def mine(self,tuple_size,num_tuples,weights=None,expand=False,table_size=2**19):
if not weights and not expand:
ldb=sa.sampledmh_mine(self.ldb,tuple_size,num_tuples,table_size)
elif expand and not weights:
max_freq=sa.mh_get_cumulative_frequency(self.ldb,expand.ldb)
ldb_=sa.mh_expand_listdb(self.ldb,max_freq)
ldb=sa.sampledmh_mine(ldb_,tuple_size,num_tuples,table_size)
elif not expand and weights:
ldb=sa.sampledmh_mine_weighted(self.ldb,ifindex, tuple_size,num_tuples,table_size,weights.weights)
elif expand and weights:
max_freq=sa.mh_get_cumulative_frequency(self.ldb,expand.ldb)
ldb_=sa.mh_expand_listdb(self.ldb, max_freq)
weights_=sa.mh_expand_weights(expand.ldb.size, max_freq, weights.weights)
ldb=sa.sampledmh_mine_weighted(ldb_,tuple_size,num_tuples,table_size,weights_)
return SMH(ldb=ldb)
def mine(self,
listdb,
weights = None,
expand = None):
"""
samples inverted file to mine sets of highly co-occurring items
"""
if not weights and not expand:
mined = sa.sampledmh_mine(listdb.ldb,
self.tuple_size_,
self.number_of_tuples_,
self.table_size_,
self.min_set_size_)
elif expand and not weights:
max_freq = sa.mh_get_cumulative_frequency(listdb.ldb, expand.ldb)
ldb_ = sa.mh_expand_listdb(listdb.ldb, max_freq)
mined = sa.sampledmh_mine(ldb_,
self.tuple_size_,
self.number_of_tuples_,
self.table_size_,
self.min_set_size_)
sa.listdb_destroy(ldb_)
elif not expand and weights:
mined = sa.sampledmh_mine_weighted(listdb.ldb,
self.tuple_size_,
self.number_of_tuples_,
self.table_size_,
weights.weights,
self.min_set_size_)
elif expand and weights:
max_freq = sa.mh_get_cumulative_frequency(listdb.ldb, expand.ldb)
ldb_ = sa.mh_expand_listdb(listdb.ldb, max_freq)
weights_ = sa.mh_expand_weights(expand.ldb.size, max_freq, weights.weights)
mined = sa.sampledmh_mine_weighted(ldb_,
self.tuple_size_,
self.number_of_tuples_,
self.table_size_,
weights_,
self.min_set_size_)
sa.listdb_destroy(ldb_)
sa.listdb_delete_smallest(mined, self.min_set_size_)
return ListDB(ldb = mined)
def mine(self,
listdb,
weights = None,
expand = None):
"""
samples nverted file to mine sets of highly co-occurring items
"""
if not weights and not expand:
mined = sa.sampledmh_mine(listdb.ldb,
self.tuple_size_,
self.number_of_tuples_,
self.table_size_,
self.min_set_size_)
elif expand and not weights:
max_freq = sa.mh_get_cumulative_frequency(listdb.ldb, expand.ldb)
ldb_ = sa.mh_expand_listdb(listdb.ldb, max_freq)
mined = sa.sampledmh_mine(ldb_,
self.tuple_size_,
self.number_of_tuples_,
self.table_size_,
self.min_set_size_)
sa.listdb_destroy(ldb_)
elif not expand and weights:
mined = sa.sampledmh_mine_weighted(listdb.ldb,
self.tuple_size_,
self.number_of_tuples_,
self.table_size_,
weights.weights,
self.min_set_size_)
elif expand and weights:
max_freq = sa.mh_get_cumulative_frequency(listdb.ldb, expand.ldb)
ldb_ = sa.mh_expand_listdb(listdb.ldb, max_freq)
weights_ = sa.mh_expand_weights(expand.ldb.size, max_freq, weights.weights)
mined = sa.sampledmh_mine_weighted(ldb_,
self.tuple_size_,
self.number_of_tuples_,
self.table_size_,
weights_,
self.min_set_size_)
sa.listdb_destroy(ldb_)
sa.listdb_delete_smallest(mined, self.min_set_size_)
return ListDB(ldb = mined)
def invert(self):
ldb=sa.sampledmh_mine(self.ldb)
ldb._inverted=True
ldb._original=self
return SMH(ldb=ldb)
