Fabricator

class managers.operatorsTranscribe.fabricator.Fabricator

Available Methods:

Method name	Method type
build_nwbfile()	class method
build_nwbseries()	class method
affix_nwbseries_to_nwbfile()	class method
build_nwbepochs()	class method
write_nwbfile()	class method
construct_nwbseries_nostimulus()	class method
generic_timeseries()	static method
link_nwbseriesresponses_to_nwbfile()	static method
strip_out_stimulus_from_nwbseries()	static method
indices_tseries_for_epoch()	static method
tseries_for_epoch()	class method
insert_a_nwbepoch()	class method

classmethod affix_nwbseries_to_nwbfile(nwbts=None, nwbfile=None)

Adds time-series (response time-series with or without the stimulus time-series).

Keyword Arguments:

Key	Value type
nwbts	dictionary of NWB time-series object its keys are the keys in chosenmodel.regions = {"soma": 0.0, "axon", 0.0} - the key "stimulus" is optional - value for each key is a pynwb.base.TimeSeries object, obtained using build_nwbseries method()
nwbfile	the built NWB file of type pynwb.file.NWBFile obtained using build_nwbfile method()

Returned value: This is the NWB file fed as an argument but updated by adding the time-series, say, updated_nwbfile. The TimeSeries object can be extracted as

>> ts_of_key = updated_nwbfile.get_acquisition(nwbseries[key].name)

where key is the region. For example,

>> ts_soma = updated_nwbfile.get_acquisition(nwbseries["soma"].name)

Then you can get all the available attributes as usual

ts_soma.name, ts_soma.data,
ts_soma.timestamps, ts_soma.unit,
ts_soma.resolution, ts_soma.converstion,
ts_soma.starting_time, ts_soma.rate,
ts_soma.comment, ts_soma.description

But if "stimulus" is one of the availabe keys then do

>> ts_stim = updated_nwbfile.get_stimulus(nwbseries["stimulus"].name)

Now all the aforementioned attributes are available as usual.

NOTE: Unlike the returned value for build_nwbseries() the time-series here are for a particular key therefore it is no longer a dictionary.

classmethod build_nwbepochs(epochmd=None, nwbfile=None, nwbts=None)

method for contructing epochs into the built nwbfile

Keyword Arguments:

Key	Value type
epochmd	dictionary of time-series metadata considering the case chosenmodel.regions = {'soma': 0.0, 'axon': 0.0} and the number of epochs per region = 2 - when there is a stimulation the dictionary will be of the form {"epoch0soma": {"source": "soma", "start_time": float, "stop_time": float, "description": string, "tags": tuple} "epoch1soma": {"source": "soma", "start_time": float, "stop_time": float, "description": string, "tags": tuple} "epoch0axon": {"source": "axon", "start_time": float, "stop_time": float, "description": string, "tags": tuple} "epoch1axon": {"source": "axon", "start_time": float, "stop_time": float, "description": string, "tags": tuple}} but when there is no stimulation the dictionary will look as {"epoch0soma": {"source": "soma", "start_time": float, "stop_time": float, "description": string, "tags": tuple} "epoch0axon": {"source": "axon", "start_time": float, "stop_time": float, "description": string, "tags": tuple}}
nwbfile	the built NWB file of type pynwb.file.NWBFile obtained using build_nwbfile method()
nwbts	dictionary of NWB time-series object its keys are the keys in chosenmodel.regions = {"soma": 0.0, "axon", 0.0} - the key "stimulus" is optional - value for each key is a pynwb.base.TimeSeries object, obtained using build_nwbseries method()

NOTE:

• the epoch metadata has the key "source". This represents the region from which epoch is considered this should not be confused with any attribute of the NWB file. Besides, since NWB2.0 source is no longer an attribute of NWB file.

