trace, partial trace, and superoperators #55

apkille · 2024-06-26T15:12:51Z

Added trace and partial trace functionalities with examples and tests. Also created a symbolic superoperator SSuperOperator, including an option to express instances of this composite type in the Kraus representation with the function kraus. Future work can be done to create additional superoperator representations and conversions between each of them.

In the process of developing this branch, I encountered a few bugs mentioned in #53 and #54, which I fix in this PR.

apkille · 2024-06-26T15:59:23Z

The failed tests seem to come from the recent bumped compat for TermInterface #52, although most recent docs (https://juliasymbolics.github.io/TermInterface.jl/dev/) say that implementing sorted_arguments is optional.

 MethodError: no method matching sorted_arguments(::QuantumSymbolics.STrace)
│    
│    Closest candidates are:
│      sorted_arguments(::SymbolicUtils.PolyForm)
│       @ SymbolicUtils ~/.julia/packages/SymbolicUtils/dtCid/src/polyform.jl:235
│      sorted_arguments(::SymbolicUtils.BasicSymbolic)
│       @ SymbolicUtils ~/.julia/packages/SymbolicUtils/dtCid/src/types.jl:119

codecov · 2024-06-26T17:18:21Z

Codecov Report

Attention: Patch coverage is 89.05109% with 15 lines in your changes missing coverage. Please review.

Project coverage is 74.23%. Comparing base (c9d8eed) to head (9f05adb).

Files	Patch %	Lines
src/QSymbolicsBase/basic_superops.jl	59.09%	9 Missing ⚠️
src/QSymbolicsBase/linalg.jl	91.54%	6 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #55      +/-   ##
==========================================
+ Coverage   70.46%   74.23%   +3.77%     
==========================================
  Files          17       18       +1     
  Lines         650      753     +103     
==========================================
+ Hits          458      559     +101     
- Misses        192      194       +2

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

Krastanov

great start! There are a few things I want us to discuss before considering for merging

src/QSymbolicsBase/QSymbolicsBase.jl

src/QSymbolicsBase/basic_ops_homogeneous.jl

Krastanov · 2024-06-28T05:05:25Z

src/QSymbolicsBase/predefined.jl

+function ptrace(x::STensorOperator, s)
+    terms = arguments(x)
+    sys_op = terms[s]
+    new_terms = deleteat!(copy(terms), s)
+    isone(length(new_terms)) ? tr(sys_op)*first(new_terms) : tr(sys_op)*STensorOperator(new_terms)
+end


This is actually a bit worse than this. You can have @op A SpinBasis(1//2)^2, i.e. the symbolic operator is already over a composite basis. The number of subsystem would be the number of subsystems in basis, which might be different from the number of terms in the tensor product.

Also, what if flattening has not happened. E.g ptrace(A⊗(B⊗C+D⊗E), 2)?

Ah, these are both fair comments. Will make changes.

Just found another small bug:

julia> @op A; @op B; @op C; julia> tp = A⊗B⊗C ((A⊗B)⊗C) julia> QuantumSymbolics.arguments(tp) 2-element Vector{Any}: (A⊗B) C

We would want tp to show (A⊗B⊗C) and QuantumSymbolics.arguments(tp) to be a 3-element vector of the symbolic operators.

Weird. I wonder why this doesn't happen for multiplication of operators:

julia> mp = A*B*C ABC julia> QuantumSymbolics.arguments(mp) 3-element Vector{Any}: A B C

Also, what if flattening has not happened. E.g ptrace(A⊗(B⊗C+D⊗E), 2)?

A question somewhat related to this: should we have a function in the package that expands expressions? For instance, if a user for some reason wanted to convert A⊗(B⊗C+D⊗E) into A⊗B⊗C+A⊗D⊗E, shouldn't they have a way to do that? Maybe we could call it qexpand, using the q prefix the same we do for qsimplify.

yup, let's have a qexpand for now. Maybe at some point we will just hook into the standard expand and at that point it can be just an alias without making a breaking release

concerning the difference between mul and tensor, I think it is because prefactorscalings were created for tensor, and it has been kinda hackishly applied to mul. We probably need to clean prefactorscalings to work on both. I will make a separate issue for that.

