Changelog
List of changes in-between PyJibe releases.
version 0.15.0
BREAKING CHANGE: Due to the upgrade of nanite, the rating values, e.g. for the “zef18” dataset might change slightly (on average by about 0.1). If you are excluding curves from your analysis based on rating, please rerun your analysis.
setup: bump nanite from 3.7.0 to 4.0.0 (support training sets that contain nan-values for zero responses and infinity values; minor other fixes)
version 0.14.3
setup: bump afmformats from 0.16.7 to 0.18.0 (support for specifying custom metadata when opening JPK files)
ci: maintenance
version 0.14.2
enh: support mirroring the input directory structure in the data conversion tool (#29)
docs: add info about data conversion tool
version 0.14.1
maintenance release
version 0.14.0
feat: support baseline corrections with linear slope (drift or shift)
ref: generalize the view of preprocessing insights
ref: support multiple options for nanite preprocessors
setup: bump nanite from 3.5.2 to 3.7.0
setup: bump Python version to 3.10
version 0.13.4
ref: don’t use deprecated matplotlib.cm.get_cmap
ref: use packaging.version.parse to compare versions
ref: fix matplotlib deprecation warning for set_data
tests: don’t be so strict
version 0.13.3
setup: bump afmformats from 0.16.4 to 0.16.7
setup: bump nanite from 3.5.0 to 3.5.2
setup: require Python 3.9
ci: update build pipelines
tests: loosen allclose call
version 0.13.2
fix: model parameters defined via expressions were not exported
fix: only export hidden parameters in developer mode
enh: export ancillary parameters in autosave
version 0.13.1
fix: make sure fit parameters are identified by their label instead of by an index (which can be wrong, now that we have hidden parameters)
fix: do not show “sneddon_spher” model in expert mode
version 0.13.0
feat: allow to specify the geometrical correction factor for non-single-contact experiments (#16)
fix: export fitting method and method kwargs with initial parameters
enh: add expert mode which is just a parital developer mode (#22)
enh: add links to docs in two places (#19)
enh: show “Axes” in fitting tab only in developer mode (#21)
setup: bump nanite from 3.4.0 to 3.5.0 (#16)
version 0.12.1
build: fixed testing pipeline
version 0.12.0
feat: support loading nanite model files as extensions
feat: added automatic update check during startup
ref: add preferences dialog
setup: bump nanite from 3.3.1 to 3.4.0
version 0.11.2
maintenance release
version 0.11.1
fix: check for identical label when updating parameters (previously, a parameter that starts with the label of another parameter got assigned the wrong boundaries)
enh: allow to hide model parameters via key names (#20)
enh: show hidden parameters in developer mode
ref: code-cleanup and simplification using latest nanite version
setup: bump nanite from 3.2.1 to 3.3.1
version 0.11.0
feat: allow custom fit methods with custom keyword arguments in developer mode (#18)
fix: correctly handle dependencies in preprocessing tab
fix: set POC method when loading recommended preprocessing (#17)
setup: bump nanite from 3.1.3 to 3.2.1 (new POC estimation methods: Frechet distance and linear+polynomial fit)
version 0.10.1
enh: use human-readable units in missing-metadata entry dialogs
enh: change default unit for spring constant to mN/m
ref: IndentationPreprocessor is deprecated in nanite
setup: bump nanite to version 3.1.3 (ask for missing spring constant for tip position computation)
version 0.10.0
BREAKING CHANGE: The default contact point estimation method changed. This means that your fitted contact points and any E(delta)-related results will change!
