import collections
from itertools import chain
from functools import partial, wraps
import os.path
import warnings
[docs]class Copyable(object):
"""
Efficient copying of classes containing numpy arrays.
Interface that provides a single method for copying classes very
efficiently.
"""
[docs] def copy(self):
r"""
Generate an efficient copy of this object.
Note that Numpy arrays and other :map:`Copyable` objects on ``self``
will be deeply copied. Dictionaries and sets will be shallow copied,
and everything else will be assigned (no copy will be made).
Classes that store state other than numpy arrays and immutable types
should overwrite this method to ensure all state is copied.
Returns
-------
``type(self)``
A copy of this object
"""
new = self.__class__.__new__(self.__class__)
for k, v in self.__dict__.items():
try:
new.__dict__[k] = v.copy()
except AttributeError:
new.__dict__[k] = v
return new
[docs]class Vectorizable(Copyable):
"""
Flattening of rich objects to vectors and rebuilding them back.
Interface that provides methods for 'flattening' an object into a
vector, and restoring from the same vectorized form. Useful for
statistical analysis of objects, which commonly requires the data
to be provided as a single vector.
"""
@property
def n_parameters(self):
r"""The length of the vector that this object produces.
:type: `int`
"""
return (self.as_vector()).shape[0]
[docs] def as_vector(self, **kwargs):
"""
Returns a flattened representation of the object as a single
vector.
Returns
-------
vector : (N,) ndarray
The core representation of the object, flattened into a
single vector. Note that this is always a view back on to the
original object, but is not writable.
"""
v = self._as_vector(**kwargs)
v.flags.writeable = False
return v
def _as_vector(self, **kwargs):
"""
Returns a flattened representation of the object as a single
vector.
Returns
-------
vector : ``(n_parameters,)`` `ndarray`
The core representation of the object, flattened into a
single vector.
"""
raise NotImplementedError()
[docs] def from_vector_inplace(self, vector):
"""
Deprecated. Use the non-mutating API, :map:`from_vector`.
For internal usage in performance-sensitive spots,
see `_from_vector_inplace()`
Parameters
----------
vector : ``(n_parameters,)`` `ndarray`
Flattened representation of this object
"""
warnings.warn('the public API for inplace operations is deprecated '
'and will be removed in a future version of Menpo. '
'Use .from_vector() instead.', MenpoDeprecationWarning)
return self._from_vector_inplace(vector)
def _from_vector_inplace(self, vector):
"""
Update the state of this object from a vector form.
Parameters
----------
vector : ``(n_parameters,)`` `ndarray`
Flattened representation of this object
"""
raise NotImplementedError()
[docs] def from_vector(self, vector):
"""
Build a new instance of the object from it's vectorized state.
``self`` is used to fill out the missing state required to
rebuild a full object from it's standardized flattened state. This
is the default implementation, which is which is a ``deepcopy`` of the
object followed by a call to :meth:`from_vector_inplace()`. This method
can be overridden for a performance benefit if desired.
Parameters
----------
vector : ``(n_parameters,)`` `ndarray`
Flattened representation of the object.
Returns
-------
object : ``type(self)``
An new instance of this class.
"""
new = self.copy()
new._from_vector_inplace(vector)
return new
[docs] def has_nan_values(self):
"""
Tests if the vectorized form of the object contains ``nan`` values or
not. This is particularly useful for objects with unknown values that
have been mapped to ``nan`` values.
Returns
-------
has_nan_values : `bool`
If the vectorized object contains ``nan`` values.
"""
import numpy as np
return np.any(np.isnan(self.as_vector()))
[docs]class Targetable(Copyable):
"""Interface for objects that can produce a target :map:`PointCloud`.
This could for instance be the result of an alignment or a generation of a
:map:`PointCloud` instance from a shape model.
Implementations must define sensible behavior for:
- what a target is: see :attr:`target`
- how to set a target: see :meth:`set_target`
- how to update the object after a target is set:
see :meth:`_sync_state_from_target`
- how to produce a new target after the changes:
see :meth:`_new_target_from_state`
Note that :meth:`_sync_target_from_state` needs to be triggered as
appropriate by subclasses e.g. when :map:`from_vector_inplace` is
called. This will in turn trigger :meth:`_new_target_from_state`, which each
subclass must implement.
"""
@property
def n_dims(self):
r"""The number of dimensions of the :attr:`target`.
