cvxpy_problem_data_structure

 problem = cvxpy_problem_data_structure(network, q_info, prior, data, bounds, true_model, cmb_options)

 Convert a model balancing problem (data structures: network, q_info, true_model, prior, data) 
 into a single data structure, exported to json and then used by python model balancing tool (based on CVXpy)

 Structure, default values, and matrix sizes for a network with nr reactions and nm metabolites (and ns metabolic states)
   .standard_concentration: '1 mM'
   .state_names:            {ns x 1 cell}
   .use_pseudo_values:      Boolean

   .network
   .network.metabolite_names:      {nm x 1 cell}
   .network.reaction_names:        {nr x 1 cell}
   .network.stoichiometric_matrix: [nm x nr double]  (sparse matrix)
   .network.activation_matrix:     [nm x nr double]  (sparse matrix)
   .network.inhibition_matrix:     [nm x nr double]  (sparse matrix)

   .kinetic_constants
   .kinetic_constants.Keq
   .kinetic_constants.Kcatf
   .kinetic_constants.Kcatr
   .kinetic_constants.KM
   .kinetic_constants.KA
   .kinetic_constants.KI

     .. where each subfield of .kinetic_constants is a struct of the form
       .unit:              string, typically 'mM' and '1/s'
       .true:              (column vector or matrix; true values in the case of artificial data; otherwise empty)
       .pprior_ln.mean     pseudoprior, mean vector of log values
       .pprior_ln.cov      pseudoprior, covariance matrix of log values
       .bounds_ln.min      lower bounds on log values
       .bounds_ln.max      upper bounds on log values
       .data_ln.mean       data mean vector of log values
       .data_ln.cov        data covariance matrix of log values
       .bounds.min         lower bounds on log values
       .bounds.max         upper bounds on log values
       .combined.geom_mean preposterior geometric mean vector
       .combined.mean_ln   preposterior mean vector of log values
       %.combined.cov_ln   preposterior covariance matrix of log values

   .kinetic_constants.all
   .kinetic_constants.all.names      parameter names, list of strings
   .kinetic_constants.all.geom_mean  preposterior geometric mean vector
   .kinetic_constants.all.mean_ln    preposterior mean vector of log values
   .kinetic_constants.all.prec_ln    preposterior precision matrix of log values

   .metabolite_concentrations
   .metabolite_concentrations.unit:              string, default 'mM'
   .metabolite_concentrations.true:              matrix of true values (only in models with artificial data; otherwise [])
   .metabolite_concentrations.prior_ln.mean      matrix of prior mean values for metabolite log-concentrations
   .metabolite_concentrations.prior_ln.std       matrix of prior std dev for metabolite log-concentrations
   .metabolite_concentrations.bounds_ln.min      matrix of lower bounds on metabolite log-concentrations
   .metabolite_concentrations.bounds_ln.max      matrix of upper bounds on metabolite log-concentrations
   .metabolite_concentrations.data_ln.mean       matrix of data  mean values for metabolite log-concentrations
   .metabolite_concentrations.data_ln.std        matrix of data  std dev for metabolite log-concentrations
   .metabolite_concentrations.bounds.min         matrix of preposterior geom mean for metabolite concentrations
   .metabolite_concentrations.bounds.max         matrix of preposterior geom std  for metabolite concentrations
   .metabolite_concentrations.combined.geom_mean matrix of preposterior geom mean for metabolite concentrations
   .metabolite_concentrations.combined.geom_std  matrix of preposterior geom std  for metabolite concentrations

   .enzyme_concentrations
    ...
    (structure analogous to .metabolite_concentrations; note that lower bounds for log(e) are typically -inf)

   .reaction_fluxes
   .reaction_fluxes.unit:          string, default 'mM/s'
   .reaction_fluxes.true:          (matrix of true values in the case of artificial data; otherwise empty)
   .reaction_fluxes.data.mean      matrix of flux data mean values
   .reaction_fluxes.data.std       matrix of flux data std values

 
 Note:
   o activations and inhibitions are described by separate (positive) matrices of the shape #metabolites x #reactions
   o convention for order of KM, KA, and KI entries (in vectors) + their relation to entries in the matrix:
     vector elements are ordered by their appearance in the respective (stoichiometric or regulation) matrix, in the format #reactions x # metabolites,
     and following matlab's convention for indices in matrix elements (first column from top to bottom, then second column etc)
     in python this corrresponds to matrix formatted as #metabolites x #reactions, and using the opposite convetion for indices 
     (first row, then second row, etc)
   fields called "combined" contain the preposterior distribution of the respective quantities