Use case:

epoch_meta_data = {
  "epoch0soma": {
       "source": "soma",
       "start_time": 0.0, "stop_time": 5.0,
       "description": "first epoch",
       "tags": ('1_epoch_responses', '0', 'soma', 'DummyTest', 'cells', "epoch0soma")},
  "epoch0axon": {
       "source": "axon",
       "start_time": 5.0, "stop_time": 10.0,
       "description": "first epoch",
       "tags": ('1_epoch_responses', '0', 'axon', 'DummyTest', 'cells', "epoch0axon")}
       }
updated_nwbfile = build_nwbepochs(nwbfile=nwbfile, epochmd=epoch_meta_data, nwbts=nwbts)

Returned Value: The updated file, say, nwbfile is returned such that it now has .epochs attributes. For example,

>> len(nwbfile.epochs)

returns 2, reflecting the number of epochs. Thus epoch in first row print(nwbfile.epochs[0]) will appear as

[(0, # => epochmd["epoch0soma"]["start_time"] 
  5, # => epochmd["epoch0soma"]["stop_time"]
  DummyTest_soma <class 'pynwb.base.TimeSeries'>
  Fields:
    comments: DummyTest_soma
    conversion: 1000.0
    data: [0.73745691 0.16092811 0.83702311 ... 0.78635451]
    description: whole single array of voltage response from soma of DummyTest
    interval: 1
    num_samples: 501
    resolution: 0.01
    timestamps: [0.   0.01 0.02 ... 5.0]
    timestamps_unit: Seconds
    unit: mV
)]

The epoch in second row print(nwbfile.epochs[1]) will return

[(5, # => epochmd["epoch0axon"]["start_time"] 
  10, # => epochmd["epoch0axon"]["stop_time"]
  DummyTest_axon <class 'pynwb.base.TimeSeries'>
  Fields:
    comments: DummyTest_axon
    conversion: 1000.0
    data: [0.73745691 0.16092811 0.83702311 ... 0.78635451]
    description: whole single array of voltage response from axon of DummyTest
    interval: 1
    num_samples: 501
    resolution: 0.01
    timestamps: [5.   5.01 5.02 ... 10.0]
    timestamps_unit: Seconds
    unit: mV
)]

For all practical purposes we are interested in the created epoch data, in particular the time-series data associated with a particular epoch. Let us assume our epoch of interest is in the first row of nwbfile.epochs, i.e, nwbfile.epochs[0].

Since the last column of the table (i.e, 5th column) is labelled "timeseries" doing nwbfile.epochs[0][4] returns

[(0, # => epochmd["epoch0soma"]["start_time"] 
  5, # => epochmd["epoch0soma"]["stop_time"]
  DummyTest_soma <class 'pynwb.base.TimeSeries'>
  Fields:
    comments: DummyTest_soma
    conversion: 1000.0
    data: [0.73745691 0.16092811 0.83702311 ... 0.78635451]
    description: whole single array of voltage response from soma of DummyTest
    interval: 1
    num_samples: 501
    resolution: 0.01
    timestamps: [0.   0.01 0.02 ... 5.0]
    timestamps_unit: Seconds
    unit: mV
)]

• therefore, root data for i’th epoch will be nwbfile.epochs[i][4]

Doing nwbfile.epochs.[0][4][0][2] returns the TimeSeries class

DummyTest_soma <class 'pynwb.base.TimeSeries'>
 Fields:
   comments: DummyTest_soma
   conversion: 1000.0
   data: [0.73745691 0.16092811 0.83702311 ... 0.78635451]
   description: whole single array of voltage response from soma of DummyTest
   interval: 1
   num_samples: 501
   resolution: 0.01
   timestamps: [0.   0.01 0.02 ... 5.0]
   timestamps_unit: Seconds
   unit: mV

• the TimeSeries object is the 3rd column (index = 2), that is, nwbfile.epochs[i][4][0][2]

Finally to retrieve the time-series data and timestamps associated with i’th epoch is achieved by

nwbfile.epochs.[i][4][i][2].timestamps

nwbfile.epochs.[i][4].[i][2].data

Refer:

• insert_a_nwbepoch() documentation.

