Respect list privacy when searching

4 files changed, 37 insertions, 9 deletions

diff --git a/kittystore/storm/model.py b/kittystore/storm/model.py
index eb89186..29cd546 100644
--- a/kittystore/storm/model.py
+++ b/kittystore/storm/model.py
@@ -162,6 +162,7 @@ class Thread(Storm):
                 order_by=Email.thread_order
              )
     category_obj = Reference(category_id, "Category.id")
+    mlist = Reference(list_name, "List.name")
     _starting_email = None
 
     def __init__(self, list_name, thread_id, date_active=None):
diff --git a/kittystore/storm/search.py b/kittystore/storm/search.py
index 3098aa6..65a79aa 100644
--- a/kittystore/storm/search.py
+++ b/kittystore/storm/search.py
@@ -18,9 +18,11 @@ import os
 import shutil
 
 from whoosh.index import create_in, exists_in, open_dir
-from whoosh.fields import Schema, ID, TEXT, DATETIME, KEYWORD
+from whoosh.fields import Schema, ID, TEXT, DATETIME, KEYWORD, BOOLEAN
 from whoosh.analysis import StemmingAnalyzer
 from whoosh.qparser import MultifieldParser
+from whoosh.query import Term
+from mailman.interfaces.archiver import ArchivePolicy
 
 from .model import Email
 
@@ -28,6 +30,7 @@ from .model import Email
 def email_to_search_doc(email):
     if not isinstance(email, Email):
         raise ValueError("not an instance of the Email class")
+    private_list = (email.mlist.archive_policy == ArchivePolicy.private)
     search_doc = {
             "list_name": email.list_name,
             "message_id": email.message_id,
@@ -36,6 +39,7 @@ def email_to_search_doc(email):
             "subject": email.subject,
             "content": email.content,
             "date": email.date, # UTC
+            "private_list": private_list,
     }
     attachments = [a.name for a in email.attachments]
     if attachments:
@@ -61,6 +65,7 @@ class SearchEngine(object):
                 date=DATETIME(),
                 attachments=TEXT,
                 tags=KEYWORD(commas=True, scorable=True),
+                private_list=BOOLEAN(),
             )
 
     @property
@@ -86,7 +91,8 @@ class SearchEngine(object):
         else:
             writer.commit()
 
-    def search(self, query, page=None, limit=10, sortedby=None, reverse=False):
+    def search(self, query, list_name=None, page=None, limit=10,
+               sortedby=None, reverse=False):
         """
         TODO: Should the searcher be shared?
         http://pythonhosted.org/Whoosh/threads.html#concurrency
@@ -94,16 +100,22 @@ class SearchEngine(object):
         query = MultifieldParser(
                 ["sender", "subject", "content", "attachments"],
                 self.index.schema).parse(query)
+        if list_name:
+            results_filter = Term("list_name", list_name)
+        else:
+            # When searching all lists, only the public lists are searched
+            results_filter = Term("private_list", False)
         return_value = {"total": 0, "results": []}
         with self.index.searcher() as searcher:
             if page:
                 results = searcher.search_page(
                         query, page, pagelen=limit, sortedby=sortedby,
-                        reverse=reverse)
+                        reverse=reverse, filter=results_filter)
                 return_value["total"] = results.total
             else:
                 results = searcher.search(
-                        query, limit=limit, sortedby=sortedby, reverse=reverse)
+                        query, limit=limit, sortedby=sortedby,
+                        reverse=reverse, filter=results_filter)
                 # http://pythonhosted.org/Whoosh/searching.html#results-object
                 if results.has_exact_length():
                     return_value["total"] = len(results)
@@ -152,3 +164,10 @@ class SearchEngine(object):
             shutil.rmtree(self.location)
             self._index = None
             self.initialize_with(store)
+        new_schema = self._get_schema()
+        writer = self.index.writer()
+        for field_name, field_type in new_schema.items():
+            if field_name not in self.index.schema:
+                print "Adding field %s to the search index" % field_name
+                writer.add_field(field_name, field_type)
+        writer.commit(optimize=True)
diff --git a/kittystore/storm/store.py b/kittystore/storm/store.py
index e2e439c..b640229 100644
--- a/kittystore/storm/store.py
+++ b/kittystore/storm/store.py
@@ -20,6 +20,7 @@ from urllib2 import HTTPError
 
 from zope.interface import implements
 from mailman.interfaces.messages import IMessageStore
+from mailman.interfaces.archiver import ArchivePolicy
 from storm.locals import Desc
 from storm.expr import And, Or, Count, Alias
 from dateutil.tz import tzutc
@@ -97,6 +98,9 @@ class StormStore(object):
         l.display_name = mlist.display_name
         l.subject_prefix = mlist.subject_prefix
         l.archive_policy = mlist.archive_policy
+        if l.archive_policy == ArchivePolicy.never:
+            print "Archiving disabled by list policy for %s" % list_name
+            return None
         if not message.has_key("Message-Id"):
             raise ValueError("No 'Message-Id' header in email", message)
         msg_id = unicode(unquote(message['Message-Id']))
@@ -297,8 +301,10 @@ class StormStore(object):
 
     def search(self, query, list_name=None, page=None, limit=10,
                sortedby=None, reverse=False):
-        """ Returns a list of email containing the specified keyword in
-        their content.
+        """
+        Returns a list of email corresponding to the query string. The
+        sender, subject, content and attachment names are searched. If
+        list_name is None, all public lists are searched.
 
         :param query: the query string to execute.
         :param list_name: name of the mailing list in which this email
@@ -310,10 +316,9 @@ class StormStore(object):
             by match score.
         :param reverse: reverse the order of the results.
         """
-        if list_name is not None:
-            query += " list_name:%s" % list_name
         results = self.search_index.search(
-                query, page, limit, sortedby=sortedby, reverse=reverse)
+                query, list_name, page, limit, sortedby=sortedby,
+                reverse=reverse)
         results["results"] = [ self.get_message_by_id_from_list(
                                     r["list_name"], r["message_id"])
                                for r in results["results"] ]
diff --git a/kittystore/test/__init__.py b/kittystore/test/__init__.py
index 170473e..0270307 100644
--- a/kittystore/test/__init__.py
+++ b/kittystore/test/__init__.py
@@ -2,6 +2,8 @@
 
 import os
 
+from mailman.interfaces.archiver import ArchivePolicy
+
 
 def get_test_file(*fileparts):
     return os.path.join(os.path.dirname(__file__), "testdata", *fileparts)
@@ -15,6 +17,7 @@ class FakeList(object):
         self.fqdn_listname = name
         self.display_name = None
         self.subject_prefix = None
+        self.archive_policy = ArchivePolicy.public
 
 class SettingsModule:
     KITTYSTORE_URL = "sqlite:"