v2.28.0

Announcements

Improvements and Changes

Bugfixes

• Fix parsing error for parameterized DEFCIRUCITs (@ameyer-rigetti, #1295)

v2.27.0

Announcements

• Switched to Github Actions.

Improvements and Changes

• Bump RPCQ dependency to 3.6.0 (@notmgsk, #1286).
• Tests can be run in parallel (@notmgsk, #1289).

Bugfixes

• Fix hanging test due to ZMQ bug (@notmgsk).
• Fix unitary comparison in Quil compilation test (@notmgsk).
• Fix parsing comments in Lark grammar (@notmgsk, #1290).

v2.26.0

Announcements

• Quil-T brings the dimension of time to your quantum programs! Quil-T is an extension of
  Quil which allows one to develop quantum programs at the level of pulses and waveforms
  and brings an unprecedented level of fine-grained control over the QPU.

Improvements and Changes

• Unpacking bitstrings is significantly faster (@mhodson-rigetti, @notmgsk, #1276).
• Parsing is now performed using Lark rather than ANTLR, often allowing a 10x improvement
  in parsing large and complex programs (@notmgsk, #1278).
• Gates now generally allow a "formal" qubit label as in DEFCIRCUIT, rather than
  requiring a numeric index (#1257).
• Program objects come with additional Quil-T related properties, such as
  calibrations, waveforms, and frames (#1257).
• The AbstractCompiler classes come with tools for performing calibration of
  programs. Namely, get_calibration_program provides a program for calibrating against
  recent QPU settings (#1257).
• rewrite_arithmetic now converts phase angle from radians to revolutions (#1257).
• Readout is more permissive, and does not require the destination to be named "ro"
  (#1257).
• The default value for QPU_COMPILER_URL has been updated to point to Rigetti's
  translation service. This changes allows one to use the translation service to translate
  a Quil-T program and receive the binary payload without having a QPU reservation
  (#1257).

v2.25.0

Announcements

Improvements and Changes

• Timeout configuration has been revamped. get_qc now accepts a compiler_timeout
  option, and QVMCompiler and QPUCompiler provide a set_timeout method, which should
  greatly simplify the task of changing the default timeout. QVMCompiler also provides a
  quilc_client property so that it shares the same interface as
  QPUCompiler. Documentation has been updated to reflect these changes (@notmgsk,
  @kalzoo, #1273).

Bugfixes

v2.24.0

Announcements

Improvements and Changes

• run_and_measure now only measures the qubits that are used in a program
  (rather than all qubits on the device) when the target QAM is a QVM without
  noise. This prevents the QVM from exhausting memory when it tries to allocate
  for e.g. 32 qubits when only e.g. 2 qubits are used in the program (@notmgsk,
  #1252)

• Include a py.typed so that libraries that depend on pyquil can validate
  their typing against it (@notmgsk, #1256)

• Fix Apache 2 license (@mpharrigan, #1255)

• Removed warnings expected in normal workflows that cannot be avoided
  programmatically. This included the warning about passing native Quil to
  native_quil_to_executable. Documentation has been updated to clarify
  expected behavior (@mhodson-rigetti, gh-1267).

Bugfixes

• Fixed incorrect return type hint for the exponential_map function, which
  now accepts both float and MemoryReference types for exponentiation
  (@mhodson-rigetti, gh-1243).

v2.23.2

This release has no material changes to the pyquil library, but is being made to trigger some build pipelines. Apologies for the spam.

v2.23.1

Announcements

Improvements and Changes

• Push new pyquil versions to pypi as part of CI/CD pipelines (@notmgsk, gh-1249)

Bugfixes

• Allow np.ndarray in DefPermutationGate (@notmgsk, gh-1248)

v2.23.0

Bump version to 2.23.0 (#1247)

v2.22.0

Bump version to 2.22.0 (#1239)

v2.21.1

Bump version to 2.21.1

This is just a cosmetic updated, to trigger a new docker build.