.. _howto:

ParSeq-XAS How-tos
------------------

Pipeline launch and command line options
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

The ParSeq-XAS pipeline starts by ``python XAS_start.py``.

.. hint::

    Use the key ``-h`` to discover the start options. Two particularly useful
    options are ``-p {filename}`` to load an existing project file and
    ``-v 100`` to troubleshoot errors.

Data loading
~~~~~~~~~~~~

.. imagezoom:: _images/filemenu.png
   :align: right
   :loc: upper-right-corner
   :alt:  A popup menu over a column file in the file tree.

.. |icoLast| image:: _images/last.png
   :width: 12

Please see |formats|.

Use the button |icoLast| at the top of the file tree to go to the last used
file location and to use the format definition of that data file.

The content of a column file can be displayed in the ParSeq application in the
Metadata splitter. Use the right button popup menu for this command.

If a beamline produces data files with a variable data format that depends on
the used instruments, the data format definitions can be automated in ParSeq
if the files have a header line with a description of the columns. ParSeq-XAS
implements `auto_format()` methods of a few data nodes (see the module
`XAS_nodes`) that can be tweaked for specific needs.

Data range
~~~~~~~~~~

The data format widget has a tab "conversion". Read its tooltip panel. In
particular, one can use the function `lim(Emin, Emax)` in the top-most edit
line to select a desirable energy range.

Data deglitching
~~~~~~~~~~~~~~~~

.. imagezoom:: _images/XAS-glitch.gif
   :align: right
   :alt:  A demonstration of glitch removal by scaling.

Please see |corrections|.

Although data corrections can be done in any transformation node, the removal
of monochromator glitches is the easiest in the χ(k) node.

**Note**: When deglitching in the µd(E) node, make sure the pre-edge
subtraction and edge normalization are switched off.

.. raw:: html

   <div class="clearer"> </div>

Data combinations
~~~~~~~~~~~~~~~~~

In addition to average, sum and rms deviation, ParSeq offers PCA analysis and
Target Transformation. Unlike many other implementations of PCA analysis, in
ParSeq it is *cumulative*, signifying that each PCA spectrum has a meaning of
the parental spectra. For instance, PCA spectra of a collection of XANES
spectra are also a kind of XANES spectra, not just ever decreasing weak
oscillatory curves. More description with definitions and examples will be
added later.

Project files and data saving
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

ParSeq-XAS comes with a few example project files located in the `saved`
folder. Use the slider in the preview panel to browse through the project node
images:

.. imagezoom:: _images/load-proj.gif
   :align: center
   :alt: &ensp;Preview in a ParSeq project file.

.. note::

   Project files and the involved data files are usually movable to a new
   location (they are not when on network locations) without losing the
   reference from the project file to the data.

.. note::

   When saving a project file, pay attention to the current data selection.
   Only the selected data items will be exported.

Make publication plots
~~~~~~~~~~~~~~~~~~~~~~

.. imagezoom:: _images/save-proj.png
   :align: right
   :alt: &ensp;Saving a ParSeq project file.

1. ParSeq plot windows (based on silx plots) have a Save button that can export
   the plot view to a graphics format.

2. The save project dialog of ParSeq has an option of saving a plotting script
   together with the involved data. The scripts have a few commented lines that
   set energy range, custom colors etc. to help the user tweak the plots.

3. The launch script ``XAS_start.py`` can be started with the ``-p`` key to
   load a project file *and* with the key ``-nG`` to run the pipeline with no
   GUI. The launch script has a short section for this case (no GUI) that does
   plotting with `matplotlib`. This section can be adjusted at will.