This project aims to provide instructions for install and execute cTAKES. Apache cTAKES is a natural language processing system for extraction of information from electronic medical record clinical free-text. cTAKES stands for clinical Text Analysis and Knowledge Extraction System.
This image shows different types of nodes that cTakes can identify.
These types of nodes are defined in a dictionary Unified Medical Language System (UMLS).
It has different functionalities:
- Entity recognition
- Boundary detection
- Tokenization
- Normalization
- Part-of-speech tagging
- Extract entity properties (negated/subject)
- Pipeline: sequence of cTAKES annotators performing a comprehensive NLP task.
- Analysis engine: A single cTAKES annotator used in a pipeline.
- Piper file: A plaintext file describing a pipeline
- CAS (Common Analysis Structure): This is the data structure through which the annotators in a pipeline communicate.
The Default Clinical Pipeline produces the most commonly desired output from cTAKES. You only need to run a bash script, setting the input and output directories.
You can create your own pipeline and edit the code of the componentes of the pipeline in order to improve the extraction according to your data or type of problem.
Create user at https://uts.nlm.nih.gov/license.html and copy the API key
Build the container:
docker build --build-arg ctakes_key={KEY} --rm -t rootstrap/ctakes:latest .Run the container:
Run the container including the environment variables CTAKES_KEY, INPUT_DIR and OUTPUT_DIR a volume for the input files and a volume for the output files
docker run -ti --env CTAKES_KEY={KEY} --env INPUT_DIR=/input --env OUTPUT_DIR=/output -v $(pwd)/input:/input -v $(pwd)/output:/output rootstrap/ctakes:latestThe results files will be at output directory.
You can use XML Viewer to inspect the result files.
- cTAKES can be downloaded from http://ctakes.apache.org/downloads.html
- The dictionaries can be downloaded from http://ctakes.apache.org/downloads.html