MoVIZ - Data Extraction is a KNIME workflow for extracting structured information from scientific literature and general PDF files with large language models (LLMs). It was developed for the publication Automating Data Extraction from Scientific Literature and General PDF Files Using Large Language Models and KNIME: An Application in Toxicology.
The current release is version 1.1.0.
Version 1.1.0 introduces the MoVIZ - Data Extraction KNIME extension for easier installation, fixes API-call issues, and adds updated LLM options. The extension includes the Python dependencies required by its nodes, so a separate Conda environment is no longer required.
- Install KNIME Analytics Platform 5.12.0 or newer.
- Open the illustrated installation guide.
- Download the workflow and extension using the links in the guide.
- Install the extension, import the workflow archive, and run the workflow.
The installation guide is also available as a printable PDF.
| Version | Status | Installation | Release |
|---|---|---|---|
| 1.1.0 | Current | Illustrated guide | Downloads and release notes |
| 1.0.0 | Legacy | Conda-based guide | Source snapshot |
GROBID is required only for the scientific-literature mode. It is not required for the other workflow modes. GROBID is not officially supported on Windows; see the version-specific installation guide for details.
The KNIME archives and the version 1.1.0 extension package are managed with Git LFS. For normal installation, use the named links on the GitHub release page instead of GitHub's automatically generated Source code archives.
To clone all files, including the LFS-managed extension archive:
git lfs install
git clone https://github.com/Moreira-Filho/Data_extraction_workflow.gitversions/v1.1.0/: current workflow, extension, and illustrated guideversions/v1.0.0/: legacy workflow, environment, and installation guide- Git tags
v1.0.0andv1.1.0: immutable snapshots of the published versions