Wait, no that is overly complicated. Why not just have scalar=true always. Is there a situation that benefits from scalar=false?

yes about a new file

It's my understanding that prefactorscalings was originally created for tensor product properties. For instance, |k1⟩⊗(A*|k2⟩) would give A|k1⟩|k2⟩.

Regardless, I will clean up the scaling and flattening functions.

prefactorscaling was always meant only for scalars. The operator identity you are giving is wrong because A is defined to act on the Hilbert space of k2, not on the composite hilbert space of k1k2.

Krastanov · 2024-06-28T05:07:56Z

Just so that I do not get lost, could you point out in this comment thread what are the fixes for #53 and #54? And the corresponding test?

Krastanov · 2024-06-28T05:08:42Z

For the spellchecker issue, BA can be whitelisted in https://github.com/QuantumSavory/QuantumSymbolics.jl/blob/main/.typos.toml

Krastanov · 2024-06-28T05:23:37Z

src/QSymbolicsBase/literal_objects.jl

@@ -92,4 +106,4 @@ symbollabel(x::SZero) =  "𝟎"
 basis(x::SZero) = nothing

 Base.show(io::IO, x::SZero) = print(io, symbollabel(x))
-Base.iszero(x::SymQObj) = isa(x, SZero)
+Base.iszero(x::Union{SymQObj, Symbolic{Number}, Symbolic{Complex}}) = isa(x, SZero)


I think the new iszero implementations here are type piracy. Where are they needed?

This fixes a bug mentioned in #54 as well: #54 (comment).

src/QSymbolicsBase/rules.jl

src/QSymbolicsBase/basic_ops_homogeneous.jl

apkille · 2024-06-29T01:38:24Z

I marked this PR as a draft because I first want to clean up the scaling and flattening functions, and also implement a qexpand function for expanding quantum expressions. Since this PR adds a lot of features, I would like to submit these changes in a separate PR.

Krastanov · 2024-07-05T04:19:34Z

The downstream buildkite error is known and unrelated.

Krastanov

I did a quick review, I have not had a chance to go through everything yet. I think there are a few minor issues with basis checks.

src/QSymbolicsBase/QSymbolicsBase.jl

src/QSymbolicsBase/basic_superops.jl

src/QSymbolicsBase/linalg.jl

test/test_expand.jl

test/test_trace.jl

Krastanov · 2024-07-05T12:32:46Z

Just FYI, QuantumInterface.jl provides some basis checking foundations and error types: https://github.com/qojulia/QuantumInterface.jl/blob/main/src/bases.jl#L140-L180

src/QSymbolicsBase/linalg.jl

Krastanov

looks good!

I left a bunch of comments related to an annoyance with QuantumInterface and created an issue to track it qojulia/QuantumInterface.jl#25

I would suggest merging this without addressing that issue, as otherwise there will be too much delay (and it is not a correctness issue).

I left a few other minor comments. Feel free to merge this when you have chance to address them.

Thanks for setting all of this up!

src/QSymbolicsBase/basic_ops_homogeneous.jl

src/QSymbolicsBase/basic_ops_inhomogeneous.jl

src/QSymbolicsBase/linalg.jl

src/QSymbolicsBase/predefined.jl

test/test_basis_consistency.jl

Krastanov

I forgot to check the coverage. There are a couple of lines that would be nice to test (and one of them is related to a broken doctest)

Krastanov · 2024-07-19T12:16:54Z

src/QSymbolicsBase/basic_superops.jl

+head(x::KrausRepr) = :kraus
+children(x::KrausRepr) = [:kraus; x.krausops]
+kraus(xs::Symbolic{AbstractOperator}...) = KrausRepr(collect(xs))
+basis(x::KrausRepr) = basis(first(x.krausops))


