Add Support for Querying Agents via SQL #8856
Conversation
| @@ -63,6 +67,52 @@ def predict(self, model_name: str, data, version=None, params=None): | |||
| ml_handler = self.integration_controller.get_ml_handler(model_metadata['engine_name']) | |||
| return ml_handler.predict(model_name, data, project_name=self.project.name, version=version, params=params) | |||
|
|
|||
| def _query_agent(self, query: ast.ASTNode, existing_agent: db.Agents, session): | |||
There was a problem hiding this comment.
maybe move this function to agent controller? it is about how agent works
There was a problem hiding this comment.
I would prefer to keep it here since there are a good amount of SQL operations. IMO would be best to keep filtering SQL targets, filtering conditions, etc out of agents_controller. Open to more discussion if you feel strongly though
|
Also I think for better design this function (or the whole class) should be moved to agent controller:
|
100% agree that we should have a more central location for agent completions such as |
* agent behaves as model * flake8 * don't add one more line to response
Description
Related to #8821
This PR adds support for querying agents using MindsDB SQL. Example:
Type of change
Verification Process
To ensure the changes are working as expected:
./tests/interfaces/agents/test_agents_controller.pyAdditional Media:
Checklist: