Translators API Reference

Convert LP file or string to LunaModel.

Reads an optimization model from an LP file or parses an LP-formatted string and converts it to a LunaModel Model object.

Parameters:

• file (str or Path) –

  Either a path to an LP file or an LP-formatted string. If the string contains newlines or looks like LP content, it will be parsed as a string. Otherwise, it will be treated as a file path.

Returns:

• Model –

  The parsed model in LunaModel format.

Raises:

• TranslationError –

  If the LP file is malformed or contains unsupported constructs.

• FileNotFoundError –

  If the specified file path does not exist.

Examples:

From a file:

>>> model = LpTranslator.to_lm("model.lp")

From a string:

>>> lp_str = '''
... Minimize
...   obj: 3 x + 2 y
... Subject To
...   c1: x + y <= 1
... Bounds
...   0 <= x <= 1
...   0 <= y <= 1
... Binary
...   x y
... End
... '''
>>> model = LpTranslator.to_lm(lp_str)

Convert LunaModel to LP file or string.

Converts a LunaModel Model to LP format, either writing to a file or returning the LP-formatted string.

Parameters:

• model (Model) –

  The LunaModel to convert.

• filepath (Path, default: None ) –

  If provided, writes the LP representation to this file path. If None, returns the LP string without writing to a file.

Returns:

• str or None –

  If filepath is None, returns the LP-formatted string. If filepath is provided, writes to file and returns None.

Raises:

• TranslationError –

  If the model contains constructs that cannot be represented in LP format (e.g., higher-order terms).

Examples:

Write to file:

>>> from luna_model import Model
>>> model = Model(name="example")
>>> # ... build model ...
>>> LpTranslator.from_lm(model, "output.lp")

Get LP string:

>>> lp_string = LpTranslator.from_lm(model)
>>> print(lp_string)
\ Model example
\ Problem name: example

Minimize

Subject To

Bounds


End

Convert QUBO matrix to LunaModel.

Creates a LunaModel Model from a QUBO matrix representation.

Parameters:

• qubo (NDArray) –

  QUBO matrix where Q[i,j] represents the coefficient for x[i]*x[j]. Diagonal elements Q[i,i] are linear coefficients. If the matrix is not symmetric, it will be made symmetric by summing Q[i,j] and Q[j,i].

• offset (float, default: None ) –

  Constant offset term to add to objective. Default is 0.

• variable_names (list[str], default: None ) –

  Names for variables. If None, generates names like "x0", "x1", etc.

• name (str, default: None ) –

  Name for the model.

• vtype (Vtype, default: None ) –

  Variable type (BINARY or SPIN). Default is BINARY.

Returns:

• Model –

  LunaModel with objective function representing the QUBO.

Examples:

Basic usage:

>>> import numpy as np
>>> Q = np.array([[-2, 1], [0, -1]])
>>> model = QuboTranslator.to_lm(Q)

With custom names and offset:

>>> model = QuboTranslator.to_lm(Q, offset=5.0, variable_names=["x1", "x2"], name="MyQUBO")

Convert LunaModel to QUBO format.

Converts a LunaModel Model to QUBO representation.

Parameters:

• model (Model) –

  The model to convert. Must be: - Unconstrained (no constraints) - Binary or spin variables only - At most quadratic (degree ≤ 2)

Returns:

• Qubo –

  QUBO representation with matrix, offset, and metadata.

Raises:

• TranslationError –

  If the model has constraints, non-binary variables, or higher-order terms.

Examples:

>>> from luna_model import Model
>>> model = Model()
>>> x = model.add_variable("x")
>>> y = model.add_variable("y")
>>> model.objective = x * y - 2 * x + y
>>> qubo = QuboTranslator.from_lm(model)
>>> print(qubo.matrix)
[[-2.   0.5]
 [ 0.5  1. ]]

Convert D-Wave BQM to LunaModel.

Converts a D-Wave Ocean SDK BinaryQuadraticModel to a LunaModel Model.

Parameters:

• bqm (BinaryQuadraticModel) –

  D-Wave BQM to convert. All variable names must be strings.

• name (str, default: None ) –

  Name for the resulting model. If None, uses default name.

Returns:

• Model –

  LunaModel representation with objective function matching the BQM.

Raises:

• RuntimeError –

  If dimod package is not installed.

• TypeError –

  If bqm is not a BinaryQuadraticModel or if variable names are not strings.

Examples:

>>> from dimod import BinaryQuadraticModel
>>> from luna_model.translator import BqmTranslator
>>> bqm = BinaryQuadraticModel({"x": -1, "y": 2}, {("x", "y"): 1}, 0.5, "BINARY")
>>> model = BqmTranslator.to_lm(bqm, name="example")
>>> print(model.objective)
x y - x + 2 y + 0.5

Convert LunaModel to D-Wave BQM.

Converts a LunaModel Model to a D-Wave BinaryQuadraticModel.

Parameters:

• model (Model) –

  The model to convert. Must be: - Unconstrained (no constraints) - Binary or spin variables only - At most quadratic (degree ≤ 2)

Returns:

• BinaryQuadraticModel –

  D-Wave BQM ready for use with D-Wave solvers.

Raises:

• RuntimeError –

  If dimod package is not installed.

• TranslationError –

  If the model has constraints, non-binary/spin variables, or higher-order terms.

