Note: This project was developed while working on the article Universal Knowledge Graph Embeddings (repository: ([https://github.com/dice-group/Universal_Embeddings], publication link)).
This repository contains code to run embeddings.cc and embeddings.cs.upb.de (also without TLS).
- The API is documented at embeddings.cc/api
- Additional examples are provided in Python examples and JavaScript HTML form
- Use the Index API to create Elasticsearch indexes and to add data.
- It is only available in UPB network (use VPN).
- It can easily accessed using the methods in API python file.
Usage examples are provided in the files for simple examples, in the adding CSV and in the adding UniKGE data. - Important: Create an alias for each index to be available in public. Only aliases can be accessed by webservices.
Webservice | Method | Parameters |
---|---|---|
/ping | GET | - |
/count | GET | index |
/get_embeddings | GET | index, entity |
/get_cpu_usage | POST | password |
/get_indexes | POST | password |
/create_index | POST | password, index, dimensions, shards |
/create_index_usagelog | POST | password |
/delete_index | POST | password, index |
/add | POST | password, index, docs |
/alias_put | POST | password, index, alias |
/alias_delete | POST | password, index, alias |
- How to install on your system
- Development (External documentation of integrated components)
- Virtual machine (Installation and deployment)
- VM nginx (Webserver configuration)
- Deployment (How to publish a new version)
- Note: This is an extension of kg-embedding-service