How NWQLib works

Every NWQLib calculation passes through five objects. You build the Problem and the Method, and NWQLib creates the other three.

Object	What it is	Example
Problem	The scientific question with its input data and requested output, independent of any algorithm	LinearSystem(A=A, b=b) holds A and b
Method	A configured algorithm with its options. It plans, executes and analyzes one Problem	LCHS(), QLS(epsilon_inv=0.01)
Plan	The choices a Method makes once for one Problem: the construction, its parameters, the random-stream snapshot and the data needed to reconstruct the output. It exists before any circuit is built	plan(problem, method=LCHS())
Run	One execution of a Plan. It records the work it reserves and submits and the observations that come back, and it can be saved and reopened	prepared.run after submit(prepared)
Result	The analyzed scientific output, with its Plan and the observations (RunData) it was computed from. Reanalysis produces a new Result from the same observations	result.solution, result.eigenvalue

Choose a problem and output lists the Problem types and their outputs, and each algorithm guide describes its Method's options.

How the objects connect

Problem + Method --plan--> Plan
Plan --estimate--> resource counts, before any circuit is built
Plan --solve--> Result
Plan --prepare--> Prepared --submit--> Run --wait--> Result
Result --save--> folder --load_result--> Result

Step	Call	What it does
Plan	plan(problem, method=...)	The Method chooses its construction and computes the numerical data it needs. No measurement is taken
Compare	compare(problem, methods=(...))	Plans and estimates each Method for the same Problem without measuring anything. Every row stays visible, and the comparison chooses no winner
Estimate	estimate(plan)	Adds up the resource formulas of the Plan's construction, such as qubits, CX gates and shots, and labels each count as exact, an upper bound, an estimate or unavailable
Solve	solve(problem, method=...) or solve(plan)	Plans when given a Problem, executes, and returns the Result
Prepare and submit	prepare(plan, backend=...), submit(prepared), run.wait()	Builds the circuits, submits them in a Run and waits for the Result. Use these calls for a provider backend or to control execution yourself
Save and load	result.save(path), load_result(path)	Saves a completed Result and reopens it for inspection and for reanalysis of its saved observations
Continue	load_run(path, backend=...), run.resume()	Reopens a saved Run and collects its pending work

Plan, compare and solve walks through these calls for two eigenvalue methods. Run on a backend covers prepare and submit, Save, load and reanalyze results covers saved Results, and Continue an interrupted run covers saved Runs. The API reference gives every signature.