def _old_savepm():
def _saveattr(attr):
with unmask_all_peakmatrix(pm) as m:
dm = m.attr_matrix(attr, flagged_only = False)
dt = np.float64 if dm.dtype.kind == 'f' else \
np.int64 if dm.dtype.kind in ('i', 'u') else \
('S%d' % np.max([list(map(len,l)) for l in dm]))
ds = f.create_dataset(attr, dm.shape, dtype=dt)
ds[...] = dm.astype(dt)
ds.attrs['dtype'] = dm.dtype.str
for a in pm.attributes: _saveattr(a)
dset = f['mz'] # must exists in pm
dset.attrs['class'] = 'PeakMatrix'
dset.attrs['attributes'] = np.string_(pm.attributes)
dset.attrs['mask'] = pm.mask
with unmask_all_peakmatrix(pm):
dset.attrs['peaklist_ids'] = np.string_(pm.peaklist_ids)
for i, tags in enumerate(pm.peaklist_tags):
dset.attrs['peaklist_tags_%d' % i] = np.string_([(t or 'None', v) for v,t in tags.to_list()])
dset.attrs['flag_names'] = np.string_(pm.flag_names)
for fn in pm.flag_names:
dset.attrs[fn] = pm.flag_values(fn)
def _saveattr(attr):
with unmask_all_peakmatrix(pm) as m:
dm = m.attr_matrix(attr, flagged_only = False)
dt = np.float64 if dm.dtype.kind == 'f' else \
np.int64 if dm.dtype.kind in ('i', 'u') else \
('S%d' % np.max([list(map(len,l)) for l in dm]))
ds = f.create_dataset(attr, dm.shape, dtype=dt)
ds[...] = dm.astype(dt)
ds.attrs['dtype'] = dm.dtype.str
def _saveattr(attr):
if attr in f.keys():
raise IOError('attribute [%s] already exists' % attr)
with unmask_all_peakmatrix(pm) as m:
dm = m.attr_matrix(attr, flagged_only=False)
dt = np.float64 if dm.dtype.kind == 'f' else \
np.int64 if dm.dtype.kind in ('i', 'u') else \
('S%d' % np.max(map(lambda l: map(len,l), dm)))
ds = f.create_dataset(attr, dm.shape, dtype=dt)
ds[...] = dm.astype(dt)
ds.attrs['dtype'] = dm.dtype.str
def save_peak_matrix_as_txt(pm: PeakMatrix, filename: str, *args, **kwargs):
"""
Saves a peak matrix in plain text file.
:param pm: the target peak matrix object
:param filename: path to a new text file
:param args: arguments to be passed to PeakMatrix.to_str
:param kwargs: keyword arguments to be passed to PeakMatrix.to_str
"""
if os.path.isfile(filename):
logging.warning('plain text file [%s] already exists, override' % filename)
with open(filename, 'w') as f:
with unmask_all_peakmatrix(pm) as m: f.write(m.to_str(*args, **kwargs))
def save_peak_matrix_as_hdf5(pm, filename):
"""
Saves a peak matrix object to a HDF5 file.
:param pm: the target peak matrix object
:param filename: path to a new HDF5 file
The order of the attributes and flags will be retained.
"""
if os.path.isfile(filename):
logging.warning('HDF5 database [%s] already exists and override' %
filename)
f = h5py.File(filename, 'w')
def _saveattr(attr):
if attr in f.keys():
raise IOError('attribute [%s] already exists' % attr)
with unmask_all_peakmatrix(pm) as m:
dm = m.attr_matrix(attr, flagged_only=False)
dt = np.float64 if dm.dtype.kind == 'f' else \
np.int64 if dm.dtype.kind in ('i', 'u') else \
('S%d' % np.max(map(lambda l: map(len,l), dm)))
ds = f.create_dataset(attr, dm.shape, dtype=dt)
ds[...] = dm.astype(dt)
ds.attrs['dtype'] = dm.dtype.str
map(_saveattr, pm.attributes)
dset = f['mz'] # must exists in pm
dset.attrs['class'] = 'PeakMatrix'
dset.attrs['attributes'] = pm.attributes
dset.attrs['mask'] = pm.mask
with unmask_all_peakmatrix(pm):
dset.attrs['peaklist_ids'] = pm.peaklist_ids
for i, tags in enumerate(pm.peaklist_tags):
dset.attrs['peaklist_tags_%d' % i] = [(t or 'None', v)
for v, t in tags.to_list()]
dset.attrs['flag_names'] = pm.flag_names
for fn in pm.flag_names:
fvals = pm.flag_values(fn)
dset.attrs[fn] = fvals if fvals.nbytes < 64000 else _packBool(
fvals) if fvals.dtype.kind == 'b' else _packMeta(fvals)
