/indexer

DeCS Indexer frontend and backend for MESINESP task.

Primary LanguagePython

Indexer web tool

Assign controlled labels to different types of documents, mainly in Spanish language.

Supported labels:

  • DeCS (Descriptores en Ciencias de la Salud), analog Spanish version of English MeSH terms
  • HPO (Human Phenotype Ontology) (future feature)

Supported types of documents:

  • Scientific articles from IBECS (Índice Bibliográfico Español en Ciencias de la Salud)
  • Scientific articles from LILACS (Literatura Latinoamericana y del Caribe en Ciencias de la Salud)
  • Clinical studies from REec (Registro Español de estudios clínicos)
  • Clinical research projects from Portal FIS (Fondo de Investigación en Salud)
  • Patents in Spanish from Google Patents

Applications

The main objective of indexing documents is to obtain a gold-standard set (i.e. annotated by humans) that maps the text within each document to some of those controlled labels.

Implementation

The technologies that build this tool are:

Usage

$ git clone https://github.com/TeMU-BSC/indexer.git
$ cd indexer

Development

$ docker-compose up

Production

$ docker-compose -f docker-compose.prod.yml up --build -d