Entropy Data

Workflows

Typical patterns for interacting with Entropy Data through the MCP server.

Discover and query data

The most common workflow: find data, get access, and run queries.

  1. Search — Use search to find relevant data products by keyword
  2. Inspect — Use fetch to get full details including output ports, schemas, and access status
  3. Get access — Use request_access if you don't have active access to the output port
  4. Query — Use execute_query to run SQL queries against the data

Example prompts

  • "Find all data products related to customer orders"
  • "Show me the schema of the sales data product"
  • "Query the top 10 customers by revenue from the customer-analytics data product"

Manage data contracts

Create, update, and test data contracts through MCP.

  1. Get schema — Use datacontract_schema to retrieve the ODCS JSON Schema
  2. Get contract — Use datacontract_get to retrieve an existing data contract
  3. Update contract — Use datacontract_save to create or update a data contract
  4. Test — Use datacontract_test to run tests against a server
  5. Review results — Use datacontract_test_results to check test outcomes

Example prompts

  • "Get the data contract for the inventory data product and check if all tests pass"
  • "Update the description field in the customer-events data contract"
  • "Show me the latest test results for the orders data contract"

Manage data products

Create and update data products through MCP.

  1. Get schema — Use dataproduct_schema to retrieve the DPS or ODPS JSON Schema
  2. Get product — Use fetch to retrieve an existing data product
  3. Save product — Use dataproduct_save to create or update a data product

Example prompts

  • "Create a new data product called orders owned by the data-platform team"
  • "Add an output port to the customer-events data product"
  • "Show me the data product schema so I can draft a valid specification"

Explore the semantic layer

Navigate the organization's ontology and connect business concepts to the data products that implement them.

  1. List namespaces — Use semantics_list_namespaces to see the available ontologies
  2. Browse or search concepts — Use semantics_list_concepts or semantics_search_concepts to find entities, metrics, and properties
  3. Inspect a concept — Use semantics_get_concept and semantics_list_relationships to understand a concept and how it relates to others
  4. Trace to data — Use semantics_find_data_products_for_concept and semantics_concepts_for_data_product to map between concepts and data products
  5. Ask in natural language — Use semantics_nl_query or semantics_sparql_query to query the graph directly

Authoring the ontology (creating or updating concepts and relationships via semantics_save_concept, semantics_save_relationship, and semantics_save_ontology) requires the organization owner role. Always call semantics_get_schema first so the YAML you produce conforms to the expected shape.

Example prompts

  • "Which data products provide data for the Customer entity?"
  • "List all metrics in the main namespace classified as financial"
  • "What semantic concepts does the customer-analytics data product reference?"