def get_structures(self, file_name=None):
"""
Read the list of adult mouse structures and return a Pandas DataFrame.
Parameters
----------
file_name: string
File name to save/read the structures table. If file_name is None,
the file_name will be pulled out of the manifest. If caching
is disabled, no file will be saved. Default is None.
"""
file_name = self.get_cache_path(file_name, self.STRUCTURES_KEY)
if os.path.exists(file_name):
structures = pd.DataFrame.from_csv(file_name)
else:
structures = OntologiesApi(base_uri=self.api.api_url).get_structures(1)
structures = pd.DataFrame(structures)
if self.cache:
Manifest.safe_make_parent_dirs(file_name)
structures.to_csv(file_name)
structures.set_index(["id"], inplace=True, drop=False)
return structures
def get_template_volume(self, file_name=None):
"""
Read the template volume. Download it first if it doesn't exist.
Parameters
----------
file_name: string
File name to store the template volume. If it already exists,
it will be read from this file. If file_name is None, the
file_name will be pulled out of the manifest. Default is None.
"""
file_name = self.get_cache_path(file_name, self.TEMPLATE_KEY,
self.resolution)
if file_name is None:
raise Exception("No save file provided for annotation volume.")
if os.path.exists(file_name):
annotation, info = nrrd.read(file_name)
else:
Manifest.safe_make_parent_dirs(file_name)
annotation, info = self.api.download_template_volume(
self.resolution, file_name)
return annotation, info
def write_volume(volume,
name,
prefix=None,
specify_resolution=None,
extension='.nrrd',
paths=None):
if prefix is None:
path = name
else:
path = os.path.join(prefix, name)
if specify_resolution is not None:
if isinstance(specify_resolution, (float, np.floating)) and \
specify_resolution % 1.0 == 0:
specify_resolution = int(specify_resolution)
path = path + '_{0}'.format(specify_resolution)
path = path + extension
logging.info('writing {0} volume to {1}'.format(name, path))
Manifest.safe_make_parent_dirs(path)
volume.SetOrigin([0, 0, 0])
sitk.WriteImage(volume, str(path), True)
if paths is not None:
paths.append(path)
def debug(container_id, local=False, plots=None):
SCRIPT = "/data/informatics/CAM/analysis/allensdk/allensdk/internal/pipeline_modules/run_observatory_container_thumbnails.py"
SDK_PATH = "/data/informatics/CAM/analysis/allensdk/"
OUTPUT_DIR = "/data/informatics/CAM/analysis/containers"
container_dir = os.path.join(OUTPUT_DIR, str(container_id))
input_data = []
for exp in get_container_info(container_id):
exp_data = robsth.get_input_data(exp['id'])
exp_input_json = os.path.join(exp_data["output_directory"],
"input.json")
input_data.append(
dict(input_json=exp_input_json,
output_json=os.path.join(exp_data["output_directory"],
"output.json")))
Manifest.safe_make_parent_dirs(exp_input_json)
ju.write(exp_input_json, exp_data)
run_module(SCRIPT,
input_data,
container_dir,
sdk_path=SDK_PATH,
pbs=dict(vmem=32,
job_name="cthumbs_%d" % container_id,
walltime="10:00:00"),
local=local,
optional_args=['--types=' + ','.join(plots)] if plots else None)
def save_reconstruction_markers(self, specimen_id, file_name):
"""
Save the marker file for the morphological reconstruction of a cell. These are
comma-delimited files indicating points of interest in a reconstruction (truncation
points, early tracing termination, etc).
Parameters
----------
specimen_id: int
ID of the specimen, from the Specimens database model in the Allen Institute API.
file_name: str
Path to save the marker file.
"""
Manifest.safe_make_parent_dirs(file_name)
criteria = '[id$eq%d],neuron_reconstructions(well_known_files)' % specimen_id
includes = 'neuron_reconstructions(well_known_files(well_known_file_type[name$eq\'%s\']))' % self.MARKER_FILE_TYPE
results = self.model_query('Specimen',
criteria=criteria,
include=includes,
num_rows='all')
try:
file_url = results[0]['neuron_reconstructions'][
0]['well_known_files'][0]['download_link']
except:
raise LookupError("Specimen %d has no marker file" % specimen_id)
self.retrieve_file_over_http(self.api_url + file_url, file_name)
def save_reconstruction(self, specimen_id, file_name):
"""
Save the morphological reconstruction of a cell as an SWC file.
Parameters
----------
specimen_id: int
ID of the specimen, from the Specimens database model in the Allen Institute API.
file_name: str
Path to save the SWC file.
"""
Manifest.safe_make_parent_dirs(file_name)
criteria = '[id$eq%d],neuron_reconstructions(well_known_files)' % specimen_id
includes = 'neuron_reconstructions(well_known_files(well_known_file_type[name$eq\'%s\']))' % self.SWC_FILE_TYPE
results = self.model_query('Specimen',
criteria=criteria,
include=includes,
num_rows='all')
try:
file_url = results[0]['neuron_reconstructions'][
0]['well_known_files'][0]['download_link']
except:
raise Exception("Specimen %d has no reconstruction" % specimen_id)
self.retrieve_file_over_http(self.api_url + file_url, file_name)
def save_ephys_data(self, specimen_id, file_name):
"""
Save the electrophysology recordings for a cell as an NWB file.
Parameters
----------
specimen_id: int
ID of the specimen, from the Specimens database model in the Allen Institute API.
file_name: str
Path to save the NWB file.