feat: allow to select point of contact method (#15)
enh: restructure preprocessing tab and avoid situations where preprocessing dependencies are not met (#15)
enh: visualize contact point estimation algorithms in preprocessing tab
setup: bump nanite from 2.0.0 to 3.1.1 (improved preprocessing)
version 0.9.4
enh: only show exact sneddon model “sneddon_spher” in developer mode to avoid confusion
setup: bump afmformats from 0.15.0 to 0.16.0 (improve support for AFM workshop data)
setup: bump nanite from 1.7.6 to 2.0.0
version 0.9.3
fix: possible fix for TypeError in InfDoubleSpinBox
setup: bump afmformats from 0.14.3 to 0.15.0 as well as nanite from 1.7.6 to 1.7.8 (requirement for #14)
version 0.9.2
ref: improve speed when displaying QMap data
enh: add developer mode for loading creep-compliance data (#14)
setup: bump afmformats from 0.13.3 to 0.14.3 (mostly speed, support AFM workshop maps)
setup: bump nanite from 1.7.4 to 1.7.6 (mostly speed)
ref: speed-up QMap visualization by caching nanite.QMap instances
version 0.9.1
docs: minor update
fix: regression where left fitting range could not be set
fix: improved checks for InfDoubleSpinBox which acted up when entering absolutely sane floating point values
version 0.9.0
feat: allow users to enter missing metadata during loading of data files (#5)
fix: automatically append .h5 suffix when rating datasets
enh: add busy cursor where applicable
ref: directory for importing training datasets for rating changed (but I think this currently affects nobody)
setup: dump appdirs dependency (moved to QSettings)
setup: bump afmformats from 0.12.5 to 0.13.3 (mostly #5 and speed)
setup: bump nanite from 1.7.3 to 1.7.4 (mostly #5)
tests: increase coverage
version 0.8.8
fix: handle bad curves properly when running bulk analysis (#13)
ref: make sure there are no duplicate files when loading data (possibly fixes #12)
ref: cleanup - call overloaded processEvents
ref: properly decorate pyqtSlots
tests: add helper function for setting up datasets
version 0.8.7
fix: add workaround for macOS where the non-native PyQt5 dialogs did not allow browsing to “/Volumes” (it is now added to the sidebar)
build: bump afmformats from 0.12.4 to 0.12.5 (OSError: “Too many files open” with the JPK file format)
build: bump nanite from 1.6.2 to 1.7.3
build: pin PyQt5 for GitHub releases
tests: don’t used deprecated weight_cp function from nanite
version 0.8.6
build: bump afmformats from 0.12.2 to 0.12.4 (AttributeError when reading HDF5 files)
tests: deprecate setup.py test
version 0.8.5
build: migrate from travisCI to GitHub Actions
build: overhauled build pipeline (#9)
version 0.8.4
build: add pyinstaller hooks for afmformats
version 0.8.3
fix: main window not focused after startup
fix: Windows installation location was confusing and did not coincide configuration file location
fix: files are sorted before batch loading
fix: change visualization of fitting region by gray wraparound instead of yellow highlight (not visible on all screens)
setup: bump afmformats from 0.11.0 to 0.12.2 (improved JPK file format speed)
version 0.8.2
setup: bump matplotlib to >=3 (NavigationToolbar2QT modifications)
version 0.8.1
fix: correctly handle lmfit models with expressions
setup: bump nanite from 1.6.0 to 1.6.2 (improve contact point retrieval for bad data, check models during registration)
version 0.8.0
fix: exclude misplaced ‘available’ method from preprocessors
setup: bump nanite from 1.5.5 to 1.6.0 (improved contact point estimation - this may introduce small changes in the fit result which should not be significant, but which justify a new minor version)
version 0.7.6
maintenance release (fix macOS build)
setup: bump afmformats from 0.10.2 to 0.11.0
setup: bump nanite from 1.5.4 to 1.5.5
version 0.7.5
maintenance release (fix build)
version 0.7.4
setup: bump afmfomats from 0.10.0 to 0.10.2
setup: bump nanite from 1.5.2 to 1.5.4
build: add command-line option –version to just print the version and exit
version 0.7.3
fix: update initial parameters when the user edits the left fitting interval boundary
fix: conical indenter model did not work, because of a wrong (and unnecessary) entry in the human_units scheme.
