The options and command set of the gnocl::tree widget allow for easy creation and editing of data arranged and displayed in a tree structure. What is lacking in the Gtk api is an equivalent of the GtkTreeView serialize and deserialize functionality.
Creating a similar offering for the gnocl::tree is not altogether difficult to implement because it is possible to sequentially extract data from a tree which can then be saved for later use (serialize). When the data is next needed for use, it can be sequentially added to a new tree widget (deserialize).
The sampe code below will produce the following, the tree on the left is created through the direct addtion of data, whereas the tree on the right (with treelines) is created by deserializing the data set of the first tree.
When data contained in a tree was serialzed the following list was returned and save to disk. Each line of list represents a row in the tree where the first item is the node path, and the second item, the data itself. More specifically, the data contains the valued held in each cel on each row. Hence, 2 values for 2 columns per row.
0 {Mamals 0}
{0 0} {Cat 4}
{0 1} {Dog 4}
{0 2} {Human 2}
1 {ABC 0}
{1 0} {Parrots 0}
{1 0 0} {Cockatoo 2}
{1 0 1} {Amazone 2}
{1 1} {Predator 0}
{1 1 0} {Hawk 2}
{1 1 1} {Eagle 2}
2 {Bugs 0}
{2 0} {Beetle 6}
{2 1} {Spider 8}
{2 2} {Worm 0}
The path ids in the above form provide a means of identifying a specific row in the tree. The number of digits in the path id indicate the level of depth within the tree. In this example there are in fact three tree structures displayed simulaneously, namely Mammals (0), Birds (1) and Insects (2).
Whenever an entry is added, the target doesn't exist and so, when deserializing the final element of the path id is ignored. In the case of the tree roots, (Mammals, Birds and Insects in the above example), the first and last nodes are the same, hence the path id set to "".
#---------------
# simple-tree.tcl
#---------------
#!/bin/sh
#\
exec tclsh "$0" "$@"
package require Gnocl
# TODO: Perhaps these two procs can be embedded into the C code?
proc serialize { wid } {
foreach node [$wid nodeList] {
::lappend nodes $node [$wid getRow $node]
}
foreach key [::dict keys $nodes] {
append res "[::list $key] [::list [::dict get $nodes $key]]\n"
}
return $res
}
proc deserialize {wid data} {
foreach { pos val } $data {
if { [::llength $pos] == 1 } { set pos "" }
$wid add [::lrange $pos 0 end-1] [::list $val]
}
}
#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# its necessary to have similar trees, i.e. same number of columns and data types.
set tree(1) [gnocl::tree \
-ruleHint 1 \
-onSelectionChanged {
# use join to safely convert the list to a string.
puts [ %w get [join %p] 0]
# to get an entire row:
puts [join %p]~~~[%w getRow [join %p] ]
} \
-types {string string} \
-titles {"Type" "# of Legs"}]
set tree(2) [gnocl::tree \
-treeLines 1 \
-onSelectionChanged {
# use join to safely convert the list to a string.
puts [ %w get [join %p] 0]
# to get an entire row:
puts [join %p]~~~[%w getRow [join %p] ]
} \
-types {string string} \
-titles {"Type" "# of Legs"}]
# add some items in a hap-hazard sort of way ;-)
$tree(1) add {} {Mammals Birds Insects}
$tree(1) add 0 {{Cat 4} {Dog 4} {Human 2}}
foreach {paro pred} [$tree(1) add 1 {Parrots Raptor}] {}
$tree(1) add $paro {{Cockatoo 2} {Sparrow 2}}
$tree(1) add $pred {{Hawk 2} {Eagle 2}}
$tree(1) add 2 {{Beetle 6} {Spider 8} {Centipede ?}}
set fp [open treeData.dat w]
puts $fp [serialize $tree(1)]
close $fp
set fp [open treeData.dat r]
deserialize $tree(2) [read $fp]
close $fp
set box1 [gnocl::vBox -label TreeView-1 -children $tree(1) -fill 1 -expand 1]
set box2 [gnocl::vBox -label TreeView-2 -children $tree(2) -fill 1 -expand 1]
set box [gnocl::hBox -children [list $box1 $box2] -fill 1 -expand 1]
foreach pid {0 1 2} {
$tree(1) expand -path $pid -recursive 1
$tree(2) expand -path $pid -recursive 1
}
gnocl::window -child $box -setSize 0.35
Creating a similar offering for the gnocl::tree is not altogether difficult to implement because it is possible to sequentially extract data from a tree which can then be saved for later use (serialize). When the data is next needed for use, it can be sequentially added to a new tree widget (deserialize).
