FrameTree

FrameTree is Python framework that is used to map categorical data organised into trees (e.g. subject data organised in file-system directory) onto virtual "data frames". Cell in these data frames can be scalars, arrays or a set of files and/or directories stored at each node across a level in the given tree. Metrics extracted from the data in these frames are stored alongside the original data, and are able to be fed into statistical analysis.

FrameTree_ manages all interactions with data stores. Support for specific specific repository software or data structures (e.g. XNAT or BIDS). Intermediate outputs are stored, along with the parameters used to derive them, back into the store for reuse by subsequent analysis steps.

Analysis workflows are constructed and executed using the Pydra_ dataflow API, and can either be run locally or submitted to cloud or HPC clusters using Pydra_'s various execution plugins. For a requested output, FrameTree determines the required processing steps by querying the store to check for missing intermediate outputs and parameter changes before constructing the required workflow graph.

Documentation

Detailed documentation on FrameTree can be found at https://frametree.readthedocs.io

Quick Installation

FrameTree can be installed for Python 3 using pip::

$ python3 -m pip install frametree

Extensions

The core FrameTree package only supports directory data trees, however, it is designed to be extended to support in-place analysis of data within data repository platforms such as XNAT and formalised data structures such as Brain Imaging Data Structure (BIDS).

License

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License

Acknowledgements

The authors acknowledge the facilities and scientific and technical assistance of the National Imaging Facility, a National Collaborative Research Infrastructure Strategy (NCRIS) capability.