"""
Manifest.safe_make_parent_dirs(file_name)
criteria = '[id$eq%d],ephys_result(well_known_files(well_known_file_type[name$eq%s]))' % (
specimen_id, self.NWB_FILE_TYPE)
includes = 'ephys_result(well_known_files(well_known_file_type))'
results = self.model_query('Specimen',
criteria=criteria,
include=includes,
num_rows='all')
try:
file_url = results[0]['ephys_result'][
'well_known_files'][0]['download_link']
except Exception as _:
raise Exception("Specimen %d has no ephys data" % specimen_id)
self.retrieve_file_over_http(self.api_url + file_url, file_name)
def get_structures(self, file_name=None):
"""
Read the list of adult mouse structures and return a Pandas DataFrame.
Parameters
----------
file_name: string
File name to save/read the structures table. If file_name is None,
the file_name will be pulled out of the manifest. If caching
is disabled, no file will be saved. Default is None.
"""
file_name = self.get_cache_path(file_name, self.STRUCTURES_KEY)
if os.path.exists(file_name):
structures = pd.DataFrame.from_csv(file_name)
else:
structures = OntologiesApi(
base_uri=self.api.api_url).get_structures(1)
structures = pd.DataFrame(structures)
if self.cache:
Manifest.safe_make_parent_dirs(file_name)
structures.to_csv(file_name)
structures.set_index(['id'], inplace=True, drop=False)
return structures
def get_injection_fraction(self, experiment_id, file_name=None):
"""
Read an injection fraction volume for a single experiment. Download it
first if it doesn't exist. Injection fraction is the proportion of
pixels in the injection site in a grid voxel in [0,1].
Parameters
----------
experiment_id: int
ID of the experiment to download/read. This corresponds to
section_data_set_id in the API.
file_name: string
File name to store the template volume. If it already exists,
it will be read from this file. If file_name is None, the
file_name will be pulled out of the manifest. Default is None.
"""
file_name = self.get_cache_path(file_name, self.INJECTION_FRACTION_KEY, experiment_id, self.resolution)
if file_name is None:
raise Exception("No file name to save volume.")
if not os.path.exists(file_name):
Manifest.safe_make_parent_dirs(file_name)
self.api.download_injection_fraction(file_name, experiment_id, self.resolution)
return nrrd.read(file_name)
def get_data_mask(self, experiment_id, file_name=None):
"""
Read a data mask volume for a single experiment. Download it
first if it doesn't exist. Data mask is a binary mask of
voxels that have valid data. Only use valid data in analysis!
Parameters
----------
experiment_id: int
ID of the experiment to download/read. This corresponds to
section_data_set_id in the API.
file_name: string
File name to store the template volume. If it already exists,
it will be read from this file. If file_name is None, the
file_name will be pulled out of the manifest. Default is None.
"""
file_name = self.get_cache_path(file_name, self.DATA_MASK_KEY, experiment_id, self.resolution)
if file_name is None:
raise Exception("No file name to save volume.")
if not os.path.exists(file_name):
Manifest.safe_make_parent_dirs(file_name)
self.api.download_data_mask(file_name, experiment_id, self.resolution)
return nrrd.read(file_name)
def save_ephys_data(self, specimen_id, file_name):
"""
Save the electrophysology recordings for a cell as an NWB file.
Parameters
----------
specimen_id: int
ID of the specimen, from the Specimens database model in the Allen Institute API.
file_name: str
Path to save the NWB file.
"""
Manifest.safe_make_parent_dirs(file_name)
criteria = '[id$eq%d],ephys_result(well_known_files(well_known_file_type[name$eq%s]))' % (
specimen_id, self.NWB_FILE_TYPE)
includes = 'ephys_result(well_known_files(well_known_file_type))'
results = self.model_query('Specimen',
criteria=criteria,
include=includes,
num_rows='all')
try:
file_url = results[0]['ephys_result']['well_known_files'][0][
'download_link']
except Exception as _:
raise Exception("Specimen %d has no ephys data" % specimen_id)
self.retrieve_file_over_http(self.api_url + file_url, file_name)
def save_reconstruction_markers(self, specimen_id, file_name):
"""
Save the marker file for the morphological reconstruction of a cell. These are
comma-delimited files indicating points of interest in a reconstruction (truncation
points, early tracing termination, etc).
Parameters
----------
specimen_id: int
ID of the specimen, from the Specimens database model in the Allen Institute API.
file_name: str
Path to save the marker file.
"""
Manifest.safe_make_parent_dirs(file_name)
criteria = '[id$eq%d],neuron_reconstructions(well_known_files)' % specimen_id
includes = 'neuron_reconstructions(well_known_files(well_known_file_type[name$eq\'%s\']))' % self.MARKER_FILE_TYPE
results = self.model_query('Specimen',
criteria=criteria,
include=includes,
num_rows='all')
try:
file_url = results[0]['neuron_reconstructions'][
0]['well_known_files'][0]['download_link']
except:
raise LookupError("Specimen %d has no marker file" % specimen_id)
self.retrieve_file_over_http(self.api_url + file_url, file_name)
def get_template_volume(self, file_name=None):
"""
Read the template volume. Download it first if it doesn't exist.
Parameters
----------
file_name: string
File name to store the template volume. If it already exists,
it will be read from this file. If file_name is None, the
file_name will be pulled out of the manifest. Default is None.
"""
file_name = self.get_cache_path(file_name, self.TEMPLATE_KEY, self.resolution)
if file_name is None:
raise Exception("No save file provided for annotation volume.")
if os.path.exists(file_name):
annotation, info = nrrd.read(file_name)
else:
Manifest.safe_make_parent_dirs(file_name)
annotation, info = self.api.download_template_volume(self.resolution, file_name)
return annotation, info
def get_injection_fraction(self, experiment_id, file_name=None):
"""
Read an injection fraction volume for a single experiment. Download it
first if it doesn't exist. Injection fraction is the proportion of
pixels in the injection site in a grid voxel in [0,1].
