def test_indexing():
con = sqlite3.connect(":memory:")
_fill_exons(con)
con.commit()
query = genomicsqlite.genomic_range_rowids_sql(con, "exons")
query = "SELECT id FROM exons WHERE exons._rowid_ IN\n" + query
print("\n" + query)
# The query should only search the relevant GRI levels
indexed = 0
for expl in con.execute("EXPLAIN QUERY PLAN " + query,
("chr17", 43044294, 43048294)):
print(expl[3])
if "USING INDEX exons__gri" in expl[3]:
indexed += 1
assert indexed == 3
# The query should be covered by the index except for one final fetch of exons.id
opcodes = list(
con.execute("EXPLAIN " + query, ("chr17", 43044294, 43048294)))
# for expl in opcodes:
# if expl[1] in {"OpenRead", "OpenPseudo", "Column"}:
# print(expl)
accessed_cursors = list(opcode[2] for opcode in opcodes
if opcode[1] == "Column")
table_rootpages = set(row[0] for row in con.execute(
"SELECT rootpage FROM sqlite_master WHERE type='table'"))
table_cursors = set(
opcode[2] for opcode in opcodes
if opcode[1] == "OpenRead" and opcode[3] in table_rootpages)
assert sum(1 for cursor in accessed_cursors
if cursor not in table_cursors) > 1 # get from idx
assert sum(1 for cursor in accessed_cursors
if cursor in table_cursors) == 1 # get from table
# (Also, it should produce correct results)
control_query = "SELECT id FROM exons NOT INDEXED WHERE rid = ? AND NOT (end < ? OR beg > ?) ORDER BY _rowid_"
random.seed(0xBADF00D)
total_results = 0
for _ in range(50000):
beg = random.randint(43044294 - 10000, 43044294 + 10000)
end = beg + random.randint(1, random.choice([10, 100, 1000, 10000]))
ids = list(row[0] for row in con.execute(query, ("chr17", beg, end)))
control_ids = list(
row[0] for row in con.execute(control_query, ("chr17", beg, end)))
assert ids == control_ids
total_results += len(control_ids)
assert total_results == 189935
with pytest.raises(sqlite3.OperationalError):
genomicsqlite.genomic_range_rowids_sql(con, "nonexistent_table")
def test_refseq():
con = sqlite3.connect(":memory:")
create_assembly = genomicsqlite.put_reference_assembly_sql(
con, "GRCh38_no_alt_analysis_set")
lines = create_assembly.strip().split("\n")
print("\n".join([line for line in lines if "INSERT INTO" in line][:24]))
assert len([line for line in lines if "INSERT INTO" in line]) == 195
print("\n".join([line for line in lines if "INSERT INTO" not in line]))
assert len([line for line in lines if "INSERT INTO" not in line]) == 2
con.executescript(create_assembly)
_fill_exons(con, floor=2)
con.commit()
refseq_by_rid = genomicsqlite.get_reference_sequences_by_rid(con)
refseq_by_name = genomicsqlite.get_reference_sequences_by_name(con)
for refseq in refseq_by_rid.values():
assert refseq_by_rid[refseq.rid] == refseq
assert refseq_by_name[refseq.name] == refseq
if refseq.name == "chr17":
assert refseq.rid == 17
assert refseq.length == 83257441
assert refseq.assembly == "GRCh38_no_alt_analysis_set"
assert refseq.refget_id == "f9a0fb01553adb183568e3eb9d8626db"
assert len(refseq_by_rid) == 195
query = (
"SELECT _gri_refseq._gri_rid, rid, beg, end, id FROM exons, _gri_refseq WHERE exons.rid = gri_refseq_name AND exons._rowid_ IN "
