1 files changed, 191 insertions, 0 deletions

diff --git a/libmat/office.py b/libmat/office.py
new file mode 100644
index 0000000..0ca1ff1
--- /dev/null
+++ b/libmat/office.py
@@ -0,0 +1,191 @@
+''' Care about office's formats
+
+'''
+
+import logging
+import os
+import shutil
+import tempfile
+import xml.dom.minidom as minidom
+import zipfile
+
+try:
+    import cairo
+    from gi.repository import Poppler
+except ImportError:
+    logging.info('office.py loaded without PDF support')
+    pass
+
+import parser
+import archive
+
+
+class OpenDocumentStripper(archive.TerminalZipStripper):
+    ''' An open document file is a zip, with xml file into.
+        The one that interest us is meta.xml
+    '''
+
+    def get_meta(self):
+        ''' Return a dict with all the meta of the file by
+            trying to read the meta.xml file.
+        '''
+        metadata = super(OpenDocumentStripper, self).get_meta()
+        zipin = zipfile.ZipFile(self.filename, 'r')
+        try:
+            content = zipin.read('meta.xml')
+            dom1 = minidom.parseString(content)
+            elements = dom1.getElementsByTagName('office:meta')
+            for i in elements[0].childNodes:
+                if i.tagName != 'meta:document-statistic':
+                    nodename = ''.join(i.nodeName.split(':')[1:])
+                    metadata[nodename] = ''.join([j.data for j in i.childNodes])
+                else:
+                    # thank you w3c for not providing a nice
+                    # method to get all attributes of a node
+                    pass
+        except KeyError:  # no meta.xml file found
+            logging.debug('%s has no opendocument metadata' % self.filename)
+        zipin.close()
+        return metadata
+
+    def remove_all(self):
+        ''' Removes metadata
+        '''
+        return super(OpenDocumentStripper, self).remove_all(ending_blacklist=['meta.xml'])
+
+    def is_clean(self):
+        ''' Check if the file is clean from harmful metadatas
+        '''
+        clean_super = super(OpenDocumentStripper, self).is_clean()
+        if clean_super is False:
+            return False
+
+        zipin = zipfile.ZipFile(self.filename, 'r')
+        try:
+            zipin.getinfo('meta.xml')
+        except KeyError:  # no meta.xml in the file
+            return True
+        zipin.close()
+        return False
+
+
+class OpenXmlStripper(archive.TerminalZipStripper):
+    ''' Represent an office openxml document, which is like
+        an opendocument format, with some tricky stuff added.
+        It contains mostly xml, but can have media blobs, crap, ...
+        (I don't like this format.)
+    '''
+    def remove_all(self):
+        return super(OpenXmlStripper, self).remove_all(
+                beginning_blacklist=('docProps/'), whitelist=('.rels'))
+
+    def is_clean(self):
+        ''' Check if the file is clean from harmful metadatas.
+            This implementation is faster than something like
+            "return this.get_meta() == {}".
+        '''
+        clean_super = super(OpenXmlStripper, self).is_clean()
+        if clean_super is False:
+            return False
+
+        zipin = zipfile.ZipFile(self.filename, 'r')
+        for item in zipin.namelist():
+            if item.startswith('docProps/'):
+                return False
+        zipin.close()
+        return True
+
+    def get_meta(self):
+        ''' Return a dict with all the meta of the file
+        '''
+        metadata = super(OpenXmlStripper, self).get_meta()
+
+        zipin = zipfile.ZipFile(self.filename, 'r')
+        for item in zipin.namelist():
+            if item.startswith('docProps/'):
+                metadata[item] = 'harmful content'
+        zipin.close()
+        return metadata
+
+
+class PdfStripper(parser.GenericParser):
+    ''' Represent a PDF file
+    '''
+    def __init__(self, filename, parser, mime, backup, is_writable, **kwargs):
+        super(PdfStripper, self).__init__(filename, parser, mime, backup, is_writable, **kwargs)
+        self.uri = 'file://' + os.path.abspath(self.filename)
+        self.password = None
+        try:
+            self.pdf_quality = kwargs['low_pdf_quality']
+        except KeyError:
+            self.pdf_quality = False
+
+        self.meta_list = frozenset(['title', 'author', 'subject',
+            'keywords', 'creator', 'producer', 'metadata'])
+
+    def is_clean(self):
+        ''' Check if the file is clean from harmful metadatas
+        '''
+        document = Poppler.Document.new_from_file(self.uri, self.password)
+        for key in self.meta_list:
+            if document.get_property(key):
+                return False
+        return True
+
+    def remove_all(self):
+        ''' Opening the PDF with poppler, then doing a render
+            on a cairo pdfsurface for each pages.
+
+            http://cairographics.org/documentation/pycairo/2/
+
+            The use of an intermediate tempfile is necessary because
+            python-cairo segfaults on unicode.
+            See http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=699457
+        '''
+        document = Poppler.Document.new_from_file(self.uri, self.password)
+        try:
+            output = tempfile.mkstemp()[1]
+            page = document.get_page(0)
+            # assume that every pages are the same size
+            page_width, page_height = page.get_size()
+            surface = cairo.PDFSurface(output, page_width, page_height)
+            context = cairo.Context(surface)  # context draws on the surface
+            logging.debug('PDF rendering of %s' % self.filename)
+            for pagenum in range(document.get_n_pages()):
+                page = document.get_page(pagenum)
+                context.translate(0, 0)
+                if self.pdf_quality:
+                    page.render(context)  # render the page on context
+                else:
+                    page.render_for_printing(context)  # render the page on context
+                context.show_page()  # draw context on surface
+            surface.finish()
+            shutil.move(output, self.output)
+        except:
+            logging.error('Something went wrong when cleaning %s.' % self.filename)
+            return False
+
+        try:
+            import pdfrw  # For now, poppler cannot write meta, so we must use pdfrw
+            logging.debug('Removing %s\'s superficial metadata' % self.filename)
+            trailer = pdfrw.PdfReader(self.output)
+            trailer.Info.Producer = None
+            trailer.Info.Creator = None
+            writer = pdfrw.PdfWriter()
+            writer.trailer = trailer
+            writer.write(self.output)
+            self.do_backup()
+        except:
+            logging.error('Unable to remove all metadata from %s, please install pdfrw' % self.output)
+            return False
+        return True
+
+    def get_meta(self):
+        ''' Return a dict with all the meta of the file
+        '''
+        document = Poppler.Document.new_from_file(self.uri, self.password)
+        metadata = {}
+        for key in self.meta_list:
+            if document.get_property(key):
+                metadata[key] = document.get_property(key)
+        return metadata