Refactor to installable MVC/MIS package with unified solver API

[Copilot]

This PR converts the repository from script/notebook-first research code into a pip-installable Python package with a stable runtime API. It adds a single class-based interface to solve both Minimum Vertex Cover (MVC) and Maximum Independent Set (MIS), plus focused documentation for usage and method context from the referenced paper.

• Package standardization

  • Introduces src/ layout and pyproject.toml for packaging and dependency management.
  • Defines runtime, optional (quantum, cplex, all), and dev extras.
  • Exposes a CLI entrypoint: qgmvc-solve.

• Unified user API (MVC + MIS)

  • Adds QuantumGreedySolver and SolveResult as the public API.
  • Supports one-call solving for problem="mvc" and problem="mis" with consistent input/output contracts.
  • Normalizes graph/weight validation and error handling (node-key matching, non-negative weights, undirected graphs).

• Problem adapters and reusable internals

  • Extracts classical and quantum logic into internal modules with stable interfaces.
  • Implements MIS as a first-class solve path via complement-of-cover mapping with explicit metadata.
  • Decouples runtime package usage from experiment plotting/SLURM/notebook flows.

• Algorithm robustness and API hardening

  • Improves quantum cover decoding behavior for undecided beta states and adds feasibility repair to guarantee edge coverage.
  • Aligns Pauli indexing to Qiskit qubit ordering.
  • Clarifies unweighted vs weighted greedy-degree behavior.

• Documentation and migration clarity

  • Replaces missing root docs with install/quickstart/API/problem overview.
  • Documents what the package solves, high-level method rationale, and paper link.
  • Adds migration guidance from legacy MVC/QGMVC.py / MVC/QGMVC_parallel.py workflows to class-based usage.

import networkx as nx
from quantum_greedy_mvc import QuantumGreedySolver

G = nx.cycle_graph(6)
solver = QuantumGreedySolver(method="greedy_degree")

mvc = solver.solve_mvc(G)   # minimum vertex cover
mis = solver.solve_mis(G)   # maximum independent set

print(mvc.solution, mvc.objective)
print(mis.solution, mis.objective)