termsuite-http is an HTTP API for TermSuite.

Currently, only preprocessing is supported. Terminology extraction and bilingual alignment ar enot supported.

Getting started

From sources

Prerequisites:

• Gradle
• Git
• TreeTagger installed locally for TermSuite. See instructions.

Clone the repo:

$ git clone git@github.com:termsuite/termsuite-http.git

Build the sources:

$ cd termsuite-http
$ gradle clean jar

Configure termsuite-http:

$ cp termsuite-http.properties.sample termsuite-http.properties
$ vi termsuite-http.properties

Edit the termsuite-http.properties according to your needs and to your local TreeTagger install.

Finally, start termsuite-http:

$ java -Xmx512m -Dlogback.configurationFile=logback.xml \
    -cp build/libs/termsuite-http-1.0.0.jar \
    fr.univnantes.termsuite.http.TermSuiteHttp

With docker

See TermSuite HTTP Docker container

Using the API

POST /preprocess

Applies TermSuite preprocessings on an input text file that is passed as request body, and returns the preprocessed annotations in JSON as response body:

curl http://0.0.0.0:4567/preprocess \
    --data-binary @myfile.txt \
    -H "Content-Type: text/plain \
    > myfile-preprocessed.json

GET /preprocess

Applies TermSuite preprocessings on an input text passed as parameter, and returns the preprocessed annotations in JSON as response body:

curl -G http://0.0.0.0:4567/preprocess?data \
  --data-urlencode "text=The black cat is eating the mouse."

Configuration

Available config parameters in termsuite-http.properties are:

• tagger.path: the path to local TreeTagger,
• termsuite.lang: the language of TermSuite preprocessor, (default: en)
• termsuite.preprocessor.max_requests: the max number of preprocessing requests before inner preprocessor service is restarted,
• http.port: customizes the listening port.

Example

termsuite.lang=fr
termsuite.preprocessor.max_requests=100
tagger.path=/opt/treetagger
http.port=4567