only decode ids for found positions

merge document.view.get_positions into item.utils.get_positions
add flag to decode id before looking up in ids

Followup to 09ebbc9cc6

pandora/document/views.py

@@ -165,33 +165,6 @@ def parse_query(data, user):
     return query
 
 
-def get_positions(qs, query_positions):
-    '''
-    qs: a QuerySet
-    query_positions: a list of AZ ids
-
-    TODO: merge this with item.utils.get_positions. The win is to fetch
-    only the integer IDs and convert the (smaller) set of query_positions to
-    ints, rather than fetch all keys for everything in qs (expected to be many
-    orders of magnitude larger), ignore most of it, and convert those ids to
-    strings.
-
-    Returns:
-        {
-            i: index of i in qs
-            for i in query_positions
-        }
-    '''
-    ids = list(qs.values_list('id', flat=True))
-    ret = {}
-    for i in query_positions:
-        try:
-            ret[i] = ids.index(ox.fromAZ(i))
-        except:
-            pass
-    return ret
-
-
 def findDocuments(request, data):
     '''
     Finds documents for a given query
@@ -225,7 +198,8 @@ def findDocuments(request, data):
         #FIXME: actually implement position requests
         response['data']['position'] = 0
     elif 'positions' in data:
-        response['data']['positions'] = get_positions(qs, query['positions'])
+        ids = list(qs.values_list('id', flat=True))
+        response['data']['positions'] = utils.get_positions(ids, query['positions'], decode_id=True)
     else:
         r = qs.aggregate(
             Sum('size')