NOTE: If we are interested in extracting an epoch in a more organized manner for instance epoch of interest may be a particular epoch for recording done from a particular region, then using the above method of extraction based on picking a row of the VectorTable is not sufficient. The more improved approach of extraction will be done by the manager.

classmethod build_nwbfile(filemd)

Builds the nwbfile

Argument:

Argument	Value type
only one	dictionary of the file metadata such that { "session_description": "How was the data generated?", "identifier": "a unique ID", "session_start_time": datetime(time when recording began), "experimenter": "name of the experimenter", "experiment_description": "described experiment", "session_id": "collab ID", "institution": "name of the institution", "lab": "name of the lab" }

Returned Value: The returned nwbfile has the following attributes of interests: nwbfile.session_description, nwbfile.identifier, nwbfile.session_start_time, nwbfile.experimenter, nwbfile.experiment_description, nwbfile.session_id, nwbfile.lab and nwbfile.institution.

Use case:

>> file_to_write = build_nwbfile( file_metadata )

NOTE: The overall available attributes are however

pynwb.file.NWBFile(
     session_description, identifier, session_start_time,
     file_create_date=None, experimenter=None, experiment_description=None,
     session_id=None, institution=None, notes=None, pharmacology=None,
     protocol=None, related_publications=None, slices=None,
     source_script=None, source_script_file_name=None, data_collection=None,
     surgery=None, virus=None, stimulus_notes=None, lab=None,
     acquisition=None, stimulus=None, stimulus_template=None, epochs=None,
     epoch_tags=set(), trials=None, modules=None, ec_electrodes=None,
     ec_electrode_groups=None, ic_electrodes=None, imaging_planes=None,
     ogen_sites=None, devices=None, subject=None )

Refer to pynwb documentation.

classmethod build_nwbseries(chosenmodel=None, tsmd=None)

Builds an NWB time-series object.

Keyword Arguments:

Key	Value type
chosenmodel tsmd	instantiated model - dictionary of time-series metadata
	key	value type
	"name"	string
	"data"	list/tuple/array
	"unit"	string
	"resolution"	string/float
	"conversion"	string/float
	"timestamps"	list/tuple/array
	"comments"	string
	"description"	string
	"stimulus" (optional)

Returned Value: The NWB time-series object, say nwbts will have the following useful attributes

nwbts[key].name, nwbts[key].data,
nwbts[key].timestamps, nwbts[key].unit,
nwbts[key].resolution, nwbts[key].converstion,
nwbts[key].starting_time, nwbts[key].rate,
nwbts[key].comment, nwbts[key].description

such that a key are the keys in chosemodel.regions = {'soma':0.0, 'axon':0.0} and with or without "stimulus" key.

NOTE: The overall available attributes are however

pynwb.base.TimeSeries(
                    name, data=None, unit=None, resolution=0.0,
                    conversion=1.0, timestamps=None, starting_time=None,
                    rate=None, comments='no comments',
                    description='no description', control=None,
                    control_description=None, parent=None )

Refer to pynwb documentation.

classmethod construct_nwbseries_nostimulus(chosenmodel, tsmd)

Creates NWB time-series object without stimulus.

Arguments:

Arguments	Value type
first	instantiated model
second	dictionary of time-series metadata it must have the following keys and their values
	key	value type
	"name"	string
	"data"	list/tuple/array
	"unit"	string
	"resolution"	string/float
	"conversion"	string/float
	"timestamps"	list/tuple/array
	"comments"	string
	"description"	string

Returned Value: The NWB time-series object, say nwbts will have the following useful attributes

nwbts[key].name, nwbts[key].data,
nwbts[key].timestamps, nwbts[key].unit,
nwbts[key].resolution, nwbts[key].converstion,
nwbts[key].starting_time, nwbts[key].rate,
nwbts[key].comment, nwbts[key].description

such that a key are the keys in chosemodel.regions = {'soma':0.0, 'axon':0.0}.

