hyperspy.misc.utils module

class hyperspy.misc.utils.DictionaryTreeBrowser(dictionary=None, double_lines=False, lazy=True)

Bases: object

A class to comfortably browse a dictionary using a CLI.

In addition to accessing the values using dictionary syntax the class enables navigating a dictionary that constains nested dictionaries as attribures of nested classes. Also it is an iterator over the (key, value) items. The __repr__ method provides pretty tree printing. Private keys, i.e. keys that starts with an underscore, are not printed, counted when calling len nor iterated.

export : saves the dictionary in pretty tree printing format in a text


keys : returns a list of non-private keys.
as_dictionary : returns a dictionary representation of the object.
set_item : easily set items, creating any necessary node on the way.
add_node : adds a node.


>>> tree = DictionaryTreeBrowser()
>>> tree.set_item("Branch.Leaf1.color", "green")
>>> tree.set_item("Branch.Leaf2.color", "brown")
>>> tree.set_item("Branch.Leaf2.caterpillar", True)
>>> tree.set_item("Branch.Leaf1.caterpillar", False)
>>> tree
└── Branch
    ├── Leaf1
    │   ├── caterpillar = False
    │   └── color = green
    └── Leaf2
        ├── caterpillar = True
        └── color = brown
>>> tree.Branch
├── Leaf1
│   ├── caterpillar = False
│   └── color = green
└── Leaf2
    ├── caterpillar = True
    └── color = brown
>>> for label, leaf in tree.Branch:
...     print("%s is %s" % (label, leaf.color))
Leaf1 is green
Leaf2 is brown
>>> tree.Branch.Leaf2.caterpillar
>>> "Leaf1" in tree.Branch
>>> "Leaf3" in tree.Branch

When creating a DictionaryTreeBrowser lazily, the dictionary is added to the _lazy_attributes attribute. The first time a lazy attribute is called or the DictionaryTreeBrowser is printed, the DictionaryTreeBrowser processes the lazy attributes with the process_lazy_attributes method. DictionaryTreeBrowser is lazy by default, using non-lazy instances can be useful for debugging purposes.

_get_html_print_items(padding='', max_len=78, recursive_level=0)

Recursive method that creates a html string for fancy display of metadata.

_get_print_items(padding='', max_len=78)

Prints only the attributes that are not methods

_process_dictionary(dictionary, double_lines)

Process the provided dictionary to set the attributes

add_dictionary(dictionary, double_lines=False)

Add new items from dictionary.


Adds all the nodes in the given path if they don’t exist.


node_path (str) – The nodes must be separated by full stops (periods).


>>> dict_browser = DictionaryTreeBrowser({})
>>> dict_browser.add_node('First.Second')
>>> dict_browser.First.Second = 3
>>> dict_browser
└── First
    └── Second = 3

Returns its dictionary representation.

export(filename, encoding='utf8')

Export the dictionary to a text file

  • filename (str) – The name of the file without the extension that is txt by default

  • encoding (valid encoding str) –

get_item(item_path, default=None)

Given a path, return it’s value if it exists, or default value if missing.

The nodes of the path are separated using periods.

  • item_path (Str) – A string describing the path with each item separated by full stops (periods)

  • default – The value to return if the path does not exist.


>>> dict = {'To' : {'be' : True}}
>>> dict_browser = DictionaryTreeBrowser(dict)
>>> dict_browser.get_item('To')
└── be = True
>>> dict_browser.get_item('To.be')
>>> dict_browser.get_item('To.be.or', 'default_value')

Given a path, return True if it exists.

The nodes of the path are separated using periods.


item_path (Str) – A string describing the path with each item separated by full stops (periods)


>>> dict = {'To' : {'be' : True}}
>>> dict_browser = DictionaryTreeBrowser(dict)
>>> dict_browser.has_item('To')
>>> dict_browser.has_item('To.be')
>>> dict_browser.has_item('To.be.or')

Returns a list of non-private keys.