Parameters
----------
experiment_id: int
ID of the experiment to download/read. This corresponds to
section_data_set_id in the API.
file_name: string
File name to store the template volume. If it already exists,
it will be read from this file. If file_name is None, the
file_name will be pulled out of the manifest. Default is None.
"""
file_name = self.get_cache_path(file_name, self.INJECTION_FRACTION_KEY,
experiment_id, self.resolution)
if file_name is None:
raise Exception("No file name to save volume.")
if not os.path.exists(file_name):
Manifest.safe_make_parent_dirs(file_name)
self.api.download_injection_fraction(file_name, experiment_id,
self.resolution)
return nrrd.read(file_name)
def save_reconstruction(self, specimen_id, file_name):
"""
Save the morphological reconstruction of a cell as an SWC file.
Parameters
----------
specimen_id: int
ID of the specimen, from the Specimens database model in the Allen Institute API.
file_name: str
Path to save the SWC file.
"""
Manifest.safe_make_parent_dirs(file_name)
criteria = '[id$eq%d],neuron_reconstructions(well_known_files)' % specimen_id
includes = 'neuron_reconstructions(well_known_files(well_known_file_type[name$eq\'%s\']))' % self.SWC_FILE_TYPE
results = self.model_query('Specimen',
criteria=criteria,
include=includes,
num_rows='all')
try:
file_url = results[0]['neuron_reconstructions'][
0]['well_known_files'][0]['download_link']
except:
raise Exception("Specimen %d has no reconstruction" % specimen_id)
self.retrieve_file_over_http(self.api_url + file_url, file_name)
def get_data_mask(self, experiment_id, file_name=None):
"""
Read a data mask volume for a single experiment. Download it
first if it doesn't exist. Data mask is a binary mask of
voxels that have valid data. Only use valid data in analysis!
Parameters
----------
experiment_id: int
ID of the experiment to download/read. This corresponds to
section_data_set_id in the API.
file_name: string
File name to store the template volume. If it already exists,
it will be read from this file. If file_name is None, the
file_name will be pulled out of the manifest. Default is None.
"""
file_name = self.get_cache_path(file_name, self.DATA_MASK_KEY,
experiment_id, self.resolution)
if file_name is None:
raise Exception("No file name to save volume.")
if not os.path.exists(file_name):
Manifest.safe_make_parent_dirs(file_name)
self.api.download_data_mask(file_name, experiment_id,
self.resolution)
return nrrd.read(file_name)
def from_file_name(cls, file_name, cache=True, **kwargs):
'''Alternative constructor using cache path file_name.
Parameters
----------
file_name : string
Path where storage_directories will be saved.
**kwargs
Keyword arguments to be supplied to __init__
Returns
-------
cls : instance of GridDataApiPrerelease
'''
if os.path.exists(file_name):
storage_directories = json_utilities.read(file_name)
else:
storage_directories = _get_grid_storage_directories(
cls.GRID_DATA_DIRECTORY)
if cache:
Manifest.safe_make_parent_dirs(file_name)
json_utilities.write(file_name, storage_directories)
return cls(storage_directories, **kwargs)
def get_experiments(self, dataframe=False, file_name=None, cre=None, injection_structure_ids=None):
"""
Read a list of experiments that match certain criteria. If caching is enabled,
this will save the whole (unfiltered) list of experiments to a file.
Parameters
----------
dataframe: boolean
Return the list of experiments as a Pandas DataFrame. If False,
return a list of dictionaries. Default False.
file_name: string
File name to save/read the structures table. If file_name is None,
the file_name will be pulled out of the manifest. If caching
is disabled, no file will be saved. Default is None.
cre: boolean or list
If True, return only cre-positive experiments. If False, return only
cre-negative experiments. If None, return all experients. If list, return
all experiments with cre line names in the supplied list. Default None.
injection_structure_ids: list
Only return experiments that were injected in the structures provided here.
If None, return all experiments. Default None.
"""
file_name = self.get_cache_path(file_name, self.EXPERIMENTS_KEY)
if os.path.exists(file_name):
experiments = json_utilities.read(file_name)
else:
experiments = self.api.experiment_source_search(injection_structures="root")
# removing these elements because they are specific to a particular
# resolution
for e in experiments:
del e["num-voxels"]
del e["injection-volume"]
del e["sum"]
del e["name"]
if self.cache:
Manifest.safe_make_parent_dirs(file_name)
json_utilities.write(file_name, experiments)
# filter the read/downloaded list of experiments
experiments = self.filter_experiments(experiments, cre, injection_structure_ids)
if dataframe:
experiments = pd.DataFrame(experiments)
experiments.set_index(["id"], inplace=True, drop=False)
return experiments
def get_experiments(self,
dataframe=False,
file_name=None,
cre=None,
injection_structure_ids=None,
age=None,
gender=None,
workflow_state=None,
workflows=None,
project_code=None):
"""Read a list of experiments.
If caching is enabled, this will save the whole (unfiltered) list of
experiments to a file.
Parameters
----------
dataframe: boolean
Return the list of experiments as a Pandas DataFrame. If False,
return a list of dictionaries. Default False.
file_name: string
File name to save/read the structures table. If file_name is None,
the file_name will be pulled out of the manifest. If caching
is disabled, no file will be saved. Default is None.
