Results Service
===============

Introspects simulation results and stores them in a searchable database. Modern replacement for ``registerSquidpgSimtool``.

Quick Start
-----------

.. code-block:: bash

    # Start service
    python -m sim2l.services.results_service --backend sqlite --port 8003

.. code-block:: python

    from sim2l.database import ResultsClient, get_session_manager

    session = get_session_manager().create_anonymous_session()
    client = ResultsClient("http://localhost:8003", session_id=session.session_id)

    # Register result (introspects automatically)
    client.register_result("exec-2024-001")

    # Search by parameters
    results = client.search(
        simulation_name="thermal_sim",
        input_filters={'temperature': 350},
        output_filters={'max_stress': 1.5e8}
    )

    # Get statistics
    stats = client.get_parameter_stats("thermal_sim", "max_stress", param_class="output")

What Gets Introspected
----------------------

From each run database:

- **Parameter Schemas** - Types, units, min/max, defaults
- **Input Values** - Actual input parameter values
- **Output Values** - Computed output values
- **Metadata** - Simulation name, version, status, duration

REST API
--------

- ``POST /register`` - Register a result
- ``POST /search`` - Search by parameters
- ``GET /results/{execution_id}`` - Get specific result
- ``GET /stats/{simulation_name}/{param_name}`` - Get parameter statistics
- ``GET /health`` - Health check

API Reference
-------------

.. autoclass:: sim2l.database.ResultsClient
   :members: