PDFDocument.py 10.6 KB
Newer Older
1
# -*- coding: utf-8 -*-
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
##############################################################################
#
# Copyright (c) 2002-2006 Nexedi SARL and Contributors. All Rights Reserved.
#
# WARNING: This program as such is intended to be used by professional
# programmers who take the whole responsability of assessing all potential
# consequences resulting from its eventual inadequacies and bugs
# End users who are looking for a ready-to-use solution with commercial
# garantees and support are strongly adviced to contract a Free Software
# Service Company
#
# This program is Free Software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License
# as published by the Free Software Foundation; either version 2
# of the License, or (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.
#
##############################################################################

29
import tempfile, os
30

31
from AccessControl import ClassSecurityInfo
32 33
from Products.CMFCore.utils import getToolByName, _setCacheHeaders,\
    _ViewEmulator
34

35
from Products.ERP5Type import Permissions, PropertySheet
36
from Products.ERP5.Document.Image import Image
37 38
from Products.ERP5.Document.Document import ConversionError
from Products.ERP5.mixin.cached_convertable import CachedConvertableMixin
39

40
class PDFDocument(Image, CachedConvertableMixin):
41
  """
42 43 44
  PDFDocument is a subclass of Image which is able to
  extract text content from a PDF file either as text
  or as HTML.
45 46
  """
  # CMF Type Definition
47
  meta_type = 'ERP5 PDF Document'
48 49 50 51 52 53 54 55
  portal_type = 'PDF'

  # Declarative security
  security = ClassSecurityInfo()
  security.declareObjectProtected(Permissions.AccessContentsInformation)

  # Default Properties
  property_sheets = ( PropertySheet.Base
56
                    , PropertySheet.XMLObject
57 58 59 60 61 62
                    , PropertySheet.CategoryCore
                    , PropertySheet.DublinCore
                    , PropertySheet.Version
                    , PropertySheet.Reference
                    , PropertySheet.Document
                    , PropertySheet.Data
63 64 65
                    , PropertySheet.ExternalDocument
                    , PropertySheet.Url
                    , PropertySheet.Periodicity
66 67
                    )

68 69 70 71
  searchable_property_list = ('asText', 'title', 'description', 'id', 'reference',
                              'version', 'short_title',
                              'subject', 'source_reference', 'source_project_title',)

72
  security.declareProtected(Permissions.View, 'index_html')
73 74
  def index_html(self, REQUEST, RESPONSE, display=None, format='', quality=75, 
                                          resolution=None, frame=0):
75
    """
76 77 78
      Returns data in the appropriate format (graphical)
      it is always a zip because multi-page pdfs are converted into a zip
      file of many images
79
    """
80 81 82 83 84 85 86 87
    _setCacheHeaders(_ViewEmulator().__of__(self), {'format' : format})
    if format is '':
      if self.getSourceReference() is not None:
        filename = self.getSourceReference()
      else:
        filename = self.getId()
      RESPONSE.setHeader('Content-Disposition',
                         'attachment; filename="%s"' % filename)
88
      RESPONSE.setHeader('Content-Type', 'application/pdf')
89
      return str(self.data)
90 91 92 93 94 95 96
    if format in ('html', 'txt', 'text'):
      mime, data = self.convert(format)
      RESPONSE.setHeader('Content-Length', len(data))
      RESPONSE.setHeader('Content-Type', '%s;charset=UTF-8' % mime)
      RESPONSE.setHeader('Accept-Ranges', 'bytes')
      return data
    return Image.index_html(self, REQUEST, RESPONSE, display=display,
97 98
                            format=format, quality=quality,
                            resolution=resolution, frame=frame)
99 100

  # Conversion API
101
  security.declareProtected(Permissions.AccessContentsInformation, 'convert')
102 103 104 105
  def convert(self, format, **kw):
    """
    Implementation of conversion for PDF files
    """
106
    if format == 'html':
107 108 109
      try:
        return self.getConversion(format=format)
      except KeyError:
110
        mime = 'text/html'
111
        data = self._convertToHTML()
112 113
        self.setConversion(data, mime=mime, format=format)
        return (mime, data)
114
    elif format in ('txt', 'text'):
115 116 117
      try:
        return self.getConversion(format='txt')
      except KeyError:
118
        mime = 'text/plain'
119
        data = self._convertToText()
120 121
        self.setConversion(data, mime=mime, format='txt')
        return (mime, data)
122 123 124 125 126
    else:
      return Image.convert(self, format, **kw)

  security.declareProtected(Permissions.ModifyPortalContent, 'populateContent')
  def populateContent(self):
127
    """
128 129 130
      Convert each page to an Image and populate the
      PDF directory with converted images. May be useful
      to provide online PDF reader
131
    """
132
    raise NotImplementedError
133 134

  security.declarePrivate('_convertToText')
135
  def _convertToText(self):
136
    """
137
      Convert the PDF text content to text with pdftotext
138
    """
139 140
    if not self.data:
      return ''
141
    tmp = tempfile.NamedTemporaryFile()
142
    tmp.write(str(self.data))
143 144 145 146 147 148
    tmp.seek(0)
    cmd = 'pdftotext -layout -enc UTF-8 -nopgbrk %s -' % tmp.name
    r = os.popen(cmd)
    h = r.read()
    tmp.close()
    r.close()
149 150 151 152 153 154 155 156 157 158 159 160
    
