PDF: use cache conversion to cache PDF content information

product/ERP5/Document/PDFDocument.py

@@ -27,6 +27,7 @@
 ##############################################################################
 
 import tempfile, os, pickle
+import json
 
 import zope.interface
 from AccessControl import ClassSecurityInfo
@@ -278,18 +279,19 @@ class PDFDocument(Image):
   def getContentInformation(self):
     """Returns the information about the PDF document with pdfinfo.
     """
+    content_information_pseudo_format = 'content_information_pseudo_format'
     if not self.hasData():
       return {}
-    try:
-      return self._content_information.copy()
-    except AttributeError:
-      pass
+    # XXX we reuse the conversion infrastructure for cache
+    if self.hasConversion(format=content_information_pseudo_format):
+      mime, data = self.getConversion(format=content_information_pseudo_format)
+      return json.loads(data)
+
     tmp = tempfile.NamedTemporaryFile()
     tmp.write(self.getData())
     tmp.seek(0)
     command_result = None
     try:
-
       # First, we use pdfinfo to get standard metadata
       command = ['pdfinfo', '-meta', '-box', tmp.name]
       try:
@@ -341,16 +343,10 @@ class PDFDocument(Image):
     finally:
       tmp.close()
 
-    # Store cache as an instance of document. FIXME: we usually try to avoid this
-    # pattern and cache the result of methods using content md5 as a cache key.
-    self._content_information = result
+    self.setConversion(json.dumps(result), format=content_information_pseudo_format)
     return result.copy()
 
   def _setFile(self, data, precondition=None):
-    try:
-      del self._content_information
-    except (AttributeError, KeyError):
-      pass
     # Use File._setFile because Image._setFile will guess image size & content
     # type
     File._setFile(self, data, precondition=precondition)