GrEBI
by EBISPOT
Overview
HPC pipeline for aggregating biomedical knowledge graphs from EMBL-EBI and other resources using ontologies and LLM embeddings, enabling integrative queries for humans and machines.
Installation
cd webapp && docker compose upEnvironment Variables
- PUBLIC_URL
- REACT_APP_APIURL
- REACT_APP_ENV
- GREBI_NEO4J_HOSTS
- GREBI_SOLR_HOST
- GREBI_RESOLVER_HOST
- GREBI_METADATA_HOST
- GREBI_PREFIX_HOST
- GREBI_CONTEXT_PATH
- GREBI_QUERY_TEMPLATES_PATH
- GREBI_NEO_DATA_PATH
- GREBI_SOLR_PATH
- GREBI_SQLITE_SEARCH_PATH
- GREBI_METADATA_JSON_SEARCH_PATH
- GREBI_PREFIX_MAP_PATH
Security Notes
The Neo4j database instances (`grebi-neo4j` in `docker-compose.yml` and in Kubernetes deployments) are configured with `NEO4J_AUTH=none`, disabling all authentication. This is a critical security vulnerability for any publicly accessible service and should be addressed immediately (e.g., by enabling authentication with appropriate credentials). Other services (prefix, resolver, metadata) expose file paths via environment variables, which could lead to path traversal if the underlying volumes are not properly secured or if user-controlled input affects these paths.
Similar Servers
mcp-neo4j
Provides a Model Context Protocol (MCP) server for interacting with Neo4j graph databases, enabling Cypher query execution, schema introspection, and generating insights for AI models.
sparql-llm
This project provides tools for Large Language Model (LLM)-powered SPARQL query generation over specific knowledge graphs, including a chat interface for user interaction and a Model Context Protocol (MCP) server for tool exposure.
ols4
A comprehensive Ontology Lookup Service (OLS) that provides a web interface, REST API, and an OpenAI-compliant MCP (Microservice-Coded-Prompt) server for biomedical ontologies, supporting search, graph traversal, and AI-driven entity understanding through embeddings.
pyomop
Facilitates OHDSI OMOP Common Data Model (CDM) database interactions, including schema management, data loading (FHIR, Eunomia), querying, and export, with optional LLM-based natural language interfaces.