Skip to content

Commit

Permalink
changelog
Browse files Browse the repository at this point in the history
  • Loading branch information
@isaacdevlugt
isaacdevlugt committed Jan 9, 2025
1 parent dbadfba commit 2f67f7f
Showing 1 changed file with 93 additions and 62 deletions.
155 changes: 93 additions & 62 deletions doc/releases/changelog-0.40.0.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -168,94 +168,118 @@ qubit operators.
[(#6573)](https://github.com/PennyLaneAI/pennylane/pull/6573)
[(#6611)](https://github.com/PennyLaneAI/pennylane/pull/6611)

* Added `qml.qchem.symmetry_shift` function to perform the
[block-invariant symmetry shift](https://arxiv.org/pdf/2304.13772) on the electronic integrals.
* A new function for performing the [block-invariant symmetry shift](https://arxiv.org/pdf/2304.13772)
on electronic integrals has been added with `qml.qchem.symmetry_shift`.
[(#6574)](https://github.com/PennyLaneAI/pennylane/pull/6574)

* Added JAX support for the differentiable Hartree-Fock workflow.
* The differentiable Hartree-Fock workflow is now compatible with JAX.
[(#6096)](https://github.com/PennyLaneAI/pennylane/pull/6096)
[(#6707)](https://github.com/PennyLaneAI/pennylane/pull/6707)

<h4>Transform for combining GlobalPhase instances</h4>

* Added a new `qml.transforms.combine_global_phases` transform to combine all `qml.GlobalPhase` gates in a circuit into a single one applied at the end.
This can be useful for circuits that include a lot of `qml.GlobalPhase` gates, which can be introduced directly during circuit creation,
decompositions that include `qml.GlobalPhase` gates, etc.
* A new transform called `qml.transforms.combine_global_phases` has been added. It combines all
`qml.GlobalPhase` gates in a circuit into a single one applied at the end. This can be useful for
circuits that include a lot of `qml.GlobalPhase` gates that are introduced directly during
circuit creation, decompositions that include `qml.GlobalPhase` gates, etc.
[(#6686)](https://github.com/PennyLaneAI/pennylane/pull/6686)

<h4>Better drawing functionality</h4>

* Added support for the `wire_options` dictionary to customize wire line formatting in `qml.draw_mpl` circuit
visualizations, allowing global and per-wire customization with options like `color`, `linestyle`, and `linewidth`.
* `qml.draw_mpl` now has a `wire_options` keyword argument, which allows for global- and per-wire
customization with options like `color`, `linestyle`, and `linewidth`.
[(#6486)](https://github.com/PennyLaneAI/pennylane/pull/6486)

Here is an example that would make all wires cyan and bold except for wires 2 and 6, which are
dashed and a different colour.

```python
@qml.qnode(qml.device("default.qubit"))
def circuit(x):
for w in range(5):
qml.Hadamard(w)
return qml.expval(qml.PauliZ(0) @ qml.PauliY(1))

wire_options = {"color": "cyan",
"linewidth": 5,
2: {"linestyle": "--", "color": "red"},
6: {"linestyle": "--", "color": "orange"}
}
print(qml.draw_mpl(circuit, wire_options=wire_options)(0.52))
```

<h4>New device capabilities 💾</h4>

* Two new methods: `setup_execution_config` and `preprocess_transforms` are added to the `Device`
class. Device developers are encouraged to override these two methods separately instead of the
* Two new methods, `setup_execution_config` and `preprocess_transforms`, have been added to the
`Device` class. Device developers are encouraged to override these two methods separately instead of the
`preprocess` method. For now, to avoid ambiguity, a device is allowed to override either these
two methods or `preprocess`, but not both. In the long term, we will slowly phase out the use of
`preprocess` in favour of these two methods for better separation of concerns.
[(#6617)](https://github.com/PennyLaneAI/pennylane/pull/6617)

* Developers of plugin devices now have the option of providing a TOML-formatted configuration file
to declare the capabilities of the device. See [Device Capabilities](https://docs.pennylane.ai/en/latest/development/plugins.html#device-capabilities) for details.
to declare the capabilities of the device. See
[Device Capabilities](https://docs.pennylane.ai/en/latest/development/plugins.html#device-capabilities)
for details.

