Data class update

[clausia]

Update Data Class as specified in Issue #27

This PR refactors the Data class to align with the specifications outlined in the corresponded issue. The following modifications have been made:

Main Changes:

• Attributes:

  • qubits, depth, and activation_matrices now store and manage the circuit's parameters and activation matrices as described.
  • The class now loads and processes matrices directly from data.json and supports the storage of converted matrices.

• Methods:

  • get_data(): Loads qubits, depth, and activation_matrices from data.json. Returns error codes if data.json or train.json are missing.
  • return_matrices(): Validates and returns activation matrices with the expected properties; returns an error code if validation fails.
  • convert_matrices(): Converts each activation matrix by combining values from three pages into a flattened format, following the required structure.
  • check_final_matrix(): Checks matrix dimensions to ensure they match the required (qubits, depth * 3) format.
  • create_training_matrix(): Generates a training matrix with randomized values, validates dimensions, and saves it to train.json.

Additional Changes:

• Created comprehensive unit tests for each method using pytest to ensure functionality and handle edge cases:
  • Tests verify error handling, matrix conversion, dimension validation, and file operations.
  • Includes checks for missing files, correct and incorrect dimensions, and successful data loading and storage.

Fixes #27