Reference#

IO#

Elastica IO Pipelines for data deserialization.

elastica_pipelines.io.series(*, file_pattern=None, metadata=None, transforms=None)#

Make a Series from pattern or metadata file.

Parameters:

file_pattern (str, Optional) – Naming pattern of time-series files.
metadata (str | Path | None) – Metadata file.
transforms (Callable, Optional) – A function/transform that takes in an array data-structure and returns a transformed version. E.g, transforms.ToArray

Returns:

Series object with temporal system evolution.

Return type:

Series

Example

>>> from elastica_pipelines.io import series
>>> metadata_fn = "tests/io/data/elastica_metadata.h5"
>>> for t, snapshot in series(metadata=metadata_fn).iterations():
>>>     print("Iteration: {0} at time {1}".format(t.iterate, t.time))

Raises:

RuntimeError – If none or both pattern and metadata is simultaneously specified.
NotImplementedError – For pattern-based iteration.

Parameters:

file_pattern (str | None) –
metadata (str | Path | None) –
transforms (Callable[[...], Any] | None) –

Return type:

Series

Temporal#

Temporal IO types.

class elastica_pipelines.io.temporal.Series(node, transforms=None)#

Temporally evolving data-series.

Has the same interface as a dictionary, and hence lookup, items() work.

Parameters:

node (Node) – Node with series information.
transforms (Callable, Optional) – A function/transform that takes in an array data-structure and returns a transformed version. E.g, transforms.ToArray

Example

>>> from elastica_pipelines.io import series
>>> from elastica_pipelines.io.temporal import Series
>>>
>>> # ``series`` returns a ``Series`` object
>>> metadata_filename = "tests/io/data/elastica_metadata.h5"
>>> s : Series = series(metadata=metadata_filename)

>>> s[50] # lookup the data at iteration 50, iterates may not be continuous!

>>> for t in s.keys(): # obtain all iteration values
>>>     print(t.iterate, t.time, t.dt)
>>>     print(s[t])

iterations()#

Obtain temporal iterations.

Returns:: Temporal iteration
Return type:: ItemsView[int | SeriesKey, Snapshot]

temporal_select(indices)#

Obtain temporal evolution for a select subset of systems.

Parameters:: indices (SystemIndices) – indices (with Traits) for system selection.
Returns:: SeriesSelection with the same interface.
Return type:: SeriesSelection

Example

>>> from elastica_pipelines.io import series
>>> from elastica_pipelines.io.temporal import Series
>>> from elastica_pipelines.io import CosseratRodRecordIndex
>>>
>>> metadata_filename = "tests/io/data/elastica_metadata.h5"
>>> s : Series = series(metadata=metadata_filename)
>>>
>>> # Select Cosserat rod with ID 1
>>> subset = s.temporal_select(CosseratRodRecordIndex(1))
>>> # Select Cosserat rod with list of indices
>>> subset = s.temporal_select(CosseratRodRecordIndex([1, 3]))
>>> # Select Cosserat rod range with indices
>>> subset = s.temporal_select(CosseratRodRecordIndex(slice(1,5,2)))
>>>
>>> # Subset has same interface as Series, see SeriesSelection
>>> for t in subset.keys(): # obtain all iteration values
>>>     print(t.iterate, t.time, t.dt)
>>>     print(subset[t])

class elastica_pipelines.io.temporal.SeriesKey(iterate, time, dt)#

Key for a temporal series.

Parameters:

iterate (int) – Unique iteration value
time (float) – Time for current iteration
dt (float) – Timestep for current iteration

class elastica_pipelines.io.temporal.SeriesKeys#: Union of integer and SeriesKey

class elastica_pipelines.io.temporal.SeriesSelection(parent, indices)#

Temporally evolving data-series restricted to a subset of systems.

Parameters:

parent (Series) – Series from which the selection is made.
indices (SystemIndices) – Selection of index subsets

Example

>>> from elastica_pipelines.io import series
>>> from elastica_pipelines.io.temporal import Series, SeriesSelection
>>> from elastica_pipelines.io import CosseratRodRecordIndex as RodIndex
>>>
>>> metadata_filename = "tests/io/data/elastica_metadata.h5"
>>> s : Series = series(metadata=metadata_filename)
>>>
>>> # Select Cosserat rod with list of indices
>>> subset : SeriesSelection = s.temporal_select(RodIndex([1, 4, 9]))
>>>
>>> # Subset has same interface as Series
>>> subset[50] # lookup the data at iteration 50
>>>
>>> for t in subset.keys(): # obtain all iteration values for subsets
>>>     print(t.iterate, t.time, t.dt)
>>>     print(subset[t])

iterations()#

Obtain temporal iterations.

