equal_metadata#

metatensor.equal_metadata(tensor_1: TensorMap, tensor_2: TensorMap, check: List[str] | str = 'all') → bool[source]#

Checks if two TensorMap objects have the same metadata, returning a bool.

The equivalence of the keys of the two TensorMap objects is always checked.

If check is 'all' (the default), all metadata (i.e. the samples, components, and properties of each block) is checked to contain the same values in the same order. Passing check as a list of strings will only check the specified metadata. Allowed values to pass are 'samples', 'components', and 'properties'.

Parameters:

  • tensor_1 (TensorMap) – The first TensorMap.

  • tensor_2 (TensorMap) – The second TensorMap to compare to the first.

  • check (List[str] | str) – Which parts of the metadata to check. This can be a list containing any of 'samples', 'components', and 'properties'; or the string 'all' to check everything. Defaults to 'all'.

Returns:

True if the metadata of the two TensorMap objects are equal, False otherwise.

Return type:

bool

Examples#

>>> import numpy as np
>>> import metatensor
>>> from metatensor import Labels, TensorBlock, TensorMap
>>> tensor_1 = TensorMap(
...     keys=Labels(
...         names=["key_1", "key_2"],
...         values=np.array([[1, 0], [2, 2]]),
...     ),
...     blocks=[
...         TensorBlock(
...             values=np.full((4, 3, 1), 4.0),
...             samples=Labels(["samples"], np.array([[0], [1], [4], [5]])),
...             components=[Labels.range("components", 3)],
...             properties=Labels(["p_1", "p_2"], np.array([[0, 1]])),
...         ),
...         TensorBlock(
...             values=np.full((4, 3, 1), 4.0),
...             samples=Labels(["samples"], np.array([[0], [1], [4], [5]])),
...             components=[Labels.range("components", 3)],
...             properties=Labels(["p_1", "p_2"], np.array([[0, 1]])),
...         ),
...     ],
... )
>>> tensor_2 = TensorMap(
...     keys=Labels(
...         names=["key_1", "key_2"],
...         values=np.array([[1, 0], [2, 2]]),
...     ),
...     blocks=[
...         TensorBlock(
...             values=np.full((4, 3, 1), 4.0),
...             samples=Labels(["samples"], np.array([[0], [1], [4], [5]])),
...             components=[Labels.range("components", 3)],
...             properties=Labels(["p_3", "p_4"], np.array([[0, 1]])),
...         ),
...         TensorBlock(
...             values=np.full((4, 3, 1), 4.0),
...             samples=Labels(["samples"], np.array([[0], [1], [4], [5]])),
...             components=[Labels.range("components", 3)],
...             properties=Labels(["p_3", "p_4"], np.array([[0, 1]])),
...         ),
...     ],
... )
>>> metatensor.equal_metadata(tensor_1, tensor_2)
False
>>> metatensor.equal_metadata(
...     tensor_1,
...     tensor_2,
...     check=("samples", "components"),
... )
True
metatensor.equal_metadata_block(block_1: TensorBlock, block_2: TensorBlock, check: List[str] | str = 'all') → bool[source]#

Checks if two TensorBlock objects have the same metadata, returning a bool.

If check is 'all' (the default), all metadata (i.e. the samples, components, and properties of each block) is checked to contain the same values in the same order. Passing check as a list of strings will only check the specified metadata. Allowed values to pass are 'samples', 'components', and 'properties'.

Parameters:

  • block_1 (TensorBlock) – The first TensorBlock.

  • block_2 (TensorBlock) – The second TensorBlock to compare to the first.

  • check (List[str] | str) – Which parts of the metadata to check. This can be a list containing any of 'samples', 'components', and 'properties'; or the string 'all' to check everything. Defaults to 'all'.

Returns:

True if the metadata of the two TensorBlock objects are equal, False otherwise.

Return type:

bool

Examples#

>>> import numpy as np
>>> import metatensor
>>> from metatensor import Labels, TensorBlock
>>> block_1 = TensorBlock(
...     values=np.full((4, 3, 1), 4.0),
...     samples=Labels(["samples"], np.array([[0], [1], [4], [5]])),
...     components=[Labels.range("components", 3)],
...     properties=Labels(["p_1", "p_2"], np.array([[0, 1]])),
... )
>>> block_2 = TensorBlock(
...     values=np.full((4, 3, 1), 4.0),
...     samples=Labels(["samples"], np.array([[0], [1], [4], [5]])),
...     components=[Labels.range("components", 3)],
...     properties=Labels(["p_3", "p_4"], np.array([[0, 1]])),
... )
>>> equal_metadata_block(block_1, block_2)
False
>>> equal_metadata_block(
...     block_1,
...     block_2,
...     check=("samples", "components"),
... )
True
metatensor.equal_metadata_raise(tensor_1: TensorMap, tensor_2: TensorMap, check: List[str] | str = 'all')[source]#

Raise a NotEqualError if two TensorMap have unequal metadata.