version 0.7.2
fix: setting bad fit range lead to ValueError during plotting
version 0.7.1
fix: wrong SI units assigend to fit results parameters
setup: bump afmformats from 0.9.0 to 0.10.0 (new formats: JPK (.jpk-qi-data), AFM workshop (.csv), NT-MDT (.txt exported by Nova))
setup: bump nanite from 1.4.1 to 1.5.2
version 0.7.0
feat: add data conversion tool
enh: allow to select multiple files when loading data
ref: changed order and clarified meaning of options in file menu
setup: bump afmformats from 0.7.0 to 0.9.0 (fix HDF5 export, support new file formats)
version 0.6.8
setup: bump afmformats from 0.5.1 to 0.7.0 (improved metadata view in info tab)
setup: bump nanite from 1.4.1 to 1.5.1 (compatibility to afmformats)
enh: fit results and parameter export dialog now supports new groups in afmformats 0.7.0
enh: autosave now only stores fit results and rating parameters
enh: remember scroll position in info tab
version 0.6.7
setup: bump nanite from 1.4.0 to 1.4.1 (baseline is now a free parameter by default)
docs: add first tutorial (PAAm gels)
version 0.6.6
fix: typo lead to error when using rating threshold
version 0.6.5
build: macOS build failed due to PyInstaller issue 4626
version 0.6.4
fix: info tab did not display QMap metadata
fix: do not fit the entire dataset when applying a rating threshold but take the values from previous ratings
fix: deselecting curves of a qmap resulted in ValueError if QMap tab was selected
fix: set correct display range for residuals
docs: UI section for basics FD analysis
build: Windows build broken since 0.5.6
version 0.6.3
fix: support JPK data recorded in the “force-modulation” feedback mode (bump afmformats from 0.5.0 to 0.5.1)
fix: make sure ancillary parameters are computed from the initial parameters set in the user interface (not from the default model parameters)
version 0.6.2
enh: allow to select which metadata is exported
fix: do not apply and fit to all before exporting metadata (user may have performed individual fits)
setup: bump nanite from 1.3.0 to 1.4.0
version 0.6.1
build: workaround for Pyinstaller issue 4626
version 0.6.0
feat: initial support for ancillary parameters
fix: always display parameter units
fix: other minor UI bugs
setup: bump afmformats from 0.4.1 to 0.5.0
setup: bump nanite from 1.2.4 to 1.3.0
version 0.5.7
enh: allow to also set right part of fitting range individually
fix: improve layout of FD fitting panel
setup: bump nanite from 1.2.3 to 1.2.4 (improved default params)
version 0.5.6
setup: bump nanite from 1.2.2 to 1.2.3 (fixes issue with fits not being redone when the user changes an initial parameter)
fix: IndexError when editing the fitting range
version 0.5.5
maintenance release
version 0.5.4
enh: write maximum indentation to statistics file (#3)
fix: unicode characters were not read correctly from statistics file by libre office (added UTF-8 BOM)
version 0.5.3
enh: improved visualization of meta data parameters in the Info tab
setup: bump afmformats from 0.3.0 to 0.4.1
setup: bump nanite from 1.2.0 to 1.2.2
version 0.5.2
ref: bump nanite to 1.2.0
ref: migrate to afmformats (0.3.0)
version 0.5.1
fix: missing title in FD window and missing FD number in export menu
enh: add help menu with link to docs, about, and software used (#1)
docs: add quick guide for importing a nanite training set
docs: add black/white logo
version 0.5.0
feat: allow to import training sets produced with nanite
ref: major code refactoring of force-distance interface and cleanup
setup: bump nanite dependency to >=1.1.1
version 0.4.4
ci: pin ‘joblib==0.11.0’ on travis-CI (workaround for infinite loop in macOS build, https://github.com/pyinstaller/pyinstaller/issues/4067)
docs: add installation instructions
version 0.4.3
maintenance release (migrate to GitHub infrastructure)
experimental macOS builds
version 0.4.2
enh: merge training set and regressor selection into scheme selection (prevents confusion for users that are unsure what to choose)
enh: user rating is saved during rating, not afterwards
enh: rating containers for user rating can be reused explicitly
enh: qmap plot now better distinguishes between unavailable and invalid data
ref: migrate user rating output to nanite
version 0.4.1
fix: autoscaling according to fitting range did not work
fix: not possible to load .jpk-force-map files
ref: migrate qmap generation to nanite
version 0.4.0
migration to GitHub