This repository belongs to our workshop publication at AI4LAC@JCDL2024. You find a short version of our article on the workshop webpage and a technical report (long version) at arXiv. The links will be added as soon as they are available.
This repository contains graph-based ranking methods for our Narrative Service. Please note that the repository has a dependency to our Narrative Service whose code is publicly available on GitHub. Unfortunately, we cannot publish the pre-processed document data (the extracted graph information).
The implemented code is available in the src directory. It contains:
The evaluation reports for all benchmarks (TREC Precision Medicine 2017, 2018, 2019, 2020 and TREC Covid) are available in our evaluation directory.
This project builds upon our previously published projects. That is why we also need to check out their code. Currently, we rely on the dev branches.
- KGExtractionToolbox: Basic entity linking methods / information extraction / pipelines for automation
- NarrativeAnnotation: Pharmaceutical specific entity linking / text classification / statement extraction logic
- NarrativeIntelligence: Code and scripts for PubPharm's Narrative Service
Please read the toolbox setup documentation.
To use this project, clone this project and its submodules via:
git clone --recurse-submodules git@github.com:HermannKroll/RankingNarrativeQueryGraphs.git
We used a Conda Environment and Python 3.8
conda create -n ranking python=3.8
Install the requirements of this project and of the submodules via:
pip install -r requirements.txt
pip install -r lib/KGExtractionToolbox/requirements.txt
pip install -r lib/NarrativeIntelligence/requirements.txt
pip install -r lib/NarrativeAnnotation/requirements.txt
Set the Python path:
export PYTHONPATH="/home/USER/RankingNarrativeQueryGraphs/src/:/home/USER/RankingNarrativeQueryGraphs/lib/KGExtractionToolbox/src/:/home/USER/RankingNarrativeQueryGraphs/lib/NarrativeAnnotation/src/:/home/USER/RankingNarrativeQueryGraphs/lib/NarrativeIntelligence/src/"