Anonymize input before sending to the Classifier

[vaunapuu]

AS A Product Owner I WANT TO be confident that some input must always be anonymized before it could be used by the Classifier SO THAT the risks of sensitive data leakage incidents could be mitigated

Acceptance Criteria

• [ ] A third-party data anonymization tool is integrated into Classifier

  E.g. https://github.com/buerokratt/Data-Anonymizer can be used for data anonymization Devleopment of a new data anonimyzer is not in scope of this issue

• [ ] Apply mandatory data anonymization based on data gropus (see #83 for reference)
• [ ] Anonymizer and Classifier use the same data structure on sample data

DSL

• [ ] REST [POST] service /classifier/anonymize with Classifier-specific restrictions (e.g. permissions) apply
• [ ] Standard Ruuter DSL
• [ ] Only JSON in pre-defined format is used as input
• [ ] Sample data structure confirmed

  {"input": "lkdöla kdölkdölakd ölka dölköad nao"}

Misc

• [ ] The third-party data anonymization tool is deployed as a stand-alone service, with specific endpoints consumed by the Classifier