Rename works when a basedata or output object are selected.

When a specific output item (eg richness or PD) is selected, the Delete and 
Export buttons are disabled because these actions can't be done for a single 
item.

Rename also can't be done in this case, so the button should be disabled.

This has been corrected for version 0.12

The scree plot below the dendrogram (cluster and view labels tabs) occupies
most of the vertical space in that section of the window.  It needs to be
smaller to begin with.

What steps will reproduce the problem?
1.  Delete the selected basedata object in the GUI
2.  The selection list defaults to a blank entry


What is the expected output? What do you see instead?
It should choose the next on the list, or 'none' if there are no more.  

Clicking on another basedata object corrects it, but choosing another from
the drop down list causes issues with any subsequent analyses.

The export system currently uses the file extension to determine which
format to use.  This needs to be converted to a chooser/combobox to make
the choices more obvious, with associated greying out or hiding of
irrelevant options.

The randomisation export methods fail in the GUI, leaving a window with the
name 'label120'.

sp_match_text does not work when used in a definition query.

Does not match the axes.

Metadata for subroutines is used to build much of the GUI, determining what
options are available at which time.  

This is currently stored as an if-block within each subroutine.  This is a
fragile approach as calls made to subs without metadata can have
undesirable results.  In many cases the metadata is also longer than the
analysis code.

Moving the metadata into separate subs will simplify things, make the code
less cluttered, and should speed up a number of the processes where subs
are called repeatedly.

All main code has been ported to this system for version 0.12.  

The cluster analysis export window is initially not properly sized, causing
some text to be cut off at the bottom. Also, a dividing line near the
bottom crosses over some of the text.  The window may have been originally
sized properly, but new options were added causing the window to be too short.

Could be "fixed" by adding a scroll bar to the window but separating it
into more than one window would be more effective, as then arguments
relevant to the file type can be specified.  

The endemism analyses (calc_endemism_central and calc_endemism_whole) both
return additional lists to allow exploration of species ranges and weights
used in the calculations.  

These are very useful, but can use large amounts of memory for large data
sets.  

They should be optional rather than default.

When displaying a tree object, it is initially shown plotted by node
length, but under the options button the "depth" mode is shown as being
selected. This corrects itself when reselecting the display mode. 

The help menu options currently open a window with a hyperlink that needs
to be copied across to a browser.  

The browser should insted be launched automatically.

What steps will reproduce the problem?
1. Re-run an analysis (possibly one that was cancelled), overwriting the
existing output.  

What is the expected output? What do you see instead?
The outputs tab lists the output twice.  The upper one points to the old
version, the lower one to the new version.  This will continue as analyses
are overwritten. 

Doing this will mean the new outputs are out of synch for any additional
iterations of the existing randomisations.

Warn the user and let them decide.  

The system consumes all the available memory over a series of iterations
(how fast depends on the size of the data set).  

Eventually it gets to the point that there is not enough memory to save the
data set or run any further analyses.  This occurs on WinXP, but is likely
to occur on all systems.

A workaround is to run the analysis for some number of iterations, save,
close and restart the application (this clears the memory), and then resume
the analysis.  Repeat solution as needed. 

The load_and_randomise_wrapper.pl script in the Biodiverse/bin folder has
been developed as a means of automating the close and restart problem but
will likely need modification for some problems.

Nexus files with a ; instead of a , after the last taxon name in the 
translate block are not loaded correctly.

Instead, the terminal node corresponding to the last name, is named as an 
internal node.

See attached file, line 126

This seems to be the format which Mesquite produces.  I'm not sure what 
other programs also do.

Solution: read the last name of a translate block, even if a ; is reached 
instead of the comma which follows other lines.

Workaround: in the meantime, open the tree in a text editor, and add a , 
after the last name in the translate block.

What steps will reproduce the problem?
1. Set both neighbour sets to be self_only() in a spatial analysis
2. Run the analysis.


What is the expected output? What do you see instead?

Warnings are raised 
[INDICES] DOUBLE COUNTING OF ELEMENTS IN calc_abc, 1 + 1 > 1


This is likely due to an optimisation in Spatial.pm not setting the
exclusions properly.

What steps will reproduce the problem?
1. My collaborator is color-blind. For instance red and blue as chosen in
Biodiversity are almost indistinguishable for him.
2. I realize that it is going to be very low priority, but to be able to
choose either grayscale or symbols instead of colors would be useful
3.

What is the expected output? What do you see instead?


What version of the product are you using? On what operating system?


Please provide any additional information below.

