.. _model_cl_test:

===========================
Model-Specific Closure Test 
===========================

Colibri can be used to run a model-specific closure tests. Here we show how to 
do so for the Les Houches parametrisation (implemented :ref:`here <lh_model>`).

Step 1: runcard
---------------

You can adapt the runcard presented :ref:`here <lh-closure-test>` to run a closure test so that it includes
the following:

.. code-block:: bash
    
    # MODEL-SPECIFIC CLOSURE TEST:
    closure_test_pdf: colibri_model
    closure_test_model_settings:
        model: les_houches_example
        # The parameters of the model.
        parameters:
            beta_gluon: 10.9      
            alpha_gluon: 0.356    
            alpha_up: 0.718       
            beta_up: 3.81         
            epsilon_up: -1.56     
            gamma_up: 3.30        
            alpha_down: 1.71      
            beta_down: 10.0       
            epsilon_down: -3.83  
            gamma_down: 4.64      
            norm_sigma: 0.211     
            alpha_sigma: -0.048   
            beta_sigma: 2.20      

And remove or comment out ``closure_test_pdf: LH_PARAM_20250519``.
Within the closure_test_model_settings, one needs to specify any of the hyperparameters relevant to build
the model instance. The example model here shown does not necessitate further settings, but in principle
one could have additional hyperparameters to tune.

Note that, unlike for a standard closure test, in this case we need to
specify the initial values for each parameter of the PDF model. For the values above, we 
have taken the best-fit values, taken from Ref. :cite:`Alekhin:2005xgg`. 

Step 2: running the fit
-----------------------

You can produce the fit by running the following command from the 
``colibri/les_houches_example`` directory:

.. code-block:: bash

    les_houches_exe runcards/lh_fit_closure_test.yaml

where the runcard path or name should be changed as needed.

Running the fit will create a directory called ``lh_fit_closure_test``, 
where you will find the output of the fit. You can read more about it in
:ref:`this tutorial <bayes_fit_folders>`.

You can evolve it by following the instructions given in 
:ref:`this tutorial <evolution>`.