Run the DictionaryTreeBrowser machinery for the lazy attributes.

set_item(item_path, value)

Given the path and value, create the missing nodes in the path and assign to the last one the value


item_path (Str) – A string describing the path with each item separated by a full stops (periods)


>>> dict_browser = DictionaryTreeBrowser({})
>>> dict_browser.set_item('First.Second.Third', 3)
>>> dict_browser
└── First
   └── Second
        └── Third = 3
hyperspy.misc.utils.add_key_value(key, value)

Returns the metadata value as a html string

hyperspy.misc.utils.attrsetter(target, attrs, value)

Sets attribute of the target to specified value, supports nested attributes. Only creates a new attribute if the object supports such behaviour (e.g. DictionaryTreeBrowser does)

  • target (object) –

  • attrs (string) – attributes, separated by periods (e.g. ‘metadata.Signal.Noise_parameters.variance’ )

  • value (object) –


First create a signal and model pair:

>>> s = hs.signals.Signal1D(np.arange(10))
>>> m = s.create_model()
>>> m.signal.data
array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])

Now set the data of the model with attrsetter >>> attrsetter(m, ‘signal1D.data’, np.arange(10)+2) >>> self.signal.data array([2, 3, 4, 5, 6, 7, 8, 9, 10, 10])

The behaviour is identical to >>> self.signal.data = np.arange(10) + 2

hyperspy.misc.utils.check_long_string(value, max_len)

Checks whether string is too long for printing in html metadata

hyperspy.misc.utils.create_map_objects(function, nav_size, iterating_kwargs, **kwargs)

To be used in _map_iterate of BaseSignal and LazySignal.

Moved to a separate method to reduce code duplication.

hyperspy.misc.utils.find_subclasses(mod, cls)

Find all the subclasses in a module.

  • mod (module) –

  • cls (class) –


Return type

dictonary in which key, item = subclass name, subclass

hyperspy.misc.utils.fsdict(nodes, value, dictionary)

Populates the dictionary ‘dic’ in a file system-like fashion creating a dictionary of dictionaries from the items present in the list ‘nodes’ and assigning the value ‘value’ to the innermost dictionary.

‘dic’ will be of the type: dic[‘node1’][‘node2’][‘node3’]…[‘nodeN’] = value where each node is like a directory that contains other directories (nodes) or files (values)

hyperspy.misc.utils.generate_axis(origin, step, N, index=0)

Creates an axis given the origin, step and number of channels

Alternatively, the index of the origin channel can be specified.

  • origin (float) –

  • step (float) –

  • N (number of channels) –

  • index (int) – index of origin


Return type

Numpy array


Get info about object package


dic – Dictionary containing package and package_version (if available)

Return type


hyperspy.misc.utils.guess_output_signal_size(test_signal, function, ragged, **kwargs)

This function is for guessing the output signal shape and size. It will attempt to apply the function to some test signal and then output the resulting signal shape and datatype.

Parameters test_signal: BaseSignal

A test signal for the function to be applied to. A signal with 0 navigation dimensions

function: function

The function to be applied to the dataset

ragged: bool

If the data is ragged then the output signal size is () and the data type is ‘object’

**kwargs: dict

Any other keyword arguments passed to the function.


Check if an object is a Hyperspy Signal


input_object (object) – Object to be tests


If true the object is a subclass of hyperspy.signal.BaseSignal

Return type



Return product of sequence of numbers.

Equivalent of functools.reduce(operator.mul, iterable, 1).

>>> product([2**8, 2**30])
>>> product([])
hyperspy.misc.utils.nested_dictionary_merge(dict1, dict2)

Merge dict2 into dict1 recursively


Converts zero or a postive integer (or their string representations) to an ordinal value.

>>> for i in range(1,13):
...     ordinal(i)
>>> for i in (100, '111', '112',1011):
...     ordinal(i)


Author: Serdar Tumgoren http://code.activestate.com/recipes/576888-format-a-number-as-an-ordinal/ MIT license

hyperspy.misc.utils.parse_quantity(quantity, opening='(', closing=')')

Parse quantity of the signal outputting quantity and units separately. It looks for the last matching opening and closing separator.

  • quantity (string) –

  • opening (string) – Separator used to define the beginning of the units

  • closing (string) – Separator used to define the end of the units


  • quantity_name (string)

  • quantity_units (string)

hyperspy.misc.utils.print_html(f_text, f_html)

Print html version when in Jupyter Notebook

hyperspy.misc.utils.process_function_blockwise(data, *args, function, nav_indexes=None, output_signal_size=None, block_info=None, arg_keys=None, **kwargs)

Function for processing the function blockwise… This gets passed to map_blocks so that the function only gets applied to the signal axes.

data: np.array

The data for one chunk

*args: tuple

Any signal the is iterated alongside the data in. In the form ((key1, value1),(key2,value2)…)

function: function

The function to applied to the signal axis

nav_indexes: tuple

The indexes of the navigation axes for the dataset.

output_signal_shape: tuple

The shape of the output signal. For a ragged signal this is equal to 1.

block_info: dict

The block info as described by the dask.array.map_blocks function

arg_keys: tuple

The list of keys for the passed arguments (args). Together this makes a set of key:value pairs to be passed to the function.