+ genomicsqlite.genomic_range_rowids_sql(con, "exons"))
print("\n" + query)
assert len([line for line in query.split("\n") if "BETWEEN" in line]) == 2
assert len(list(con.execute(query, ("chr17", 43115725, 43125370)))) == 56
def test_attach(tmp_path):
dbfile = str(tmp_path / "test.gsql")
con = genomicsqlite.connect(dbfile, unsafe_load=True)
_fill_exons(con, gri=False)
con.commit()
attach_script = genomicsqlite.attach_sql(con,
str(tmp_path /
"test_attached.gsql"),
"db2",
unsafe_load=True)
con.executescript(attach_script)
con.executescript("CREATE TABLE db2.exons2 AS SELECT * FROM exons")
con.executescript(
genomicsqlite.create_genomic_range_index_sql(con, "db2.exons2", "rid",
"beg", "end"))
ref_script = genomicsqlite.put_reference_assembly_sql(
con, "GRCh38_no_alt_analysis_set", schema="db2")
con.executescript(ref_script)
query = (
"SELECT exons.id, db2.exons2.id FROM exons LEFT JOIN db2.exons2 ON db2.exons2._rowid_ IN\n"
+ genomicsqlite.genomic_range_rowids_sql(
con, "db2.exons2", "exons.rid", "exons.beg", "exons.end") +
" AND exons.id != db2.exons2.id ORDER BY exons.id, db2.exons2.id")
results = list(con.execute(query))
assert len(results) == 5191
refseq_by_name = genomicsqlite.get_reference_sequences_by_name(
con, schema="db2")
assert len(refseq_by_name) > 24
def test_join():
for len_gri in (False, True):
con = sqlite3.connect(":memory:")
con.executescript(
genomicsqlite.put_reference_assembly_sql(
con, "GRCh38_no_alt_analysis_set"))
_fill_exons(con, table="exons")
_fill_exons(con, floor=2, table="exons2", len_gri=len_gri)
con.commit()
query = (
"SELECT exons.id, exons2.id FROM exons LEFT JOIN exons2 ON exons2._rowid_ IN\n"
+ genomicsqlite.genomic_range_rowids_sql(
con, "exons2", "exons.rid", "exons.beg", "exons.end") +
" AND exons.id != exons2.id ORDER BY exons.id, exons2.id")
print(query)
indexed = 0
for expl in con.execute("EXPLAIN QUERY PLAN " + query):
print(expl[3])
if ("((_gri_rid,_gri_lvl,_gri_beg)>(?,?,?) AND (_gri_rid,_gri_lvl,_gri_beg)<(?,?,?))"
in expl[3]):
indexed += 1
assert indexed == 2
results = list(con.execute(query))
assert len(results) == 5191
assert len([result for result in results if result[1] is None]) == 5
control = "SELECT exons.id, exons2.id FROM exons LEFT JOIN exons2 NOT INDEXED ON NOT (exons2.end < exons.beg OR exons2.beg > exons.end) AND exons.id != exons2.id ORDER BY exons.id, exons2.id"
control = list(con.execute(control))
assert results == control
def test_web():
conn = genomicsqlite.connect(
"https://github.com/mlin/sqlite_zstd_vfs/releases/download/web-test-db-v1/TxDb.Hsapiens.UCSC.hg38.knownGene.vacuum.genomicsqlite",
read_only=True,
)
results = list(t[0] for t in conn.execute(
"SELECT tx_name FROM transcript WHERE _rowid_ IN " +
genomicsqlite.genomic_range_rowids_sql(conn, "transcript") +
" ORDER BY tx_name",
("chr12", 111803912, 111804012),
))
print(results)
sys.stdout.flush()
assert results == sorted([
"ENST00000416293.7", "ENST00000261733.7", "ENST00000548536.1",
"ENST00000549106.1"
])
exon_cds_counts = """