NOTE: The overall available attributes are however

pynwb.base.TimeSeries(
                    name, data=None, unit=None, resolution=0.0,
                    conversion=1.0, timestamps=None, starting_time=None,
                    rate=None, comments='no comments',
                    description='no description', control=None,
                    control_description=None, parent=None )

Refer to pynwb documentation.

static generic_timeseries(metadata)

Creates a generic NWB time-series object.

Argument:

Argument	Value type
only one	dictionary of time-series metadata it must have the following keys and their value type
	key	value type
	"name"	string
	"data"	list/tuple/array
	"unit"	string
	"resolution"	string/float
	"conversion"	string/float
	"timestamps"	list/tuple/array
	"comments"	string
	"description"	string

Returned Value: The NWB time-series object, say nwbts will have the following useful attributes nwbts.name, nwbts.data, nwbts.timestamps, nwbts.unit, nwbts.resolution, nwbts.converstion, #nwbts.starting_time, #nwbts.rate, nwbts.comment, nwbts.description.

NOTE: The overall available attributes are however

pynwb.base.TimeSeries(
                   name, data=None, unit=None, resolution=0.0,
                   conversion=1.0, timestamps=None, starting_time=None,
                   rate=None, comments='no comments',
                   description='no description', control=None,
                   control_description=None, parent=None )

Refer to pynwb documentation.

Potential Bug:

• due to the bug I reported starting_time and rate attributes are not included thus defaults to None
• pynwb developer says “.. is because you either enter timestamps or rate+starting_time, not all three. We should catch this and raise an error.”
• so to avoid error in future pynwb version the two attributes are taken out.

static indices_tseries_for_epoch(epochmd_i_cellregion, nwbts)

Returns list of indices of the timestamps from the given time-series that is between start_time and stop_time in an epoch of a given region.

Arguments:

Argument	Value type
first	dictionary of time-series metadata for a particular region of an epoch considering the case chosenmodel.regions = {'soma': 0.0, 'axon': 0.0} and the number of epochs per region = 2 - when there is a stimulation the dictionary will be of the form {"epoch0soma": {"source": "soma", "start_time": float, "stop_time": float, "description": string, "tags": tuple} "epoch1soma": {"source": "soma", "start_time": float, "stop_time": float, "description": string, "tags": tuple} "epoch0axon": {"source": "axon", "start_time": float, "stop_time": float, "description": string, "tags": tuple} "epoch1axon": {"source": "axon", "start_time": float, "stop_time": float, "description": string, "tags": tuple}} but when there is no stimulation the dictionary will look as {"epoch0soma": {"source": "soma", "start_time": float, "stop_time": float, "description": string, "tags": tuple} "epoch0axon": {"source": "axon", "start_time": float, "stop_time": float, "description": string, "tags": tuple}}
second	dictionary of NWB time-series object its keys are the keys in chosenmodel.regions = {"soma": 0.0, "axon", 0.0} - the key "stimulus" is optional - value for each key is a pynwb.base.TimeSeries object, obtained using build_nwbseries method()

NOTE:

• the whole NWB TimeSeries is not passed here
• only time series that corresponds to the respective region is passed into this function
• the epoch metadata has the key "source". This represents the region from which epoch is considered this should not be confused with any attribute of the NWB file. Besides, since NWB2.0 source is no longer an attribute of NWB file.

classmethod insert_a_nwbepoch(epoch_i_cellregion, epochmd, nwbfile, nwbts)

Creates an epoch for a given region and updates the already created parent NWB file. This method called by construct_nwbepochs().