The endemism calculations allow the extraction of lists containing the
labels and their ranges and weights (see wiki page
Indices#Endemism_and_Rarity ).  These need to also be available for the
rarity indices, as they are calculated using the same algorithm but with
sample counts insted of ranges.

The process of implementing this will also require the calc_rarity subs to
be modfied in line with the calc_endemism subs (as per issue #46).  

In the spatial analysis and cluster map panes, changing the colour scheme
from “hue” to “sat” or vice versa results in the colours disappearing.  

The view is “fixed” by switching to another index and then back (or
adjusting the slider on the cluster plot), indicating that it's likely some
sort of display refresh problem.

Import a data file with one ore more records containing undefined values
for the group fields.  

The import fails when it reaches the first such record.  


line 102 of file C:\filepath\blah.csv
*** unhandled exception in callback:
***    at /<C:\biodiverse\bin\BiodiverseGUI_win.exe>Biodiverse/GUI/Bas
edataImport.pm line 265
***  ignoring at BiodiverseGUI.pl line 111.
Can't use an undefined value as an ARRAY reference at
/<C:\biodiverse\bin\BiodiverseGUI_win.exe>Biodiverse/GUI/Grid.pm
line 973.
 at
/<C:\\biodiverse\bin\BiodiverseGUI_win.exe>Biodiverse/GUI/Tabs/Outputs.pm
line 174
User cancelled grid initialisation, closing

The GUI currently does not support the deletion of randomisation outputs.  

It will involve extra searching of other outputs to remove the lists that
were added to them.

Rather than making this an option, the ideal would be to recognise and handle 
this file format as part of the default behaviour.

A sample file in this format is a attached.

If the map of a spatial analysis is open, and a new analysis is specified
and overwrites the previous analysis, the display will not change (even if
“display results” is selected for the new analysis), and the tab must be
closed and then reopened for it to display the new analysis properly.

The display code needs cleaning up so events are not triggered by certain.

Currently the system will return quietly if many of the analyses fail (as
far as the GUI is concerned).  It does, however, print warnings to the log
window.

This information should be sent to the GUI to make it obvious.

The method used to build a tree from a matrix passed as an argument crashes
in many cases.  

What steps will reproduce the problem?
1. In the GUI, import a matrix using the normal methods.
2. Use the Matrix -> Convert Matrix to Tree menu option.
3. The conversion will fail, complaining of too many root nodes. 

This is an issue with the underlying conversion code which has not been
updated to conform to the current BaseData cluster analysis methods in
library Biodiverse::Cluster.

The GUI will currently plot only one shapefile overlay at a time.  This
should be extended to allow multiple files. 

Add an option to the GUI to give the list of available spatial conditions
and their syntax.  

This will also require an example option in the function metadata to save
user guesswork.

What steps will reproduce the problem?
1.  Run a spatial analysis and display it.
2.  Now run a randomisation.
3.  Redisplay the spatial analysis.  

What is the expected output? What do you see instead?
The randomisation list results should be available in the combo box at the
bottom left.  It is not.  The list can still be accessed from the popup
windows.

This is due to caching of the lists to help with larger data sets when
definition queries are used.  Running the randomisation adds new lists and
these are not updated in the cached set.  

calc_mx_rao_qe currently returns a zero value if there are no values in the
neighbour sets that are on the matrix.  It should return undef instead so
there is no ambiguity. 

BaseStruct objects now have generic export methods instead of being based
on file extensions.  See issue #33.

The same needs to be done for the trees and matrices.  

Analyses in Biodiverse specify a set of global and local precalculations
required for each subroutine.  This saves processing time overall, as
several analyses might depend on a single subroutine.  Using this method,
the dependency is run once and then its results are passed as arguments to
the subs that need them.

Having a post_calc option would also be useful, for example to run some
cleanup where subs use local caching and some memory could be saved by
clearing it after each window.  

[deleted issue]

What steps will reproduce the problem?
1. Set both the neighbourhood definitions to be empty strings
2. Run the analysis
3. The system fails, complaining about undefined values. The spatial tab
closes, or does not close and cannot be closed.  

Renaming a basedata object while it is also open in the view labels tab
does not rename the tab.

Add capability to the GUI to rename an output without needing to open the
tab, and therefore avoiding the need to display large data sets in order to
rename them.  Also add capability to rename a basedata object, for which
there is currently no option in the GUI.

Often species observation data are presented in a matrix form, where the
columns are species and the locations or their IDs are on the rows. 
Biodiverse exports to this format, but does not import them back in.  

An annulus will allow calculations in concentric rings and is used as the
standard neighbourhood function in many spatial analyses, eg semivariograms.

