Skip to content

Latest commit

 

History

History
235 lines (169 loc) · 12 KB

RELEASE_NOTES.rst

File metadata and controls

235 lines (169 loc) · 12 KB
Raw

qupulse 0.10 (2024-04-04)

Features

  • Move HDAWG driver to qupulse-hdawg-legacy to disentangle driver version from qupulse version. The new HDAWG driver will be published under qupulse-hdawg. (#779)
  • Add the ProgramBuilder interface pattern to make the generated program of PulseTemplate.create_program easily customizable. (#781)
  • Measurement windows can now automatically shrank in case of overlap to counteract small numeric errors. (#791)

Bugfixes

  • ``ConstantPulseTemplate``s from all versions can now be deserialized. (#696)
  • Fixed that single segment tables where always interpreted to be constant. (#707)
  • Add missing pulse registry support to ArithmeticPT. (#775)

qupulse 0.9 (2023-11-08)

Features

  • Add __pow__ as a repetition shortcut. This means you can do my_pulse_template ** 5 or my_pulse_template ** 'my_repetition_count'. (#692)

  • Promote qupulse.expression to a subpackage and create qupulse.expression.protocol with protocol classes that define the expression interface that is supposed to be used by qupulse. The `sympy based implementation is moved to qupulse.expressions.sympy and imported in qupulse.expressions.

    The intended use is to be able to use less powerful but faster implementations of the Expression protocol where appropriate. In this first iteration, qupulse still relies on internals of the sympy based implementation in many places which is to be removed in the future. (#750)

  • Promote parts of the private subpackage qupulse._program to the public subpackage qupulse.program, i.e. loop, volatile, transformation and waveforms. This allows external packages/drivers to rely on stability of the Loop class. (#779)

  • Add PulseTemplate.pad_to method to help padding to minimal lengths or multiples of given durations. (#801)

Misc

qupulse 0.8 (2023-03-28)

Features

  • New two dimensional plotting function qupulse.pulses.plotting.plot_2d. (#703)
  • Add support for time dependent expressions for arithmetics with atomic pulse templates i.e. ParallelChannelPT and ArithmeticPT support time dependent expressions if used with atomic pulse templates. Rename ParallelConstantChannelPT to ParallelChannelPT to reflect this change. (#709)
  • Add with_ family of helper methods to PulseTemplate to allow convinient and easily discoverable pulse template combination. (#710)
  • The plotting module is now located at qupulse.plotting. There is a legacy alias at qupulse.pulses.plotting. (#735)

Deprecations and Removals

  • Remove the Parameter, MappedParameter and ConstantParameter classes that where deprecated in version 0.5. (#512)
  • Drop support for python version 3.7. (#760)

qupulse 0.7 (2022-10-05)

Features

  • Add optional numba uses in some cases. (#501)

  • Add initial_values and final_values attributes to PulseTemplate.

    This allows pulse template construction that depends on features of arbitrary existing pulses i.e. like extension until a certain length. (#549)

  • Support sympy 1.9 (#615)

  • Add option to automatically reduce the sample rate of HDAWG playback for piecewise constant pulses. Use qupulse._program.seqc.WaveformPlayback.ENABLE_DYNAMIC_RATE_REDUCTION to enable it. (#622)

  • Add a TimeReversalPT. (#635)

  • Add specialied parameter Scope for ForLoopPT. This increases performance by roughly a factor of 3 for long ranges! (#642)

  • Add sympy 1.10 support and make ExpressionVector hashable. (#645)

  • Serializable is now comparable via it's get_serialized_data. PulseTemplate implements Hashable via the same. (#653)

  • Add an interface that uses atsaverage.config2. (#686)

Bugfixes

  • floor will now return an integer in lambda expressions with numpy to allow usage in ForLoopPT range expression. (#612)

Deprecations and Removals

  • Drop cached_property dependency for python>=3.8. (#638)
  • Add frozendict dependency to replace handwritten solution. Not having it installed will break in a future release when the old implementation is removed. (#639)
  • Drop python 3.6 support. (#656)

qupulse 0.6 (2021-07-08)

Features

  • Add evaluate_with_exact_rationals method to ExpressionScalar (#546)
  • New feature based AWG abstraction. Can be found in qupulse.hardware.feature_awg. Allows easier code reuse across awg drivers. (#557)
  • Add ConstantPulseTemplate (#565)
  • Add interface to use atsaverage auto rearm (#566)
  • Adds the methods is_constant, constant_value_dict and constant_value to Waveform class to allow more efficient AWG usage. (#588)

