.. include:: include/links.rst

.. highlight:: python
   :linenothreshold: 3




.. _fit:

Quick Start Guide
====================

Installation
-------------


**csst-ifs-elfo** can be installed as follows:

.. code-block:: console

  git clone https://csst-tb.bao.ac.cn/code/csst-pipeline/ifs/csst_ifs_elfo.git
  cd csst-ifs-elfo
  python -m pip install .


Prepare the IFS data
----------------------
The IFS simulated observation data from CSST can be downloaded from the following link: `Download <http://202.127.29.3/~shen/NGC3359/>`_.

Import Required Modules and Functions

--------------------------

Import the  functions from **csst-ifs-elfo** to perform the fitting:


   
.. code-block:: python

  from elfo import process_i_refit
  from elfo import process_j_refit

.. note:: 
  Both functions share the same fitting logic, fitting the spectra column by column.  
  The only difference lies in the direction of progression  ``process_i_refit`` proceeds along rows,  
  while ``process_j_refit`` proceeds along columns.

The fitting of the IFS data is performed using the two imported functions. 
In the following, we take ``process_i_refit`` as an example, which fits all spectra using the fitting results of adjacent rows.  
For a detailed description of the fitting strategy, see the :ref:`workflow` section below, where we explain how ELFO's ``process_i_refit`` 
function specifically fits the IFS data.



**Parameters**

- ``i_1``: The starting row index for the first fitting. Type: ``int``.
- ``i_2``: The starting row index for the second fitting. Type: ``int``.
- ``fits_file``: The path to the input FITS file. Type: ``str``.
- ``z``: Redshift of the target object. Type: ``float``.
- ``scale_factor`` *(optional)*: The scale factor for rebinning the IFU spectra. Default is ``1``. Type: ``int``.
- ``flux_cube_path`` *(optional)*: The file path to the flux cube of the IFU data. (for pre-binned spectra)Default is ``None``. Type: ``str``.
- ``var_cube_path`` *(optional)*: The file path to the variance cube of the IFU data. (for pre-binned spectra)Default is ``None``. Type: ``str``.
- ``format`` *(optional)*: The format of the input FITS file. Default is ``'muse'``. Type: ``str``.

**Returns**

- ``str``: The path to the output directory where the results are saved.


Set up the model input parameters
----------------------------------
See an example in the `example`_ notbook.



Run the Fitting
----------------

**Usage Example**

Once the IFS data is ready and the functions are imported, we can run the fitting.

.. code-block:: python

    path_out = process_i_refit(i_1=30, i_2=20, fits_file='CSST_IFS_CUBE.fits', z=0.003373, format='csst')
    # Using optional parameters for custom flux/var cubes and scale factor (for pre-binned spectra)
    path_out = process_i_refit(i_1=80, i_1=90, fits_file='example.fits', z=0.02, scale_factor=2, flux_cube_path='reduced2_flux.npy', var_cube_path='reduced2_var.npy', format='csst')

The function will automatically save the fitting results to the returned path (``path_out``).  
For each spectrum,  `PyQSOFit`_ generates a ``.fits`` file with parameters and a ``.pdf`` showing the fitting result.
More details on the fitting process can be found in the :ref:`workflow` section below.

|

----------

|

Here is an example using MUSE spectral fitting of the quasar *MR 2251−178* to demonstrate the improvement 
brought by **ELFO** in emission-line fitting. The upper panel shows the velocity and flux maps of the :math:`\text{H}\alpha` narrow 
component obtained by fitting each spectrum individually using the default initial parameters in `PyQSOFit`_. 
In contrast, **ELFO** utilizes the best-fitting result from neighboring points as the initial guess for each fit. 


.. figure:: figures/initial.png
    :height: 200px
    :align: center

    The result obtained by fitting each spectrum with the default parameters

.. figure:: figures/elfo.png
    :height: 200px
    :align: center

    The result obtained by using the best fit from neighboring points as the initial guess for each spectrum.