- Import as maven project.
- Download the MetaMap Web API and add to the build path.
- Download the Stanford CoreNLP models and add to the build path. Note: a UMLS Terminlogy Service (UTS) account is required to use MetaMap and OntologyBuilder.
This program takes 4 command line arguments: UTS account name, password, email, and file name (the file must be in plain text format).
This program takes 5 command line arguments: UTS account name, password, email, input xml file (from clinical trials), and output json file name.
- Modify main method of GraphBuilder, provide required neo4j authentication information.
- Create a directory named xml and place all xml files that need to be annotated there.
- Run the program, converted json files are going to be created under the build path.
Since MetaMap does not support non-ASCII characters, you will get an error if the input file contains any non-ASCII characters that are not yet handled by the program (see Issues tab for details). GraphBuilder will not crash in this case, but you will get a untokenizable warning from Stanford NLP.