"""
file_name = self.get_cache_path(file_name,
self.EXPERIMENTS_PRERELEASE_KEY)
if os.path.exists(file_name):
experiments = json_utilities.read(file_name)
else:
experiments = self.api.get_experiments()
if self.cache:
Manifest.safe_make_parent_dirs(file_name)
json_utilities.write(file_name, experiments)
# filter the read/downloaded list of experiments
experiments = self.filter_experiments(experiments, cre,
injection_structure_ids, age,
gender, workflow_state,
workflows, project_code)
if dataframe:
experiments = pd.DataFrame(experiments)
experiments.set_index(['id'], inplace=True, drop=False)
return experiments
def save_ophys_experiment_data(self, ophys_experiment_id, file_name):
Manifest.safe_make_parent_dirs(file_name)
data = self.template_query('brain_observatory_queries',
'ophys_experiment_data',
ophys_experiment_id=ophys_experiment_id)
try:
file_url = data[0]['download_link']
except Exception as _:
raise Exception("ophys experiment %d has no data file" %
ophys_experiment_id)
self._log.warning(
"Downloading ophys_experiment %d NWB. This can take some time." % ophys_experiment_id)
self.retrieve_file_over_http(self.api_url + file_url, file_name)
def get_deformation_field(self,
section_data_set_id,
header_path=None,
voxel_path=None):
''' Extract the local alignment parameters for this dataset. This a 3D vector image (3 components) describing
a deformable local mapping from CCF voxels to this section data set's affine-aligned image stack.
Parameters
----------
section_data_set_id : int
Download the deformation field for this data set
header_path : str, optional
If supplied, the deformation field header will be downloaded to this path.
voxel_path : str, optiona
If supplied, the deformation field voxels will be downloaded to this path.
Returns
-------
numpy.ndarray :
3D X 3 component vector array (origin 0, 0, 0; 25-micron isometric resolution) defining a
deformable transformation from CCF-space to affine-transformed image space.
'''
if self.resolution not in self.DFMFLD_RESOLUTIONS:
warnings.warn(
'deformation fields are only available at {} isometric resolutions, but this is a '\
'{}-micron cache'.format(self.DFMFLD_RESOLUTIONS, self.resolution)
)
header_path = self.get_cache_path(header_path,
self.DEFORMATION_FIELD_HEADER_KEY,
section_data_set_id)
voxel_path = self.get_cache_path(voxel_path,
self.DEFORMATION_FIELD_VOXEL_KEY,
section_data_set_id)
if not (os.path.exists(header_path) and os.path.exists(voxel_path)):
Manifest.safe_make_parent_dirs(header_path)
Manifest.safe_make_parent_dirs(voxel_path)
self.api.download_deformation_field(section_data_set_id,
header_path=header_path,
voxel_path=voxel_path)
return sitk.GetArrayFromImage(sitk.ReadImage(str(
header_path))) # TODO the str call here is only necessary in 2.7
def get_structure_mask(self,
structure_id,
file_name=None,
annotation_file_name=None):
"""
Read a 3D numpy array shaped like the annotation volume that has non-zero values where
voxels belong to a particular structure. This will take care of identifying substructures.
Parameters
----------
structure_id: int
ID of a structure.
file_name: string
File name to store the structure mask. If it already exists,
it will be read from this file. If file_name is None, the
file_name will be pulled out of the manifest. Default is None.
annotation_file_name: string
File name to store the annotation volume. If it already exists,
it will be read from this file. If file_name is None, the
file_name will be pulled out of the manifest. Default is None.
"""
file_name = self.get_cache_path(file_name, self.STRUCTURE_MASK_KEY,
structure_id)
if os.path.exists(file_name):
return nrrd.read(file_name)
else:
ont = self.get_ontology()
structure_ids = ont.get_descendant_ids([structure_id])
annotation, _ = self.get_annotation_volume(annotation_file_name)
mask = self.make_structure_mask(structure_ids, annotation)
if self.cache:
Manifest.safe_make_parent_dirs(file_name)
nrrd.write(file_name, mask)
return mask, None
def get_deformation_field(self, section_data_set_id, header_path=None, voxel_path=None):
''' Extract the local alignment parameters for this dataset. This a 3D vector image (3 components) describing
a deformable local mapping from CCF voxels to this section data set's affine-aligned image stack.
Parameters
----------
section_data_set_id : int
Download the deformation field for this data set
header_path : str, optional
If supplied, the deformation field header will be downloaded to this path.
voxel_path : str, optiona
If supplied, the deformation field voxels will be downloaded to this path.
Returns
-------
numpy.ndarray :
3D X 3 component vector array (origin 0, 0, 0; 25-micron isometric resolution) defining a
deformable transformation from CCF-space to affine-transformed image space.