SELECT exon._rowid_ AS exon_id, COUNT(cds._rowid_) AS contained_cds
FROM genomic_range_index_levels('cds'), exon LEFT JOIN cds
ON cds._rowid_ IN genomic_range_rowids('cds', exon_chrom, exon_start, exon_end, _gri_ceiling, _gri_floor)
AND (exon_start = cds_start AND exon_end >= cds_start OR exon_start <= cds_start AND exon_end = cds_end)
GROUP BY exon._rowid_
"""
for expl in conn.execute("EXPLAIN QUERY PLAN " + exon_cds_counts):
print(expl[3])
sys.stdout.flush()
cds_exon_count_hist = list(
conn.execute(
f"SELECT contained_cds, count(exon_id) AS exon_count FROM ({exon_cds_counts}) GROUP BY contained_cds ORDER BY contained_cds"
))
assert cds_exon_count_hist[:2] == [(0, 270532), (1, 310059)]
def test_abutment():
# Test GRI query correctness with various cases of features abutting the query range
con = sqlite3.connect(":memory:")
con.executescript("CREATE TABLE features(rid INTEGER, beg INTEGER, end INTEGER)")
pos0 = 10000000000
for lvl in range(9):
for ofs in (-1, 0, 1):
for tup in ((pos0 - 16 ** lvl, pos0 + ofs), (pos0 + 123 + ofs, pos0 + 123 + 16 ** 9)):
con.execute("INSERT INTO features VALUES(42,?,?)", tup)
con.executescript(
genomicsqlite.create_genomic_range_index_sql(con, "features", "rid", "beg", "end")
)
query = genomicsqlite.genomic_range_rowids_sql(
con, "features", "42", str(pos0), str(pos0 + 123)
)[1:-1]
control_query = f"SELECT _rowid_ FROM features NOT INDEXED WHERE rid = 42 AND NOT (end < {pos0} OR beg > {pos0+123}) ORDER BY _rowid_"
assert list(con.execute(query)) == list(con.execute(control_query))
def test_gnomad_sites_small(tmp_path):
dbfile = str(tmp_path / "test.gsql")
vcf_into_sqlite(
os.path.join(HERE, "data/gnomad.r3.0.sites.ALDH2.vcf.bgz"),
str(dbfile),
"--table-prefix",
"gnomad_",
)
con = genomicsqlite.connect(dbfile, read_only=True)
query = (
"SELECT variant_rowid, id_jsarray FROM (SELECT _gri_rid AS rid FROM _gri_refseq WHERE gri_refseq_name=?1) AS query, gnomad_variants WHERE variant_rowid IN"
+ genomicsqlite.genomic_range_rowids_sql(con, "gnomad_variants",
"query.rid"))
rs671 = ("chr12", 111803912, 111804012)
print(query)
indexed = 0
for expl in con.execute("EXPLAIN QUERY PLAN " + query, rs671):
print(expl[3])
if "USING INDEX gnomad_variants__gri" in expl[3]:
indexed += 1
assert indexed == 3
opcodes = list(con.execute("EXPLAIN " + query, rs671))
accessed_cursors = list(opcode[2] for opcode in opcodes
if opcode[1] == "Column")
table_rootpages = set(row[0] for row in con.execute(
"SELECT rootpage FROM sqlite_master WHERE type='table'"))
table_cursors = set(
opcode[2] for opcode in opcodes
if opcode[1] == "OpenRead" and opcode[3] in table_rootpages)
assert sum(
1 for cursor in accessed_cursors if cursor not in table_cursors) > 1
assert sum(1 for cursor in accessed_cursors
if cursor in table_cursors) == 1
results = list(con.execute(query, rs671))
results_rowids = set(vt[0] for vt in results)
assert next(vt for vt in results if vt[1] and "rs671" in vt[1])
control = "SELECT variant_rowid FROM gnomad_variants NATURAL JOIN _gri_refseq WHERE gri_refseq_name = ? AND NOT ((pos+rlen) < ? OR pos > ?)"
