def test_is_run_id_in_db(empty_temp_db):
conn = mut_db.connect(get_DB_location())
mut_queries.new_experiment(conn, 'test_exp', 'no_sample')
for _ in range(5):
ds = DataSet(conn=conn, run_id=None)
# there should now be run_ids 1, 2, 3, 4, 5 in the database
good_ids = [1, 2, 3, 4, 5]
try_ids = [1, 3, 9999, 23, 0, 1, 1, 3, 34]
sorted_try_ids = np.unique(try_ids)
expected_dict = {tid: (tid in good_ids) for tid in sorted_try_ids}
acquired_dict = mut_queries.is_run_id_in_database(conn, *try_ids)
assert expected_dict == acquired_dict
def extract_runs_into_db(source_db_path: str,
target_db_path: str, *run_ids: int,
upgrade_source_db: bool = False,
upgrade_target_db: bool = False) -> None:
"""
Extract a selection of runs into another DB file. All runs must come from
the same experiment. They will be added to an experiment with the same name
and ``sample_name`` in the target db. If such an experiment does not exist, it
will be created.
Args:
source_db_path: Path to the source DB file
target_db_path: Path to the target DB file. The target DB file will be
created if it does not exist.
run_ids: The ``run_id``'s of the runs to copy into the target DB file
upgrade_source_db: If the source DB is found to be in a version that is
not the newest, should it be upgraded?
upgrade_target_db: If the target DB is found to be in a version that is
not the newest, should it be upgraded?
"""
# Check for versions
(s_v, new_v) = get_db_version_and_newest_available_version(source_db_path)
if s_v < new_v and not upgrade_source_db:
warn(f'Source DB version is {s_v}, but this function needs it to be'
f' in version {new_v}. Run this function again with '
'upgrade_source_db=True to auto-upgrade the source DB file.')
return
if os.path.exists(target_db_path):
(t_v, new_v) = get_db_version_and_newest_available_version(target_db_path)
if t_v < new_v and not upgrade_target_db:
warn(f'Target DB version is {t_v}, but this function needs it to '
f'be in version {new_v}. Run this function again with '
'upgrade_target_db=True to auto-upgrade the target DB file.')
return
source_conn = connect(source_db_path)
# Validate that all runs are in the source database
do_runs_exist = is_run_id_in_database(source_conn, *run_ids)
if False in do_runs_exist.values():
source_conn.close()
non_existing_ids = [rid for rid in run_ids if not do_runs_exist[rid]]
err_mssg = ("Error: not all run_ids exist in the source database. "
"The following run(s) is/are not present: "
f"{non_existing_ids}")
raise ValueError(err_mssg)
# Validate that all runs are from the same experiment
source_exp_ids = np.unique(get_exp_ids_from_run_ids(source_conn, run_ids))
if len(source_exp_ids) != 1:
source_conn.close()
raise ValueError('Did not receive runs from a single experiment. '
f'Got runs from experiments {source_exp_ids}')
# Fetch the attributes of the runs' experiment
# hopefully, this is enough to uniquely identify the experiment
exp_attrs = get_experiment_attributes_by_exp_id(source_conn, source_exp_ids[0])
# Massage the target DB file to accommodate the runs
# (create new experiment if needed)
target_conn = connect(target_db_path)
# this function raises if the target DB file has several experiments
# matching both the name and sample_name
try:
with atomic(target_conn) as target_conn:
target_exp_id = _create_exp_if_needed(target_conn,
exp_attrs['name'],
exp_attrs['sample_name'],
exp_attrs['format_string'],
exp_attrs['start_time'],
exp_attrs['end_time'])
# Finally insert the runs
for run_id in run_ids:
_extract_single_dataset_into_db(DataSet(run_id=run_id,
conn=source_conn),
target_conn,
target_exp_id)
finally:
source_conn.close()
target_conn.close()
