Wildcard lookup! #89

Test suite no longer reloads the entire database.  Takes too long.

Factored out some magic numbers in lookup().

diff --git a/pokedex/lookup.py b/pokedex/lookup.py
index cb6d5f6..35c98e2 100644 (file)
--- a/pokedex/lookup.py
+++ b/pokedex/lookup.py
@@ -20,6 +20,9 @@ from pokedex.roomaji import romanize
 
 __all__ = ['open_index', 'lookup']
 
 
 __all__ = ['open_index', 'lookup']
 

+INTERMEDIATE_LOOKUP_RESULTS = 25
+MAX_LOOKUP_RESULTS = 10
+

 # Dictionary of table name => table class.
 # Need the table name so we can get the class from the table name after we
 # retrieve something from the index
 # Dictionary of table name => table class.
 # Need the table name so we can get the class from the table name after we
 # retrieve something from the index


@@ -242,10 +245,13 @@ def lookup(input, valid_types=[], session=None, indices=None, exact_only=False):
             # provided
             valid_types = prefixes
 
             # provided
             valid_types = prefixes
 

-    # If the input provided is a number, match it as an id.  Otherwise, name.
-    # Term objects do an exact match, so we don't have to worry about a query
-    # parser tripping on weird characters in the input
-    if rx_is_number.match(name):
+    # Do different things depending what the query looks like
+    # Note: Term objects do an exact match, so we don't have to worry about a
+    # query parser tripping on weird characters in the input
+    if '*' in name or '?' in name:
+        exact_only = True
+        query = whoosh.query.Wildcard(u'name', name)
+    elif rx_is_number.match(name):

         # Don't spell-check numbers!
         exact_only = True
         query = whoosh.query.Term(u'row_id', name)
         # Don't spell-check numbers!
         exact_only = True
         query = whoosh.query.Term(u'row_id', name)


@@ -285,14 +291,14 @@ def lookup(input, valid_types=[], session=None, indices=None, exact_only=False):
     searcher.weighting = LanguageWeighting()  # XXX kosher?  docs say search()
                                               # takes a weighting kw but it
                                               # certainly does not
     searcher.weighting = LanguageWeighting()  # XXX kosher?  docs say search()
                                               # takes a weighting kw but it
                                               # certainly does not

-    results = searcher.search(query)
+    results = searcher.search(query, limit=INTERMEDIATE_LOOKUP_RESULTS)

 
     # Look for some fuzzy matches if necessary
     if not exact_only and not results:
         exact = False
         results = []
 
 
     # Look for some fuzzy matches if necessary
     if not exact_only and not results:
         exact = False
         results = []
 

-        for suggestion in speller.suggest(name, 25):
+        for suggestion in speller.suggest(name, INTERMEDIATE_LOOKUP_RESULTS):

             query = whoosh.query.Term('name', suggestion)
             results.extend(searcher.search(query))
 
             query = whoosh.query.Term('name', suggestion)
             results.extend(searcher.search(query))
 


@@ -319,4 +325,4 @@ def lookup(input, valid_types=[], session=None, indices=None, exact_only=False):
     # should have more than 10 here and lost a few.  The speller returns 25 to
     # give us some padding, and should avoid that problem.  Not a big deal if
     # we lose the 25th-most-likely match anyway.
     # should have more than 10 here and lost a few.  The speller returns 25 to
     # give us some padding, and should avoid that problem.  Not a big deal if
     # we lose the 25th-most-likely match anyway.

-    return objects[:10]
+    return objects[:MAX_LOOKUP_RESULTS]

diff --git a/pokedex/tests/__init__.py b/pokedex/tests/__init__.py
index 29b7a56..7565cc7 100644 (file)
--- a/pokedex/tests/__init__.py
+++ b/pokedex/tests/__init__.py
@@ -7,9 +7,8 @@ from pokedex.db.load import load
 def setup():
     # Reload data just in case
     session = connect()
 def setup():
     # Reload data just in case
     session = connect()

-    load(session, verbose=False, drop_tables=True)

     open_index(session=session, recreate=True)
 
 
 def teardown():
     open_index(session=session, recreate=True)
 
 
 def teardown():

-    print "teardown"
+    pass

diff --git a/pokedex/tests/test_lookup.py b/pokedex/tests/test_lookup.py
index b6bef60..703cde9 100644 (file)
--- a/pokedex/tests/test_lookup.py
+++ b/pokedex/tests/test_lookup.py
@@ -89,3 +89,16 @@ def test_fuzzy_lookup():
     top_names = [_.object.name for _ in results[0:2]]
     assert_true(u'Nidoran♂' in top_names,       u'Nidoran♂ is a top result for "Nidoran"')
     assert_true(u'Nidoran♀' in top_names,       u'Nidoran♀ is a top result for "Nidoran"')
     top_names = [_.object.name for _ in results[0:2]]
     assert_true(u'Nidoran♂' in top_names,       u'Nidoran♂ is a top result for "Nidoran"')
     assert_true(u'Nidoran♀' in top_names,       u'Nidoran♀ is a top result for "Nidoran"')

+
+def test_wildcard_lookup():
+    tests = [
+        (u'pokemon:*meleon',    u'Charmeleon'),
+        (u'item:master*',       u'Master Ball'),
+        (u'ee?ee',              u'Eevee'),
+    ]
+
+    for wildcard, name in tests:
+        results = pokedex.lookup.lookup(wildcard)
+        first_result = results[0]
+        assert_equal(first_result.object.name, name,
+                                                u'Wildcards work correctly')