v1.1 (03 Jan 2013)
******************

This document explains the changes made to Iris for this release
(:doc:`View all changes <index>`.)


Features
========

With the release of Iris 1.1, we are introducing support for Mac OS X.
Version 1.1 also sees the first batch of performance enhancements, with
some notable improvements to netCDF/PP import.

* Support for Mac OS X.

* GRIB1 import now supports time units of "3 hours".

* Fieldsfile import now supports unpacked and "CRAY" 32-bit packed data
  in 64-bit Fieldsfiles.

* PP file import now supports "CRAY" 32-bit packed data.

* Various performance improvements, particularly for netCDF import,
  PP import, and constraints.

* GRIB2 export now supports level types of altitude and height
  (codes 102 and 103).

* iris.analysis.cartography.area_weights now supports non-standard
  dimension orders.

* PP file import now adds the "forecast_reference_time" for fields
  where LBTIM is 11, 12, 13, 31, or 32.

* PP file import now supports LBTIM values of 1, 2, and 3.

* Fieldsfile import now has some support for ancillary files.

* Coordinate categorisation functions added for day-of-year and
  user-defined seasons.

* GRIB2 import now has partial support for probability data defined
  with product template 4.9.


Coordinate Categorisation
-------------------------

An :func:`~iris.coord_categorisation.add_day_of_year` categorisation
function has been added to the existing suite in
:mod:`iris.coord_categorisation`.


Custom Seasons
~~~~~~~~~~~~~~

The conventional seasonal categorisation functions have been
complemented by two groups of functions which handle user-defined,
custom seasons.

The first group of functions is:

* :func:`iris.coord_categorisation.add_custom_season`
* :func:`iris.coord_categorisation.add_custom_season_number`
* :func:`iris.coord_categorisation.add_custom_season_year`

These functions mimic their non-custom versions, but with the addition
of a ``seasons`` parameter which is used to define the custom seasons.
These seasons are defined by concatenating the single letter
abbreviations of the relevant, consecutive months.

For example, to categorise a Cube based on "winter" and "summer" months,
one might do::

    >>> seasons = ['mamjja', 'sondjf']
    >>> iris.coord_categorisation.add_custom_season(cube, 'time', seasons)
    >>> print(cube.coord('season').points)
    ['ondjfm' 'ondjfm' 'mamjja' 'mamjja' 'mamjja' 'mamjja' 'mamjja' 'mamjja'
     'ondjfm' 'ondjfm' 'ondjfm' 'ondjfm']

The other custom season function is:

* :func:`iris.coord_categorisation.add_custom_season_membership`.

This function adds a coordinate containing True/False values determined
by membership of a single custom season.


Bugs Fixed
==========

* PP export no longer attempts to set/overwrite the STASH code based on
  the standard_name.

* Cell comparisons now work consistently, which fixes a bug where
  bounded_cell > point_cell compares the point to the bounds but,
  point_cell < bounded_cell compares the points.

* Fieldsfile import now correctly recognises pre v3.1 and post v5.2
  versions, which fixes a bug where the two were interchanged.

* iris.analysis.trajectory.interpolate now handles hybrid-height.