'''
if self.resolution not in self.DFMFLD_RESOLUTIONS:
warnings.warn(
'deformation fields are only available at {} isometric resolutions, but this is a '\
'{}-micron cache'.format(self.DFMFLD_RESOLUTIONS, self.resolution)
)
header_path = self.get_cache_path(header_path, self.DEFORMATION_FIELD_HEADER_KEY, section_data_set_id)
voxel_path = self.get_cache_path(voxel_path, self.DEFORMATION_FIELD_VOXEL_KEY, section_data_set_id)
if not (os.path.exists(header_path) and os.path.exists(voxel_path)):
Manifest.safe_make_parent_dirs(header_path)
Manifest.safe_make_parent_dirs(voxel_path)
self.api.download_deformation_field(
section_data_set_id,
header_path=header_path,
voxel_path=voxel_path
)
return sitk.GetArrayFromImage(sitk.ReadImage(str(header_path))) # TODO the str call here is only necessary in 2.7
def get_session_data(self,
session_id: int,
filter_by_validity: bool = True,
**unit_filter_kwargs):
""" Obtain an EcephysSession object containing detailed data for a single session
"""
path = self.get_cache_path(None, self.SESSION_NWB_KEY, session_id,
session_id)
def read(_path):
session_api = self._build_nwb_api_for_session(
_path, session_id, filter_by_validity, **unit_filter_kwargs)
return EcephysSession(api=session_api, test=True)
Manifest.safe_make_parent_dirs(path)
return one_file_call_caching(path,
partial(self.s3fs.get,
self._get_s3_path(path), path),
lambda *a, **k: None,
read,
num_tries=self.fetch_tries)
def test_get_grid_storage_directories(storage_dirs, query_result, fn_temp_dir):
# ------------------------------------------------------------------------
# test dirs only have grid/ subdirectory
with mock.patch('allensdk.internal.core.lims_utilities.query',
new=lambda a: query_result):
obtained = _get_grid_storage_directories(
GridDataApiPrerelease.GRID_DATA_DIRECTORY)
assert not obtained
# ------------------------------------------------------------------------
# test returns storage_dirs
for path in storage_dirs.values():
Manifest.safe_make_parent_dirs(os.path.join(path, 'grid'))
with mock.patch('allensdk.internal.core.lims_utilities.query',
new=lambda a: query_result):
obtained = _get_grid_storage_directories(
GridDataApiPrerelease.GRID_DATA_DIRECTORY)
for key, value in obtained:
assert storage_dirs[key] == value
def download_section(savepath, section_id, downsample):
# Downloading all of the images from a section data set
image_api = ImageDownloadApi()
input_directory = str(section_id) + '_input'
output_directory = str(section_id) + '_output'
format_str = '.jpg'
section_images = image_api.section_image_query(section_id)
section_image_ids = [si['id'] for si in section_images]
# You have probably noticed that the AllenSDK has a logger which notifies you of file downloads.
# Since we are downloading ~300 images, we don't want to see messages for each one.
# The following line will temporarily disable the download logger.(optional)
logging.getLogger(
'allensdk.api.api.retrieve_file_over_http').disabled = True
for section_image_id in section_image_ids:
file_name = str(section_image_id) + format_str
input_file_path = os.path.join(savepath, input_directory, file_name)
output_file_path = os.path.join(savepath, output_directory, file_name)
Manifest.safe_make_parent_dirs(input_file_path)
image_api.download_section_image(section_image_id,
file_path=input_file_path,
downsample=downsample,
expression=0)
Manifest.safe_make_parent_dirs(output_file_path)
image_api.download_section_image(section_image_id,
file_path=output_file_path,
downsample=downsample,
expression=1)
# re-enable the logger (optional)
logging.getLogger(
'allensdk.api.api.retrieve_file_over_http').disabled = False
file_names = os.listdir(os.path.join(savepath, input_directory))
print(len(file_names))
def get_structure_mask(self, structure_id, file_name=None, annotation_file_name=None):
"""
Read a 3D numpy array shaped like the annotation volume that has non-zero values where
voxels belong to a particular structure. This will take care of identifying substructures.
Parameters
----------
structure_id: int
ID of a structure.
file_name: string
File name to store the structure mask. If it already exists,
it will be read from this file. If file_name is None, the
file_name will be pulled out of the manifest. Default is None.
annotation_file_name: string
File name to store the annotation volume. If it already exists,
it will be read from this file. If file_name is None, the
file_name will be pulled out of the manifest. Default is None.
"""
file_name = self.get_cache_path(file_name, self.STRUCTURE_MASK_KEY, structure_id)
if os.path.exists(file_name):
return nrrd.read(file_name)
else:
ont = self.get_ontology()
structure_ids = ont.get_descendant_ids([structure_id])
annotation, _ = self.get_annotation_volume(annotation_file_name)
mask = self.make_structure_mask(structure_ids, annotation)
if self.cache:
Manifest.safe_make_parent_dirs(file_name)
nrrd.write(file_name, mask)
return mask, None
def copy_file_entry(source, dest, use_rsync, make_parent_dirs, chmod=None):
leftmost = None
if make_parent_dirs:
leftmost = Manifest.safe_make_parent_dirs(dest)
if use_rsync:
sp.check_call(['rsync', '-a', source, dest])
else:
if Path(source).is_dir():
shutil.copytree(source, dest)
else:
shutil.copy(source, dest)
if chmod is not None:
chmod_target = leftmost if leftmost is not None else dest
apply_permissions = lambda path: path.chmod(int(f"0o{chmod}", 0))
walk_fs_tree(chmod_target, apply_permissions)
logging.info(f"copied from {source} to {dest}")
def get_experiment_structure_unionizes(self,
experiment_id,
file_name=None,
is_injection=None,
structure_ids=None,
include_descendants=False,
hemisphere_ids=None):
"""
Retrieve the structure unionize data for a specific experiment. Filter by
structure, injection status, and hemisphere.
Parameters
----------
experiment_id: int
ID of the experiment of interest. Corresponds to section_data_set_id in the API.
file_name: string
File name to save/read the experiments list. If file_name is None,
the file_name will be pulled out of the manifest. If caching
is disabled, no file will be saved. Default is None.
is_injection: boolean
If True, only return unionize records that disregard non-injection pixels.
If False, only return unionize records that disregard injection pixels.
If None, return all records. Default None.
structure_ids: list
Only return unionize records for a specific set of structures.
If None, return all records. Default None.
include_descendants: boolean
Include all descendant records for specified structures. Default False.
hemisphere_ids: list
Only return unionize records that disregard pixels outside of a hemisphere.
or set of hemispheres. Left = 1, Right = 2, Both = 3. If None, include all
records [1, 2, 3]. Default None.
"""
file_name = self.get_cache_path(file_name,
self.STRUCTURE_UNIONIZES_KEY,
experiment_id)
if os.path.exists(file_name):
unionizes = pd.DataFrame.from_csv(file_name)
else:
unionizes = self.api.get_structure_unionizes([experiment_id])
unionizes = pd.DataFrame(unionizes)
# rename section_data_set_id column to experiment_id
unionizes.columns = [
'experiment_id' if c == 'section_data_set_id' else c
for c in unionizes.columns
]
if self.cache:
Manifest.safe_make_parent_dirs(file_name)
unionizes.to_csv(file_name)
return self.filter_structure_unionizes(unionizes, is_injection,
structure_ids,
include_descendants,
hemisphere_ids)
def download_brain_slice(df):
# create an image download API
image_api = ImageDownloadApi()
format_str = ".jpg"
# You have probably noticed that the AllenSDK has a logger which notifies you of file downloads.
# Since we are downloading ~300 images, we don't want to see messages for each one.
# The following line will temporarily disable the download logger.
logging.getLogger("allensdk.api.api.retrieve_file_over_http").disabled = True
# get parameters
path, downsample, indices = ask_parameters_for_downloading(df)
print(
"Downloads initiated", end="...", file=sys.stderr, flush=True,
)
for index in indices:
# from indices, get experiment id and gene symbol from df
exp_id = df["Experiment"][index]
# set the dirname as the gene symbol
dirname = df["Gene Symbol"][index]
plane = df["Plane"][index]
section_data_set_id = exp_id
section_image_directory = os.path.join(path, dirname)
# get the image ids for all of the images in this data set
section_images = image_api.section_image_query(
section_data_set_id
) # Should be a dicionary of the features of section images
section_image_ids = [
si["id"] for si in section_images
] # Take value of 'id' from the dictionary
# Create a progress bar
pbar_image = tqdm(total=len(section_image_ids), desc=dirname + " " + plane)
for section_image_id in section_image_ids:
file_name = str(section_image_id) + format_str
file_path = os.path.join(section_image_directory, file_name)
Manifest.safe_make_parent_dirs(file_path)
# Check if the file is already downloaded, which happens if the downloads have been interrupted.
saved_file_names = os.listdir(section_image_directory)
if file_name in saved_file_names:
pass
else:
image_api.download_section_image(
section_image_id, file_path=file_path, downsample=downsample
)
pbar_image.update()
pbar_image.close()
# re-enable the logger
logging.getLogger("allensdk.api.api.retrieve_file_over_http").disabled = False
print(
"Downloads completed.", file=sys.stderr, flush=True,
)
def write_ecephys_nwb(output_path,
session_id,
session_start_time,
stimulus_table_path,
invalid_epochs,
probes,
running_speed_path,
session_sync_path,
eye_tracking_rig_geometry,
eye_dlc_ellipses_path,
eye_gaze_mapping_path,
pool_size,
optotagging_table_path=None,
session_metadata=None,
**kwargs):
nwbfile = pynwb.NWBFile(
session_description='Data and metadata for an Ecephys session',
identifier=f"{session_id}",
session_id=f"{session_id}",
session_start_time=session_start_time,
institution="Allen Institute for Brain Science")
if session_metadata is not None:
nwbfile = add_metadata_to_nwbfile(nwbfile, session_metadata)
stimulus_columns_to_drop = [
"colorSpace", "depth", "interpolate", "pos", "rgbPedestal", "tex",
"texRes", "flipHoriz", "flipVert", "rgb", "signalDots"
]
stimulus_table = read_stimulus_table(
stimulus_table_path, columns_to_drop=stimulus_columns_to_drop)
nwbfile = add_stimulus_timestamps(
nwbfile, stimulus_table['start_time'].values
) # TODO: patch until full timestamps are output by stim table module
nwbfile = add_stimulus_presentations(nwbfile, stimulus_table)
nwbfile = add_invalid_times(nwbfile, invalid_epochs)
if optotagging_table_path is not None:
optotagging_table = pd.read_csv(optotagging_table_path)
nwbfile = add_optotagging_table_to_nwbfile(nwbfile, optotagging_table)
nwbfile = add_probewise_data_to_nwbfile(nwbfile, probes)
running_speed, raw_running_data = read_running_speed(running_speed_path)
add_running_speed_to_nwbfile(nwbfile, running_speed)
add_raw_running_data_to_nwbfile(nwbfile, raw_running_data)
add_eye_tracking_rig_geometry_data_to_nwbfile(nwbfile,
eye_tracking_rig_geometry)
# Collect eye tracking/gaze mapping data from files
eye_tracking_frame_times = su.get_synchronized_frame_times(
session_sync_file=session_sync_path,
sync_line_label_keys=Dataset.EYE_TRACKING_KEYS)
eye_dlc_tracking_data = read_eye_dlc_tracking_ellipses(
Path(eye_dlc_ellipses_path))
if eye_gaze_mapping_path:
eye_gaze_data = read_eye_gaze_mappings(Path(eye_gaze_mapping_path))
else:
eye_gaze_data = None
add_eye_tracking_data_to_nwbfile(nwbfile, eye_tracking_frame_times,
eye_dlc_tracking_data, eye_gaze_data)
Manifest.safe_make_parent_dirs(output_path)
with pynwb.NWBHDF5IO(output_path, mode='w') as io:
logging.info(f"writing session nwb file to {output_path}")
io.write(nwbfile, cache_spec=True)
probes_with_lfp = [p for p in probes if p["lfp"] is not None]
probe_outputs = write_probewise_lfp_files(probes_with_lfp,
session_id,
session_metadata,
session_start_time,
pool_size=pool_size)
return {'nwb_path': output_path, "probe_outputs": probe_outputs}
def get_experiment_structure_unionizes(
self,
experiment_id,
file_name=None,
is_injection=None,
structure_ids=None,
include_descendants=False,
hemisphere_ids=None,
):
"""
Retrieve the structure unionize data for a specific experiment. Filter by
structure, injection status, and hemisphere.
