This plugin 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 following technologies are used:
- DBPedia Spotlight (https://www.dbpedia-spotlight.org/)
- SPARQL Language (https://www.w3.org/TR/rdf-sparql-query/)
- Spacy (https://spacy.io/)
- PyOrient (http://www.orientdb.com)
- Pyphen (https://pypi.org/project/Pyphen/)
- Morph_IT (https://github.com/giodegas/morphit-lemmatizer/tree/master/master)
The API is documented by using Swagger2: https://api.openreq.eu/#/services/prs-improving-requirements-quality-features
We aim to carry out a platform which may help to:
- make requirements features explicit, in order to make identification and requirement quality classification straightforward
- provide requirement structure
Our approach consists in:
- Identifying requirements candidate paragraphs keeping title/section dependencies.
- Using machine learning-based tools to build a dependency logical structure of each paragraphs (subject, actions, complementary objects).
- Defining very weak dependency trees among concept, allowing to extend it or include dependecies from external ontologies.
- Giving a set of metrics to classify existing requirements using feature-based classifiers
- Conversion in free text
- Parsing of word document in Blocks keeping dependendeces (title/section)
- Paragraph enrichment
- Building of OpenReq Ontology
The workflow is based on the following steps:
- Conversion in free text
- Parsing of word document in Blocks, keeping dependendences (title/section)
- Paragraph enrichment
- Production of OpenReq Ontology
Make sure that docker-compose is installed (not included in normal docker installation)! Then go to the root folder of task_3_3 and type docker-compose up. Docker will then automatically build and run the two containers (spotlight and openreq_t_33). Also after building, the containers can be run at any time by docker-compose up.
Clone spotlight-docker repo (git clone https://github.com/dbpedia-spotlight/spotlight-docker); next cd with bash into spotlight-docker/v1.0/english and generate the spotlight container via sudo docker build -f Dockerfile -t english_spotlight .
Please note that the docker-spotlight repo has been added as a submodule to the internal DataScience gitlab.
Go inside the task_3_3 folder and generate the openreq container via sudo docker build . -t openreq_t_33. If you are behind a proxy, you have to specify an env and use sudo docker build . --network=host --build-arg "HTTPS_PROXY=$https_proxy" -t openreq_t_33.
First generate bridge to connect containers
sudo docker network create [name][subnet]
(use: sudo docker network create db_req_bridge --subnet 172.31.38.152/20)
specify docker run options
sudo docker run [net][name][ip][port][container][entrypoint]
(use: sudo docker run --net db_req_bridge -i --name db --ip 172.31.32.3 -p 2222:80 english_spotlight spotlight.sh)
(use: sudo docker run --net db_req_bridge -it --name req --ip 172.31.32.2 -p 10602:5008 openreq_t_33)
The platform comes with a set of micrososervices that are internally called by a unique entrypoint (/uploader). Therefore, altohugh all microservices have been documented, only the /uploader endpoint is responsible of making
the whole workflow running. In particular, "/prettify" service format the output of enpoint workflow according to OpenReq ontology.
Therefore, the Microservice has one entrypoint at the position:
http://217.172.12.199:10602/prs-improving-requirements-quality-features/uploader
with the following code:
files = [('file', (filename, open(filename, 'rb'), 'application/octet'))]
response = requests.post("http://217.172.12.199:10602/prs-improving-requirements-quality-features/upolader/num_par", files=files)`
being filename the name of the file to be uploaded and num_par the number of paragraph. The user should pass a document and the number of sub-blocks that should be parsed (each sub block may be represented by a title, a list of paragraph and so on). The full processing stage may require several minutes, depending on the infrastructure performance and the size of the document to be uploaded. No GUI is provided (a json-like output is provided, according to OpenReq Ontology). The output is compliant to OpenReq JSON structure.
None
None
ensure to change IP:PORT in config.json of app.py if -- and only if -- other IP/ports are used for the bridge
e.g. from "http://localhost:42001/rest/annotate", " to "http://172.31.37.91:11111/rest/annotate"
See OpenReq project contribution link: guidelines
Free use of this software is granted under the terms of the EPL version 2 (EPL2.0).