Returns:: Temporal iteration
Return type:: ItemsView[SeriesKeys, RecordLeafs]

temporal_select(indices)#

Obtain temporal evolution for a select subset of systems.

Parameters:: indices (SystemIndices) – indices (with Traits) for system selection.
Returns:: SeriesSelection with the same interface.
Raises:: TypeError – If index is out of bounds.
Return type:: SeriesSelection

Example

>>> from elastica_pipelines.io import series
>>> from elastica_pipelines.io.temporal import Series, SeriesSelection
>>> from elastica_pipelines.io import CosseratRodRecordIndex as RodIndex
>>>
>>> metadata_filename = "tests/io/data/elastica_metadata.h5"
>>> ser : Series = series(metadata=metadata_filename)
>>>
>>> s : SeriesSelection = ser.temporal_select(RodIndex([1, 2, 3]))
>>>
>>> # Select Cosserat rod with list of indices
>>> subset : SeriesSelection = s.temporal_select(RodIndex([1, 2]))
>>>
>>> # Subset has same interface as Series
>>> subset[50] # lookup the data at iteration 50
>>>
>>> for t in subset.keys(): # obtain all iteration values for subsets
>>>     print(t.iterate, t.time, t.dt)
>>>     print(subset[t])

Note

The subset of indices requested must be contained within the bounds alredy selected, else an error is thrown.

class elastica_pipelines.io.temporal.Snapshot(node, transforms=None)#

Data access for a single snapshot of an Elastica++ simulation.

Parameters:

node (Node) – Node in which to lookup the information at one time iterate.
transforms (Callable, Optional) – A function/transform that takes in an array data-structure and returns a transformed version. E.g, transforms.ToArray

Example

>>> from elastica_pipelines.io import series
>>> from elastica_pipelines.io.temporal import Snapshot
>>>
>>> # ``series`` returns a ``Series`` object
>>> metadata_filename = "tests/io/data/elastica_metadata.h5"
>>> s = series(metadata=metadata_filename)

>>> snap : Snapshot = s[50] # lookup the data at iteration 50

>>> cosserat_rod_records = snap["CosseratRod"] # dict-like interface
>>> for system_type in snap.keys(): # obtain all system types
>>>     print(system_type, snap[system_type])

spheres()#: Access sphere records

cosserat_rods()#: Access cosserat rod records

systems()#

Access all system records.

Returns:: Records across all systems.
Return type:: ChainMap[RecordsAdapterKey, RecordLeafs]

Example

>>> from elastica_pipelines.io import series
>>> from elastica_pipelines.io.temporal import Snapshot
>>>
>>> # ``series`` returns a ``Series`` object
>>> metadata_filename = "tests/io/data/elastica_metadata.h5"
>>> s = series(metadata=metadata_filename)
>>> snap : Snapshot = s[50] # lookup the data at iteration 50

>>> # Iterate over all systems regardless of the type
>>> for sys_id, system in snap.systems().items():
>>>     print(sys_id, system)

class elastica_pipelines.io.temporal.RecordsAdapterKey(sys_type, sys_id)#

Key containing both the type and id of the system, for iteration.

Parameters:

sys_type (str) – Type of the system.
sys_id (int) – Unique ID of the system.

Transforms#

Transformations to apply when reading/writing Elastica IO.

class elastica_pipelines.io.transforms.Compose(transforms)#

Composes several transforms together.

Parameters:: transforms (list of Transform objects) – list of transforms to compose.

Example

>>> from elastica_pipelines.io.transforms import Compose, ToArray
>>> Compose([
>>>     ToArray(),
>>> ])

class elastica_pipelines.io.transforms.ToArray#

Convert a HDF5 dataset or numpy.ndarray to numpy.ndarray.

Example

>>> from elastica_pipelines.io.transforms import ToArray
>>> ToArray()([1, 2, 3, 4]).shape

Protocols#

Elastica IO protocols.

class elastica_pipelines.io.protocols.SystemIndices(indices)#

Protocol for index into data-records.

Parameters:: indices (int | slice | List[int]) –

Typing#

Elastica IO typing.

class elastica_pipelines.io.typing.Node#: Mapping from a string to any.

class elastica_pipelines.io.typing.RecordLeafs#: Record objects that can be queried for physical data.