Parameters
----------
experiment_id: int
ID of the experiment of interest. Corresponds to section_data_set_id in the API.
file_name: string
File name to save/read the experiments list. If file_name is None,
the file_name will be pulled out of the manifest. If caching
is disabled, no file will be saved. Default is None.
is_injection: boolean
If True, only return unionize records that disregard non-injection pixels.
If False, only return unionize records that disregard injection pixels.
If None, return all records. Default None.
structure_ids: list
Only return unionize records for a specific set of structures.
If None, return all records. Default None.
include_descendants: boolean
Include all descendant records for specified structures. Default False.
hemisphere_ids: list
Only return unionize records that disregard pixels outside of a hemisphere.
or set of hemispheres. Left = 1, Right = 2, Both = 3. If None, include all
records [1, 2, 3]. Default None.
"""
file_name = self.get_cache_path(file_name, self.STRUCTURE_UNIONIZES_KEY, experiment_id)
if os.path.exists(file_name):
unionizes = pd.DataFrame.from_csv(file_name)
else:
unionizes = self.api.get_structure_unionizes([experiment_id])
unionizes = pd.DataFrame(unionizes)
# rename section_data_set_id column to experiment_id
unionizes.columns = ["experiment_id" if c == "section_data_set_id" else c for c in unionizes.columns]
if self.cache:
Manifest.safe_make_parent_dirs(file_name)
unionizes.to_csv(file_name)
return self.filter_structure_unionizes(
unionizes, is_injection, structure_ids, include_descendants, hemisphere_ids
)
def cacher(fn,
*args,
**kwargs):
'''make an rma query, save it and return the dataframe.
Parameters
----------
fn : function reference
makes the actual query using kwargs.
path : string
where to save the data
strategy : string or None, optional
'create' always generates the data,
'file' loads from disk,
'lazy' queries the server if no file exists,
None generates the data and bypasses all caching behavior
pre : function
df|json->df|json, takes one data argument and returns filtered version, None for pass-through
post : function
df|json->?, takes one data argument and returns Object
reader : function, optional
path -> data, default NOP
writer : function, optional
path, data -> None, default NOP
kwargs : objects
passed through to the query function
Returns
-------
Object or None
data type depends on fn, reader and/or post methods.
'''
path = kwargs.pop('path', None)
strategy = kwargs.pop('strategy', None)
pre = kwargs.pop('pre', lambda d: d)
post = kwargs.pop('post', None)
reader = kwargs.pop('reader', None)
writer = kwargs.pop('writer', None)
if strategy is None:
if writer or path:
strategy = 'lazy'
else:
strategy = 'pass_through'
if not strategy in ['lazy', 'pass_through', 'file', 'create']:
raise ValueError("Unknown query strategy: {}.".format(strategy))
if 'lazy' == strategy:
if os.path.exists(path):
strategy = 'file'
else:
strategy = 'create'
if strategy == 'pass_through':
data = fn(*args, **kwargs)
elif strategy in ['create']:
Manifest.safe_make_parent_dirs(path)
if writer:
data = fn(*args, **kwargs)
data = pre(data)
writer(path, data)
else:
data = fn(*args, **kwargs)
if reader:
data = reader(path)
# Note: don't provide post if fn or reader doesn't return data
if post:
data = post(data)
return data
try:
data
return data
except:
pass
return
section_images = image_api.section_image_query(section_data_set_id)
section_image_ids = [si['id'] for si in section_images]
print(len(section_image_ids))
# You have probably noticed that the AllenSDK has a logger which notifies you of file downloads.
# Since we are downloading ~300 images, we don't want to see messages for each one.
# The following line will temporarily disable the download logger.(optional)
logging.getLogger('allensdk.api.api.retrieve_file_over_http').disabled = True
for section_image_id in section_image_ids:
file_name = str(section_image_id) + format_str
file_path = os.path.join(section_image_directory, file_name)
Manifest.safe_make_parent_dirs(file_path)
image_api.download_section_image(section_image_id, file_path=file_path, downsample=downsample,expression = expression, view='expression', colormap='expression')
# re-enable the logger (optional)
logging.getLogger('allensdk.api.api.retrieve_file_over_http').disabled = False
file_names = os.listdir(section_image_directory)
print(len(file_names))
def get_experiments(self,
dataframe=False,
file_name=None,
cre=None,
injection_structure_ids=None):
"""
Read a list of experiments that match certain criteria. If caching is enabled,
this will save the whole (unfiltered) list of experiments to a file.
Parameters
----------
dataframe: boolean
Return the list of experiments as a Pandas DataFrame. If False,
return a list of dictionaries. Default False.
file_name: string
File name to save/read the structures table. If file_name is None,
the file_name will be pulled out of the manifest. If caching
is disabled, no file will be saved. Default is None.
cre: boolean or list
If True, return only cre-positive experiments. If False, return only
cre-negative experiments. If None, return all experients. If list, return
all experiments with cre line names in the supplied list. Default None.
injection_structure_ids: list
Only return experiments that were injected in the structures provided here.
If None, return all experiments. Default None.
