v2.7.6

• fix
  • solveLP use specified parameters during flux minimization
  • optimizeProb use default feasTol and optTol e-9 (instead of e-6), otherwise raising issues in Gecko (Light)
  • addRxns, addRxnsGenesMets, getModelFromHomology various minor bugs
  • exportModel revert 3996e6a, as it resulted in warnings when validating SBML files
  • changeGeneAssoc when genes are already in model (solves #440)
• feat
  • checkInstallation reports system architecture
  • getAllowedBounds reports progress
  • randomSampling more detailed and accurate progress reporting
• refactor
  • optimizeProb speed-up by avoid running ispref if not required

v2.7.5

• fix:
  • fatal bug in checkTasks (closes #434)

v2.7.4

• fix:
  • resolves bug in runINIT with csense vector
  • resolves bug in changeRxns if equations are provided as structure
• chore:
  • update KEGG mat files and trained HMMs to release 102
  • copied curation of getGenesFromGrRules from human-GEM (part of SysBioChalmers/Human-GEM#392)

v2.7.3

• refactor:
  • constructS is about 3-times faster
• fix:
  • set solver coherently as preference and global variable (closes #425)

v2.7.2

• feature
  • RAVEN can now also be installed via MATLAB File Exchange, or even simpler, via MATLAB Add-Ons Explorer
  • ftINIT fast tINIT algorithm
  • contractModel removes _EXP_* suffixes if present in all duplicated reactions, as these reactions were derived from expandModel
  • findGeneDeletions gives optional grRatioMuts output, with the growth ratio of the mutant vs. wild-type
• refactor
  • optimizeProb to match the new ftINIT function
  • remove MATLAB-git dependency in exportForGit
  • set solver as global variable for speed-up of multiple optimizeProb runs
• fix
  • addMets accepts single compartment if it is the same for all metabolites (see ff462fd#r74565757)
  • expandModel adds _EXP_* suffixes to all expanded reactions, also the first one (it previously skipped _EXP_1)
  • removeGenes accepts numeric and logical input

v2.7.1

• fix
  • correctly parse relative paths to model files on Mac/Unix (e.g. by importModel)
• feature
  • checkInstallation checks if a new RAVEN version is available on GitHub, and gives instructions on how to proceed

v2.7.0

• feature
  • provide libSBML and glpk binaries and force their usage (instead of others present in MATLAB path) (#408 and #409)
  • exportForGit has option to write text files with metabolite names (instead of IDs), which is more informative when used in diff-ing model changes (particularly when yaml file is also generated)
  • numerous unit-tests included (contributing to #414)
• refactor
  • format of tests reports in PRs (#411)
  • parsing of (MI)LP to solvers via optimizeProb (#408)
  • layout of checkInstallation and use unit_tests (#408)
  • allow text input as either character vector, string array, or array of either of those (solves #397)
• fix
  • various minor bugs, more informative error messages and improved documentation in sortModel, printFluxes, addRxns, addMets, compareMultipleModels, loadWorkbook

v2.6.2

• fix:
  • addRxns new subSystems should be cell arrays

v2.6.1

• fix
  • checkInstallation close version.txt after reading
  • keggRxns.mat & getRxnsFromKEGG empty subSystems in KEGG model should be cells, not strings
  • dispEM parse new line at end of error text
  • writeYaml correct indexing
  • .gitattributes specify mafft-linux64 as binary, to prevent EOL changes that can corrupt running it via WSL on Windows
  • exportModel: remove unnecessary warning regarding "absence of unconstrained metabolites" (= boundary metabolites). This is not problematic at all, and also very common for models
  • exportModel: remove useless note in SBML file about the file being exported from RAVEN. This was added to SBML files if no model.annotation field could be found, but it results in a warning in SBML Validator
  • addMets can take metNotes field
  • addRxns properly formats subSystems
  • changeGrRules does not try to add empty gene lists
• feat
  • getModelFromKEGG indicate progress during slowest step
  • new editMiriam function that can edit existing MIRIAM annotations, add new ones, replace, etc.
  • new miriamTests for the editMiriam and extractMiriam functions
• refactor
  • exportModel remove duplicate warning about missing objective function
  • getModelFromKEGG minor speed improvement
  • extractMiriam remove repetitive prefix (e.g. bigg.metabolite) from extractedMiriams output as it is always deleted after this function is called, while this info is also already in extractedMiriamNames

v2.6.0

• fix:
  • exportModel correct handling if a compartment ID is i (solves #366)
  • functions that use a species FASTA file as input (e.g. getBlast) will copy this FASTA file to a temporary directory to avoid potential issues with the original path when running external software (e.g. BLAST) (PR #368)
  • rename compareModels to compareRxnsGenesMetsComps (PR #388)
  • speed-up convertToIrrev (PR #391)
  • getKEGGModelForOrganism now checks organismID validity (PR #392)
• feat:
  • pretrained HMM sets updated to KEGG version 100. Only 90% sequence identity sets will be distributed from now on (PR #395)
  • CD-HIT and MAFFT run on Windows via Windows Subsystem for Linux (PR #395)
  • sortIdentifiers sorts metabolites, reactions, genes and compartments alphabetically by their identifier, and can be set as option (sortIds, default = false) when exporting models (PR #364)
  • allow getKEGGModelForOrganism to take a preloaded model from getModelFromKEGG, to speed up generating many models for different species (PR #369)
  • add unit tests for external software (PR #377)
  • checkModelStruct includes check for objective function (PR #379)
  • automated testing via GitHub Actions (PR #384)
• doc:
  • checkInstallation: remove reference to mosek, and output MATLAB release (PR #373)
  • updateDocumentation removes keggModel.mat (PR #389)
• chore:
  • rename stabile branch to main (PR #383)