Support multiple document sources

Onto-Med / top-backend

Spring Boot based backend of the TOP Framework

MIT License

0 stars 1 forks source link

Support multiple document sources #197

Closed ChristophB closed 6 months ago

ChristophB commented 9 months ago

Currently, it is not possible to connect to multiple document sources. Only the default adapter configuration is used, others are ignored.

It seems like this is related to using sping-boot-starter-data-elasticsearch and modifications to the configuration, which are only performed on startup of the framework.

ChristophB commented 9 months ago

Ideally, the backend does not need any Elasticsearch dependencies. Connection to document sources should be handled by top-document-query or some other module.

In contrast to Elasticsearch, it seems like we definitely need Neo4J in the backend, because clusters are stored there and are part of the TOP main business model.

ChristophB commented 9 months ago

I created an issue in top-document-query that addresses outsourcing of concept cluster related stuff. See: https://github.com/Onto-Med/top-document-query/issues/9

ChristophB commented 9 months ago

Some notes:

Because Neo4J is used to store the document related business model, we should remove Neo4J from top-document-query.
top-document-query module only returns documents or graphs.
The backend stores the graphs in a single Neo4J instance => only one configuration, no messing around with Spring properties.
As we want to store multiple graphs from different document sources, all nodes need an additional property that indicates to which data source they belong.

fmatthies commented 8 months ago

Some other todos for this issue:

[x] Remove/adapt DocumentQuery configurations
[x] Remove ElasticsearchConfigBean