IKB4Stream is a module of the smart waterleak detection system WAVES.
WAVES is a scalable and distributed platform for processing semantic streams based on open source frameworks.
WAVES is the result of a research project that aims to provide smart water management system. Thanks to its abstract level, it could be easily applied on streaming sensor data in other domains.
More info on http://www.waves-rsp.org/
IKB4Stream is a module inside WAVES. Its goal is to detect false positives among the leaks reported from the WAVES system water sensors.
- Project Manager: Vincent Heng
- Technical Experts: Valentin Michalak, Ludovic Carretti
- NLP Expert, Web Designer, Documentation: Mehmet Demir
- NLP Expert: Sandy Allibert
- Business Analyst: Boubacar Dabo
- DevOps: Valentin Michalak
- Quality Expert: Ludovic Carretti
- Security, Performance: Kevin Mollenhauer
- Metrics Analyst, Communication, Documentation: Loïck Renée
First of all, download the file archive.
Unzip it.
unzip PATH/FILE_NAME.zipMove todocker_images directory.
cd docker_imagesThen, add execution flag on each script file.
chmod +x setup.sh
chmod +x docker_images/prepare_docker_images.sh
chmod +x docker_images/build_docker_images.shJust before launch the project, change the conf.properties file. You need to declare the container name as hostname.
vim resources/conf.propertiesChange the line:
database.metrics.host = http://localhost:8086
to:
database.metrics.host = http://influx:8086
and the line:
database.host = mongodb://localhost:27017/
to:
database.host = mongodb://mongo:27017/
Now, you can launch the entire project.
./setup.shTODO
The directory resources contains all the config properties files.
-
anomaly.ttl is an anomaly request for testing
-
config.properties contains the general properties of the project
-
logback.properties contains properties for logback
-
The directory communication contains files properties for kafka and the web GUI. .
-
The directory datasource contains files properties for each ProducerConnector. If the ProducerConnector is a mock (name endind by "Mock"), the directory also contains data mock
-
The directory scoreprocessor contains files properties for each ScoreProcessor and their mock.
-
The directory nlp_model contains the binaries for the NLP functions (lemmatization, NER...) and a dictionnary for lemmatization.
All sources are in the package com.waves_rsp.ikb4stream.
This package contains all the connectors for clients applications : Kafka Connector and the Web Connector
This package contains consumers : Database Reader and the Communication Manager. It also contains the Main class for launching program.
This package contains all utilities classes.
- The sub package communication contains Functional Interfaces and define Objects for communicating with the database.
- The sub package metrics contains classes wich are used for logs metrics in the Influx database.
- The sub package model define Objects : Event, LatLong and PropertyManager
- The sub package util contains the utilities : the Geocoder, the NLP functions, the Class Manager, the Jar Loader and Rules Reader.
This package contains all the Producer Connectors which search data from differents API (stream or batch). A Producer Connector implements the IProducerConnector interface.
This package contains the Main class for the producer module.
TODO : complete
This package contains all the ScoreProcessor apply to the events created by the Producer Connectors. In the sub package event, there is the default score processor. Other subpackage contains a specific Score Processor for a specific Producer Connector. All classes in this package implement the IScoreProcessor interface.
This package contains all the unit Test (JUNIT). It takes up the architecture of the rest of the project.
See the LICENSE file for license rights and limitations (MIT).