control_rowids = set(vt[0] for vt in con.execute(control, rs671))
assert len(control_rowids) == 22
assert results_rowids == control_rowids
def test_connect(tmp_path):
dbfile = str(tmp_path / "test.gsql")
con = genomicsqlite.connect(dbfile, unsafe_load=True)
con.executescript(
genomicsqlite.put_reference_assembly_sql(con,
"GRCh38_no_alt_analysis_set"))
_fill_exons(con)
con.commit()
del con
con = genomicsqlite.connect(dbfile, read_only=True)
query = (
"WITH exons2 AS (SELECT * from exons) SELECT exons.id, exons2.id FROM exons2 LEFT JOIN exons ON exons._rowid_ IN\n"
+ genomicsqlite.genomic_range_rowids_sql(con, "exons", "exons2.rid",
"exons2.beg", "exons2.end") +
" AND exons.id != exons2.id ORDER BY exons.id, exons2.id")
results = list(con.execute(query))
assert len(results) == 5191
def test_gnomad_sites_small(tmp_path):
dbfile = str(tmp_path / "gnomad_lines.gsql")
vcf_lines_into_sqlite(
os.path.join(HERE, "data/gnomad.r3.0.sites.ALDH2.vcf.bgz"),
str(dbfile),
"--table",
"gnomad_vcf_lines",
)
con = genomicsqlite.connect(dbfile, read_only=True)
query = ("SELECT _rowid_, line FROM gnomad_vcf_lines WHERE _rowid_ IN " +
genomicsqlite.genomic_range_rowids_sql(con, "gnomad_vcf_lines"))
rs671 = ("chr12", 111803912, 111804012)
print(query)
indexed = 0
for expl in con.execute("EXPLAIN QUERY PLAN " + query, rs671):
print(expl[3])
if "USING INDEX gnomad_vcf_lines__gri" in expl[3]:
indexed += 1
assert indexed == 3
opcodes = list(con.execute("EXPLAIN " + query, rs671))
accessed_cursors = list(opcode[2] for opcode in opcodes
if opcode[1] == "Column")
table_rootpages = set(row[0] for row in con.execute(
"SELECT rootpage FROM sqlite_master WHERE type='table'"))
table_cursors = set(
opcode[2] for opcode in opcodes
if opcode[1] == "OpenRead" and opcode[3] in table_rootpages)
assert sum(
1 for cursor in accessed_cursors if cursor not in table_cursors) > 1
assert sum(1 for cursor in accessed_cursors
if cursor in table_cursors) == 1
results = list(con.execute(query, rs671))
results_rowids = set(vt[0] for vt in results)
assert next(vt for vt in results if vt[1] and "rs671" in vt[1])
control = "SELECT rowid FROM gnomad_vcf_lines WHERE NOT ((POS+rlen) < ? OR POS > ?)"
control_rowids = set(vt[0]
for vt in con.execute(control, (rs671[1], rs671[2])))
assert len(control_rowids) == 22
assert results_rowids == control_rowids
def main():
header = f"""\
-- GenomicSQLite smoke test --
timestamp: {time.asctime(time.gmtime())}
platform: {platform.platform()}
uname: {os.uname()}
python: {platform.python_implementation()} {platform.python_version()}"""
print(textwrap.dedent(header))
print("cpu", end="")
try:
if platform.system() == "Linux":
with open("/proc/cpuinfo") as cpuinfo:
modelname = (line.strip().replace("\t", "") for line in cpuinfo
if line.lower().strip().startswith("model name"))
print(" " + next(modelname, ": ???"))
elif platform.system() == "Darwin":
sysctl = subprocess.run(
["sysctl", "-n", "machdep.cpu.brand_string"],
check=True,
universal_newlines=True,
stdout=subprocess.PIPE,
stderr=subprocess.PIPE,
)
if sysctl.returncode == 0 and sysctl.stdout.strip():
print(": " + sysctl.stdout.strip())
else:
print(": ???")
else:
print(": ???")
except Exception:
print(": ???")
env_keys = [
k for k in os.environ
if ("genomicsqlite" in k.lower()) or k in ("PATH", "PYTHONPATH",
"LD_LIBRARY_PATH",
"DYLD_LIBRARY_PATH")
]
if env_keys:
print("environment:")
for k in env_keys:
print(f" {k}={os.environ[k]}")
print("sqlite3: ", end="")
import sqlite3
conn = sqlite3.connect(":memory:")
print(next(conn.execute("SELECT sqlite_version()"))[0])
print("genomicsqlite: ", end="")
try:
import genomicsqlite
except ImportError as exn:
print(f"\n\nUnable to import genomicsqlite: {exn}")
print(
"The Python genomicsqlite package may need to be installed via pip3 or conda,"
)
print("or its location may need to be added to PYTHONPATH.")