    if h != '':
      return h
    else:
      # Try to use OCR
      # As high dpi images are required, it may take some times to convert the
      # pdf. 
      # It may be required to use activities to fill the cache and at the end, 
      # to calculate the final result
      text = ''
      content_information = self.getContentInformation()
      page_count = int(content_information.get('Pages', 0))
161 162 163 164 165 166 167 168 169 170 171 172 173
      try:
        # if the dimension is too big, rasterized image can be too
        # big. so we limit the maximum of rasterized image to 4096
        # pixles.
        # XXX since the dimention can be different on each page, it is
        # better to call 'pdfinfo -f page_num -l page_num' to get the
        # size of each page.
        max_size = 4096
        size = content_information.get('Page size',
                                       '%s x %s pts' % (max_size, max_size))
        width = int(size.split(' ')[0])
        height = int(size.split(' ')[2])
        resolution = 72.0 * max_size / max(width, height)
Jérome Perrin's avatar
Jérome Perrin committed
174
      except (ValueError, ZeroDivisionError):
175
        resolution = None
176 177
      for page_number in range(page_count):
        src_mimetype, png_data = self.convert(
178
            'png', quality=100, resolution=resolution,
179 180 181 182 183 184 185 186 187 188
            frame=page_number, display='identical')
        if not src_mimetype.endswith('png'):
          continue
        content = '%s' % png_data
        mime_type = getToolByName(self, 'mimetypes_registry').\
                                    lookupExtension('name.%s' % 'txt')
        if content is not None:
          portal_transforms = getToolByName(self, 'portal_transforms')
          result = portal_transforms.convertToData(mime_type, content,
                                                   context=self,
189
                                                   filename=self.getTitleOrId(),
190 191
                                                   mimetype=src_mimetype)
          if result is None:
192 193
            raise ConversionError('PDFDocument conversion error. '
                                  'portal_transforms failed to convert to %s: %r' % (mime_type, self))
194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209
          text += result
      return text

  security.declareProtected('View', 'getSizeFromImageDisplay')
  def getSizeFromImageDisplay(self, image_display):
    """
    Return the size for this image display, or None if this image display name
    is not known. If the preference is not set, (0, 0) is returned.
    """
    # identical parameter can be considered as a hack, in order not to
    # resize the image to prevent text distorsion when using OCR.
    # A cleaner API is required.
    if image_display == 'identical':
      return (self.getWidth(), self.getHeight())
    else:
      return Image.getSizeFromImageDisplay(self, image_display)
210 211 212 213 214

  security.declarePrivate('_convertToHTML')
  def _convertToHTML(self):
    """
    Convert the PDF text content to HTML with pdftohtml
215 216 217

    NOTE: XXX check that command exists and was executed
    successfully
218
    """
219 220
    if not self.data:
      return ''
221
    tmp = tempfile.NamedTemporaryFile()
222
    tmp.write(str(self.data))
223 224 225 226 227 228 229
    tmp.seek(0)
    cmd = 'pdftohtml -enc UTF-8 -stdout -noframes -i %s' % tmp.name
    r = os.popen(cmd)
    h = r.read()
    tmp.close()
    r.close()
    h = h.replace('<BODY bgcolor="#A0A0A0"', '<BODY ') # Quick hack to remove bg color - XXX
230
    h = h.replace('href="%s.html' % tmp.name.split(os.sep)[-1], 'href="asEntireHTML') # Make links relative
231 232 233 234 235 236 237
    return h

  security.declareProtected(Permissions.AccessContentsInformation, 'getContentInformation')
  def getContentInformation(self):
    """
    Returns the information about the PDF document with
    pdfinfo.
238 239 240

    NOTE: XXX check that command exists and was executed
    successfully
241
    """
242 243 244 245
    try:
      return self._content_information.copy()
    except AttributeError:
      pass
246
    tmp = tempfile.NamedTemporaryFile()
247
    tmp.write(str(self.data))
248
    tmp.seek(0)
249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282
    try:
      # First, we use pdfinfo to get standard metadata
      cmd = 'pdfinfo -meta -box %s' % tmp.name
      r = os.popen(cmd)
      h = r.read()
      r.close()
      result = {}
      for line in h.splitlines():
        item_list = line.split(':')
        key = item_list[0].strip()
        value = ':'.join(item_list[1:]).strip()
        result[key] = value

      # Then we use pdftk to get extra metadata
      cmd = 'pdftk %s dump_data output' % tmp.name
      r = os.popen(cmd)
      h = r.read()
      r.close()
      line_list = (line for line in h.splitlines())
      while True:
        try:
          line = line_list.next()
        except StopIteration:
          break
        if line.startswith('InfoKey'):
          key = line[len('InfoKey: '):]
          line = line_list.next()
          assert line.startswith('InfoValue: '),\
              "Wrong format returned by pdftk dump_data"
          value = line[len('InfoValue: '):]
          result.setdefault(key, value)
    finally:
      tmp.close()

283 284 285 286 287 288
    self._content_information = result
    return result.copy()

  def _setFile(self, data, precondition=None):
    try:
      del self._content_information
Yusei Tahara's avatar
Yusei Tahara committed
289
    except (AttributeError, KeyError):
290 291
      pass
    Image._setFile(self, data, precondition)