The equivalence of the keys of the two TensorMap objects is always checked.

If check is 'all' (the default), all metadata (i.e. the samples, components, and properties of each block) is checked to contain the same values in the same order. Passing check as a list of strings will only check the specified metadata. Allowed values to pass are 'samples', 'components', and 'properties'.

Parameters:

  • tensor_1 (TensorMap) – The first TensorMap.

  • tensor_2 (TensorMap) – The second TensorMap to compare to the first.

  • check (List[str] | str) – Which parts of the metadata to check. This can be a list containing any of 'samples', 'components', and 'properties'; or the string 'all' to check everything. Defaults to 'all'.

Raises:

NotEqualError – If the metadata is not the same.

Examples#

>>> import numpy as np
>>> import metatensor
>>> from metatensor import Labels, TensorBlock, TensorMap
>>> tensor_1 = TensorMap(
...     keys=Labels(
...         names=["key_1", "key_2"],
...         values=np.array([[1, 0], [2, 2]]),
...     ),
...     blocks=[
...         TensorBlock(
...             values=np.full((4, 3, 1), 4.0),
...             samples=Labels(["samples"], np.array([[0], [1], [4], [5]])),
...             components=[Labels.range("components", 3)],
...             properties=Labels(["p_1", "p_2"], np.array([[0, 1]])),
...         ),
...         TensorBlock(
...             values=np.full((4, 3, 1), 4.0),
...             samples=Labels(["samples"], np.array([[0], [1], [4], [5]])),
...             components=[Labels.range("components", 3)],
...             properties=Labels(["p_1", "p_2"], np.array([[0, 1]])),
...         ),
...     ],
... )
>>> tensor_2 = TensorMap(
...     keys=Labels(
...         names=["key_1", "key_2"],
...         values=np.array([[1, 0], [2, 2]]),
...     ),
...     blocks=[
...         TensorBlock(
...             values=np.full((4, 3, 1), 4.0),
...             samples=Labels(["samples"], np.array([[0], [1], [4], [5]])),
...             components=[Labels.range("components", 3)],
...             properties=Labels(["p_3", "p_4"], np.array([[0, 1]])),
...         ),
...         TensorBlock(
...             values=np.full((4, 3, 1), 4.0),
...             samples=Labels(["samples"], np.array([[0], [1], [4], [5]])),
...             components=[Labels.range("components", 3)],
...             properties=Labels(["p_3", "p_4"], np.array([[0, 1]])),
...         ),
...     ],
... )
>>> metatensor.equal_metadata_raise(tensor_1, tensor_2)
Traceback (most recent call last):
    ...
metatensor.operations._utils.NotEqualError: inputs to 'equal_metadata_block_raise' should have the same properties, but they are not the same or not in the same order
>>> metatensor.equal_metadata_raise(
...     tensor_1,
...     tensor_2,
...     check=("samples", "components"),
... )
metatensor.equal_metadata_block_raise(block_1: TensorBlock, block_2: TensorBlock, check: List[str] | str = 'all')[source]#

Raise a NotEqualError if two TensorBlock have unequal metadata.

If check is 'all' (the default), all metadata (i.e. the samples, components, and properties of each block) is checked to contain the same values in the same order. Passing check as a list of strings will only check the specified metadata. Allowed values to pass are 'samples', 'components', and 'properties'.

Parameters:

  • block_1 (TensorBlock) – The first TensorBlock.

  • block_2 (TensorBlock) – The second TensorBlock to compare to the first.

  • check (List[str] | str) – A sequence of strings specifying which metadata of each block to check. If none, all metadata is checked. Allowed values are “samples”, “components”, and “properties”.

Raises:

NotEqualError – If the metadata is not the same.

Examples#

>>> import numpy as np
>>> import metatensor
>>> from metatensor import Labels, TensorBlock
>>> block_1 = TensorBlock(
...     values=np.full((4, 3, 1), 4.0),
...     samples=Labels(["samples"], np.array([[0], [1], [4], [5]])),
...     components=[Labels.range("components", 3)],
...     properties=Labels(["p_1", "p_2"], np.array([[0, 1]])),
... )
>>> block_2 = TensorBlock(
...     values=np.full((4, 3, 1), 4.0),
...     samples=Labels(["samples"], np.array([[0], [1], [4], [5]])),
...     components=[Labels.range("components", 3)],
...     properties=Labels(["p_3", "p_4"], np.array([[0, 1]])),
... )
>>> metatensor.equal_metadata_block_raise(block_1, block_2)
Traceback (most recent call last):
    ...
metatensor.operations._utils.NotEqualError: inputs to 'equal_metadata_block_raise' should have the same properties, but they are not the same or not in the same order
>>> metatensor.equal_metadata_block_raise(
...     block_1, block_2, check=("samples", "components")
... )