def _savepm():
def _saveattr(attr):
with unmask_all_peakmatrix(pm) as m:
dm = m.attr_matrix(a, flagged_only = False)
ds = f.create_array(f.root, a, dm)
ds.attrs.dtype = dm.dtype.str
for a in pm.attributes: _saveattr(a)
dset = f.root.mz # must exists in pm
dset.attrs.data_class = 'PeakMatrix'
dset.attrs.attributes = pm.attributes
dset.attrs.mask = pm.mask
with unmask_all_peakmatrix(pm):
dset.attrs.peaklist_ids = pm.peaklist_ids
for i, tags in enumerate(pm.peaklist_tags):
dset.attrs['peaklist_tags_%d' % i] = np.array([(t or 'None', v) for v,t in tags.to_list()])
dset.attrs.flag_names = pm.flag_names
for fn in pm.flag_names:
dset.attrs[fn] = pm.flag_values(fn)
def test_pm_mask(self):
pm = self._createPeakMatrix()
self.assertEqual(set(map(lambda x: x.value, pm.tags_of('plate'))),
{1, 2})
self.assertEqual(set(map(lambda x: x.value, pm.tags_of())),
{'sample', 'qc'})
self.assertRaises(KeyError, lambda: pm.tags_of('treatment'))
self.assertRaises(KeyError, lambda: pm.tags_of('not_exist'))
pm.mask_tags(1)
self.assertTupleEqual(pm.peaklist_ids,
('sample_1_1', 'sample_1_2', 'QC_1',
'sample_2_1', 'sample_2_2', 'QC_2'))
pm.mask_tags('qc', plate=1) # mask samples with both of the two
self.assertTupleEqual(
pm.peaklist_ids,
('sample_1_1', 'sample_1_2', 'sample_2_1', 'sample_2_2', 'QC_2'))
pm.mask = None
pm.mask_tags('qc')
self.assertTupleEqual(
pm.peaklist_ids,
('sample_1_1', 'sample_1_2', 'sample_2_1', 'sample_2_2'))
pm.mask = None
pm.mask_tags('qc').mask_tags(plate=1)
self.assertTupleEqual(pm.peaklist_ids, ('sample_2_1', 'sample_2_2'))
pm.mask = None
pm.mask_tags('not_exist')
self.assertTupleEqual(pm.peaklist_ids,
('sample_1_1', 'sample_1_2', 'QC_1',
'sample_2_1', 'sample_2_2', 'QC_2'))
pm.mask = [True] * 6
pm.unmask_tags('qc', plate=1) # unmask samples with both of the two
self.assertTupleEqual(pm.peaklist_ids, ('QC_1', ))
pm.mask = [True] * 6
pm.unmask_tags('qc')
self.assertTupleEqual(pm.peaklist_ids, ('QC_1', 'QC_2'))
pm.mask = [True] * 6
pm.unmask_tags('qc').unmask_tags(plate=1)
self.assertTupleEqual(pm.peaklist_ids,
('sample_1_1', 'sample_1_2', 'QC_1', 'QC_2'))
pm.mask = [True] * 6
pm.unmask_tags('not_exist')
self.assertTupleEqual(pm.peaklist_ids, ())
pm.unmask_tags('qc', override=True)
self.assertTupleEqual(pm.peaklist_ids, ('QC_1', 'QC_2'))
with mask_all_peakmatrix(pm) as m:
m.unmask_tags('sample')
self.assertTupleEqual(
pm.peaklist_ids,
('sample_1_1', 'sample_1_2', 'sample_2_1', 'sample_2_2'))
pm.mask_tags('qc', override=True)
self.assertTupleEqual(
pm.peaklist_ids,
('sample_1_1', 'sample_1_2', 'sample_2_1', 'sample_2_2'))
with unmask_all_peakmatrix(pm) as m:
m.mask_tags('sample')
self.assertTupleEqual(m.peaklist_ids, ('QC_1', 'QC_2'))
self.assertTupleEqual(
pm.peaklist_ids,
('sample_1_1', 'sample_1_2', 'sample_2_1', 'sample_2_2'))
pm.mask = None
with unmask_peakmatrix(pm, plate=1) as m:
self.assertTupleEqual(m.peaklist_ids,
('sample_1_1', 'sample_1_2', 'QC_1'))
self.assertTupleEqual(m.full_shape, (6, 10))
self.assertEqual(len(pm.peaklist_ids), 6)
with mask_peakmatrix(pm, plate=2) as m:
self.assertTupleEqual(m.peaklist_ids,
('sample_1_1', 'sample_1_2', 'QC_1'))
with unmask_all_peakmatrix(pm) as mm:
self.assertTupleEqual(mm.peaklist_ids,
('sample_1_1', 'sample_1_2', 'QC_1',
'sample_2_1', 'sample_2_2', 'QC_2'))
with mask_peakmatrix(pm, 'qc') as m:
m.remove_samples((1, 2))
self.assertTupleEqual(pm.peaklist_ids,
('sample_1_1', 'QC_1', 'sample_2_2', 'QC_2'))
def _saveattr(attr):
with unmask_all_peakmatrix(pm) as m:
dm = m.attr_matrix(a, flagged_only = False)
ds = f.create_array(f.root, a, dm)
ds.attrs.dtype = dm.dtype.str