What steps will reproduce the problem?
1. Run a cluster analysis and export it to nexus
2. Import it as a tree
3. It does not remove the quotes properly from the names so will not link
correctly to a transposed basedata object.

There is currently no option to remap matrix labels on import.  

If an analysis fails for some reason then the associated tab cannot be closed.

They do not come back if the application is closed and restarted.  

Need to implement better error trapping.  

What steps will reproduce the problem?
1. Run an analysis with a definition query to reduce the number of groups
containing results.
2. The lists available for display are sometimes limited to the
SPATIAL_RESULTS, none of the lists generated by indices like
calc_endemism_central are visible.


This is due to only the first group being checked to see what lists are
available.  Need to either search the first "n" groups to get a good
sample, or keep track of lists generated when creating them.

What steps will reproduce the problem?
1. try to import data
2. about 12 columns, 3 groups, 2 labels
3. 3114 lines of data

What is the expected output? What do you see instead?
What I saw is apparently a crash. The GUI never finishes loading the file,
but just continues to indicate busy. The console seems to indicate that the
system has crashed.

In addition, some buttons (like cancel) are not working, and only default
import worked. Trying to tell the GUI that space was the separator and "
the quote character gave incorrect parsing.

What version of the product are you using? On what operating system?
64 bit vista
version 0.10
Gtk+ as per the link in web page

Please provide any additional information below.

output in console window:
[Project] Basedata Model empty
[GUI] Created new Biodiverse project
[Outputs tab] Loaded tab - Outputs
Gtk-WARNING **: Attempting to read the recently used resources file at
`C:\Users
\lca\.recently-used.xbel', but the parser failed: Unexpected attribute
'modified
' for element 'application'. at
/<C:\Biodiverse\biodiverse_0.10\bin\BiodiverseGU
I_win32.exe>Biodiverse/GUI/BasedataImport.pm line 505, <DATA> line 165.
GLib-CRITICAL **: g_bookmark_file_get_size: assertion `bookmark != NULL' failed
at
/<C:\Biodiverse\biodiverse_0.10\bin\BiodiverseGUI_win32.exe>Biodiverse/GUI/Ba
sedataImport.pm line 505, <DATA> line 165.
[GUI] Discovering columns from
C:\Projects\BestRotation\Debug\biodiversitydata.t
xt
[COMMON] Guessed field separator as ' '
[COMMON] Guessed quote char as "
[GUI] Generating make columns dialog for 12 columns
[BASEDATA] Loading from files
C:\Projects\BestRotation\Debug\biodiversitydata.tx
t
[BASEDATA] INPUT FILE: C:\Projects\BestRotation\Debug\biodiversitydata.txt
[COMMON] Guessed quote char as "
[COMMON] Guessed field separator as ' '
[BASEDATA] Line number: 1
[BASEDATA]  Chunk size 3114 lines
Use of uninitialized value in concatenation (.) or string at
/<C:\Biodiverse\bio
diverse_0.10\bin\BiodiverseGUI_win32.exe>Biodiverse/BaseData.pm line 639,
<GEN0>
 line 3115.
[BASEDATA] Non-numeric cell size field 7, check your data or cellsize
arguments.

line 1 of file C:\Projects\BestRotation\Debug\biodiversitydata.txt
*** unhandled exception in callback:
***    at
/<C:\Biodiverse\biodiverse_0.10\bin\BiodiverseGUI_win32.exe>Biodiverse
/GUI/BasedataImport.pm line 265
***  ignoring at BiodiverseGUI.pl line 111.

The system currently does not allow this.  It should.

The addition of a definition query to the arguments will allow the
calculations to be run on only a subset of groups.  This should speed
things up, and also allow the user to work on spatial subsets.

The system currently allows the user to specify a Biodiverse project file
as a command line argument on startup.  

Support is needed for native format BaseData, matrix and tree files as well
to save having to then open it manually.  

What steps will reproduce the problem?
1.  Run an analysis using a recyclable spatial condition, eg sp_block(size
=> 300000)
2.  Set a definition query, eg "$x <= 3650000"


What is the expected output? What do you see instead?
The results are recycled into those groups that failed the definition query
when they should not be.

[deleted issue]

The processing for spatial outputs uses an optimisation such that it reuses
neighbour sets from another spatial output with the same conditions (if one
exists in that basedata).  

These currently do not consider the definition queries, resulting in groups
excluded by the definition query still having results. 

The system does not guard against exclusions being run when a basedata
contains outputs.  This will cause problems with existing cluster and
spatial outputs as they will refer, or be a function of, no longer existing
groups.