Arguments:

Argument	Value type
first	string for key of an epoch from a particular region, say, "epoch1soma"
second	dictionary of time-series metadata considering the case chosenmodel.regions = {'soma': 0.0, 'axon': 0.0} and the number of epochs per region = 2 - when there is a stimulation the dictionary will be of the form {"epoch0soma": {"source": "soma", "start_time": float, "stop_time": float, "description": string, "tags": tuple} "epoch1soma": {"source": "soma", "start_time": float, "stop_time": float, "description": string, "tags": tuple} "epoch0axon": {"source": "axon", "start_time": float, "stop_time": float, "description": string, "tags": tuple} "epoch1axon": {"source": "axon", "start_time": float, "stop_time": float, "description": string, "tags": tuple}} but when there is no stimulation the dictionary will look as {"epoch0soma": {"source": "soma", "start_time": float, "stop_time": float, "description": string, "tags": tuple} "epoch0axon": {"source": "axon", "start_time": float, "stop_time": float, "description": string, "tags": tuple}}
third	the built NWB file of type pynwb.file.NWBFile obtained using build_nwbfile method()
fourth	dictionary of NWB time-series object its keys are the keys in chosenmodel.regions = {"soma": 0.0, "axon", 0.0} - the key "stimulus" is optional - value for each key is a pynwb.base.TimeSeries object, obtained using build_nwbseries method()

NOTE:

• the whole NWB TimeSeries is not passed here
• only time series that corresponds to the respective region is passed into this function
• the epoch metadata has the key "source". This represents the region from which epoch is considered this should not be confused with any attribute of the NWB file. Besides, since NWB2.0 source is no longer an attribute of NWB file.

Use case:

epochmd = {"epoch0soma": {"source": "soma", "start_time": float, "stop_time": float,
                          "description": string, "tags": tuple},
           "epoch1axon": {"source": "soma", "start_time": float, "stop_time": float,
                          "description": string, "tags": tuple},
           "epoch0axon": {"source": "axon", "start_time": float, "stop_time": float,
                          "description": string, "tags": tuple},
           "epoch1axon": {"source": "axon", "start_time": float, "stop_time": float,
                          "description": string, "tags": tuple}}
updated_nwbfile = insert_a_nwbepoch( "epoch0soma", epochmd, nwbfile )

Returned Value: The updated file, say, nwbfile is returned such that it now has .epochs attributes. For example,

>> len(nwbfile.epochs)

returns 1 because we inserted only one epoch. The number will reflect the number of epochs. See VectorData. Thus print(nwbfile.epochs[0]) returns a tuple such that

  ( 0,  # => row number, i.e, epoch number in the order of insertion
    0.0,   # => epochmd["epoch0soma"]["start_time"] 
    0.05,  # => epochmd["epoch0soma"]["stop_time"]
    ['1_epoch_responses', '0', 'axon', 'DummyTest', 'cellsepoch0axon'], # => epochmd["epoch0soma"]["tags"]
    [(0, 5,
     DummyTest_axon <class 'pynwb.base.TimeSeries'>
     Fields:
       comments: DummyTest_axon
       conversion: 1000.0
       data: [0.73745691 0.16092811 0.83702311 0.73879371 0.55916576 0.78635451]
       description: whole single array of voltage response from axon of DummyTest
       interval: 1
       num_samples: 6
       resolution: 0.01
       timestamps: [0.   0.01 0.02 0.03 0.04 0.05]
       timestamps_unit: Seconds
       unit: mV
  )])

*NOTE:*

* for a particular epoch such tuple has 5-elements
* pynwb puts all the epochs in a `DynamicTable <https://pynwb.readthedocs.io/en/latest/pynwb.core.html#pynwb.core.DynamicTable>`_ such that

  - a row corresponds to a particular epoch
  - there are 5-columns which corresponds to the 5-elements of an epoch's tuple (as seen above)
  - the first column (seen above as ``0`` NOT ``0.0``) is like the row number or epoch number.
  - the rest have names which by doing ``nwbfile.epochs.colnames`` returns ``('start_time', 'stop_time', 'tags', 'timeseries')``

* if this function was called again, say, ``updated_nwbfile = insert_a_nwbepoch( "epoch0axon", epochmd, nwbfile )`` then the same ``print(nwbfile.epochs[0])`` (as above) returns the same tuple

For all practical purposes we are interested in the created epoch data, in particular the time-series data associated with a particular epoch.

