Assigning properties from document content upon ingestion;

Portal_type extraction from document content if not supplied.

git-svn-id: https://svn.erp5.org/repos/public/erp5/trunk@11212 20353a03-c40f-0410-a6d1-a30d3c3de9de

bt5/erp5_dms/ExtensionTemplateItem/documentUtils.py

+import zipfile, cStringIO, re
+
+rx_atr='([\w]+)###([\w/]+)'
+rx_atr=re.compile(rx_atr)
+
+def extractContent(data):
+  cs=cStringIO.StringIO()
+  cs.write(data)
+  try:
+    z=zipfile.ZipFile(cs)
+  except zipfile.BadZipfile:
+    cs.close()
+    return ''
+  s=z.read('content.xml')
+  cs.close()
+  z.close()
+  return s
+
+def getAttrFromContent(data):
+  return dict(rx_atr.findall(extractContent(data)))
+
+def getDoctypeFromContent(data):
+  atrs=getAttrFromContent(data)
+  return atrs.get('doctype')
+
+# vim: syntax=python shiftwidth=2 

bt5/erp5_dms/ExtensionTemplateItem/mailUtils.py

 import re
 
 def findAddress(txt):
-  validchars='A-Za-z.\-_'
+  validchars='0-9A-Za-z.\-_'
   r=re.compile('[%s]+@[%s]+' % (validchars,validchars))
   m=r.search(txt)
   return m and m.group()

bt5/erp5_dms/SkinTemplateItem/portal_skins/erp5_dms/DMS_createObjectFromFile.xml

@@ -68,14 +68,19 @@
         </item>
         <item>
             <key> <string>_body</string> </key>
-            <value> <string># determine content type\n
-ctype=context.content_type_registry.findTypeName(fname,None,None)\n
-context.log(fname,ctype)\n
-if ctype is None:\n
+            <value> <string># determine content type if not given\n
+# first try from content\n
+if not doctype:\n
+  doctype=context.Document_getDoctypeFromContent(data)\n
+# then from filename\n
+if not doctype:\n
+  doctype=context.content_type_registry.findTypeName(fname,None,None)\n
+context.log(fname,doctype)\n
+if not doctype:\n
   raise Exception("content type for file %s not registered" % fname)\n
 \n
 # create content\n
-ob=context.document_module.newContent(portal_type=ctype)\n
+ob=context.document_module.newContent(portal_type=doctype)\n
 ob.manage_upload(data)\n
 ob.DMS_ingestFile(fname,data)\n
 return ob\n
@@ -95,7 +100,7 @@ return ob\n
         </item>
         <item>
             <key> <string>_params</string> </key>
-            <value> <string>fname,data</string> </value>
+            <value> <string>fname,data, doctype=None</string> </value>
         </item>
         <item>
             <key> <string>errors</string> </key>
@@ -115,7 +120,7 @@ return ob\n
                   <dictionary>
                     <item>
                         <key> <string>co_argcount</string> </key>
-                        <value> <int>2</int> </value>
+                        <value> <int>3</int> </value>
                     </item>
                     <item>
                         <key> <string>co_varnames</string> </key>
@@ -123,10 +128,10 @@ return ob\n
                           <tuple>
                             <string>fname</string>
                             <string>data</string>
+                            <string>doctype</string>
                             <string>_getattr_</string>
                             <string>context</string>
                             <string>None</string>
-                            <string>ctype</string>
                             <string>Exception</string>
                             <string>ob</string>
                           </tuple>
@@ -140,7 +145,9 @@ return ob\n
         <item>
             <key> <string>func_defaults</string> </key>
             <value>
+              <tuple>
                 <none/>
+              </tuple>
             </value>
         </item>
         <item>

bt5/erp5_dms/SkinTemplateItem/portal_skins/erp5_dms/DMS_handleMail.xml

@@ -85,6 +85,7 @@ if senderemail is None:\n
   return printed\n
 \n
 # find sender\n
+context.log(senderemail)\n
 r=context.portal_catalog(portal_type=\'Email\',url_string=senderemail)\n
 if len(r)==0:\n
   print noSenderMsg\n
@@ -99,6 +100,7 @@ context.log(script.getId(),\'ok, this address belongs to \'+person.getRelativeUr
 try:\n
   for fname,data in theMail[\'attachments\'].items():\n
     obj=context.DMS_createObjectFromFile(fname,data)\n
+    obj.setGroup(person.getSubordinationValue().getGroup())\n
     obj.manage_setLocalRoles(person.getReference(),[\'Owner\',])\n
     context.DMS_notifyByEmail(address=senderemail,event=\'ingest\',object=obj)\n
 except Exception,e:\n

bt5/erp5_dms/SkinTemplateItem/portal_skins/erp5_dms/DMS_ingestFile.xml

@@ -72,10 +72,20 @@
 # it can be run any time\n
 # is meant to extract metadata from file name and contents\n
 \n
+if data is None:\n
+  if hasattr(context,\'oo_data\'):\n
+    data=context.oo_data\n
+  else:\n
+    if hasattr(context,\'data\'):\n
+      data=data\n
+if data is not None:\n
+  context.setPropertyListFromContent(data)\n
+\n
 if fname is None:\n
   fname=context.getSourceReference()\n
 else:\n
   context.setSourceReference(fname)\n
+\n
 if fname is not None:\n
   context.setPropertyListFromFilename(fname)\n
   context.guessMimeType(fname)\n
@@ -137,9 +147,9 @@ if fname is not None:\n
                             <string>fname</string>
                             <string>data</string>
                             <string>None</string>
-                            <string>_getattr_</string>
-                            <string>context</string>
                             <string>hasattr</string>
+                            <string>context</string>
+                            <string>_getattr_</string>
                           </tuple>
                         </value>
                     </item>

bt5/erp5_dms/SkinTemplateItem/portal_skins/erp5_dms/Document_getDoctypeFromContent.xml

+<?xml version="1.0"?>
+<ZopeData>
+  <record id="1" aka="AAAAAAAAAAE=">
+    <pickle>
+      <tuple>
+        <tuple>
+          <string>Products.ExternalMethod.ExternalMethod</string>
+          <string>ExternalMethod</string>
+        </tuple>
+        <none/>
+      </tuple>
+    </pickle>
+    <pickle>
+      <dictionary>
+        <item>
+            <key> <string>__ac_local_roles__</string> </key>
+            <value>
+              <none/>
+            </value>
+        </item>
+        <item>
+            <key> <string>_function</string> </key>
+            <value> <string>getDoctypeFromContent</string> </value>
+        </item>
+        <item>
+            <key> <string>_module</string> </key>
+            <value> <string>documentUtils</string> </value>
+        </item>
+        <item>
+            <key> <string>_owner</string> </key>
+            <value>
+              <none/>
+            </value>
+        </item>
+        <item>
+            <key> <string>id</string> </key>
+            <value> <string>Document_getDoctypeFromContent</string> </value>
+        </item>
+        <item>
+            <key> <string>title</string> </key>
+            <value> <string></string> </value>
+        </item>
+      </dictionary>
+    </pickle>
+  </record>
+</ZopeData>

bt5/erp5_dms/bt/revision

-317
\ No newline at end of file
+321
\ No newline at end of file

bt5/erp5_dms/bt/template_extension_id_list

@@ -3,3 +3,4 @@ searchUtils
 mailUtils
 cutFound
 asSecurityGroupId
+documentUtils
\ No newline at end of file