Add update kwarg to QuantumScript.copy

[lillian542]

Context:
We often make new, modified tapes in transforms, and occasionally an attribute of the old tape is not included. It can be unclear in some cases whether this was intentional or not. For example, in many transforms, we don't copy over trainable_params when making the new tape. It can be unclear whether this is because we are intending to set trainable_params to the default None, or whether it's an oversight. We've also had at least one transform where we missed copying shots over until later.

Description of the Change:
We add an update kwarg to the tape.copy method that takes a dict with possible keys operations, measurements, shots and/or trainable_params. Passing a dict to update changes the defined attributes while copying everything else over from the original tape.

Benefits:
If we use this in transforms, we won't be able to forget to copy over tape attributes - and if we overwrite anything, it will be clearly intentional (i.e. we will have to explicitly pass None to the update method instead of leaving the attribute out when creating a new tape).

[sc-65487]

[lillian542]

[sc-65487]

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.70%. Comparing base (8f91b8a) to head (a44971c).
Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #6285   +/-   ##
=======================================
  Coverage   99.70%   99.70%           
=======================================
  Files         444      444           
  Lines       42226    42234    +8     
=======================================
+ Hits        42103    42111    +8     
  Misses        123      123           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[PietropaoloFrisoni]

Thanks, this seems helpful!

[PietropaoloFrisoni]

Thanks, this looks good to me!

I guess we just need to ensure that the latest doc version is nice and clean 👍

[astralcai]

LGTM

[albi3ro]

🚀