README.md

# About

This project is designed to extract entities (i.e., `taxa`, `phenotypes`, `habitats`, `disease names`, `hosts`, `pathogen`, `vector`, `dates` and `geographic names`) from textual data for the purpose of scientific watch.

It contains a workflow based on Framework [AlvisNLP](https://github.com/Bibliome/alvisnlp) and uses the Ontobiotope Ontology and NCBI taxonomy.


## Usage
> To run the workflow you need `GNU bash, version 4.4.x` and `singularity version 3.4.x` must be installed in yout machine. The workflow is compatible with `AlvisNLP version 0.7.x` 


1. clone the project

```
git clone https://forgemia.inra.fr/mandiayba/pesv-tm.git
cd pesv-tm
```

2. pull the singularity image of AlvisNLP. 
> `login` and `password` are required to pull the AlvisNLP singularity image from forgemia, contact maintainer if you don't have grant.

```
cd pesv-tm/softwares
singularity pull --docker-login alvisnlp.sif oras:registry.forgemia.inra.fr/migale/tm-tools-packages/sif/alvisnlp:v0.0.4
```

3. run the workflow. 
> You may run the workflow with the test corpus `corpus/pesv/Xylella-test/txt/`, results are stored into `corpus/pesv/Xylella-test/`

```
softwares/alvisnlp.sif -J-Xmx32G -verbose -cleanTmp \
-alias input corpus/pesv/Xylella-test/txt/ \
-outputDir corpus/pesv/Xylella-test/ \
-entity ontobiotope resources/BioNLP-OST+EnovFood \
-feat inhibit-syntax inhibit-syntax \
plans/PESV_workflow.plan
```

4. See results from `corpus/Xylella/visualisation_html`

## Maintainer
Mouhamadou Ba : mouhamadou.ba@inrae.fr