14.1.7. cobra.util package¶
14.1.7.1. Submodules¶
14.1.7.2. cobra.util.array module¶

cobra.util.array.
constraint_matrices
(model, array_type='dense', include_vars=False, zero_tol=1e06)[source]¶ Create a matrix representation of the problem.
This is used for alternative solution approaches that do not use optlang. The function will construct the equality matrix, inequality matrix and bounds for the complete problem.
Notes
To accomodate nonzero equalities the problem will add the variable “const_one” which is a variable that equals one.
Parameters:  model (cobra.Model) – The model from which to obtain the LP problem.
 array_type (string) – The type of array to construct. if ‘dense’, return a standard numpy.array, ‘dok’, or ‘lil’ will construct a sparse array using scipy of the corresponding type and ‘DataFrame’ will give a pandas DataFrame with metabolite indices and reaction columns.
 zero_tol (float) – The zero tolerance used to judge whether two bounds are the same.
Returns: A named tuple consisting of 6 matrices and 2 vectors:  “equalities” is a matrix S such that S*vars = b. It includes a row
for each constraint and one column for each variable.
 “b” the right side of the equality equation such that S*vars = b.
 “inequalities” is a matrix M such that lb <= M*vars <= ub. It contains a row for each inequality and as many columns as variables.
 “bounds” is a compound matrix [lb ub] containing the lower and upper bounds for the inequality constraints in M.
 “variable_fixed” is a boolean vector indicating whether the variable at that index is fixed (lower bound == upper_bound) and is thus bounded by an equality constraint.
 “variable_bounds” is a compound matrix [lb ub] containing the lower and upper bounds for all variables.
Return type:

cobra.util.array.
create_stoichiometric_matrix
(model, array_type='dense', dtype=None)[source]¶ Return a stoichiometric array representation of the given model.
The the columns represent the reactions and rows represent metabolites. S[i,j] therefore contains the quantity of metabolite i produced (negative for consumed) by reaction j.
Parameters:  model (cobra.Model) – The cobra model to construct the matrix for.
 array_type (string) – The type of array to construct. if ‘dense’, return a standard numpy.array, ‘dok’, or ‘lil’ will construct a sparse array using scipy of the corresponding type and ‘DataFrame’ will give a pandas DataFrame with metabolite indices and reaction columns
 dtype (datatype) – The desired datatype for the array. If not given, defaults to float.
Returns: The stoichiometric matrix for the given model.
Return type: matrix of class dtype

cobra.util.array.
nullspace
(A, atol=1e13, rtol=0)[source]¶ Compute an approximate basis for the nullspace of A. The algorithm used by this function is based on the singular value decomposition of A.
Parameters:  A (numpy.ndarray) – A should be at most 2D. A 1D array with length k will be treated as a 2D with shape (1, k)
 atol (float) – The absolute tolerance for a zero singular value. Singular values smaller than atol are considered to be zero.
 rtol (float) – The relative tolerance. Singular values less than rtol*smax are considered to be zero, where smax is the largest singular value.
 both atol and rtol are positive, the combined tolerance is the (If) –
:param maximum of the two; that is::: :param tol = max(atol, rtol * smax): :param Singular values smaller than tol are considered to be zero.:
Returns: If A is an array with shape (m, k), then ns will be an array with shape (k, n), where n is the estimated dimension of the nullspace of A. The columns of ns are a basis for the nullspace; each element in numpy.dot(A, ns) will be approximately zero. Return type: numpy.ndarray Notes
Taken from the numpy cookbook.
14.1.7.3. cobra.util.context module¶
14.1.7.4. cobra.util.solver module¶
Additional helper functions for the optlang solvers.
All functions integrate well with the context manager, meaning that all operations defined here are automatically reverted when used in a with model: block.
The functions defined here together with the existing model functions should allow you to implement custom flux analysis methods with ease.

exception
cobra.util.solver.
SolverNotFound
[source]¶ Bases:
Exception
A simple Exception when a solver can not be found.

cobra.util.solver.
add_absolute_expression
(model, expression, name='abs_var', ub=None, difference=0, add=True)[source]¶ Add the absolute value of an expression to the model.
Also defines a variable for the absolute value that can be used in other objectives or constraints.
Parameters:  model (a cobra model) – The model to which to add the absolute expression.
 expression (A sympy expression) – Must be a valid expression within the Model’s solver object. The absolute value is applied automatically on the expression.
 name (string) – The name of the newly created variable.
 ub (positive float) – The upper bound for the variable.
 difference (positive float) – The difference between the expression and the variable.
 add (bool) – Whether to add the variable to the model at once.
Returns: A named tuple with variable and two constraints (upper_constraint, lower_constraint) describing the new variable and the constraints that assign the absolute value of the expression to it.
Return type: namedtuple

