader=csv.CSVParser().parse_string(f)    #there was a module cvs.py in earlier inkscape that behaved differently
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                except:
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                    reader=csv.reader(f.split( os.linesep ))
629aea629b26536692b4cb4a2adebecfd59c6193bryce                err.close()
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            else:
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                _,f,err = os.popen3('inkscape --query-all "%s"' % ( file ) )
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                reader=csv.reader( f )
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                err.close()
629aea629b26536692b4cb4a2adebecfd59c6193bryce
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            #build a dictionary with id as the key
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            dimen = dict()
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            for line in reader:
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                if len(line) > 0:
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                    dimen[line[0]] = map( float, line[1:])
9274e24cf014e4d132518578c221b04c60974d63JazzyNico
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            if not bsubprocess: #close file if opened using os.popen3
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                f.close
629aea629b26536692b4cb4a2adebecfd59c6193bryce
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            #find the center of all selected objects **Not the average!
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            x,y,w,h = dimen[self.selected.keys()[0]]
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            minx = x
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            miny = y
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            maxx = x + w
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            maxy = y + h
9274e24cf014e4d132518578c221b04c60974d63JazzyNico
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            for id, node in self.selected.iteritems():
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                # get the bounding box
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                x,y,w,h = dimen[id]
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                if x < minx:
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                    minx = x
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                if (x + w) > maxx:
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                    maxx = x + w
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                if y < miny:
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                    miny = y
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                if (y + h) > maxy:
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                    maxy = y + h
9274e24cf014e4d132518578c221b04c60974d63JazzyNico
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            midx = (minx + maxx) / 2
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            midy = (miny + maxy) / 2
9274e24cf014e4d132518578c221b04c60974d63JazzyNico
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            #calculate distances for each selected object
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            for id, node in self.selected.iteritems():
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                # get the bounding box
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                x,y,w,h = dimen[id]
9274e24cf014e4d132518578c221b04c60974d63JazzyNico
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                # calc the comparison coords
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                if self.options.xanchor == "l":
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                    cx = x
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                elif self.options.xanchor == "r":
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                    cx = x + w
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                else:  # middle
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                    cx = x + w / 2
9274e24cf014e4d132518578c221b04c60974d63JazzyNico
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                if self.options.yanchor == "t":
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                    cy = y
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                elif self.options.yanchor == "b":
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                    cy = y + h
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                else:  # middle
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                    cy = y + h / 2
629aea629b26536692b4cb4a2adebecfd59c6193bryce
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                #direction chosen
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                if self.options.direction == "tb":
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                    objlist.append([cy,id])
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                elif self.options.direction == "bt":
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                    objlist.append([-cy,id])
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                elif self.options.direction == "lr":
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                    objlist.append([cx,id])
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                elif self.options.direction == "rl":
9274e24cf014e4d132518578c221b04c60974d63JazzyNico                    objlist.append([-cx,id])
9274e24cf014e4d132518578c221b04c60974d63JazzyNico
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            objlist.sort()
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            #move them to the top of the object stack in this order.
9274e24cf014e4d132518578c221b04c60974d63JazzyNico            for item in objlist:
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNico                self.recurse(deepcopy(self.selected[item[1]]))
629aea629b26536692b4cb4a2adebecfd59c6193bryce
9274e24cf014e4d132518578c221b04c60974d63JazzyNico    def recurse(self, node):
9274e24cf014e4d132518578c221b04c60974d63JazzyNico        istext = (node.tag == '{http://www.w3.org/2000/svg}flowPara' or node.tag == '{http://www.w3.org/2000/svg}flowDiv' or node.tag == '{http://www.w3.org/2000/svg}text')
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNico        if node.text != None or node.tail != None:
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNico            for child in node:
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNico                if child.get('{http://sodipodi.sourceforge.net/DTD/sodipodi-0.dtd}role'):
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNico                    child.tail = "\n"
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNico            inkex.errormsg(inkex.etree.tostring(node, method='text').strip())
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNico        else:
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNico            for child in node:
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNico                self.recurse(child)
13f50632425a012e0bd5a7b31015c7895f5d1375JazzyNico
9274e24cf014e4d132518578c221b04c60974d63JazzyNico
9274e24cf014e4d132518578c221b04c60974d63JazzyNicoif __name__ == '__main__':
9274e24cf014e4d132518578c221b04c60974d63JazzyNico    e = Extract()
9274e24cf014e4d132518578c221b04c60974d63JazzyNico    e.affect()
9274e24cf014e4d132518578c221b04c60974d63JazzyNico
76afb6b15a6b74cb64d169e0afac83c6915e9dc8~suv# vim: expandtab shiftwidth=4 tabstop=8 softtabstop=4 fileencoding=utf-8 textwidth=99