sys.exit(1)
print(genomicsqlite.__version__)
print(f"genomicsqlite library: {genomicsqlite._DLL}")
dbfile = os.path.join(os.environ.get("TMPDIR", "/tmp"),
f"genomicsqlite_smoke_test.{time.monotonic()}")
print(f"\ntest database: {dbfile}")
try:
dbconn = genomicsqlite.connect(dbfile)
for (table, bed) in (("exons1", _EXONS1), ("exons2", _EXONS2)):
dbconn.execute(
f"CREATE TABLE {table}(chrom TEXT, pos INTEGER, end INTEGER, id TEXT PRIMARY KEY)"
)
for line in bed.strip().split("\n"):
line = line.split("\t")
dbconn.execute(
f"INSERT INTO {table}(chrom,pos,end,id) VALUES(?,?,?,?)",
(line[0], int(line[1]) - 1, int(line[2]), line[3]),
)
dbconn.executescript(
genomicsqlite.create_genomic_range_index_sql(
dbconn, table, "chrom", "pos", "end"))
dbconn.commit()
query = (
"SELECT exons1.id, exons2.id FROM exons1 LEFT JOIN exons2 ON exons2._rowid_ IN\n"
+ textwrap.indent(
genomicsqlite.genomic_range_rowids_sql(
dbconn, "exons2", "exons1.chrom", "exons1.pos",
"exons1.end"),
" ",
) + "\nORDER BY exons1.id, exons2.id")
print("GRI query:\n" + textwrap.indent(query, " "))
print("GRI query plan:")
gri_constraints = 0
for expl in dbconn.execute("EXPLAIN QUERY PLAN " + query):
print(" " + expl[3])
if ("((_gri_rid,_gri_lvl,_gri_beg)>(?,?,?) AND (_gri_rid,_gri_lvl,_gri_beg)<(?,?,?))"
in expl[3]):
gri_constraints += 1
results = list(dbconn.execute(query))
control = "SELECT exons1.id, exons2.id FROM exons1 LEFT JOIN exons2 NOT INDEXED ON NOT (exons2.end < exons1.pos OR exons2.pos > exons1.end) ORDER BY exons1.id, exons2.id"
control_results = list(dbconn.execute(control))
assert len(control_results) == 1139
assert results == control_results
if gri_constraints != 3:
print("GRI query opcodes:")
for expl in dbconn.execute("EXPLAIN " + query):
print(" " + str(expl))
print(
"\n** WARNING: GRI yielded correct results, but with a possibly suboptimal query plan."
)
print(
"** Please redirect this log to a file and send to maintainers @ https://github.com/mlin/GenomicSQLite\n"
)
sys.exit(2)
dbconn.close()
print("\nGenomicSQLite smoke test OK =)\n")
finally:
os.remove(dbfile)
def test_query_in_sql(tmp_path):
dbfile = str(tmp_path / "test.gsql")
con = genomicsqlite.connect(dbfile, unsafe_load=True)
_fill_exons(con)
con.commit()
query = "SELECT id FROM exons WHERE exons._rowid_ IN genomic_range_rowids('exons',?,?,?)"
results = list(con.execute(query, ("chr17", 43044294, 43048294)))
control_query = genomicsqlite.genomic_range_rowids_sql(con, "exons")
control_query = "SELECT id FROM exons WHERE exons._rowid_ IN\n" + control_query
control_results = list(
con.execute(control_query, ("chr17", 43044294, 43048294)))
assert results == control_results
for expl in con.execute(
"EXPLAIN QUERY PLAN SELECT _rowid_ FROM genomic_range_rowids('exons',?,?,?) ORDER BY _rowid_",
("chr17", 43044294, 43048294),
):
assert "USE TEMP B-TREE FOR ORDER BY" not in expl[3]
assert next((expl[3] for expl in con.execute(
"EXPLAIN QUERY PLAN SELECT _rowid_ FROM genomic_range_rowids('exons',?,?,?) ORDER BY _rowid_ DESC",
("chr17", 43044294, 43048294),
) if "USE TEMP B-TREE FOR ORDER BY" in expl[3]))
with pytest.raises(sqlite3.OperationalError, match="domain error"):
con.execute(
"SELECT * FROM genomic_range_rowids('exons', 'chr17', 43044294, 43048294, 16, 0)"
)