**kwargs: dict

Any additional key value pairs to be used by the function (Note that these are the constants that are applied.)


Escapes any &, < and > tags that would become invisible when printing html

hyperspy.misc.utils.rollelem(a, index, to_index=0)

Roll the specified axis backwards, until it lies in a given position.

  • a (list) – Input list.

  • index (int) – The index of the item to roll backwards. The positions of the items do not change relative to one another.

  • to_index (int, optional) – The item is rolled until it lies before this position. The default, 0, results in a “complete” roll.


res – Output list.

Return type


hyperspy.misc.utils.slugify(value, valid_variable_name=False)

Normalizes string, converts to lowercase, removes non-alpha characters, and converts spaces to hyphens.

Adapted from Django’s “django/template/defaultfilters.py”.

hyperspy.misc.utils.stack(signal_list, axis=None, new_axis_name='stack_element', lazy=None, stack_metadata=True, show_progressbar=None, **kwargs)

Concatenate the signals in the list over a given axis or a new axis.

The title is set to that of the first signal in the list.

  • signal_list (list of BaseSignal instances) – List of signals to stack.

  • axis ({None, int, str}) – If None, the signals are stacked over a new axis. The data must have the same dimensions. Otherwise the signals are stacked over the axis given by its integer index or its name. The data must have the same shape, except in the dimension corresponding to axis.

  • new_axis_name (str) – The name of the new axis when axis is None. If an axis with this name already exists it automatically append ‘-i’, where i are integers, until it finds a name that is not yet in use.

  • lazy ({bool, None}) – Returns a LazySignal if True. If None, only returns lazy result if at least one is lazy.

  • stack_metadata ({bool, int}) – If integer, this value defines the index of the signal in the signal list, from which the metadata and original_metadata are taken. If True, the original_metadata and metadata of each signals are stacked and saved in original_metadata.stack_elements of the returned signal. In this case, the metadata are copied from the first signal in the list. If False, the metadata and original_metadata are not copied.

  • show_progressbar (None or bool) – If True, display a progress bar. If None, the default from the preferences settings is used.



Return type

BaseSignal instance


>>> data = np.arange(20)
>>> s = hs.stack([hs.signals.Signal1D(data[:10]),
...               hs.signals.Signal1D(data[10:])])
>>> s
<Signal1D, title: Stack of , dimensions: (2, 10)>
>>> s.data
array([[ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9],
       [10, 11, 12, 13, 14, 15, 16, 17, 18, 19]])
hyperspy.misc.utils.str2num(string, **kargs)

Transform a a table in string form into a numpy array


string (string) –


Return type

numpy array

hyperspy.misc.utils.swapelem(obj, i, j)

Swaps element having index i with element having index j in object obj IN PLACE.


>>> L = ['a', 'b', 'c']
>>> spwapelem(L, 1, 2)
>>> print(L)
['a', 'c', 'b']
hyperspy.misc.utils.transpose(*args, signal_axes=None, navigation_axes=None, optimize=False)

Transposes all passed signals according to the specified options.

For parameters see BaseSignal.transpose.


>>> signal_iterable = [hs.signals.BaseSignal(np.random.random((2,)*(i+1)))
                       for i in range(3)]
>>> signal_iterable
[<BaseSignal, title: , dimensions: (|2)>,
 <BaseSignal, title: , dimensions: (|2, 2)>,
 <BaseSignal, title: , dimensions: (|2, 2, 2)>]
>>> hs.transpose(*signal_iterable, signal_axes=1)
[<BaseSignal, title: , dimensions: (|2)>,
 <BaseSignal, title: , dimensions: (2|2)>,
 <BaseSignal, title: , dimensions: (2, 2|2)>]
>>> hs.transpose(signal1, signal2, signal3, signal_axes=["Energy"])
hyperspy.misc.utils.underline(line, character='-')

Return the line underlined.