:type: `int`
"""
return self.target.n_dims
@property
def n_points(self):
r"""The number of points on the :attr:`target`.
:type: `int`
"""
return self.target.n_points
@property
def target(self):
r"""The current :map:`PointCloud` that this object produces.
:type: :map:`PointCloud`
"""
raise NotImplementedError()
[docs] def set_target(self, new_target):
r"""
Update this object so that it attempts to recreate the ``new_target``.
Parameters
----------
new_target : :map:`PointCloud`
The new target that this object should try and regenerate.
"""
self._target_setter_with_verification(new_target) # trigger the update
self._sync_state_from_target() # and a sync
def _target_setter_with_verification(self, new_target):
r"""Updates the target, checking it is sensible, without triggering a
sync.
Should be called by :meth:`_sync_target_from_state` once it has
generated a suitable target representation.
Parameters
----------
new_target : :map:`PointCloud`
The new target that should be set.
"""
self._verify_target(new_target)
self._target_setter(new_target)
def _verify_target(self, new_target):
r"""Performs sanity checks to ensure that the new target is valid.
This includes checking the dimensionality matches and the number of
points matches the current target's values.
Parameters
----------
new_target : :map:`PointCloud`
The target that needs to be verified.
Raises
------
ValueError
If the ``new_target`` has differing ``n_points`` or ``n_dims`` to
``self``.
"""
# If the target is None (i.e. on construction) then dodge the
# verification
if self.target is None:
return
if new_target.n_dims != self.target.n_dims:
raise ValueError(
"The current target is {}D, the new target is {}D - new "
"target has to have the same dimensionality as the "
"old".format(self.target.n_dims, new_target.n_dims))
elif new_target.n_points != self.target.n_points:
raise ValueError(
"The current target has {} points, the new target has {} "
"- new target has to have the same number of points as the"
" old".format(self.target.n_points, new_target.n_points))
def _target_setter(self, new_target):
r"""Sets the target to the new value.
Does no synchronization. Note that it is advisable that
:meth:`_target_setter_with_verification` is called from
subclasses instead of this.
Parameters
----------
new_target : :map:`PointCloud`
The new target that will be set.
"""
raise NotImplementedError()
def _sync_target_from_state(self):
new_target = self._new_target_from_state()
self._target_setter_with_verification(new_target)
def _new_target_from_state(self):
r"""Generate a new target that is correct after changes to the object.
Returns
-------
object : ``type(self)``
"""
raise NotImplementedError()
def _sync_state_from_target(self):
r"""Synchronizes the object state to be correct after changes to the
target.
Called automatically from the target setter. This is called after the
target is updated - only handle synchronization here.
"""
raise NotImplementedError()
[docs]def menpo_src_dir_path():
r"""The path to the top of the menpo Python package.
Useful for locating where the data folder is stored.
Returns
-------
path : ``pathlib.Path``
The full path to the top of the Menpo package
"""
from pathlib import Path # to avoid cluttering the menpo.base namespace
return Path(os.path.abspath(__file__)).parent
[docs]class MenpoDeprecationWarning(Warning):
r"""
A warning that functionality in Menpo will be deprecated in a future major
release.
"""
pass
class MenpoMissingDependencyError(Exception):
r"""
An exception that a dependency required for the requested functionality
was not detected.
"""
def __init__(self, package_name):
super(MenpoMissingDependencyError, self).__init__()
self.message = "You need to install the '{pname}' package in order " \
"to use this functionality. We recommend that you " \
"use conda to achieve this - try the command " \
"'conda install -c menpo {pname}' " \
"in your terminal.".format(pname=package_name)
def __str__(self):
return self.message
[docs]def name_of_callable(c):
r"""
Return the name of a callable (function or callable class) as a string.
Recurses on partial function to attempt to find the wrapped
methods actual name.
Parameters
----------
c : `callable`
A callable class or function, or any valid Python object that can
be wrapped with partial.
Returns
-------
name : `str`
The name of the passed object.
"""
try:
if isinstance(c, partial): # partial
# Recursively call as partial may be wrapping either a callable
# or a function (or another partial for some reason!)
return name_of_callable(c.func)
else:
return c.__name__ # function
except AttributeError:
return c.__class__.__name__ # callable class
class doc_inherit(object):
"""
Docstring inheriting method descriptor.
