Wired input and output of function "c.amplitude"

[FelixXu35]

Issue Description

For a quantum circuit with n qubits, function "c.amplitude" returns the amplitude of a given quantum state. It will work when an n-bit binary string is passed into. An error will raise if the length of the string is not n, except n+1. If an (n+1)-bit string is passed into, "c.amplitude" will return the amplitude of a state corresponding to the first n bits, while the last bit controls the "phase". For example, if a circuit has 3 qubits, and one amplitude is
"000" -> 0.5
then
"0000" -> 0.5
"0001" -> 0.5*j
and j is the imaginary unit

How to Reproduce

Error Output

Environment Context

macOS 13.3.1(a)
tensorcircuit 0.8.0
BTW, why tc.about() gives me "module 'tensorcircuit' has no attribute 'about'"?

[refraction-ray]

BTW, why tc.about() gives me "module 'tensorcircuit' has no attribute 'about'"?

you can upgrade you tensorcircuit version: pip install -U tensorcircuit.

Thanks for report the n+1 qubit input case for c.amplitude, can be reproduced, will have a look

[refraction-ray]

turns out it is a feature related to tensornetwork, when there is only one disentangling edge (one extra bit string), the contractor doesn't raise error since unspecified output edge is unique.

See code below:

for n in [2, 3, 4]:
    ns = []
    for _ in range(n):
        ns.append(tc.gates.Gate(np.array([1.0, 0.0])))
    
    ns[0][0]^ns[1][0]
    try:
        print(n, tc.contractor(ns).tensor)
    except ValueError:
        print(n, "ValueError")

The output will be:

2 1.0
3 [1. 0.]
4 ValueError

If you really want to fix it (nothing really wrong if the use specify the correct bitstring), adding a length check at the beginning of amplitude method is enough. Welcome to PR on this error check protection

[FelixXu35]

Thanks for your reply.

I agree. Nothing is really wrong.

But I believe it is worthy to add a length check, because the error message of a (<n) binary string is:

"list index out of range"

and of a (>n+1) binary string is:

"The final node after contraction has more than one remaining edge. In this case output_edge_order has to be provided."

Neither of them is clear for a beginner.

I will give a try.