.. include:: _config.rst

.. _arrays_page:


The Named Arrays Page
====================================

The Named Arrays page will contain a list of arrays saved in the Sitka
session, and will look like

.. image:: images/sitka_arrayspage.png
   :width: 98 %


Each of the named Array can be highlighted, or multiple arrays can be
selected at a time.  The **Plot Current** button will show the
highlighted array (as a simple plot or image display).

Note that this list may contain arrays named `_imgdat`, `_ydat`, and
`_tabledat`.  These arrays are automatically written as the last data
used by the Image Display page, the XY Display page and the Table
Display page.

You can use **Delete Selected Arrays** to remove selected named arrays.
You can use the **Export Selected Arrays to HDF5** to export
selected arrays to a simple HDF5 file (and so easily read into later
Sitka sessions.

The arrays selected will be written to a new HDF5 file, and all arrays
will be put into a Group name "sitka_arrays".  If you read in this
file in a later session and select thie "sitka_arrays' Group, the
**Import Named Arrays"** at the upper right of the main Sitka window
will be enables.  Hitting that button will import all of these named
arrays into the Named Arrays panel.  This allows you to extract data
from complex datasets and easily retrieve them.


At the bottom, you can also create a new array directly.  Here you
give a name (which must follow the valid conventions for a Python
variable: A letter or underscore followed by any number of letters,
numbers, or underscore), and an mathematical expression to evaluate
for the array value.  This uses an `asteval`_ Interpreter, and so uses
Python conventions, the built-in math functions from the `math` module
and `numpy`, and also contains the other named arrays.  Example
expressions would include

   * '0.01 * arange(4096)', maybe to set an X-axis for XYPlots
   * 'linspace(-2, 2, 501)', as another form to set an X-axis for XYPlots
   * 'image1 - image2' to take the difference of two images.
   * 'sin(ydat/10)'  you can use any trig functions from numpy.
   * 'gradient(ydat)/gradient(xdat)' to take the derivative 'dy/dx'.