Skip to content

Releases: matthiaskoenig/sbmlsim

0.2.2

27 Sep 12:39
Compare
Choose a tag to compare

Release notes for sbmlsim 0.2.2

  • bugfix release

0.2.0

22 Mar 11:38
Compare
Choose a tag to compare

Release notes for sbmlsim 0.2.0

  • breaking changes in units handling
  • bugfix plotting dimensionless units in labels
  • remove deprecated distrib and sampling functionality.
  • fixed typo in sensitivity

parameter fitting

  • refactoring parameter fitting
  • improved weight and options handling
  • documentation
  • reports for parameter fitting (fit mapping contribution, #79)
  • storing and combination of parameter fitting results (#73)
  • better cost plots (#72)
  • improved parameter fitting results plots (#96)
  • loss functions implemented (#99)
    • ‘linear’ (default) : rho(z) = z. Gives a standard least-squares problem.
    • ‘soft_l1’ : rho(z) = 2 * ((1 + z)**0.5 - 1). The smooth approximation of l1 (absolute value) loss. Usually a good choice for robust least squares.
    • ‘cauchy’ : rho(z) = ln(1 + z). Severely weakens outliers influence, but may cause difficulties in optimization process.
    • ‘arctan’ : rho(z) = arctan(z). Limits a maximum loss on a single residual, has properties similar to ‘cauchy’.
  • figures closing in reports (#88)
  • improved parameter fitting plots and results (#32)
  • bugfix all NaN errors (#101)

0.1.14

01 Mar 00:02
Compare
Choose a tag to compare

Release notes for sbmlsim 0.1.14

  • Fix #81, feature timecourse concatenation and reuse in timecoursesim

This allows simplified creation of multiple dosing timecourse simulations and reuse
of timecourse definitions in multiple simulations. E.g. repeated timecourse [tc]*3

    simulator = Simulator(MODEL_REPRESSILATOR)
    tc = Timecourse(
                    start=0,
                    end=50,
                    steps=100,
                    changes={"X": 10})

    s = simulator._timecourse(
        simulation=TimecourseSim(
            [tc]*3
        )
    )
  • Fix #82, bugfix; remove time shift from discarded simulations
  • Fix #83, allow empty timecourses in timecoursesim
  • Fix #86, dependency update ray>=1.2.0
  • Fix #87, weighting by counts in parameter fitting
  • Fix #88, bugfix figure closing in parameter fitting
  • adding mypy configuration and initial mypy fixes
  • many flake8 fixes and type annotations
  • Fix #91, support hex colors in plot
  • Fix #94, update xarray>=0.17.0

0.1.13

04 Feb 21:47
Compare
Choose a tag to compare

Release notes for sbmlsim 0.1.13

  • Fix #51, feature discarding pre-simulation
  • Fix #76, Package templates in pip packages
  • dependency update
	python-libsbml-experimental>=5.19.0
	python-libsedml>=2.0.15
	python-libcombine>=0.2.7
	python-libnuml>=1.1.3
	libroadrunner>=2.0.4
    ray>=1.1.0
  • Fix #77, updated github actions
  • Fix #78, contributing information and depinfo

0.1.12

01 Feb 12:06
Compare
Choose a tag to compare

Release notes for sbmlsim 0.1.12

  • Fix #51, feature discarding pre-simulation
  • Fix #76, Package templates in pip packages
  • dependency update
	python-libsbml-experimental>=5.19.0
	python-libsedml>=2.0.15
	python-libcombine>=0.2.7
	python-libnuml>=1.1.3
	libroadrunner>=2.0.4
    ray>=1.1.0

0.1.11

11 Nov 22:00
Compare
Choose a tag to compare

Release notes for sbmlsim 0.1.11

  • update to ray ray>=1.0
  • working on
  • minor bugfixes
  • support for model changes with initial assignments

0.1.10

05 Nov 11:48
Compare
Choose a tag to compare

Release notes for sbmlsim 0.1.10

  • better handling of model changes and model manipulations
  • working on
  • Minor bugfixes

0.1.9

03 Nov 09:07
Compare
Choose a tag to compare

Release notes for sbmlsim 0.1.9

  • Fixed regression in unit serialization
  • Minor bugfixes

0.1.8

26 Oct 15:47
Compare
Choose a tag to compare

Release notes for sbmlsim 0.1.8

  • Fix #64 support of model changes (with unit normalization)
  • minor bugfixes

0.1.7

04 Oct 22:35
Compare
Choose a tag to compare

Release notes for sbmlsim 0.1.7

  • fixed packaging of data
  • support for count from DataSets (#50)
  • improved plotting serialization, adding order on curves
  • plotting bugfixes (better handling of labels and legends, default colors serialized, #56)