The sampe code below will produce the following, the tree on the left is created through the direct addtion of data, whereas the tree on the right (with treelines) is created by deserializing the data set of the first tree.
When data contained in a tree was serialzed the following list was returned and save to disk. Each line of list represents a row in the tree where the first item is the node path, and the second item, the data itself. More specifically, the data contains the valued held in each cel on each row. Hence, 2 values for 2 columns per row.
0 {Mamals 0}
{0 0} {Cat 4}
{0 1} {Dog 4}
{0 2} {Human 2}
1 {ABC 0}
{1 0} {Parrots 0}
{1 0 0} {Cockatoo 2}
{1 0 1} {Amazone 2}
{1 1} {Predator 0}
{1 1 0} {Hawk 2}
{1 1 1} {Eagle 2}
2 {Bugs 0}
{2 0} {Beetle 6}
{2 1} {Spider 8}
{2 2} {Worm 0}
The path ids in the above form provide a means of identifying a specific row in the tree. The number of digits in the path id indicate the level of depth within the tree. In this example there are in fact three tree structures displayed simulaneously, namely Mammals (0), Birds (1) and Insects (2).
Whenever an entry is added, the target doesn't exist and so, when deserializing the final element of the path id is ignored. In the case of the tree roots, (Mammals, Birds and Insects in the above example), the first and last nodes are the same, hence the path id set to "".
#---------------
# simple-tree.tcl
#---------------
#!/bin/sh
#\
exec tclsh "$0" "$@"
package require Gnocl
# TODO: Perhaps these two procs can be embedded into the C code?
proc serialize { wid } {
foreach node [$wid nodeList] {
::lappend nodes $node [$wid getRow $node]
}
foreach key [::dict keys $nodes] {
append res "[::list $key] [::list [::dict get $nodes $key]]\n"
}
return $res
}
proc deserialize {wid data} {
foreach { pos val } $data {
if { [::llength $pos] == 1 } { set pos "" }
$wid add [::lrange $pos 0 end-1] [::list $val]
}
}
#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# its necessary to have similar trees, i.e. same number of columns and data types.
set tree(1) [gnocl::tree \
-ruleHint 1 \
-onSelectionChanged {
# use join to safely convert the list to a string.
puts [ %w get [join %p] 0]
# to get an entire row:
puts [join %p]~~~[%w getRow [join %p] ]
} \
-types {string string} \
-titles {"Type" "# of Legs"}]
set tree(2) [gnocl::tree \
-treeLines 1 \
-onSelectionChanged {
# use join to safely convert the list to a string.
puts [ %w get [join %p] 0]
# to get an entire row:
puts [join %p]~~~[%w getRow [join %p] ]
} \
-types {string string} \
-titles {"Type" "# of Legs"}]
# add some items in a hap-hazard sort of way ;-)
$tree(1) add {} {Mammals Birds Insects}
$tree(1) add 0 {{Cat 4} {Dog 4} {Human 2}}
foreach {paro pred} [$tree(1) add 1 {Parrots Raptor}] {}
$tree(1) add $paro {{Cockatoo 2} {Sparrow 2}}
$tree(1) add $pred {{Hawk 2} {Eagle 2}}
$tree(1) add 2 {{Beetle 6} {Spider 8} {Centipede ?}}
set fp [open treeData.dat w]
puts $fp [serialize $tree(1)]
close $fp
set fp [open treeData.dat r]
deserialize $tree(2) [read $fp]
close $fp
set box1 [gnocl::vBox -label TreeView-1 -children $tree(1) -fill 1 -expand 1]
set box2 [gnocl::vBox -label TreeView-2 -children $tree(2) -fill 1 -expand 1]
set box [gnocl::hBox -children [list $box1 $box2] -fill 1 -expand 1]
foreach pid {0 1 2} {
$tree(1) expand -path $pid -recursive 1
$tree(2) expand -path $pid -recursive 1
}
gnocl::window -child $box -setSize 0.35
Comments