Since the last column of the table (i.e, 5th column) is labelled "timeseries" doing nwbfile.epochs[0][4] returns

[(0, 5,
  DummyTest_axon <class 'pynwb.base.TimeSeries'>
  Fields:
    comments: DummyTest_axon
    conversion: 1000.0
    data: [0.73745691 0.16092811 0.83702311 0.73879371 0.55916576 0.78635451]
    description: whole single array of voltage response from axon of DummyTest
    interval: 1
    num_samples: 6
    resolution: 0.01
    timestamps: [0.   0.01 0.02 0.03 0.04 0.05]
    timestamps_unit: Seconds
    unit: mV
)]

Since this is a list with tuple elements and we are interested in the tuple this can be extracted as nwbfile.epochs[0][4][0].

• the first index (i.e, [0]) is the row index thus representing a particular epoch
  • here the index = 0 because there is only one epoch
• the data (i.e, timeseries) is always index = 4, i.e, 5th column
• therefore, data for i’th epoch will be nwbfile.epochs[i][4]

Since nwbfile.epochs.[0][4] returns returned a list of tuple elements (above we see a list of only one because only one epoch was inserted), doing nwbfile.epochs.[0][4][0] returns the tuple containing the NWB TimeSeries class.

Doing nwbfile.epochs.[0][4][0][2] returns the TimeSeries class

• the third index corresponds to the row index
• for i’th epoch, nwbfile.epochs[i][4] its corresponding tuple containing the TimeSeries is nwbfile.epochs[i][4][0]
• the TimeSeries object is the 3rd column (index = 2), that is, nwbfile.epochs[i][4][0][2]

Finally to retrieve the time-series data and timestamps associated with i’th epoch is achieved by

nwbfile.epochs.[i][4][i][2].timestamps

nwbfile.epochs.[i][4].[i][2].data

• len(nwbfile.epochs) = len(epoch_metadata), i.e, the total number of rows and hence the total number of epochs in the created NWB file.
  • nwbfile.epochs is a VectorTable whose length is equal to the number of epochs.
  • hence len(nwbfile.epochs[0]) = len(nwbfile.epochs[i]) = 1
  • and each print(nwbfile.epochs[i]) is a tuple as seen in the beginning of the above example
• the root tuple of any i’th epoch nwbfile.epochs[i] is always a tuple of five elements.
  • nwbfile.epochs[i] is a DynamicTable with five columns.
  • its fifth element nwbfile.epochs[i][4] is a list with one tuple
• the only child tuple of any i’th epoch nwbfile.epochs.[i][4][0] is always a tuple of three elements
  • nwbfile.epochs[i][4][0] is a DynamicTable with three columns.
  • the third element nwbfile.epochs[i][4][0][2] within the tuple is the TimeSeries class
• (in other words) for any i (over all epochs) nwbfile.epochs.[i][4] is the same for all the epochs.
  • this is because nwbfile.epochs[i][4] is a DynamicTable
• recall that nwbfile.epochs is also a table however, they are not the same table types
  • nwbfile.epochs is a VectorTable whose length is equal to the number of epochs.
  • nwbfile.epochs[i][4] is a DynamicTable with five columns.
  • but in both tables its rows correspond to respective epoch
  • therefore, nwbfile.epochs[i][4][i] for i’th epoch

Refer:

• http://pynwb.readthedocs.io/en/latest/pynwb.epoch.html#pynwb.epoch.Epochs
• https://github.com/AllenInstitute/nwb-api/blob/master/ainwb/nwb/nwbep.py
• https://pynwb.readthedocs.io/en/latest/pynwb.core.html#pynwb.core.VectorData
• https://pynwb.readthedocs.io/en/latest/pynwb.core.html#pynwb.core.DynamicTable

static link_nwbseriesresponses_to_nwbfile(nwbseries, nwbfile)

Adds response related time-series NOT the stimulus signal (time-series). This is called by affix_nwbseries_to_nwbfile().