This uses some Python magic in order to create a decorator that implements
the descriptor protocol that allows functions to inherit documentation.
This is particularly useful for methods that directly override methods
on their base class and simply alter the implementation but not the
effective behaviour. Usage of this decorator is as follows:
@doc_inherit()
def foo():
# Do something, but inherit the documentation from the method
# called 'foo' found on the super() chain.
@doc_inherit(name="foo2")
def foo():
# Do something, but inherit the documentation from the method
# called 'foo2' found on the super() chain.
When no argument is passed the name of the method being decorated is
looked up on the ``super`` call chain.
Parameters
----------
name : `str`
The name of the method to copy documentation from that exists somewhere
on the ``super`` inheritance hierarchy.
"""
def __init__(self, name=None):
self.name = name
def __call__(self, mthd):
# Implementing the call method on a decorator allows the decorator
# to recieve arguments in the constructor (__init__). Therefore,
# the argument to the call method is always the method being wrapped.
self.mthd = mthd
# If name is None then default to the name of the method being wrapped.
if self.name is None:
self.name = self.mthd.__name__
return self
def __get__(self, obj, cls):
# Implement the descriptor protocol. There are two different calling
# strategies that involve whether the wrapped method has been passed
# an instance or not.
if obj:
return self._get_with_instance(obj, cls)
else:
return self._get_with_no_instance(cls)
def _get_with_instance(self, obj, cls):
# An instance was passed, so lookup the name on the super chain
overridden = getattr(super(cls, obj), self.name, None)
# Return the wrapped method, passing through the arguments and the
# object instance.
@wraps(self.mthd, assigned=('__name__', '__module__'))
def f(*args, **kwargs):
return self.mthd(obj, *args, **kwargs)
return self._use_parent_doc(f, overridden)
def _get_with_no_instance(self, cls):
# This case is more complicated (than when an instance is passed). Here
# we use reflection to try and lookup the method. When found, we drop
# out the loop.
for parent in cls.__mro__[1:]:
overridden = getattr(parent, self.name, None)
if overridden:
break
# Return the wrapped method, passing through the arguments and the
# object instance.
@wraps(self.mthd, assigned=('__name__', '__module__'))
def f(*args, **kwargs):
return self.mthd(*args, **kwargs)
return self._use_parent_doc(f, overridden)
def _use_parent_doc(self, func, source):
# Attach the documentation (unless the method was not found on the
# super chain).
if source is None:
raise NameError("Can't find '{}' in parents".format(self.name))
func.__doc__ = source.__doc__
return func
[docs]class LazyList(collections.Sequence, Copyable):
r"""
An immutable sequence that provides the ability to lazily access objects.
In truth, this sequence simply wraps a list of callables which are then
indexed and invoked. However, if the callable represents a function that
lazily access memory, then this list simply implements a lazy list
paradigm.
When slicing, another `LazyList` is returned, containing the subset
of callables.
Parameters
----------
callables : list of `callable`
A list of `callable` objects that will be invoked if directly indexed.
"""
def __init__(self, callables):
self._callables = callables
def __getitem__(self, slice_):
if isinstance(slice_, int) or hasattr(slice_, '__index__'):
# PEP 357 and single integer index access - returns element
return self._callables[slice_]()
elif isinstance(slice_, collections.Iterable):
# An iterable object is passed - return a new LazyList
return LazyList([self._callables[s] for s in slice_])
else:
# A slice or unknown type is passed - let List handle it
return LazyList(self._callables[slice_])
def __len__(self):
return len(self._callables)
@classmethod
[docs] def init_from_iterable(cls, iterable, f=None):
r"""
Create a lazy list from an existing iterable (think Python `list`) and
optionally a `callable` that expects a single parameter which will be
applied to each element of the list. This allows for simply
creating a `LazyList` from an existing list and if no `callable` is
provided the identity function is assumed.
Parameters
----------
iterable : `collections.Iterable`
An iterable object such as a `list`.
f : `callable`, optional
Callable expecting a single parameter.
Returns
-------
lazy : `LazyList`
A LazyList where each element returns each item of the provided
iterable, optionally with `f` applied to it.
"""
if f is None:
# The identity function
def f(i):
return i
return cls([partial(f, x) for x in iterable])
@classmethod
[docs] def init_from_index_callable(cls, f, n_elements):
r"""
Create a lazy list from a `callable` that expects a single parameter,
the index into an underlying sequence. This allows for simply
creating a `LazyList` from a `callable` that likely wraps
another list in a closure.
