PyGFF

A Python package to read and write Grace Format files (.gff / .segff).

Easiest way to install PyGFF

$ pip install pygff

Basic usage

• Loading a .gff file:

   >>> from pygff import load
   >>> data = load("image.gff")
  

• Saving a numpy ndarray np_array as a .gff file:

   >>> from pygff import GFF, save
   >>> save("image.gff", GFF(np_arr))
  

• Saving a numpy array seg_array as a .segff (segmentation) file with class names, indices, and colors:

   >>> from pygff import GFF, save
   >>> segff = GFF(seg_array)
   >>> segff.info.set_class_dict(
       {"Heart": {"index": 1, "color": (255, 0, 0, 255)},
         "Liver": {"index": 2, "color": (0, 255, 0, 255)}})
   >>> save("segmentation.segff", segff)
  

  Note that seg_array is converted to np.uint8 before saving. Colors are specified as RGBA, and class index 0 is reserved for "unclassified" data.

Tutorials

Tutorial notebooks can be found in the /tutorials/ directory of this repository. They are not included with the PyGFF package. We recommed that you to start with 01_load_and_save.ipynb to learn more about loading, saving, and GFF objects. More tutorials will be added in the future.

Running the examples requires the packages jupyter, matplotlib, numpy, and scipy to be installed. Also, please download the required example datasets if you have not cloned the repository yet.

What is GFF?

GFF is an open source file format for multimodal biomedical images (.gff) and segmentations (.segff). The format supports datasets with up to five dimensions (three spatial dimensions, time-variant, and multi-channel) and a rich set of metadata key-value pairs. By default, the implementation uses a lossless compression algorithm to reduce file size and cryptographic hashing for secure writing. Multithreading is also used if possible to speed up reading and writing of GFF files.

The PyGFF package is developed by Gremse-IT GmbH (Aachen, Germany) as a Python interface for Imalytics Preclinical 3.0 which utilizes .gff by default for underlay, overlay, segmentation (.segff), and project files (.imaproj).

For more details, please refer to this publication:

Yamoah, Grace Gyamfuah et al. “Data Curation for Preclinical and Clinical Multimodal Imaging Studies.” Molecular imaging and biology vol. 21,6 (2019): 1034-1043. doi:10.1007/s11307-019-01339-0

Full text: https://pubmed.ncbi.nlm.nih.gov/30868426/

How to build the package yourself

1. Clone the repository

   $ git clone git@bitbucket.org:felixgremse/gff_file_format.git
   

2. (Optional) Create a virtual environment, e.g. with venv

   $ python -m venv env
   ...
   $ source env/bin/activate
   

   or using conda

   $ conda env create --file environment.yml
   ...
   $ conda activate pygff
   

3. Then, install pygff in editable mode, e.g. using build

   $ python -m pip install --upgrade build
   ...
   $ python -m pip install -e .
   

   or conda-build

   $ conda develop .
   

How to run package tests

Make sure that pytest is installed and then simply call it from a shell in the main directory.

$ pytest

License

The PyGFF package is licensed under the terms of the MIT license.

All .gff files, .segff files, and Jupyter notebooks contained in the /tutorials/ directory of the repository are licensed under CC BY-NC-SA 4.0.