Support svdvals, generalize norm, and introduce tracedist for Qobj

[ytdHuang]

The original LinearAlgebra.norm calculates standard vector norm no matter the input array is vector or matrix. But usually we calculate Schatten p-norm for matrices (density operators), which need to first calculate the singular values before taking the standard vector norm.

Summary for this PR:

• Support LinearAlgebra.svdvals (singular values) for Qobj
• Generalize LinearAlgebra.norm(A::QuantumObject, p::Real=2):
  • if Qobj is Ket, Bra, OperatorKet, or OperatorBra, calculate the standard vector p-norm
  • if Qobj is Operator or SuperOperator, calculate the Schatten p-norm
• Since the definition of negativity is related to the trace-norm (Schatten 1-norm), use the above method directly can slightly increase the efficiency.
• Introduce tracedist (also relates to the trace-norm) between two given Qobj

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 91.28%. Comparing base (82dc55f) to head (3fe2450).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #102      +/-   ##
==========================================
+ Coverage   91.24%   91.28%   +0.03%     
==========================================
  Files          20       20              
  Lines        1737     1744       +7     
==========================================
+ Hits         1585     1592       +7     
  Misses        152      152              

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