* An internal module `pennylane.devices.capabilities` is added that defines a new `DeviceCapabilites`
data class, as well as functions that load and parse the TOML-formatted configuration files.
* An internal module called `qml.devices.capabilities` has been added that defines a new
`DeviceCapabilites` data class, as well as functions that load and parse the TOML-formatted
configuration files.
[(#6407)](https://github.com/PennyLaneAI/pennylane/pull/6407)

```pycon
>>> from pennylane.devices.capabilities import DeviceCapabilities
>>> capabilities = DeviceCapabilities.from_toml_file("my_device.toml")
>>> isinstance(capabilities, DeviceCapabilities)
True
>>> from pennylane.devices.capabilities import DeviceCapabilities
>>> capabilities = DeviceCapabilities.from_toml_file("my_device.toml")
>>> isinstance(capabilities, DeviceCapabilities)
True
```

* Devices that extends `qml.devices.Device` now has an optional class attribute `capabilities`
that is an instance of the `DeviceCapabilities` data class, constructed from the configuration
file if it exists. Otherwise, it is set to `None`.
* Devices that extend `qml.devices.Device` now have an optional class attribute called
`capabilities`, which is an instance of the `DeviceCapabilities` data class constructed from the
configuration file if it exists. Otherwise, it is set to `None`.
[(#6433)](https://github.com/PennyLaneAI/pennylane/pull/6433)

```python
from pennylane.devices import Device

class MyDevice(Device):

config_filepath = "path/to/config.toml"

...
```

```pycon
>>> isinstance(MyDevice.capabilities, DeviceCapabilities)
True
```

* Default implementations of `Device.setup_execution_config` and `Device.preprocess_transforms`
are added to the device API for devices that provides a TOML configuration file and thus have
a `capabilities` property.
have been added to the device API for devices that provide a TOML configuration file and, thus,
have a `capabilities` property.
[(#6632)](https://github.com/PennyLaneAI/pennylane/pull/6632)
[(#6653)](https://github.com/PennyLaneAI/pennylane/pull/6653)

<h4>Capturing and representing hybrid programs</h4>

* Support is added for `if`/`else` statements and `for` and `while` loops in circuits executed with `qml.capture.enabled`, via Autograph.
Autograph conversion is now used by default in `make_plxpr`, but can be skipped with the keyword arg `autograph=False`.
* Support has been added for `if`/`else` statements and `for` and `while` loops in circuits executed
with `qml.capture.enabled`, via Autograph. Autograph conversion is now used by default in
`make_plxpr`, but can be skipped with `autograph=False`.
[(#6406)](https://github.com/PennyLaneAI/pennylane/pull/6406)
[(#6413)](https://github.com/PennyLaneAI/pennylane/pull/6413)
[(#6426)](https://github.com/PennyLaneAI/pennylane/pull/6426)
[(#6645)](https://github.com/PennyLaneAI/pennylane/pull/6645)
[(#6685)](https://github.com/PennyLaneAI/pennylane/pull/6685)

* `qml.transform` now accepts a `plxpr_transform` argument. This argument must be a function that can transform plxpr.
Note that executing a transformed function will currently raise a `NotImplementedError`. To see more details, check
out the documentation of `qml.transform`.
* `qml.transform` now accepts a `plxpr_transform` argument. This argument must be a function that
can transform plxpr. Note that executing a transformed function will currently raise a
`NotImplementedError`. To see more details, check out the
:func:`documentation of qml.transform <pennylane.transform>`.
[(#6633)](https://github.com/PennyLaneAI/pennylane/pull/6633)
[(#6722)](https://github.com/PennyLaneAI/pennylane/pull/6722)

