CHANGES

# CHANGELOG: (started at 2016-08-15 16:30:22)
#
# NOTE: 2016-08-15 16:30:27 -- created xmlutils.py + a new XmlViewerWindow class in 
# pict.py
# very basic read-only support for OME XML metadata in image files 
# based on the out-going QtXml module


#NOTE: TODO 2017-03-21 21:52:50 implement drag and drop for opening files:
#NOTE: this will quite likely require a custom item model subclassing the default 
#NOTE: QAbstractItemModel that is used by QTreeWidget
#NOTE:
#NOTE: 2017-03-21 22:31:07
#NOTE: until then, I fallback on a half-baked approach: on the drop event I test 
#NOTE: for the type of the drop event source; if it is the ScipyenWindow's history 
#NOTE: tree type (i.e. QtWidgets.QTreeWidget) then I just call ScipyenWindow's
#NOTE: _rerunCommand method asynchronously (by means of Qt signal/slot mechanism)
#NOTE: whereby I emit a custom signal in EmbedIPython's dropEvent method which
#NOTE: should be connected to the ScipyenWindow's _rerunCommand slot
#NOTE:
#NOTE: 2017-03-21 22:31:16
#NOTE: this half-baked approach works pretty well in term of excuting the dropped
#NOTE: command(s) via %rerun line magic; however, it does not allow actually "pasting"
#NOTE: the command to the console input (and thus prevents the dropped command(s) from
#NOTE: beign edited by the user before being actually executed)
#NOTE:
#NOTE: 2017-03-21 22:57:59 
#NOTE: finally I decided to just use the signal/slot mechanism such that ScipyenWindow
#NOTE: code will COPY the string data (the command(s)) on the system's clipboard then use 
#NOTE: that to paste it into the console's text widget

# NOTE: 2017-06-30 23:44:07
#implemented a unified file handling mechanism, using python mimetypes module and 
    #a simple dispatch mechanism
    
#started first steps to uniformize the viewers API -- building towards a dispatch paradigm

#implemented linkable cursors in both SignalViewer and ImageViewer logic

#the two viewers still rely on different kits: SignalViewer on matplotlib; ImageViewer
#almost entirely on pyqt5, except for colormaps which are taken from matplotlib

# NOTE: 2017-06-30 23:48:37 TODO:
# use quantities as values for AxisInfo objects (can one do this?)
#   cannot use quantities: AxisInfo.resolution expects a float value, not a numpy array
#   or a quantity (which is a numpy array subclass) -- Pity!

# NOTE: 2019-09-09 22:06:46
# DONE: partial implementation of Drag & Drop to load files and change working 
# directory
# DONE: using drag & drop to run or load scripts in the console buffer + script manager
#
# FIXME SignalViewer switched to pyqtgraph - seems faster, but there are some bugs
# when reusing plotitems
#
# TODO: 2019-09-09 22:13:50
# streamline the viewing logic; untangle viewing NDarrays e.g. use keyboard mods
# to determine if numpy arrays plotted in SignalViewer or matplotlib figure, versus
# displaying their contents in a tableditor.
# likewise, for neo and datatypes signal objects and vigra Kernel objects, use
# keyboard mods to determine if theyr are ot be plotted in signal viewer or
# have their values displayed in a table editor
# NOTE here extra option to plot vigra Kernel 1D/2D in matplotlib figure