def insert_hashes(self, sid, hashes):
"""
Insert series of hash => song_id, offset
values into the database.
"""
values = []
for hash, offset in hashes:
values.append((hash, sid, offset))
with self.cursor() as cur:
for split_values in grouper(values, 1000):
cur.executemany(self.INSERT_FINGERPRINT, split_values)
def return_matches(self, hashes):
"""
Return the (song_id, offset_diff) tuples associated with
a list of (sha1, sample_offset) values.
"""
# Create a dictionary of hash => offset pairs for later lookups
mapper = {}
for hash, offset in hashes:
mapper[hash] = offset
# Get an iteratable of all the hashes we need
values = mapper.keys()
for split_values in grouper(values, 900):
results = self._DBSession.query(Fingerprint.hash, Fingerprint.song_id,
Fingerprint.offset).filter(Fingerprint.hash.in_(split_values))
for hash, sid, offset in results:
# (sid, db_offset - song_sampled_offset)
yield (sid, offset - mapper[hash])
def return_matches(self, hashes):
"""
Return the (song_id, offset_diff) tuples associated with
a list of (sha1, sample_offset) values.
"""
# Create a dictionary of hash => offset pairs for later lookups
mapper = {}
for hash, offset in hashes:
mapper[hash] = offset
# Get an iteratable of all the hashes we need
values = mapper.keys()
with self.cursor() as cur:
for split_values in grouper(values, 1000):
# Create our IN part of the query
query = self.SELECT_MULTIPLE
query = query % ', '.join(['UNHEX(%s)'] * len(split_values))
cur.execute(query, split_values)
for hash, sid, offset in cur:
# (sid, db_offset - song_sampled_offset)
yield (sid, offset - mapper[hash])