* Users can now apply transforms with program capture enabled. Transformed functions cannot be executed by default. To apply
the transforms (and be able to execute the function), it must be decorated with the new `qml.capture.expand_plxpr_transforms`
function, which accepts a callable as input and returns a new function to which all present transforms have been applied.
* Users can now apply transforms with program capture enabled. Transformed functions cannot be
executed by default. To apply the transforms (and to be able to execute the function), it must be
decorated with the new `qml.capture.expand_plxpr_transforms` function, which accepts a callable as
input and returns a new function for which all present transforms have been applied.
[(#6722)](https://github.com/PennyLaneAI/pennylane/pull/6722)

```python
Expand Down Expand Up @@ -300,29 +324,31 @@ qubit operators.
in (d,) }
```

* The `qml.iterative_qpe` function can now be compactly captured into jaxpr.
* The `qml.iterative_qpe` function can now be compactly captured into plxpr.
[(#6680)](https://github.com/PennyLaneAI/pennylane/pull/6680)

* Functions and plxpr can now be natively transformed using the new `qml.capture.transforms.CancelInterpreter`
when program capture is enabled. This class cancels operators appearing consecutively that are adjoints of each
other, and follows the same API as `qml.transforms.cancel_inverses`.
[(#6692)](https://github.com/PennyLaneAI/pennylane/pull/6692)
* Three new plxpr interpreters have been added that allow for functions and plxpr to be natively
transformed with the same API as the corresponding existing transforms in PennyLane when program
capture is enabled:

* `qml.capture.transforms.CancelInterpreter`:this class cancels operators appearing consecutively
that are adjoints of each other following the same API as `qml.transforms.cancel_inverses`.
[(#6692)](https://github.com/PennyLaneAI/pennylane/pull/6692)

* Functions and plxpr can now be natively transformed using the new `qml.capture.transforms.DecomposeInterpreter`
when program capture is enabled. This class decomposes pennylane operators following the same API as
`qml.transforms.decompose`.
[(#6691)](https://github.com/PennyLaneAI/pennylane/pull/6691)
* `qml.capture.transforms.DecomposeInterpreter`: this class decomposes pennylane operators
following the same API as `qml.transforms.decompose`.
[(#6691)](https://github.com/PennyLaneAI/pennylane/pull/6691)

* Implemented a `MapWiresInterpreter` class that can be used as a quantum transform to map
operator and measurement wires with capture enabled.
[(#6697)](https://github.com/PennyLaneAI/pennylane/pull/6697)
* `qml.capture.transforms.MapWiresInterpreter`: this class maps wires to new values following the
same API as `qml.map_wires`.
[(#6697)](https://github.com/PennyLaneAI/pennylane/pull/6697)

* A `qml.tape.plxpr_to_tape` function can now convert plxpr to a tape.
* A `qml.tape.plxpr_to_tape` function is now available that converts plxpr to a tape.
[(#6343)](https://github.com/PennyLaneAI/pennylane/pull/6343)

* Execution with capture enabled now follows a new execution pipeline and natively passes the
captured jaxpr to the device. Since it no longer falls back to the old pipeline, execution
only works with a reduced feature set.
captured plxpr to the device. Since it no longer falls back to the old pipeline, execution only
works with a reduced feature set.
[(#6655)](https://github.com/PennyLaneAI/pennylane/pull/6655)
[(#6596)](https://github.com/PennyLaneAI/pennylane/pull/6596)

Expand All @@ -334,54 +360,59 @@ qubit operators.
[(#6422)](https://github.com/PennyLaneAI/pennylane/pull/6422)
[(#6668)](https://github.com/PennyLaneAI/pennylane/pull/6668)

* `qml.capture.PlxprInterpreter` base class has been added for easy transformation and execution of
pennylane variant jaxpr.
* A `qml.capture.PlxprInterpreter` base class has been added for easy transformation and execution
of plxpr.
[(#6141)](https://github.com/PennyLaneAI/pennylane/pull/6141)