coverage review: could you add a test for this line

src/QSymbolicsBase/basic_superops.jl

src/QSymbolicsBase/literal_objects.jl

Permission to merge

apkille added 10 commits June 24, 2024 11:59

superop and trace additions

353d689

ptrace additions

19e1b3f

tests

1af6b2a

examples and redef

45a4224

tests

49411c9

Merge branch 'main' into channels

c397478

minor fix

e3d2a49

another test

ac731e2

hashing for trace and inner product

5066d34

printing fix

1e3c956

sorted_arguments for trace

4793dc7

apkille added 3 commits June 26, 2024 13:40

superop tests

b7aeab1

change doctest

871b1c3

JET fixes

343108b

Krastanov requested changes Jun 28, 2024

View reviewed changes

Krastanov reviewed Jun 28, 2024

View reviewed changes

apkille commented Jun 28, 2024

View reviewed changes

src/QSymbolicsBase/rules.jl Outdated Show resolved Hide resolved

apkille commented Jun 28, 2024

View reviewed changes

src/QSymbolicsBase/basic_ops_homogeneous.jl Outdated Show resolved Hide resolved

apkille commented Jun 28, 2024

View reviewed changes

src/QSymbolicsBase/basic_ops_homogeneous.jl Outdated Show resolved Hide resolved

cleanup

7c25134

apkille marked this pull request as draft June 29, 2024 01:34

apkille mentioned this pull request Jun 29, 2024

Add qexpand functionality and cleanup scalings #59

Merged

apkille added 3 commits July 2, 2024 10:52

Merge branch 'main' into channels

057a837

doctest and piracy fixes

e41dc7e

Merge branch 'main' into channels

194ebba

update ptrace and added tests

5997dfb

apkille marked this pull request as ready for review July 2, 2024 19:24

apkille requested a review from Krastanov July 2, 2024 19:24

Krastanov requested changes Jul 5, 2024

View reviewed changes

apkille added 3 commits July 5, 2024 17:05

fix basis consistency, ptrace, and superops

3e9f5e7

more codecov

e9c4c91

typo fix

e3d7eb0

apkille commented Jul 6, 2024

View reviewed changes

src/QSymbolicsBase/linalg.jl Show resolved Hide resolved

boson op change

ffd9c26

Krastanov self-requested a review July 17, 2024 20:04

Krastanov requested changes Jul 19, 2024

View reviewed changes

Krastanov previously requested changes Jul 19, 2024

View reviewed changes

apkille added 3 commits July 19, 2024 13:55

review changes

bcddbc5

Merge branch 'main' into channels

77f737d

update changelog and project.toml

9f05adb

apkille merged commit 68fbe00 into QuantumSavory:main Jul 19, 2024
16 of 17 checks passed

apkille deleted the channels branch July 19, 2024 18:30

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

trace, partial trace, and superoperators #55

trace, partial trace, and superoperators #55

apkille commented Jun 26, 2024 •

edited

Loading

apkille commented Jun 26, 2024 •

edited

Loading

codecov bot commented Jun 26, 2024 •

edited

Loading

Krastanov left a comment

Krastanov Jun 28, 2024

apkille Jun 28, 2024

apkille Jun 28, 2024

apkille Jun 28, 2024

Krastanov Jun 28, 2024

Krastanov Jun 28, 2024

Krastanov Jun 28, 2024

apkille Jun 28, 2024 •

edited

Loading

apkille Jun 28, 2024

Krastanov Jun 28, 2024

Krastanov commented Jun 28, 2024

Krastanov commented Jun 28, 2024

Krastanov Jun 28, 2024

apkille Jun 28, 2024

apkille commented Jun 29, 2024

Krastanov commented Jul 5, 2024

Krastanov left a comment

Krastanov commented Jul 5, 2024

Krastanov left a comment

Krastanov left a comment

Krastanov Jul 19, 2024

trace, partial trace, and superoperators #55

trace, partial trace, and superoperators #55

Conversation

apkille commented Jun 26, 2024 • edited Loading

apkille commented Jun 26, 2024 • edited Loading

codecov bot commented Jun 26, 2024 • edited Loading

Codecov Report

Krastanov left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

apkille Jun 28, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Krastanov commented Jun 28, 2024

Krastanov commented Jun 28, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

apkille commented Jun 29, 2024

Krastanov commented Jul 5, 2024

Krastanov left a comment

Choose a reason for hiding this comment

Krastanov commented Jul 5, 2024

Krastanov left a comment

Choose a reason for hiding this comment

Krastanov left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

apkille commented Jun 26, 2024 •

edited

Loading

apkille commented Jun 26, 2024 •

edited

Loading

codecov bot commented Jun 26, 2024 •

edited

Loading

apkille Jun 28, 2024 •

edited

Loading