Examples:

>>> from luna_model import Model
>>> from luna_model.translator import BqmTranslator
>>> model = Model()
>>> x = model.add_variable("x")
>>> y = model.add_variable("y")
>>> model.objective = x * y - 2 * x + y
>>> bqm = BqmTranslator.from_lm(model)

Convert D-Wave CQM to LunaModel.

Converts a D-Wave ConstrainedQuadraticModel to a LunaModel Model.

Parameters:

• cqm (ConstrainedQuadraticModel) –

  D-Wave CQM to convert.

• name (str, default: None ) –

  Name for the resulting model. If None, uses the CQM's name or a default name.

Returns:

• Model –

  LunaModel representation with objective and constraints matching the CQM.

Raises:

• RuntimeError –

  If dimod package is not installed.

• TypeError –

  If cqm is not a ConstrainedQuadraticModel.

Examples:

>>> from dimod import ConstrainedQuadraticModel, Binary, Integer
>>> from luna_model.translator import CqmTranslator
>>> cqm = ConstrainedQuadraticModel()
>>> x = Binary("x")
>>> y = Integer("y", lower_bound=0, upper_bound=10)
>>> cqm.set_objective(x + 2 * y)
>>> _ = cqm.add_constraint(x + y >= 3, label="min_sum")
>>> model = CqmTranslator.to_lm(cqm, name="example")
>>> print(model.objective)
x + 2 y
>>> print(len(model.constraints))
1

Convert LunaModel to D-Wave CQM.

Converts a LunaModel Model to a D-Wave ConstrainedQuadraticModel.

Parameters:

• model (Model) –

  The model to convert. Should have: - At most quadratic objective and constraints

Returns:

• ConstrainedQuadraticModel –

  D-Wave CQM ready for use with D-Wave hybrid solvers.

Raises:

• RuntimeError –

  If dimod package is not installed.

• TranslationError –

  If the model contains constructs not supported by CQM format.

Examples:

>>> from luna_model import Model, Variable, Vtype
>>> from luna_model.translator import CqmTranslator
>>> model = Model(name="knapsack")
>>> x = [model.add_variable(f"x{i}", vtype=Vtype.BINARY) for i in range(5)]
>>> values = [10, 15, 20, 25, 30]
>>> weights = [2, 3, 4, 5, 6]
>>> # Maximize value
>>> model.objective = sum(v * x[i] for i, v in enumerate(values))
>>> # Weight constraint
>>> model.constraints += sum(w * x[i] for i, w in enumerate(weights)) <= 10
>>> cqm = CqmTranslator.from_lm(model)

Convert NumPy arrays to LunaModel solution.

Parameters:

• result (NDArray) –

  2D array of solution samples, shape (n_samples, n_variables).

• energies (NDArray) –

  1D array of energy/objective values, shape (n_samples,).

• timing (Timing, default: None ) –

  Timing information for the solution process.

• env (Environment, default: None ) –

  Environment for mapping array indices to variable names. Required either as parameter or active context.

Returns:

• Solution –

  LunaModel Solution with samples and energies.

Convert D-Wave SampleSet to LunaModel solution.

Parameters:

• sample_set (SampleSet) –

  D-Wave SampleSet returned by a sampler.

• timing (Timing, default: None ) –

  Timing information for the solution process.

• env (Environment, default: None ) –

  Environment for variable mapping. Required either as parameter or active context.

Returns:

• Solution –

  LunaModel Solution with aggregated samples.

Raises:

• RuntimeError –

  If dimod package is not installed.

Convert IBM Qiskit result to LunaModel solution.

Parameters:

• result (PrimitiveResult[PubResult]) –

  Qiskit primitive result containing measurement outcomes.

• quadratic_program (QuadraticProgram) –

  Original QuadraticProgram used to evaluate objective values.

• timing (Timing, default: None ) –

  Timing information for the solution process.

• env (Environment, default: None ) –

  Environment for variable mapping. Required either as parameter or active context.

Returns:

• Solution –

  LunaModel Solution with bitstring counts and energies.

Raises:

• RuntimeError –

  If qiskit or qiskit-optimization packages are not installed.

Convert Amazon Braket result to LunaModel solution.

Parameters:

• aws_result (dict[str, Any]) –

  Amazon Braket result dictionary containing 'samples' and 'energies', default None.

• timing (Timing, default: None ) –

  Timing information for the solution process.

• env (Environment, default: None ) –

  Environment for variable mapping. Required either as parameter or active context.

Returns:

• Solution –

  LunaModel Solution with samples and energies.

Convert Q-Ctrl result to LunaModel solution.

Parameters:

• result (dict[str, Any]) –

  Q-Ctrl result dictionary containing 'final_bitstring_distribution' and 'variables_to_bitstring_index_map'.

• timing (Timing, default: None ) –

  Timing information for the solution process.

• env (Environment, default: None ) –

  Environment for variable mapping. Required either as parameter or active context.

Returns:

• Solution –

  LunaModel Solution with bitstring counts.

Convert SCIP solution to LunaModel solution.

Parameters:

• model (Model) –

  A solved SCIP model (from pyscipopt).

• timing (Timing, default: None ) –

  Timing information for the solution process.

• env (Environment, default: None ) –

  Environment for variable filtering. Required either as parameter or active context. Only variables in the environment are included in the solution.

Returns:

• Solution –

  LunaModel Solution with variable values from SCIP.

Raises:

• RuntimeError –

  If pyscipopt package is not installed.