* A `DefaultQubitInterpreter` class has been added to provide plxpr execution using python based tools,
and the `DefaultQubit.eval_jaxpr` method is now implemented.
* A `DefaultQubitInterpreter` class has been added to provide plxpr execution using python based
tools, and the `DefaultQubit.eval_jaxpr` method has been implemented.
[(#6594)](https://github.com/PennyLaneAI/pennylane/pull/6594)
[(#6328)](https://github.com/PennyLaneAI/pennylane/pull/6328)

* An optional method `eval_jaxpr` is added to the device API for native execution of plxpr programs.
* An optional method, `eval_jaxpr`, has been added to the device API for native execution of plxpr
programs.
[(#6580)](https://github.com/PennyLaneAI/pennylane/pull/6580)

* `qml.capture.qnode_call` has been made private and moved to the `workflow` module.
[(#6620)](https://github.com/PennyLaneAI/pennylane/pull/6620/)

<h4>Other Improvements</h4>

* `qml.math.grad` and `qml.math.jacobian` added to differentiate a function with inputs of any
interface in a jax-like manner.
* `qml.math.grad` and `qml.math.jacobian` have been added to differentiate a function with inputs of
any interface in a JAX-like manner.
[(#6741)](https://github.com/PennyLaneAI/pennylane/pull/6741)

* `qml.GroverOperator` now has a `work_wires` property.
[(#6738)](https://github.com/PennyLaneAI/pennylane/pull/6738)

* `Wires` object usage across Pennylane source code has been tidied up.
* The `Wires` object's usage across Pennylane source code has been tidied up for internal
consistency.
[(#6689)](https://github.com/PennyLaneAI/pennylane/pull/6689)

* `qml.equal` now supports `PauliWord` and `PauliSentence` instances.
* `qml.equal` now supports `qml.PauliWord` and `qml.PauliSentence` instances.
[(#6703)](https://github.com/PennyLaneAI/pennylane/pull/6703)

* Remove redundant commutator computations from `qml.lie_closure`.
* Redundant commutator computations from `qml.lie_closure` have been removed.
[(#6724)](https://github.com/PennyLaneAI/pennylane/pull/6724)

* Raises a comprehensive error when using `qml.fourier.qnode_spectrum` with standard numpy
* A comprehensive error is now raised when using `qml.fourier.qnode_spectrum` with standard Numpy
arguments and `interface="auto"`.
[(#6622)](https://github.com/PennyLaneAI/pennylane/pull/6622)

* Added Pauli String representations for the gates X, Y, Z, S, T, SX, SWAP, ISWAP, ECR, SISWAP. Fixed a shape error in the matrix conversion of `PauliSentence`s with list or array input.
* Pauli string representations for the gates `{X, Y, Z, S, T, SX, SWAP, ISWAP, ECR, SISWAP}` have
been added, and a shape error in the matrix conversion of `qml.PauliSentence`s with `list` or
`array` inputs has been fixed.
[(#6562)](https://github.com/PennyLaneAI/pennylane/pull/6562)
[(#6587)](https://github.com/PennyLaneAI/pennylane/pull/6587)

* `QNode` and `qml.execute` now forbid certain keyword arguments from being passed positionally.
* `qml.QNode` and `qml.execute` now forbid certain keyword arguments from being passed positionally.
[(#6610)](https://github.com/PennyLaneAI/pennylane/pull/6610)

* Shortened the string representation for the `qml.S`, `qml.T`, and `qml.SX` operators.
* The string representations for the `qml.S`, `qml.T`, and `qml.SX` have been shortened.
[(#6542)](https://github.com/PennyLaneAI/pennylane/pull/6542)

* Added functions and dunder methods to add and multiply Resources objects in series and in parallel.
* Internal class functions and dunder methods have been added to allow for multiplying Resources
objects in series and in parallel.
[(#6567)](https://github.com/PennyLaneAI/pennylane/pull/6567)

* The `diagonalize_measurements` transform no longer raises an error for unknown observables. Instead,
Expand Down

0 comments on commit 2f67f7f

Please sign in to comment.