Arguments:

Arguments	Value type
first	dictionary of NWB time-series object its keys are the keys in chosenmodel.regions = {"soma": 0.0, "axon", 0.0} - value for each key is a pynwb.base.TimeSeries object, obtained using build_nwbseries method()
second	the built NWB file of type pynwb.file.NWBFile obtained using build_nwbfile method()

Returned value: This is the NWB file fed as an argument but updated by adding the time-series, say, updated_nwbfile. The TimeSeries object can be extracted as

>> ts_of_key = updated_nwbfile.get_acquisition(nwbseries[key].name)

where key is the region. For example,

>> ts_soma = updated_nwbfile.get_acquisition(nwbseries["soma"].name)

Then you can get all the available attributes as usual

ts_soma.name, ts_soma.data,
ts_soma.timestamps, ts_soma.unit,
ts_soma.resolution, ts_soma.converstion,
ts_soma.starting_time, ts_soma.rate,
ts_soma.comment, ts_soma.description

NOTE:

• unlike the returned value for build_nwbseries() the time-series here are for a particular key therefore it is no longer a dictionary.

static strip_out_stimulus_from_nwbseries(nwbseries)

Extracts from the root time-series object all the time-series objects with the exception of the stimulus time-series. This method is called by affix_nwbseries_to_nwbfile().

classmethod tseries_for_epoch(epochmd_i_cellregion, nwbts)

Returns the NWB TimeSeries object with data and timestamps occuring between start_time and stop_time in an epoch of a given region. This method is called by :py:meth::.insert_a_nwbepoch.

Arguments:

Argument	Value type
first	dictionary of time-series metadata for a particular region of an epoch considering the case chosenmodel.regions = {'soma': 0.0, 'axon': 0.0} and the number of epochs per region = 2 - when there is a stimulation the dictionary will be of the form {"epoch0soma": {"source": "soma", "start_time": float, "stop_time": float, "description": string, "tags": tuple} "epoch1soma": {"source": "soma", "start_time": float, "stop_time": float, "description": string, "tags": tuple} "epoch0axon": {"source": "axon", "start_time": float, "stop_time": float, "description": string, "tags": tuple} "epoch1axon": {"source": "axon", "start_time": float, "stop_time": float, "description": string, "tags": tuple}} but when there is no stimulation the dictionary will look as {"epoch0soma": {"source": "soma", "start_time": float, "stop_time": float, "description": string, "tags": tuple} "epoch0axon": {"source": "axon", "start_time": float, "stop_time": float, "description": string, "tags": tuple}}
second	dictionary of NWB time-series object its keys are the keys in chosenmodel.regions = {"soma": 0.0, "axon", 0.0} - the key "stimulus" is optional - value for each key is a pynwb.base.TimeSeries object, obtained using build_nwbseries method()

NOTE:

• the whole NWB TimeSeries is not passed here
• only time series that corresponds to the respective region is passed into this function
• the epoch metadata has the key "source". This represents the region from which epoch is considered this should not be confused with any attribute of the NWB file. Besides, since NWB2.0 source is no longer an attribute of NWB file.

classmethod write_nwbfile(nwbfile=None, filepath=None)

Writes the created NWB file into a HDF5 file in the given filepath.

Keyword Arguments:

Key	Value type
nwbfile	the built NWB file of type pynwb.file.NWBFile obtained using build_nwbfile method()
filepath	string; eg, “/path/to/desired/directory/”

Return value: The filename of the saved file.