dbfile2 = str(tmp_path / "test2.gsql")
con2 = genomicsqlite.connect(dbfile2, unsafe_load=True)
_fill_exons(con2)
con2.commit()
con2.close()
con.executescript(
genomicsqlite.attach_sql(con, dbfile2, "db2", immutable=True))
query = """
SELECT main.exons.id, db2.exons.id
FROM main.exons LEFT JOIN db2.exons ON
db2.exons._rowid_ IN genomic_range_rowids('db2.exons', main.exons.rid, main.exons.beg, main.exons.end)
AND main.exons.id != db2.exons.id ORDER BY main.exons.id, db2.exons.id
"""
results = list(con.execute(query))
assert len(results) == 5191
with pytest.raises(sqlite3.OperationalError, match="no such table"):
con.execute(
"SELECT * FROM genomic_range_rowids('nonexistent', 'chr17', 43044294, 43048294)"
)
con.executescript("CREATE TABLE empty(rid TEXT, beg INTEGER, end INTEGER)")
with pytest.raises(sqlite3.OperationalError, match="no such index"):
con.execute(
"SELECT * FROM genomic_range_rowids('empty', 'chr17', 43044294, 43048294)"
)
con.executescript(
genomicsqlite.create_genomic_range_index_sql(con, "empty", "rid",
"beg", "end"))
results = list(
con.execute(
"SELECT * FROM genomic_range_rowids('empty', 'chr17', 43044294, 43048294)"
))
assert results == []
def test_level_detection():
# test corner cases for the bit of genomic_range_rowids() which detects the level range of
# extant features
con = sqlite3.connect(":memory:")
con.executescript(
"CREATE TABLE features(rid INTEGER, beg INTEGER, end INTEGER)")
con.executescript(
genomicsqlite.create_genomic_range_index_sql(con, "features", "rid",
"beg", "end"))
def fanout(query):
return sum(
1 for expl in con.execute("EXPLAIN QUERY PLAN " +
query, (None, None, None)) if
"((_gri_rid,_gri_lvl,_gri_beg)>(?,?,?) AND (_gri_rid,_gri_lvl,_gri_beg)<(?,?,?))"
in expl[3])
assert fanout(
genomicsqlite.genomic_range_rowids_sql(con, "features")[1:-1]) == 16
con.executescript(
"INSERT INTO features VALUES(NULL, NULL, NULL); INSERT INTO features VALUES(NULL, 0, 10000000000)"
)
assert fanout(
genomicsqlite.genomic_range_rowids_sql(con, "features")[1:-1]) == 16
assert not list(
con.execute(
genomicsqlite.genomic_range_rowids_sql(con, "features")[1:-1],
(None, 123, 456)))
con.executescript("INSERT INTO features VALUES(42, 1048568, 1048584)")
query = genomicsqlite.genomic_range_rowids_sql(con, "features")[1:-1]
print("\n" + query)
assert "-1" in query and "-0x10)" in query
assert not list(
con.execute(
genomicsqlite.genomic_range_rowids_sql(con, "features")[1:-1],
(42, None, 1048584)))
assert fanout(query) == 1
con.executescript("""
INSERT INTO features VALUES(44, 1048568, 1048584);
INSERT INTO features VALUES(44, 0, 64000)
""")
query = genomicsqlite.genomic_range_rowids_sql(con, "features")[1:-1]
print("\n" + query)
assert "-4," in query and "-0x10000)" in query
assert "-3," in query and "-0x1000)" in query
assert "-2," in query and "-0x100)" in query
assert "-1," in query and "-0x10)" in query
assert fanout(query) == 4
assert fanout(
genomicsqlite.genomic_range_rowids_sql(con, "features",
ceiling=6)[1:-1]) == 7
assert (fanout(
genomicsqlite.genomic_range_rowids_sql(con,
"features",
ceiling=6,
floor=3)[1:-1]) == 4)
con.executescript("""
INSERT INTO features VALUES(43, NULL, 10000000000);
INSERT INTO features VALUES(44, 0, NULL)
""")
assert fanout(