Parameters
----------
f : `callable`
Callable expecting a single integer parameter, index. This is an
index into (presumably) an underlying sequence.
n_elements : `int`
The number of elements in the underlying sequence.
Returns
-------
lazy : `LazyList`
A LazyList where each element returns the underlying indexable
object wrapped by ``f``.
"""
return cls([partial(f, i) for i in range(n_elements)])
[docs] def map(self, f):
r"""
Create a new LazyList where the passed callable ``f`` wraps
each element.
``f`` should take a single parameter, ``x``, that is the result
of the underlying callable - it must also return a value. Note that
mapping is lazy and thus calling this function should return
immediately.
Alternatively, ``f`` may be a list of `callable`, one per entry
in the underlying list, with the same specification as above.
Parameters
----------
f : `callable` or `iterable` of `callable`
Callable to wrap each element with. If an iterable of callables
(think list) is passed then it **must** by the same length as
this LazyList.
Returns
-------
lazy : `LazyList`
A new LazyList where each element is wrapped by (each) ``f``.
"""
# We need this delayed helper function in order to ensure that f
# is passed the actual instantiated object and not the callable itself.
def delayed(delay_f, delay_x):
return delay_f(delay_x())
if isinstance(f, collections.Iterable) and callable(f):
raise ValueError('It is ambiguous whether the provided argument '
'is an iterable object or a callable.')
new = self.copy()
if isinstance(f, collections.Iterable):
if len(f) != len(new):
raise ValueError('A callable per element of the LazyList must '
'be passed.')
new._callables = [partial(delayed, one_f, x)
for one_f, x in zip(f, new._callables)]
else:
new._callables = [partial(delayed, f, x) for x in new._callables]
return new
[docs] def repeat(self, n):
r"""
Repeat each item of the underlying LazyList ``n`` times. Therefore,
if a list currently has ``D`` items, the returned list will contain
``D * n`` items and will return immediately (method is lazy).
Parameters
----------
n : `int`
The number of times to repeat each item.
Returns
-------
lazy : `LazyList`
A LazyList where each element returns each item of the provided
iterable, optionally with `f` applied to it.
Examples
--------
>>> from menpo.base import LazyList
>>> ll = LazyList.init_from_list([0, 1])
>>> repeated_ll = ll.repeat(2) # Returns immediately
>>> items = list(repeated_ll) # [0, 0, 1, 1]
"""
new = self.copy()
new._callables = list(chain(*zip(*[new._callables] * n)))
return new
[docs] def copy(self):
r"""
Generate an efficient copy of this LazyList - copying the underlying
callables will be lazy and shallow (each callable will **not** be
called nor copied) but they will reside within in a new `list`.
Returns
-------
``type(self)``
A copy of this LazyList.
"""
new = Copyable.copy(self)
new._callables = list(self._callables)
return new
def __add__(self, other):
r"""
Create a new LazyList from this list and the given list. The passed list
items will be concatenated to the end of this list to give a new
LazyList that contains the concatenation of the two lists.
If a Python list is passed then the elements are wrapped in a function
that just returns their values to maintain the callable nature of
LazyList elements.
Parameters
----------
other : `collections.Sequence`
Sequence to concatenate with this list.
Returns
-------
lazy : `LazyList`
A new LazyList formed of the concatenation of this list and
the ``other`` list.
"""
new = self.copy()
# If the passed Sequence was not lazy then fake it being lazy by
# wrapping it in a function that just returns the value.
if not isinstance(other, LazyList):
new_callables = LazyList.init_from_iterable(other)._callables
else:
new_callables = other._callables
new._callables = new._callables + new_callables
return new
def partial_doc(func, *args, **kwargs):
r"""
Return a partial function but the __doc__ attached to the returned
partial. Note that no effort is made to correct the docstring for
any parameters that are covered by the partial.
Parameters
----------
func : `callable`
The func to partial and whose docs should be copied.
args : ...
Any arguments to partial.
kwargs : `dict`
Any keyword arguments to partial.
Returns
-------
p_func : `callable`
The partially wrapped func with __doc__ attached.
"""
p = partial(func, *args, **kwargs)
p.__doc__ = func.__doc__
return p