nmsa / tma-framework-a Goto Github PK

Implement classes as defined in the quality metamodel.

• automated regression/unit tests for this component
• based on some sample data
• @tbasso provides data and test cases
• @mrmassis implements tests

Dashboard (extends Horizon in openstack)

@felipengaia is currently working on a fast prototype based upon AngularJS.

Refactor the existing privacy tool to work as an actuator.

• add a persistence layer to quality models according to knowledge requirements

Design quality model for Isolation metric according to requirements from WP4. See deliverable for further detatil.

Validate the initial/fast prototype.
This issue includes the following:

• (@felipengaia ) provide a demonstration (screenshots?) -- See screenshot below.
• (@brenofranca83 @tbasso @montex ) validade demo against expectations and provide feedback for next sprint.
• (@brenofranca83 @tbasso @montex) Decide whether this prototype goes public or not.

Dashboard should provide a public API (REST API??) for receiving data from other sources, possibly other dashboards.
This is valid for thresholds and weights.

Mapping the knowledge database to java classes using jpa according to the database modeling created in the tma-framework-k project.

• Refactory was done at the original PRIVAaaS (BIGSEA). That is good, but new developments should occur at ATMOSPHERE fork.
• Fork at ATMOSPHERE should be synchronized with BIGSEA.

Develop API for interactions among dashboard and other TMA's components.

• Events/notifications for interaction with knowledge listener, quality model, etc.

@felipengaia gives details, please.

• A hierarchy of files is defined statically. It can be generalized to be dynamic like the privacy policy.
  We need a (simple) design for this refactory.
  @tbasso please detail.

This issue includes the following:

• install new implementation
• validate whether output is still according to expected behavior (compared to former implementation)
• discuss/define and prioritize next steps and further refinements.

Integrate Analyzer in the environment at ic.cloud where privacy probes/actuators are running.
This integrated environment is supposed to work as a demonstration showing all mape-k components working together for completing the adaptation cycle for privacy with k-anonymity.

1- Openstack + dashboard
2- Openstack + angularJS
3- dashboard in angularJS decoupled from Openstack

Third option was adopted for further analysis and implementation of a fast prototype.

• This is the second internal API for dashboard. This time, dashboard will implement many notify messages and quality model have to adapt to these many messages.
• It is necessary to develop this integration code in both sides: quality model and dashboard.

Analyze component responsible for interacting with the knowledge repository.
It works polling by requests to the knowledge repository.
due date: middle july

Design the quality model for security metric according to the requirements from WP5. See deliverable D3.1 (Section 3.3.3. Integrity and confidentiality ) for further detail. Initially, this metric is related to integrity and confidentiality.
Due date: Middle July

Design the quality model for the fairness according to requirements from WP6. See deliverable for further detail.

Design the quality model for the transparency metric according to the requirements of WP6. See deliverable for further detail
Due date: End July

• New dashboard uses metabase tool
• it has to be integrated to analyzer.
• it does not connects to data sources other than databases
• a way to integrate it is to access tables directly from knowledge

tasks related to this issue:

1. experiment in accessing tables in knowledge directly
2. design of views to be accessed by metabase, instead of actual tables
3. possible integration to the existing dashboard

@jreluiz Please, define the general structure using a Maven + SpringBoot project.
The Analyze module is a dockerized microservice.

• To simplify deployment of dashboard.
• Simplify the startup of the dashboard by integrating it with Analyzer startup. Basically, the dashboard will be integrated into the Analyzer Dockerfile.

Refactor the existing privacy tool to work as a probe.

The error described below occurred while testing the refactory of PRIVAaaS. This was due to the missing hierarchy file (the file was included in a copy to perform the tests and it worked well). We only ask for including the file in the project before releasing it.

C:\arx-poc (MA)\arx-poc>java -jar dist\arxAllInOneJar.jar -k 2 -policy test_files\param_adulto1.csv -datadb test_files\adulto1.csv -output test_files\out_adulto1.csv
Exception in thread "Character reading thread" java.lang.IllegalStateException: Error processing input
at com.univocity.parsers.common.input.concurrent.ConcurrentCharLoader.run(ConcurrentCharLoader.java:92)
at java.lang.Thread.run(Unknown Source)
Caused by: java.io.FileNotFoundException: arx-poc\hierarchy\age.csv (O sistema não pode encontrar o caminho especificado)
at java.io.FileInputStream.open0(Native Method)
at java.io.FileInputStream.open(Unknown Source)
at java.io.FileInputStream.(Unknown Source)
at org.deidentifier.arx.io.CSVDataInput$LazyFileReader.read(Unknown Source)
at com.univocity.parsers.common.input.concurrent.CharBucket.fill(CharBucket.java:70)
at com.univocity.parsers.common.input.concurrent.ConcurrentCharLoader.run(ConcurrentCharLoader.java:81)
... 1 more
java.lang.RuntimeException: Empty generalization hierarchy for attribute 'age'
at org.deidentifier.arx.framework.data.GeneralizationHierarchy.(Unknown Source)
at org.deidentifier.arx.framework.data.DataManager.(Unknown Source)
at org.deidentifier.arx.ARXAnonymizer.getDataManager(Unknown Source)
at org.deidentifier.arx.ARXAnonymizer.anonymize(Unknown Source)
at br.unicamp.ft.arx.service.Anonymizer.run(Anonymizer.java:160)
at br.unicamp.ft.arx.service.Anonymizer.main(Anonymizer.java:71)

• develop code to implement privacy QM according to the design already done.

Integrate two of Analyzer's internal components: Dashboard and
KnowledgeListener. These components have been developed independently and now is (almost) time to put them together in a single working environment.

@felipengaia defined an API for dashborad
@jreluiz implemented knowledgeListener

• Literature review
• interface design
• technologies

• dashboard have to be able to show the calculation tree for a score.
• This may require new accesses to information in knowledge.
• this require recursive accesses to database in order to build a tree

Initial design and fast implementation for knowledge Service

create repository for this for privacy component.

Adapt privacy component to offer a public interface (API) for Java applications which uses a resultsets (see JDBC) instead of files. That is, when using this new interface, the component's output is a resultset.
This adaptation would facilitate the addition of privacy controls (based upon k-anonymity) to java apps.

• Analysis of the code contained in the repository
• Creation and migration of the code to the Github of the project.
• Definition and treatment of entry parameters.
• Implementation of methods with specific objectives within the main class.
• Organization and insertion of comments in the code.
• Removal of unneeded and obsolete code.

Data structure and score calculation for trustworthiness properties.