genomicsqlite.genomic_range_rowids_sql(con, "features")[1:-1]) == 4
con.executescript("""
INSERT INTO features VALUES(43, 0, 10000000000);
INSERT INTO features VALUES(43, 32, 33)
""")
query = genomicsqlite.genomic_range_rowids_sql(con, "features")[1:-1]
print("\n" + query)
assert fanout(query) == 10
con.executescript("""
INSERT INTO features VALUES(43, 0, 10000000000);
INSERT INTO features VALUES(43, 32, 32)
""")
query = genomicsqlite.genomic_range_rowids_sql(con, "features")[1:-1]
assert fanout(query) == 10
assert len(list(con.execute(query, (43, 32, 33)))) == 4
assert len(list(con.execute(query, (43, 33, 33)))) == 3
def test_txdbquery(genomicsqlite_txdb):
conn = genomicsqlite.connect(genomicsqlite_txdb, read_only=True)
assert next(conn.execute("PRAGMA page_size"))[0] == 16384
# one query
results = list(t[0] for t in conn.execute(
"SELECT tx_name FROM transcript WHERE _rowid_ IN " +
genomicsqlite.genomic_range_rowids_sql(conn, "transcript") +
" ORDER BY tx_name",
("chr12", 111803912, 111804012),
))
print(results)
assert results == sorted([
"ENST00000416293.7", "ENST00000261733.7", "ENST00000548536.1",
"ENST00000549106.1"
])
# random queries
chroms = list(
conn.execute(
"SELECT tx_chrom, length FROM (SELECT tx_chrom, MAX(tx_end) AS length FROM transcript GROUP BY tx_chrom) WHERE length > 1000000"
))
random.seed(0xBADF00D)
for tbl in ("transcript", "cds"):
query = genomicsqlite.genomic_range_rowids_sql(conn, tbl)[1:-1]
fanout = 0
for expl in conn.execute("EXPLAIN QUERY PLAN " + query,
(None, None, None)):
print(expl[3])
if ("((_gri_rid,_gri_lvl,_gri_beg)>(?,?,?) AND (_gri_rid,_gri_lvl,_gri_beg)<(?,?,?))"
in expl[3]):
fanout += 1
assert (tbl, fanout) in (("transcript", 5), ("cds", 3))
pfx = "tx" if tbl == "transcript" else tbl
control_query = f"SELECT _rowid_ FROM {tbl} NOT INDEXED WHERE {pfx}_chrom = ? AND NOT ({pfx}_end < ? OR {pfx}_start > ?) ORDER BY _rowid_"
total_results = 0
for _ in range(2000):
chrom = random.choice(chroms)
beg = random.randint(0, chrom[1] - 65536)
end = beg + random.randint(1, random.choice([16, 256, 4096, 65536
]))
ids = list(row[0]
for row in conn.execute(query, (chrom[0], beg, end)))
control_ids = list(
row[0]
for row in conn.execute(control_query, (chrom[0], beg, end)))
assert ids == control_ids
total_results += len(control_ids)
assert total_results in (7341, 2660)
# join exon to cds ("which exons are coding?")
exon_cds_counts = (
"SELECT exon._rowid_ AS exon_id, COUNT(cds._rowid_) AS contained_cds FROM exon LEFT JOIN cds ON cds._rowid_ IN "
+ genomicsqlite.genomic_range_rowids_sql(conn, "cds", "exon_chrom",
"exon_start", "exon_end") +
" AND (exon_start = cds_start AND exon_end >= cds_start OR exon_start <= cds_start AND exon_end = cds_end) GROUP BY exon._rowid_"
)
print(exon_cds_counts)
fanout = 0
for expl in conn.execute("EXPLAIN QUERY PLAN " + exon_cds_counts):
print(expl[3])
if ("((_gri_rid,_gri_lvl,_gri_beg)>(?,?,?) AND (_gri_rid,_gri_lvl,_gri_beg)<(?,?,?))"
in expl[3]):
fanout += 1
assert fanout == 3
cds_exon_count_hist = list(
conn.execute(
f"SELECT contained_cds, count(exon_id) AS exon_count FROM ({exon_cds_counts}) GROUP BY contained_cds ORDER BY contained_cds"
))
for elt in cds_exon_count_hist:
print(elt)
assert cds_exon_count_hist[:2] == [(0, 270532), (1, 310059)]
