In older JIPipe versions, all extensions are automatically loaded during the JIPipe startup. This was changed to reduce the number of dependencies by letting the users decide which extensions should be loaded (excluding core extensions that are mandatory).
- New installations will come with the following extensions enabled: Annotations, Filesystem, Forms, ImageJ algorithms, ImageJ integration, Plots, Python, R, Strings, Table operations, Tools, Utils, ImageJ2 integration, Multi-parameter algorithms
- Existing installations will enable the following extensions: Annotations, Filesystem, Forms, ImageJ algorithms, ImageJ integration, Plots, Python, R, Strings, Table operations, Tools, Utils, ImageJ2 integration, Multi-parameter algorithms, Cellpose, CLIJ2 integration, Multi-Template matching, Weka, OMERO
In older JIPipe versions, external environments (Python, Cellpose, R, …) can be installed via prepackaged installation scripts that replicate the installation procedure that would be applied by a user. We were made aware of issues that are caused by updated to the software or differences in the system configuration. To simplify the installation of various environments, JIPipe now provides ready-to-used installation packages that can be simply downloaded and extracted via a user-friendly interface termed “EasyInstaller”.
Following EasyInstaller packages are available:
- Cellpose 2.x (Windows, Linux, macOS) CPU/GPU
- Omnipose (Windows, Linux, macOS) CPU/GPU
- Python with skimage, pandas, … (Windows, Linux, macOS)
- R (Windows)
TrackMate integration (Beta)
The Cellpose integration was rewritten to make use of the improved Cellpose CLI.
Omnipose integration (Beta)
Based on the improved Cellpose integration, Omnipose was integrated using dedicated nodes. Please note that Omnipose is based on Cellpose 1.x and requires a dedicated environment.
There were various ROI-generating nodes with different feature sets. A new set of nodes was developed that make use of modern JIPipe API features. All nodes consume two optional inputs:
- ROI: Connect an existing ROI list to append to it. If left unconnected, a new ROI list is generated
- Reference: Connect an image to position ROI according to the image dimensions. If nothing is connected, the boundaries of the ROI provided by ROI is used, otherwise the width and height is assumed to be zero.
The functionality of Coloc2 was integrated into a node “Coloc2”.
General data processing
In previous JIPipe versions, various nodes included a “Load example” button that allowed users to learn about how to utilize the node. An issue with this functionality is that it cannot be easily discovered by users, e.g., via the search box. Additionally, examples cannot be created and distributed by users due to the reliance on Java code. The new version of JIPipe introduces a standardized node example system that can dynamically load examples from files, JAR resources, and node templates.
An online-repository system was implemented that allows to download template projects from the web (if you want to add your own project as downloadable example, file a pull request to https://github.com/applied-systems-biology/JIPipe-Repositories/blob/main/project-templates/project-templates.json)
By default, JIPipe uses the repository https://raw.githubusercontent.com/applied-systems-biology/JIPipe-Repositories/main/project-templates/project-templates.json. Feel free to configure the repository list within Project > Application settings > General > Projects > Template downloader repositories
In older JIPipe versions, the help of a parameter was displayed upon hovering the item via the mouse. This lead to issues with usability, as well as a slowdown of the interface due to the involvement of performance-intensive Java functions. Since this version, the parameter documentation can be opened by clicking the
? icon next to the entry.
Cache browser / Results viewer
ImageJ alias menu
To facility the transition from ImageJ, a function was introduced to organize JIPipe functionality into different menu locations (alias). Various existing JIPipe nodes were organized into a menu “IJ” that follows the structure of the ImageJ menu. For example,
Images > Threshold > Auto Threshold 2D is also organized into
IJ > Image > Adjust > Auto threshold ....
Measurement tool / mask drawer / ROI drawer
The UI was redesigned by implementing a Ribbon interface
The plot editor now saves ZIP files instead of directories (easier to handle)
The table editor was redesigned with a Ribbon UI
On loading pipelines with missing dependencies, users are prompted with dialogs to indicate that extensions are missing.
You can now right-click entries in the node template editor
Due to the improved data exporter interface, some parameters were deleted. Please evaluate the parameters of “Export data”/“Export images”/“Export table”
Bugfixes and small improvements
Please refer to the commits in the JIPipe GitHub repository to find a list of all bugfixes and improvements.