Quick Start to GalfitS Everything
==================================

We will use GalfitS to perform a bulge-disk decomposition of a galaxy J0056-0021, using multi-band images from 10 images from GALEX, SDSS, and 2MASS. The data can be downloaded from `data.zip <https://github.com/RuancunLi/GalfitS-Public/tree/main/examples/data.zip>`_

We will use the configuration file `quickstart.lyric <https://github.com/RuancunLi/GalfitS-Public/tree/main/examples/quickstart.lyric>`_ to run GalfitS. Let's first move to the directory where the data is located:

.. code-block:: bash

    unzip ../../examples/data.zip
    cd data
    mv ../quickstart.lyric .

Now we can run GalfitS:

.. code-block:: bash

    galfits quickstart.lyric --work ./result --num_steps 15000

The result will be saved in the directory ``result``. We choose 15000 steps to ensure the result is converged.

Now we can check the result. First, in the directory ``result``, there is a summary file named ``J0056-0021.gssummary``. The first part of the file looks like this:

.. code-block:: none

    # target: J0056-0021
    # config file: quickstart.lyric
    # fitting mode: images - SED
    # fiting method: optimizer
    # chisq: 14326.755859375
    # reduced chisq: 0.5901126861572266
    # BIC: 14750.84375
    # fitting time: 11.773038339614867 mins
    ############# data summary ############## 
    # image atals: 2mass
    #    image number: 0  band: j  chisq: [479.75192]  dof: [676.]  reduced chisq: [0.7096922]
    #    image number: 1  band: h  chisq: [266.45596]  dof: [676.]  reduced chisq: [0.39416564]
    #    image number: 2  band: ks  chisq: [352.4618]  dof: [676.]  reduced chisq: [0.5213932]
    # image atals: galex
    #    image number: 0  band: galex_fuv  chisq: [235.94382]  dof: [256.]  reduced chisq: [0.92165554]
    #    image number: 1  band: galex_nuv  chisq: [550.92737]  dof: [256.]  reduced chisq: [2.15206]
    # image atals: sdss
    #    image number: 0  band: sloan_u  chisq: [2409.7078]  dof: [4356.]  reduced chisq: [0.5531928]
    #    image number: 1  band: sloan_g  chisq: [2535.5747]  dof: [4356.]  reduced chisq: [0.5820879]
    #    image number: 2  band: sloan_r  chisq: [2095.2021]  dof: [4356.]  reduced chisq: [0.48099223]
    #    image number: 3  band: sloan_i  chisq: [2613.1365]  dof: [4356.]  reduced chisq: [0.59989357]
    #    image number: 4  band: sloan_z  chisq: [2787.595]  dof: [4356.]  reduced chisq: [0.6399438]
    ############ fitting summary ############ 

This summary provides an overview of the fitting results, including the target name, configuration file, fitting mode, method, chi-square values, reduced chi-square, Bayesian Information Criterion (BIC), and the fitting time. It also includes a detailed breakdown of the chi-square values for each image in the different image atlases (GALEX, SDSS, 2MASS), showing the chi-square, degrees of freedom (dof), and reduced chi-square for each band.

The second part of the file is the summary of the parameters:

.. code-block:: none

    pname    best_value
    disk_age_value    1.5849
    disk_Z_value    0.0300
    disk_Av_value    0.4300
    logM_disk    10.2398
    disk_xcen    -0.5388
    disk_ycen    -0.3541
    disk_Re    15.0592
    disk_ang    -43.6280
    disk_axrat    0.4126
    bulge_age_value    0.3981
    bulge_Z_value    0.0200
    bulge_Av_value    1.0165
    logM_bulge    10.6108
    bulge_xcen    -0.1534
    bulge_ycen    -0.1036
    bulge_Re    3.1752
    bulge_n    4.2031
    bulge_ang    24.9169
    bulge_axrat    0.7701
    ......

This table lists the best-fit values for each parameter, such as the age, metallicity (Z), dust extinction (Av), logarithmic stellar mass (logM), star formation history (SFH, see details in :doc:`Appendix <appendix>` ), and structural parameters (e.g., effective radius Re, position angle ang, axis ratio axrat) for both the disk and bulge components.

The above file can be easily read using ``astropy.table``, for example:

.. code-block:: python

    import astropy.table as Table 
    result = Table.read('result/J0056-0021.gssummary', format='ascii')
    logM_bulge = result['best_value'][result['pname'] == 'logM_bulge'][0]

Besides the summary files, there are also two output images. The first is ``J0056-0021image_fit.png``, which displays the original image, the model image, and the residual image for each band:

.. figure:: ./fig/J0056-0021image_fit.png
   :align: center

The second output image is ``J0056-0021SED_model.png``, which shows the SED model for the bulge and disk components, along with the model points for each band and a simple photometry measurement:

.. figure:: ./fig/J0056-0021SED_model.png
   :align: center

Running the above example takes different times on different machines. Below is a table summarizing the approximate fitting times:

.. list-table::
   :header-rows: 1

   * - Machine
     - Time
   * - RTX 4090
     - 2.6 mins
   * - GTX 980Ti
     - 6.0 mins
   * - Apple M4 CPU
     - 12 mins
   * - Apple intel CPU
     - 51 mins