Containerization

[chreman]

This PR contains a deployable MVP of a new data source integration in a micro-service oriented architecture. The PR has following main elements:

• A minimum search example for TRIPLE data (examples/triple)
• new backend micro-services, running in Docker containers (server/workers)
• documentation README and example config files for each service

Following changes to the backend pipeline have been made: For predefined integrations (here TRIPLE), the pipeline now branches off for the stages of document retrieval and map data generation (function search in search.php):

• Instead of executing R-scripts on the OS, a POST request is sent through a localhost reverse proxy to the API (Apache2 setup is described in the README), which passes the request to the data source connector (in this case an ElasticSearch API)
• The connector retrieves documents according to the search query, preprocesses them, and returns input data (metadata and textual content), which is then handed over to dataprocessing
• The data processing service executes the machine learning and NLP components on the input data and returns a map representation
• The API service takes the map representation and returns it as JSON to the call in search.php.

Other functionalities in search.php such as snapshot generation and map persistence to SQLITE function as before.

For an initial setup, please follow the instructions in README.

[pkraker]

Awesome, thanks! This is a big step forward in terms of our backend development and deployment.