[Sugar-devel] Searching for datastore entries by title

[Bert Freudenberg]

On 16.09.2009, at 13:48, Tomeu Vizoso wrote:

> On Wed, Sep 16, 2009 at 12:09, Bert Freudenberg  
> <bert at freudenbergs.de> wrote:
>>
>> On 16.09.2009, at 03:11, Gary C Martin wrote:
>>
>>> Hi Bert,
>>>
>>> On 16 Sep 2009, at 00:51, Bert Freudenberg wrote:
>>>
>>>> I want to find a datastore entry with a specific title and  
>>>> mime_type.
>>>> But it returns *all* entries with that mimetype instead. Here's the
>>>> log:
>>>>
>>>> 1253058031.600170 DEBUG root: datastore.find
>>>> dbus.Dictionary({dbus.String(u'mime_type'):
>>>> dbus.String(u'application/
>>>> x-squeak-project'), dbus.String(u'title'): dbus.String(u'The  
>>>> Hole')},
>>>> signature=dbus.Signature('ss'))
>>>> 1253058031.600731 DEBUG root: parse_query
>>>> dbus.Dictionary({dbus.String(u'mime_type'):
>>>> dbus.String(u'application/
>>>> x-squeak-project'), dbus.String(u'title'): dbus.String(u'The  
>>>> Hole')},
>>>> signature=dbus.Signature('ss')) None
>>>> 1253058031.600886 WARNING root: Unknown term:
>>>> dbus.String(u'title')=dbus.String(u'The Hole')
>>>> 1253058031.601042 WARNING root: Unknown term(s):
>>>> {dbus.String(u'mime_type'): dbus.String(u'application/x-squeak-
>>>> project'), dbus.String(u'title'): dbus.String(u'The Hole')}
>>>> 1253058031.601170 DEBUG root: queries:
>>>> ['Xapian::Query(FMapplication/x-
>>>> squeak-project)']
>>>> 1253058031.602399 DEBUG org.laptop.sugar.DataStore: find():
>>>> 0.0017540454864501953
>>>>
>>>> What does "unknown term" mean? It seams to get ignored. But it's in
>>>> the log quite often, not only from my query, that does not sound
>>>> right, something seems broken:
>>>>
>>>> 1253057661.236064 DEBUG root: datastore.find
>>>> dbus.Dictionary({dbus.String(u'limit'): dbus.Int32(5,
>>>> variant_level=1), dbus.String(u'order_by'):
>>>> dbus.Array([dbus.String(u'+timestamp')],
>>>> signature=dbus.Signature('s'), variant_level=1),
>>>> dbus.String(u'activity'): dbus.String(u'org.laptop.Pippy',
>>>> variant_level=1)}, signature=dbus.Signature('sv'))
>>>> 1253057661.236569 DEBUG root: parse_query
>>>> dbus.Dictionary({dbus.String(u'activity'):
>>>> dbus.String(u'org.laptop.Pippy', variant_level=1)},
>>>> signature=dbus.Signature('sv')) None
>>>> 1253057661.236753 WARNING root: Unknown term(s):
>>>> {dbus.String(u'activity'): dbus.String(u'org.laptop.Pippy',
>>>> variant_level=1)}
>>>> 1253057661.236886 DEBUG root: queries:
>>>> ['Xapian::Query(FAorg.laptop.Pippy)']
>>>> 1253057661.240903 DEBUG org.laptop.sugar.DataStore: find():
>>>> 0.0045418739318847656
>>>>
>>>> My test program (run in Pippy) and its output was this:
>>>>
>>>> =================================
>>>> def unwrap(x):
>>>>    if isinstance(x, list): return map(unwrap, x)
>>>>    if isinstance(x, tuple): return tuple(map(unwrap, x))
>>>>    if isinstance(x, dict): return dict([(unwrap(k), unwrap(v)) for
>>>> k, v in x.iteritems()])
>>>>    for t in [unicode, str, long, int, float, bool]:
>>>>        if isinstance(x, t): return t(x)
>>>>    return x
>>>>
>>>> import dbus
>>>> datastore =
>>>> dbus.SessionBus().get_object('org.laptop.sugar.DataStore',
>>>> '/org/laptop/sugar/DataStore')
>>>> (items,count) = datastore.find(
>>>>    {
>>>>      'title':'The Hole',
>>>
>>> Hmm FWIW, from my copy-from-journal experience, I though title was
>>> no longer supported (I opened a ticket about in 0.84 and it was
>>> closed as 'not a bug' and the copy-from-journal script changed). I
>>> think you search with 'query' instead now... Might have happened as
>>> part of Tomeu's (much welcomed) re-write of the data-store for
>>> robustness, back in 0.84 I guess.
>>
>>
>> 'query' searches for (sub-) strings in all text fields. I want an
>> exact title match.
>
> Several features were dropped from the DS when it was rewritten,
> because non-fundamental features were getting in the way of
> fundamental ones (like retaining the users' data).
>
> I thought this had been discussed back then (almost a year ago).

If so I forgot that title search was removed.

> If there's interest, now should be easier to add more features but we
> should remember that the DS is a critical component and we don't have
> too many resources to make sure we don't regress in robustness.


I wasn't really complaining, just asking ;)

Where do I find what searching options are supported now?

And what about the many "Unknown term(s)" warnings? Just ignore?  
Shouldn't this not be used anymore if it is not supported?

- Bert -