def test_plot_gates(self):
fks = fk.Session(fcs_path)
fks.add_sample_group('my_group', gml_path)
group_name = 'my_group'
sample_name = 'B07'
fks.assign_sample(sample_name, group_name)
gate_ids = fks.get_gate_ids(group_name)
for gate_id in gate_ids:
gate = fks.get_gate(group_name, sample_name, gate_id)
if isinstance(gate, fk.gates.Quadrant):
# cannot plot single quadrants of a quadrant gate
continue
try:
p = fks.plot_gate('my_group', sample_name, gate_id)
except NotImplementedError:
continue
self.assertIsInstance(p, Figure)
def test_plot_gates(self):
fks = fk.Session(copy.deepcopy(test_sample))
fks.add_sample_group('my_group', copy.deepcopy(test_gating_strategy))
group_name = 'my_group'
sample_name = 'B07'
fks.assign_samples(sample_name, group_name)
gate_tuples = fks.get_gate_ids(group_name)
fks.analyze_samples(group_name, sample_id=sample_name)
for gate_name, ancestors in gate_tuples:
gate = fks.get_gate(group_name, gate_name, sample_id=sample_name)
if isinstance(gate, fk.gates.Quadrant):
# cannot plot single quadrants of a quadrant gate
continue
try:
p = fks.plot_gate('my_group', sample_name, gate_name)
except NotImplementedError:
continue
self.assertIsInstance(p, bk_Figure)
def test_plot_gated_scatter(self):
fks = fk.Session(fcs_path)
fks.add_sample_group('my_group', gml_path)
group_name = 'my_group'
sample_name = 'B07'
fks.assign_sample(sample_name, group_name)
x_dim = fk.Dimension('FL2-H',
compensation_ref='MySpill',
transformation_ref='Logicle_10000_0.5_4.5_0')
y_dim = fk.Dimension('FL3-H',
compensation_ref='MySpill',
transformation_ref='Logicle_10000_0.5_4.5_0')
p = fks.plot_scatter(sample_name,
x_dim,
y_dim,
sample_group=group_name,
gate_id='ScaleRect1')
self.assertIsInstance(p, Figure)
def test_plot_gated_scatter(self):
fks = fk.Session(copy.deepcopy(test_sample))
fks.add_sample_group('my_group', copy.deepcopy(test_gating_strategy))
group_name = 'my_group'
sample_name = 'B07'
fks.assign_samples(sample_name, group_name)
fks.analyze_samples(group_name, sample_id=sample_name)
x_dim = fk.Dimension('FL2-H', compensation_ref='MySpill', transformation_ref='Logicle_10000_0.5_4.5_0')
y_dim = fk.Dimension('FL3-H', compensation_ref='MySpill', transformation_ref='Logicle_10000_0.5_4.5_0')
p = fks.plot_scatter(
sample_name,
x_dim,
y_dim,
group_name=group_name,
gate_name='ScaleRect1',
subsample=True
)
self.assertIsInstance(p, bk_Figure)
def extract_gated_data(output_dir,
wsp_file,
fcs_dir,
sample_group,
event_columns=None,
exclude_samples=None,
verbose=False):
"""
Extracts pre-processed event data and gate membership for each of the FCS
files referenced in given FlowJo10 workspace file (wsp file). Both event
and gate membership data are saved as Feather files to the provided output
directory. Two subdirectories, 'events' and 'gates', will be created in
the output directory where 2 Feather files will be created for each FCS
file, one in the 'events' directory (prepended with 'events_') and one in
the 'gates' directory (prepended with 'gates_').
:param output_dir: File path location for output Feather files
:param wsp_file: File path of FlowJo 10 workspace (wsp file)
:param fcs_dir: Directory path containing FCS files
:param sample_group: Sample group label within the WSP file from which
gated events and gate membership data will be extracted
:param event_columns: List of channels/markers to include in the
preprocessed event data. If None, all channels are exported.
:param exclude_samples: List of FCS sample IDs to exclude from export
:param verbose: If True, prints various issues found when parsing
the workspace and FCS files (missing FCS files, missing channel
columns, missing gates, and output array shapes)
:return: None
"""
fs = fk.Session()
fs.import_flowjo_workspace(wsp_file, ignore_missing_files=True)
# find FCS files in given fcs_dir
fcs_files = [f for f in os.listdir(fcs_dir) if re.search(r'.*\.fcs$', f)]
sample_ids = fs.get_group_sample_ids(sample_group, loaded_only=False)
events_dir = os.path.join(output_dir, 'events')
gates_dir = os.path.join(output_dir, 'gates')
os.makedirs(output_dir, exist_ok=True)
os.makedirs(events_dir, exist_ok=True)
os.makedirs(gates_dir, exist_ok=True)
sample_paths_to_load = []
samples_to_process = []
for f in sample_ids:
if f in exclude_samples:
continue
if f not in fcs_files and verbose:
print('%s not found in our list' % f)
sample_paths_to_load.append(os.path.join(fcs_dir, f))
samples_to_process.append(f)
fs.add_samples(sample_paths_to_load)
fs.analyze_samples(sample_group)
for sample_id in samples_to_process:
# extract all processed events using gate ID as None
events_df = fs.get_wsp_gated_events(sample_group, [sample_id],
gate_name=None,
gate_path=None)[0]
if event_columns:
try:
events_df = events_df[event_columns]
except KeyError:
print(f'Sample to exclude: {sample_id}')
continue
# now get gate membership for all gates & all events
gates_df = pd.DataFrame()
gate_ids = fs.get_gate_ids(sample_group)
for gate_name, gate_path in gate_ids:
try:
s_gate_membership = fs.get_gate_membership(sample_group,
sample_id,
gate_name=gate_name,
gate_path=gate_path)
gates_df['/'.join([*gate_path, gate_name])] = s_gate_membership
except KeyError:
if verbose:
print("Gate %s (%s) not found in sample group" %
(gate_name, gate_path))
# add sample_id field to gate DataFrame & put it first
gates_df['sample_id'] = sample_id
sample_id_col = gates_df.pop('sample_id')
gates_df.insert(0, sample_id_col.name, sample_id_col)
# save events_df & gates_df to feather here
sample_feather_basename = sample_id.replace('fcs', 'feather')
events_filename = "events_" + sample_feather_basename
gates_filename = "gates_" + sample_feather_basename
events_path = os.path.join(events_dir, events_filename)
gates_path = os.path.join(gates_dir, gates_filename)
if not os.path.exists(events_path):
events_df.reset_index(drop=True, inplace=True)
events_df.to_feather(events_path)
print(events_df.shape)
if not os.path.exists(gates_path):
gates_df.reset_index(drop=True, inplace=True)
gates_df.to_feather(gates_path)
print(gates_df.shape)