NWB Converter#

Classes:

NWBConverter(*args, **kwargs)

ONICE extension to nwb_conversion_tools.NWBConverter

Functions:

_monkeypatch_spikeextractors()

To make NWBConverter.hail_mary() work, we have to override some __del__ methods in spikeextractors that throw errors on incomplete __init__ calls

class onice_conversion.nwbconverter.NWBConverter(*args, **kwargs)#

Bases: nwb_conversion_tools.nwbconverter.NWBConverter

ONICE extension to nwb_conversion_tools.NWBConverter

Parameters

base_dir (pathlib.Path) – The base directory of the source data, from which all paths are relative. If not provided at initialization, must be provided when calling run_conversion()
source_data (dict) – Old style source_data dictionary, kept for compatibility

Methods:

`add_container`([container_name, spec])	Add a
`hail_mary`([base_dir, interface_type])	Just try every interface on every file and see what instantiates.
`add_interface`([interface_type, device_name, ...])	Add a recording interface
`add_metadata`(spec)	Parameters spec (BaseSpec, dict) -- if an object that inherits from `BaseSpec`, then the keys and values of metadata
`convert_many`(expt_paths[, out_fns])
`from_json`(json_path[, hook, base_dir])	Reconstitute a parameterized converter from a json file created by `NWBConverter.to_json()`
`get_conversion_options_schema`()	Compile conversion option schemas from each of the data interface classes.
`get_metadata`()	Auto-fill as much of the metadata as possible.
`get_metadata_schema`()	Compile metadata schemas from each of the data interface objects.
`get_source_schema`()	Compile input schemas from each of the data interface classes.
`run_conversion`([metadata, nwbfile_path, ...])	Run the NWB conversion over all the instantiated data interfaces.
`to_json`([output_path, mode])	Save the converter parameterization from `add_metadata()` and `add_interface()` to a .json file, for use with `from_json()` to recreate conversion objects :)

Attributes:

`base_nwb_metadata`	Return descriptions for the basic file-level metadata container objects, `'NWBFile', 'Subject'`
`conversion_options_schema`
`metadata`
`metadata_schema`
`source_schema`

add_container(container_name: Optional[str] = None, spec: Optional[onice_conversion.spec.base_spec.BaseSpec] = None, **kwargs)#

Add a

Parameters

container_type (str)
container_name (str)
spec (BaseSpec) – Spec object declaring the metadata for the container
**kwargs – stored as static metadata and passed to container

Returns:

property base_nwb_metadata: Dict[str, tuple]#

Return descriptions for the basic file-level metadata container objects, 'NWBFile', 'Subject'

Returns: dict of tuples of parameter spec for each container type

hail_mary(base_dir: Optional[pathlib.Path] = None, interface_type: Optional[str] = None)#

Just try every interface on every file and see what instantiates.

Parameters

base_dir (directory to peruse. if none, then the base_dir provided on init is used.)
interface_type (if provided, only try interfaces of this type)

Returns

(interface object, path (relative to base_dir), parameter key that was used, and the instantiated object itself)

Return type

tuple of:

add_interface(interface_type: Optional[str] = None, device_name: Optional[str] = None, spec: Optional[nwb_conversion_tools.spec.base_spec.BaseSpec] = None, **kwargs)#

Add a recording interface

Specify interface either with an interface type and name, or else give the class itself as interface_class. If both are present, use the class.

Everything afterwards

Parameters

interface_type (str) – Type of interface, like ‘recording’ – a name of a package in interfaces
device_name (str) – Name of specific interface, matching the interfaces device_name
spec (BaseSpec) – Metadata specifier to parameterize interface object
kwargs – kwargs passed to data interface.

add_metadata(spec: Union[nwb_conversion_tools.spec.base_spec.BaseSpec, str])#

Parameters

spec (BaseSpec, dict) – if an object that inherits from BaseSpec, then the keys and values of metadata are resolved by the object: ie. the keys are the value of BaseSpec.specifies and BaseSpec.parse() returns a dictionary of keys and values.

if dictionary, assumes static metadata (unchanged across multiple sessions/experiments) otherwise, use spec to resolve Either a string representing a “top-level” metadata property, or a tuple of nested metadata properties like (‘NWBFile’, ‘experimenter’)

property conversion_options_schema#

convert_many(expt_paths: list, out_fns: Optional[list] = None, *args, **kwargs)#

classmethod from_json(json_path: Union[str, pathlib.Path, dict], hook: Optional[Callable] = None, base_dir: Optional[Union[str, pathlib.Path]] = None) → nwb_conversion_tools.nwbconverter.NWBConverter#

Reconstitute a parameterized converter from a json file created by NWBConverter.to_json()

Parameters

json_path (str, pathlib.Path, dict) – Path to the .json file, or else the already-loaded dict
hook (callable) – Optional callable to use with json.load’s object_hook
base_dir (str, pathlib.Path) – Optional, instantiate the converter with a base_path

Return type

Reconstituted Converter!

get_conversion_options_schema()#: Compile conversion option schemas from each of the data interface classes.

get_metadata()#: Auto-fill as much of the metadata as possible. Must comply with metadata schema.

get_metadata_schema()#: Compile metadata schemas from each of the data interface objects.

get_source_schema()#: Compile input schemas from each of the data interface classes.

property metadata#

property metadata_schema#

run_conversion(metadata: Optional[dict] = None, nwbfile_path: Optional[str] = None, overwrite: Optional[bool] = False, nwbfile: Optional[pynwb.file.NWBFile] = None, conversion_options: Optional[dict] = None, base_dir: Optional[pathlib.Path] = None)#

Run the NWB conversion over all the instantiated data interfaces.

Parameters

metadata (dict)
nwbfile_path (str, optional) – Location to save the NWBFile, if save_to_file is True. The default is None.
overwrite (bool, optional) – If True, replaces any existing NWBFile at the nwbfile_path location, if save_to_file is True. If False, appends the existing NWBFile at the nwbfile_path location, if save_to_file is True. The default is False.
nwbfile (NWBFile, optional) – A pre-existing NWBFile object to be appended (instead of reading from nwbfile_path).
conversion_options (dict, optional) – Similar to source_data, a dictionary containing keywords for each interface for which non-default conversion specification is requested.

Returns

nwbfile – the created NWBFile

Return type

NWBFile

property source_schema#

to_json(output_path: Optional[Union[str, pathlib.Path]] = None, mode: str = 'w') → dict#

Save the converter parameterization from add_metadata() and add_interface() to a .json file, for use with from_json() to recreate conversion objects :)

Parameters

output_path (str, pathlib.Path) – Path to write .json file to. if None, don’t save, just return dict.
mode (str) – Write mode, default: ‘w’

Return type

dict created and saved

onice_conversion.nwbconverter._monkeypatch_spikeextractors()#: To make NWBConverter.hail_mary() work, we have to override some __del__ methods in spikeextractors that throw errors on incomplete __init__ calls