def merge_grid_cover(final_db, other_dbs):
if not isfile(final_db):
copyfile(other_dbs[0], final_db)
other_dbs = other_dbs[1:]
final_db = GridCoverDB.load_from_filepath(final_db)
for other_db_filename in other_dbs:
other_db = GridCoverDB.load_from_filepath(other_db_filename)
final_db.load_other(other_db)
def test_get_centroids(self):
ramifier = RotatingRamifier.from_file(4, KMER_ROTATION)
db = GridCoverDB(sqlite3.connect(':memory:'),
ramifier=ramifier,
box_side_len=0.5)
db.py_add_point_to_cluster(np.array([0., 0., 0., 0.]), KMER_30 + 'A')
db.py_add_point_to_cluster(np.array([0., 0., 0., 0.]), KMER_30 + 'T')
db.py_add_point_to_cluster(np.array([1., 0., 0., 0.]), KMER_30 + 'C')
db.commit()
centroids = db.centroids()
self.assertEqual(centroids.shape, (2, 4))
def cli_dump_kmers(outfile, cluster_ids, grid_cover):
grid = GridCoverDB.load_from_filepath(grid_cover)
for centroid_index, kmer in grid.get_kmers():
if cluster_ids:
print(f'{centroid_index},{kmer}', file=outfile)
else:
print(kmer, file=outfile)
def cli_dump_kmers(outfile, grid_cover):
grid = GridCoverDB.load_from_filepath(grid_cover)
counts = {}
for centroid_index, _ in grid.get_kmers():
counts[centroid_index] = 1 + counts.get(centroid_index, 0)
for centroid_index, count in counts.items():
print(f'{centroid_index},{count}', file=outfile)
def build_grid_cover(grid_db):
db = GridCoverDB.load_from_filepath(grid_db)
start = time()
n_centers = db.centroids().shape[0]
with click.progressbar(list(range(n_centers))) as centroid_ids:
for centroid_id in centroid_ids:
db.build_and_store_bloom_grid(centroid_id)
db.close()
add_time = time() - start
click.echo(f'Built {n_centers} bloom filters in {add_time:.5}s.', err=True)
def cli_dump_kmers(grid_cover):
click.echo(grid_cover)
grid = GridCoverDB.load_from_filepath(grid_cover)
n_centers = grid.centroids().shape[0]
click.echo(f'centers\t{n_centers}')
n_kmers = len(grid.get_kmers())
click.echo(f'kmers\t{n_kmers}')
box_side = grid.box_side_len
click.echo(f'box_side\t{box_side}')
dims = grid.ramifier.d
click.echo(f'dims\t{dims}')
def test_build_grid_cover_from_fasta(self):
ramifier = RotatingRamifier.from_file(4, KMER_ROTATION)
db = GridCoverDB(sqlite3.connect(':memory:'),
ramifier=ramifier,
box_side_len=0.5)
grid = GridCoverBuilder(db)
grid.fast_add_kmers_from_fasta(KMER_FASTA)
grid.commit()
n_centers = grid.db.centroids().shape[0]
n_points = len(grid.db.get_kmers())
self.assertGreater(n_centers, 0)
self.assertLess(n_centers, 98)
self.assertEqual(n_points, 98)
def test_add_kmer(self):
ramifier = RotatingRamifier.from_file(4, KMER_ROTATION)
db = GridCoverDB(sqlite3.connect(':memory:'),
ramifier=ramifier,
box_side_len=0.5)
db.py_add_point_to_cluster(np.array([0., 0., 0., 0.]), KMER_31)
db.commit()
members = db.py_get_cluster_members(0)
self.assertEqual(len(members), 1)
self.assertIn(KMER_31,
[reverse_convert_kmer(member) for member in members])
def test_save_and_reload(self):
DB_SAVE_TEMP_FILE = join(dirname(__file__), 'temp.db_save_temp.sqlite')
ramifier = RotatingRamifier.from_file(4, KMER_ROTATION)
db = GridCoverDB(sqlite3.connect(DB_SAVE_TEMP_FILE),
ramifier=ramifier,
box_side_len=0.5)
db.py_add_point_to_cluster(np.array([0., 0., 0., 0.]), KMER_31)
db.close()
del db
db = GridCoverDB.load_from_filepath(DB_SAVE_TEMP_FILE)
members = db.py_get_cluster_members(0)
self.assertEqual(len(members), 1)
self.assertIn(KMER_31,
[reverse_convert_kmer(member) for member in members])
remove(DB_SAVE_TEMP_FILE)
def test_save(self):
DB_SAVE_TEMP_FILE = join(dirname(__file__), 'temp.db_save_temp.sqlite')
ramifier = RotatingRamifier.from_file(4, KMER_ROTATION)
db = GridCoverDB(sqlite3.connect(DB_SAVE_TEMP_FILE),
ramifier=ramifier,
box_side_len=0.5)
db.py_add_point_to_cluster(np.array([0., 0., 0., 0.]), KMER_31)
db.close()
remove(DB_SAVE_TEMP_FILE)
def cli_dump_kmers(outfile, grid_cover):
grid = GridCoverDB.load_from_filepath(grid_cover)
pd.DataFrame(grid.centroids()).to_csv(outfile, header=None, index=None)
def test_pre_build_blooms(self):
ramifier = RotatingRamifier.from_file(4, KMER_ROTATION)
db = GridCoverDB(sqlite3.connect(':memory:'),
ramifier=ramifier,
box_side_len=0.5)
db.py_add_point_to_cluster(np.array([0., 0., 0., 0.]), KMER_30 + 'A')
db.py_add_point_to_cluster(np.array([0., 0., 0., 0.]), KMER_30 + 'T')
db.py_add_point_to_cluster(np.array([1., 0., 0., 0.]), KMER_30 + 'C')
db.commit()
for centroid_id in [0, 1]:
db.build_and_store_bloom_grid(centroid_id)
bg_0 = db.retrieve_bloom_grid(0)
bg_1 = db.retrieve_bloom_grid(1)
self.assertEqual(max(bg_0.py_count_grid_contains(KMER_30 + 'A')),
32 - bg_0.col_k)
self.assertEqual(max(bg_1.py_count_grid_contains(KMER_30 + 'C')),
32 - bg_1.col_k)
self.assertRaises(IndexError, lambda: db.retrieve_bloom_grid(2))
def test_merge_dbs(self):
ramifier = RotatingRamifier.from_file(4, KMER_ROTATION)
db1 = GridCoverDB(sqlite3.connect(':memory:'),
ramifier=ramifier,
box_side_len=0.5)
db1.py_add_point_to_cluster(np.array([0., 0., 0., 0.]), KMER_30 + 'A')
db1.py_add_point_to_cluster(np.array([1., 0., 0., 0.]), KMER_30 + 'T')
db1.commit()
db2 = GridCoverDB(sqlite3.connect(':memory:'),
ramifier=ramifier,
box_side_len=0.5)
db2.py_add_point_to_cluster(np.array([0., 0., 0., 0.]), KMER_30 + 'C')
db2.py_add_point_to_cluster(np.array([1., 1., 0., 0.]), KMER_30 + 'G')
db2.commit()
db1.load_other(db2)
centroids = db1.centroids()
self.assertEqual(centroids.shape, (3, 4))
kmers = [el[1] for el in db1.get_kmers()]
self.assertEqual(len(kmers), 4)
for char in 'ATCG':
self.assertIn(KMER_30 + char, kmers)