CHANGELOG

0.4.1: 2020/09/14

  • Modify setup.py (MR !82):

    • Add opencv-python, scikit-image and silx in install_requires.

    • Add package full in extras with matplotlib, PyQt5 and orange.

    • Add package test in extras with pillow.

  • Core:
    • dataset.py:

      • Add parameter bins when partitioning data in Dataset (MR !67)

      • Create in_memory setter in Dataset (MR !68)

      • Improve methods flatten, reshape and take in Data (MR !67)

      • Add try-except when fitting data in blind source separation methods to close file in case of crash (MR !71)

      • Add methods to convert dimensions to dictionary and viceversa in AcquisitionDims (MR !74)

      • Add to_memory method in Dataset to load only part of the data into memory (MR !77)

      • Add property ndim in Data (MR !80)

  • Decomposition:

    • Fix bug in IPCA with cases where in the last iteration the chunk was smaller than the number of components (omit last iteration) (MR !73)

  • GUI:

    • Temporarily register resources directory in when used in DisplayComponentsWidget (MR !84)

    • Change colormap default normalization to log for better visualization of the data (MR !85)

    • Add widget for bins in DataPartitionWidget (MR !67)

    • Add parameter for chunks in NoiseRemovalWidget (MR !68)

    • Add abort options in dataPartitionWidget, NoiseRemovalWidget and ShiftCorrectionWidget (MR !68)

    • Get next free axis when removing dimensions in `DimensionsWidget (MR !86)

    • Fix bug when clearing stack in RoiSelectionWidget (MR !67)

    • Fix bug when getting filtered indices in DataPartitionWidget (MR !68)

  • IO:

    • New file dataset_io.py with functions to save and load data into/from json files (MR !74)

  • Orangecontrib:

    • Remove progressBar due to bugs in the workflows (MR !84)

    • Fix bug with qthreads when creating and deleting the thread using orange widgets (MR !70)

0.4.0: 2020/05/15

  • Refactors the project (MR !48)

    • Changes default colormap to cividis.

    • Core:

      • Dataset class stops inheriting from silx.qt.QObject

      • The dataset class takes control of all the operations done to the data.

      • Metadata is obtained from the frames using fabioh5 from silx and then each frame is closed to spare space in memory.

      • Creates new class Data that inherits from numpy.ndarray and contains the corresponding urls and metadata of the data.

      • Adds new methods to Dataset for image operations, image registration, region of interest, and partition by intensity (filter data), that are applied differently depending on if the data is loaded into memory or is taken from disk in chunks.

      • Data can be saved into an Hdf5 file with images in the rows.

      • Implement in Data methods shape, reshape, flatten, take, __getitem__ that work similarly to a numpy.ndarray.

      • Adds and modifies functions in imageOperations.py and imageRegistration.py to be used when data is not loaded into memory.

      • Adds class IPCA in blindSourceSeparation.py that uses IncrementalPCA from scikit-learn to apply PCA in chunks.

      • Implements methods in Dataset to apply blind source separation algorithms to the data.

      • PCA and NMF algorithms used from scikit-learn when data is loaded into memory.

    • Decomposition:

      • A new package decomposition is created.

      • A base class Base is created for decomposition methods.

      • Adds class NMF that updates the matrices H and W using NMF method without having the data in memory.

      • Adds class NICA with the already implemented algorithm in core.

      • Adds class PCA currently not used (substitued for PCA in scikit-learn).

    • GUI:

      • Modifies all widgets to only have Dataset objects and call its corresponding methods.

  • Core:

    • Adds method in_memory to Dataset to upload data from disk or remove data from memory (MR !51)

    • Adds method that returns intensity per frame (MR !58)

    • Number of bins can be passed as parameter when partitioning the data by intensity (MR !52)

    • Fix bugs when not closing Hdf5 files (MR !56)

    • Adds methos zsum in Dataset to sum images when not in memory (MR !57)

    • Fix bugs with __getitem__ in Data (MR !57)

  • Decomposition:

    • Adds IPCA to decomposition package (MR !50)

    • NICA implements IPCA for whitening when data is not in memory. Fix bugs when computing IPCA with data in the rows or in the columns (MR !53)

    • Allow matrices H and W in IPCA to be stored in disk using Hdf5 (MR !54)

  • GUI:

    • Adds ComboBox in DatasetSelectionWidget to use data from disk or load it into memory (MR !57)

    • Adds widget to partition the data by intensity (MR !58)

    • Removes ComboBox in DatasetSelectionWidget to filter the data (MR !59)

  • Orangecontrib:

    • Add corresponding data partition widget (MR !58)

    • Update example tutorial (MR !64)

0.3.0: 2020/03/03

  • Adds resources package (MR !29)

  • Adds configuration file for default parameters (MR !30)

  • Minor fixes (MR !12 !14 !15 !32 !37 !41 !46)

  • Use silx backend when available (MR !44)

  • Core:

    • Adds tolerance to dimensions that defines the uniqueness of its values (MR !17)

    • Modifies ROI performance (MR !23)

    • Implements blind source separation algorithms (MR !18 !20 !22 !28)

    • Implements algorithms for linking components between datasets: ORB, SIFT, Euclidean distance (MR !36 !44 !45)

  • GUI:

    • Creates a unified noise removal widget (MR !21 !24 !25)

    • Adds checkbox to filter the data in DataSelectionWidget (MR !19)

    • Improves DimensionWidget (MR !16 !17 !33)

    • Improves roiSelectionWidget (MR !23 !31 !43)

    • Implements widgets blindSourceSeparationWidget and displayComponentsWidget to display the components (MR !18 !22 !26 !29 !35)

    • Creates PCAWidget to show the eigenvalues of the PCA computation (MR !22)

    • Creates widget LinkComponentsWidget that implements the component matching algorithms from two sets of components (MR !36 !38)

  • IO:

    • Creates functions for reading and writing the components of the BSS into Nexus files (MR !42)

0.2.0: 2019/10/15

  • Renames project linkdick06 to darfix

  • Refactors project into core, gui, io and test packages

  • Adds unittests

  • Adds orangecontrib package

  • Adds examples package

  • Adds continous integration with gitlab-ci

  • Adds requirements.txt file with dependencies on numpy, silx, PyQt5, opencv-python, scikit-image and orange3 with version 3.22.0

  • Core:

    • Dataset class inherits from silx.qt.QObject

    • Adds threshold to split data into high intensity images and low intensity images

    • Adds Dimension class to save information about how the frames are taken through the dataset

    • Adds roi file that implements region of interest functions

    • Image registration operations moved to new file imageRegistration.py

    • Adds shift detection using OpenCV

    • Adds autofocus.py to use at the shift detection

    • Adds geneticShiftDetection file to improve shift detection using a genetic algorithm

  • GUI:

    • Adds RoiSelectionWidget

    • Adds ShiftCorrectionWidget to apply shift detection and shift correction

    • Adds widgets BackgroundSubstractionWidget and HotPixelRemovalWidget for noise removal

    • Adds widgets for data selection, metadata, and choosing dimensions

    • Adds a QThread, operationThread, to run widgets operations in a thread

    • Adds utils to create datasets from scratch

    • Removes frameInterpretationWidget and adds DimensionWidget

  • IO:

    • Adds utils

0.1.0: 2019/07/11

  • Adds project build, documentation and test structure

  • Adds continuous integration set-up for Gitlab

  • Adds EDF reader from Fabio

  • Adds operations for correction of images in imageOperations

  • Adds notebook tutorial