cobra.util.solver.
add_cons_vars_to_problem
(model, what, **kwargs)[source]¶ Add variables and constraints to a Model’s solver object.
Useful for variables and constraints that can not be expressed with reactions and lower/upper bounds. Will integrate with the Model’s context manager in order to revert changes upon leaving the context.
Parameters:  model (a cobra model) – The model to which to add the variables and constraints.
 what (list or tuple of optlang variables or constraints.) – The variables or constraints to add to the model. Must be of class model.problem.Variable or model.problem.Constraint.
 **kwargs (keyword arguments) – passed to solver.add()

cobra.util.solver.
assert_optimal
(model, message='optimization failed')[source]¶ Assert model solver status is optimal.
Do nothing if model solver status is optimal, otherwise throw appropriate exception depending on the status.
Parameters:  model (cobra.Model) – The model to check the solver status for.
 message (str (optional)) – Message to for the exception if solver status was not optimal.

cobra.util.solver.
check_solver_status
(status, raise_error=False)[source]¶ Perform standard checks on a solver’s status.

cobra.util.solver.
choose_solver
(model, solver=None, qp=False)[source]¶ Choose a solver given a solver name and model.
This will choose a solver compatible with the model and required capabilities. Also respects model.solver where it can.
Parameters:  model (a cobra model) – The model for which to choose the solver.
 solver (str, optional) – The name of the solver to be used. Optlang solvers should be prefixed by “optlang”, for instance “optlangglpk”.
 qp (boolean, optional) – Whether the solver needs Quadratic Programming capabilities.
Returns:  legacy (boolean) – Whether the returned solver is a legacy (old cobra solvers) version or an optlang solver (legacy = False).
 solver (a cobra or optlang solver interface) – Returns a valid solver for the problem. May be a cobra solver or an optlang interface.
Raises: SolverNotFound
– If no suitable solver could be found.

cobra.util.solver.
fix_objective_as_constraint
(model, fraction=1, bound=None, name='fixed_objective_{}')[source]¶ Fix current objective as an additional constraint.
When adding constraints to a model, such as done in pFBA which minimizes total flux, these constraints can become too powerful, resulting in solutions that satisfy optimality but sacrifices too much for the original objective function. To avoid that, we can fix the current objective value as a constraint to ignore solutions that give a lower (or higher depending on the optimization direction) objective value than the original model.
When done with the model as a context, the modification to the objective will be reverted when exiting that context.
Parameters:  model (cobra.Model) – The model to operate on
 fraction (float) – The fraction of the optimum the objective is allowed to reach.
 bound (float, None) – The bound to use instead of fraction of maximum optimal value. If not None, fraction is ignored.
 name (str) – Name of the objective. May contain one {} placeholder which is filled with the name of the old objective.

cobra.util.solver.
get_solver_name
(mip=False, qp=False)[source]¶ Select a solver for a given optimization problem.
Parameters: Returns: The name of feasible solver.
Return type: Raises: SolverNotFound
– If no suitable solver could be found.

cobra.util.solver.
interface_to_str
(interface)[source]¶ Give a string representation for an optlang interface.
Parameters: interface (string, ModuleType) – Full name of the interface in optlang or cobra representation. For instance ‘optlang.glpk_interface’ or ‘optlangglpk’. Returns: The name of the interface as a string Return type: string

cobra.util.solver.
linear_reaction_coefficients
(model, reactions=None)[source]¶ Coefficient for the reactions in a linear objective.
Parameters:  model (cobra model) – the model object that defined the objective
 reactions (list) – an optional list for the reactions to get the coefficients for. All reactions if left missing.
Returns: A dictionary where the key is the reaction object and the value is the corresponding coefficient. Empty dictionary if there are no linear terms in the objective.
Return type:

cobra.util.solver.
remove_cons_vars_from_problem
(model, what)[source]¶ Remove variables and constraints from a Model’s solver object.
Useful to temporarily remove variables and constraints from a Models’s solver object.
Parameters:  model (a cobra model) – The model from which to remove the variables and constraints.
 what (list or tuple of optlang variables or constraints.) – The variables or constraints to remove from the model. Must be of class model.problem.Variable or model.problem.Constraint.

cobra.util.solver.
set_objective
(model, value, additive=False)[source]¶ Set the model objective.
Parameters:  model (cobra model) – The model to set the objective for
 value (model.problem.Objective,) –
e.g. optlang.glpk_interface.Objective, sympy.Basic or dict
If the model objective is linear, the value can be a new Objective object or a dictionary with linear coefficients where each key is a reaction and the element the new coefficient (float).
If the objective is not linear and additive is true, only values of class Objective.
 additive (bool) – If true, add the terms to the current objective, otherwise start with an empty objective.
14.1.7.5. cobra.util.util module¶

class
cobra.util.util.
AutoVivification
[source]¶ Bases:
dict
Implementation of perl’s autovivification feature. Checkout http://stackoverflow.com/a/652284/280182