"""
file_name = self.get_cache_path(file_name, self.EXPERIMENTS_KEY)
if os.path.exists(file_name):
experiments = json_utilities.read(file_name)
else:
experiments = self.api.experiment_source_search(
injection_structures='root')
# removing these elements because they are specific to a particular
# resolution
for e in experiments:
del e['num-voxels']
del e['injection-volume']
del e['sum']
del e['name']
if self.cache:
Manifest.safe_make_parent_dirs(file_name)
json_utilities.write(file_name, experiments)
# filter the read/downloaded list of experiments
experiments = self.filter_experiments(experiments, cre,
injection_structure_ids)
if dataframe:
experiments = pd.DataFrame(experiments)
experiments.set_index(['id'], inplace=True, drop=False)
return experiments
def write_ecephys_nwb(output_path,
session_id,
session_start_time,
stimulus_table_path,
invalid_epochs,
probes,
running_speed_path,
session_sync_path,
eye_tracking_rig_geometry,
eye_dlc_ellipses_path,
eye_gaze_mapping_path,
pool_size,
optotagging_table_path=None,
session_metadata=None,
**kwargs):
nwbfile = pynwb.NWBFile(session_description='EcephysSession',
identifier='{}'.format(session_id),
session_start_time=session_start_time)
if session_metadata is not None:
nwbfile = add_metadata_to_nwbfile(nwbfile, session_metadata)
stimulus_table = read_stimulus_table(stimulus_table_path)
nwbfile = add_stimulus_timestamps(
nwbfile, stimulus_table['start_time'].values
) # TODO: patch until full timestamps are output by stim table module
nwbfile = add_stimulus_presentations(nwbfile, stimulus_table)
nwbfile = add_invalid_times(nwbfile, invalid_epochs)
if optotagging_table_path is not None:
optotagging_table = pd.read_csv(optotagging_table_path)
nwbfile = add_optotagging_table_to_nwbfile(nwbfile, optotagging_table)
nwbfile = add_probewise_data_to_nwbfile(nwbfile, probes)
running_speed, raw_running_data = read_running_speed(running_speed_path)
add_running_speed_to_nwbfile(nwbfile, running_speed)
add_raw_running_data_to_nwbfile(nwbfile, raw_running_data)
# --- Add eye tracking ellipse fits to nwb file ---
eye_tracking_frame_times = su.get_synchronized_frame_times(
session_sync_file=session_sync_path,
sync_line_label_keys=Dataset.EYE_TRACKING_KEYS)
eye_dlc_tracking_data = read_eye_dlc_tracking_ellipses(
Path(eye_dlc_ellipses_path))
if eye_tracking_data_is_valid(eye_dlc_tracking_data=eye_dlc_tracking_data,
synced_timestamps=eye_tracking_frame_times):
add_eye_tracking_ellipse_fit_data_to_nwbfile(
nwbfile,
eye_dlc_tracking_data=eye_dlc_tracking_data,
synced_timestamps=eye_tracking_frame_times)
# --- Append eye tracking rig geometry info to nwb file (with eye tracking) ---
append_eye_tracking_rig_geometry_data_to_nwbfile(
nwbfile, eye_tracking_rig_geometry=eye_tracking_rig_geometry)
# --- Add gaze mapped positions to nwb file ---
if eye_gaze_mapping_path:
eye_gaze_data = read_eye_gaze_mappings(Path(eye_gaze_mapping_path))
add_eye_gaze_mapping_data_to_nwbfile(nwbfile,
eye_gaze_data=eye_gaze_data)
Manifest.safe_make_parent_dirs(output_path)
io = pynwb.NWBHDF5IO(output_path, mode='w')
logging.info(f"writing session nwb file to {output_path}")
io.write(nwbfile)
io.close()
probes_with_lfp = [p for p in probes if p["lfp"] is not None]
probe_outputs = write_probewise_lfp_files(probes_with_lfp,
session_start_time,
pool_size=pool_size)
return {'nwb_path': output_path, "probe_outputs": probe_outputs}
def safe_write(data: Q):
Manifest.safe_make_parent_dirs(path)
write(path, data)
def cacher(fn,
*args,
**kwargs):
'''make an rma query, save it and return the dataframe.
Parameters
----------
fn : function reference
makes the actual query using kwargs.
path : string
where to save the data
strategy : string or None, optional
'create' always generates the data,
'file' loads from disk,
'lazy' queries the server if no file exists,
None generates the data and bypasses all caching behavior
pre : function
df|json->df|json, takes one data argument and returns
filtered version, None for pass-through
post : function
df|json->?, takes one data argument and returns Object
reader : function, optional
path -> data, default NOP
writer : function, optional
path, data -> None, default NOP
kwargs : objects
passed through to the query function
Returns
-------
Object or None
data type depends on fn, reader and/or post methods.
'''
path = kwargs.pop('path', None)
strategy = kwargs.pop('strategy', None)
pre = kwargs.pop('pre', lambda d: d)
post = kwargs.pop('post', None)
reader = kwargs.pop('reader', None)
writer = kwargs.pop('writer', None)
if strategy is None:
if writer or path:
strategy = 'lazy'
else:
strategy = 'pass_through'
if strategy not in ['lazy', 'pass_through',
'file', 'create']:
raise ValueError("Unknown query strategy: {}.".format(strategy))
if 'lazy' == strategy:
if os.path.exists(path):
strategy = 'file'
else:
strategy = 'create'
if strategy == 'pass_through':
data = fn(*args, **kwargs)
elif strategy in ['create']:
Manifest.safe_make_parent_dirs(path)
if writer:
data = fn(*args, **kwargs)
data = pre(data)
writer(path, data)
else:
data = fn(*args, **kwargs)
if reader:
data = reader(path)
# Note: don't provide post if fn or reader doesn't return data
if post:
data = post(data)
return data
try:
data
return data
except Exception:
pass
return
