This service was created as a result of the OpenReq project funded by the European Union Horizon 2020 Research and Innovation programme under grant agreement No 732463.
The ri-analytics-classification-twitter microservice is cabable of classifying English and Italien tweets into problem report, inquiry, and irrelevant.
-
Docker (→ https://www.docker.com/)
-
SentiStrength (→ http://sentistrength.wlv.ac.uk/)
ri-analytics-classification-twitter requires you to:
-
aquire the sentistrength library from http://sentistrength.wlv.ac.uk/
-
mount the SentiStrength.jar file in the Docker run environment (e.g.,
docker run --rm -d -v "<path>/SentiStrength.jar:/app/amazon-kinesis-client-python/sentistrength/SentiStrength.jar") -
set a Docker ARG GDRIVE_DL_LINK to the <DL_LINK_ID> (e.g,
docker build --build-arg "GDRIVE_DL_LINK=1fcMjYmmjY9-WMOauYle6fFLooJ0u9OlU"). This will download and unzip the classification models from Google Drive.-
the link should contain a models.zip archive. Inside it needs a folder per language models/english and models/italian.
-
in the subfolders (for each language) place all models as specified in the config/ files
-
the code runs the feature extraction exactly in the order as specified in the config files
-
Run the following commands to start the microservice:
-
docker build --rm -f "Dockerfile" --build-arg "GDRIVE_DL_LINK=<DL_LINK_ID>" -t ri-analytics-classification-twitter:latest . -
docker run --rm -d -v "<path>/SentiStrength.jar:/app/amazon-kinesis-client-python/sentistrength/SentiStrength.jar" -p 9655:9655 ri-analytics-classification-twitter
A full description of the the microservice can be found in the following swagger documentation:
The API is documented by using Swagger2:
See OpenReq project contribution Guidlines