Merge lp:~ntoll/fom/instance-pre-populate-search-812246 into lp:fom

[Nicholas Tollervey]

This branch introduces the following changes:

1) When calling the filter() method on subclasses of Object FOM will use the /values API to attempt to pre-populate the attribute values. This ensures that the results are instantiated *with* initial values and no more lazy-getting of attributes is required (thus saving time and calls to Fluidinfo).

2) Changes in the unit tests to reflect this new functionality.

This change was made in light of behaviour observed whilst writing the O'Reilly book.

[lp:~ntoll/fom/instance-pre-populate-search-812246]

151. By Nicholas Tollervey on 2011-07-18

Whitespace clean

[Jamu Kakar (jkakar)]

[1]

         self.db.add_resp(200, 'application/json',
- '{"ids": ["466"]}')
+ '''{"results":{"id": {"05eee31e-fbd1-43cc-9500-0469707a9bc3":{
+ "fluiddb/users/username":{"value": "ntoll"},
+ "fluiddb/users/name":{"value": "Nicholas"},
+ "fluiddb/about": {"value": "Object about the user ntoll"}}}}}''')

There are a few bits of code like this. I think they'd be easier to
read if the data was split up a bit and used the json module:

         objectValues = {
             'fluiddb/users/username': {'value': 'ntoll'},
             'fluiddb/users/name': {'value': 'Nicholas'},
             'fluiddb/about': {'value': 'Object about the user ntoll'}}
         resultValues = {'05eee31e-fbd1-43cc-9500-0469707a9bc3': objectData}
         self.db.add_resp(200, 'application/json',
                          dumps({'results': {'id': resultValues}}))

Nice work, +1!

review: Approve

[Jamu Kakar (jkakar)]

Er, s/objectData/objectValues/ in the example above.

[lp:~ntoll/fom/instance-pre-populate-search-812246]

152. By Nicholas Tollervey on 2011-07-19

Updated the unit tests to make the expected response slightly more clear