Alation

Run SQL Evaluation: Run SQL Evaluation on a Data Product's Evaluation Set. If an evaluation result matching the current data product's hash exists, we return that result. Otherwise, it starts an evaluation run if one is not currently running, then waits until that run completes and returns the result. Args: ctx: The run context (automatically injected) data_product_id: The data product ID to evaluate post_eval_results_back_to_chat: Whether to continue the chat asynchronously once evaluation is complete, posting the results back to the chat. ONLY for use by DP_SUGGEST_FIXES_JOB agent, since this will force the chat stream to end here. agent_config_id: The agent configuration ID (UUID) to use for evaluation. Hidden from the model, only for user overrides. auth_id: ID of the authentication credentials to use for the query. pre_exec_sql: An arbitrary SQL query to run in the session before the standard queries. data_product_version: The version of the data product to evaluate. If None, extracted from input_payload or uses the default version. Returns: ToolReturn with structured evaluation results showing passed and failed cases

SQL Execution: Execute the given SQL query or return an error message. The resulting table will be serialized and returned. The serialization may drop rows and columns if it is determined the result is too large. To see more of the result, please adjust the query by looking at fewer columns and or rows. Or apply a limit/offset to the query. Args: ctx (RunContext[AAIRunContext]): The run context containing dependencies. Automatically passed in. data_product_id: The ID of the data product to execute the query on. sql: The SQL query to execute. result_table_name: The name of the result table of the sql query after execution. This is for understanding table semantics. data_product_version: The version of the data product to use. If None, extracted from input_payload or uses the default version. pre_exec_sql: An arbitrary SQL query to execute within the session before the normal query. auth_id: ID of the authentication credentials to use for the query. auth_mode: Authentication mode ('datasource_service_account' or None for user credentials). allow_fallback_auth: Whether to allow falling back to service account creds. This is a temporary workaround to avoid having to specify an auth_id and will be removed in the near future.

Update an Existing Data Product: Update an existing data product using JSON Patch operations (RFC 6902). This tool fetches the current data product specification, applies your patches, validates the result, and saves it. Use JSON Patch operations to make targeted changes without regenerating the entire specification. Args: ctx (RunContext[AAIRunContext]): The run context containing dependencies. Automatically passed in. data_product_id: The ID of the data product to update. patches: A list of JSON Patch operations to apply. Each operation is a dict with: - op: The operation type ("add", "remove", "replace", "move", "copy", "test") - path: JSON Pointer to target location (e.g., "/product/en/description") - value: Value for add/replace/test operations - from: Source path for move/copy operations Common examples: - Update description: {"op": "replace", "path": "/product/en/description", "value": "New description"} - Add a metric: {"op": "add", "path": "/product/metadata/metrics/revenue", "value": {"displayName": "Revenue", "description": "...", "expression": "..."}} - Update column description: {"op": "replace", "path": "/product/recordSets/users/schema/0/description", "value": "User ID"} - Remove a field: {"op": "remove", "path": "/product/recordSets/orders/schema/2"} PROTECTED: Cannot modify /product/productId, /product/version, or /product/deliverySystems - these are managed by the system. data_product_version: The version of the data product to update. If None, extracted from input_payload or uses the default version. Returns: ToolReturn with the result of the update operation.