Bugfixes

  • Fix TimeType comparisons with non-finite floats (inf, -inf, NaN) (#536)
  • Improve alazar usability:
    • Do not touch the default config when arming a measurement
    • Keep current config in a seperate field
    • Extend record to a multiple of a configurable value (4KB by default) (#571)
  • Replace pytabor and teawg with tabor_control to support newer(>=1.11) pyvisa versions (#599)
  • Fix repr of ExpressionScalar when constructed from a sympy expression. Also replace Expression with ExpressionScalar in repr. (#604)

Deprecations and Removals

  • Deprecate HashableNumpyArray due to its inconsistency. (#408)
  • Drop support for python 3.5 (#504)
  • Remove deprecated external_parameters keyword argument from SequencePT and AtomicMultiChannelPT (#592)
  • Deprecate boolean duration argument of AtomicMultiChannelPulseTemplate and remove duration check in __init__. (#593)

0.5.1

  • General:
    • Unify TimeType.from_float between fractions and gmpy2 backend behaviour (fixes issue 529).

0.5

  • General:
    • Improve TimeType consistency by leveraging str(float) for rounding by default.
    • Add support for sympy==1.5
    • Add volatile parameters. Repetition counts can now be changed at runtime in some cases (useful for DNP). See volatile kwarg of create_program
  • Hardware:
    • Add a measure_program method to the DAC interface. This method is used by the QCoDeS integration.
    • Add a set_measurement_mask to DAC interface. This method is used by the QCoDeS integration.
    • Add a get_sample_times util method to share code for exact and fast sample time calculation
    • Add a driver for Tektronix AWG5000/7000
    • Add a driver for Zurich Instruments HDAWG
    • Warn the user if waveforms need to be concatenated to be compatible with hardware requirements.
  • Pulse Templates:
    • Add __repr__ and __format__ for easier inspection
    • MappingPulseTemplate:
      • allow_partial_parameter_mapping is now True as a default. The default can be changed with the class variable MappingPulseTemplate.ALLOW_PARTIAL_PARAMETER_MAPPING.
      • Add specializations for map_parameters because the auto-inference of the return type did not work for empty input.
      • Channels mapped to None are now dropped
    • Add simple arithmetic operations for pulse templates
      • offset and scaling with scalars
      • addition with atomic pulse templates
  • Expressions:
    • Expressions can now be formatted as floats if they do not have free variables
  • Parameters:
    • Replace Parameter class with Scope
    • Parameter class is now deprecated
  • Backward incompatible changes:
    • Removed deprecated classes: - Sequencer: Replaced by PulseTemplate.create_program method - Condition: Never used - InstructionBlock: Old representation of programs. Replaced by Loop - MultiChannelProgram: Was required in the instruction block framework

0.4

  • General:
    • Add utility function qupulse.utils.types.has_type_interface and use it to circumvent autoreload triggered isinstance fails
    • Add utility function qupulse.utils.time_from_fraction to make creation from numerator and denominator obvious.
  • Pulse Templates:
    • MappingPulseTemplate:
      • Raise a ValueError if more than one inner channel is mapped to the same outer channel
    • Plotting:
      • Make plotting.render behaviour and return value consistent between calls with InstructionBlock and Loop. Render now always returns 3 arguments.

0.3

  • General:
    • Introduce qupulse.utils.isclose (an alias for math.isclose if available)
    • Dropped support for Python 3.4 in setup.py due to incompatible syntax in qupulse.
    • Official support for Python 3.7 has begun.
  • Pulse Templates:
    • AtomicMultichannelPulseTemplate:
      • Add duration keyword argument & example (see MultiChannelTemplates notebook)
      • Make duration equality check approximate (numeric tolerance)
    • Plotting:
      • Add time_slice keyword argument to render() and plot()
    • Add AbstractPulseTemplate class
    • PointPulseTemplate:
      • Fixed bug in integral evaluation
    • Add ParallelConstantChannelPulseTemplate which allows adding a constant valued channel to an arbitrary pulse template
  • Expressions:
    • Make ExpressionScalar hashable
    • Fix bug that prevented evaluation of expressions containing some special functions (erfc, factorial, etc.)
  • Parameters:
    • ConstantParameter now accepts a Expression without free variables as value (given as Expression or string)

0.2

  • General:
    • officially removed support for Python 3.3 (qupulse and dependencies are not compatible anymore)
  • Serialization / Storage:
    • Added functionality to easily access available content/identifiers in PulseStorage and StorageBackend.
    • DEPRECATED list_contents() of StorageBackend (use contents property instead).
    • DEPRECATED: CachingBackend because its functionality is a subset of PulseStorage.
  • Expressions:
    • Fixed bug in Expression.evaluate_numeric if result is array of numeric sympy objects