Description

Docker Image of the QuickUMLS tools (https://github.com/Georgetown-IR-Lab/QuickUMLS)

tested With UMLS2018AA data

Maintainer : Alice ROGIER alice.rogier-ext@aphp.fr

REQUIREMENTS:

docker https://docs.docker.com/engine/install/

nextflow https://www.nextflow.io/

Dataset

located on the data/demo folder: are the result of the pubmed request from 2000 to 2019

("EHR" or "biomedical NLP" or "clinical NLP" or "clinical natural language processing" or "natural language processing" OR "NLP" OR "natural language processing") AND ("system" or "framework" or "tool" or "workflow" or "pipeline" or "architecture") and ("text mining" or "platform")

each result is put on a separated file

How to use

1) Add MRSTY.RRF and MRCONSO.RRF in the UMLS folder.

2) Build the image

To build both the client and the server image run

    make build

3) start the server

a) create a docker network named annotations

    make createNetwork

 b) start the server it will be available from port localhost:8080

    make startServer

NB: if you want to change the server port modify it in bin/startServer.sh Furthermore we create a docker network to ease the communication with the docker client.

4) Live demo with the client

a) start client

    make liveDemo

NB: It will take as input data in folder data

b) call the script

An example of client, server interaction is located on src/quickumls_main.py

    #in the container run 
    python3 /home/quickumls/src/client.py --serverName quickumlsserver --inputFile /home/quickumls/tmp/demo/input_1.txt 

4) RUN nexflow and analyse all the demo data folder

The user need to install nexflow as required

a) Generate Nexflow script

    make generate

b) run Nexflow script

    make runnf 
    #or
    nextflow run nfscript/demo_nlp.nf -c config/nextflow.config                                                                                                                                                                                                                                  

results will be output on the result folder

Customization:

If the user wants to custom the script :

nfscript/demo_nlp.nf  # for the nextflow script
config/nextflow.config #nexflow config
src/client.py # quickumls client