def lex_search(self):
"""
Update the viewer after lexicalization
"""
words = self._e.get().split()
if len(words) == 0:
return
self.clean_anchor()
self._tagset = TAGTreeSet()
self._treeview.clear()
for word in words:
self._tagset[word] = TAGTreeSet()
fset = self._tagset[word]
lex_list = word_to_features(word)
self.lexicalize_tagset(lex_list, fset)
for morph in lex_list:
if len(morph[0]) > 1:
phrases = [v[0] for v in morph[0]]
if words == phrases:
index = ''
for i in phrases:
index = index + i + ' '
if not isinstance(self._tagset[index], TAGTreeSet):
self._tagset[index] = TAGTreeSet()
fset = self._tagset[index]
self.lexicalize_tagset([morph], fset)
self._tagset.set_start_fs(self._alltrees.start_fs)
self._treeview.update(self._tagset)
self._count = {}
def update_lex_tree(self):
"""
Update the viewer with selected words
"""
words = self.phrases.get().split()
if len(words) == 0:
return
self._tagset = TAGTreeSet()
self._e.delete(0, END)
self._treeview.clear()
for word in words:
self._tagset[word] = TAGTreeSet()
fset = self._tagset[word]
self.lexicalize_tagset(lex_list, fset)
for morph in lex_list:
if len(morph[0]) > 1:
phrases = [v[0] for v in morph[0]]
if phrases == words:
index = ''
for i in phrases:
index = index + i + ' '
if not index in self._tagset:
self._tagset[index] = TAGTreeSet()
fset = self._tagset[index]
self.lexicalize_tagset([morph], fset)
self._tagset.set_start_fs(self._alltrees.start_fs)
self._treeview.update(self._tagset)
self._count = {}
def clear(self):
"""
Empty the treeview and TAG tree set.
"""
self._sfs_button['text'] = 'Add Start Features'
self._add_fs = True
x = self._tagview.get_children()
for item in x:
self._tagview.delete(item)
self._trees = TAGTreeSet()
class TAGTreeSetView(object):
"""
A window that displays a TAG Tree set. TAG Tree Set
contains a group of TAG trees, when clicking the tree
name on the tree list, the tree will be displayed on
the canvas.
"""
def __init__(self, tagtrees, parent=None):
self._trees = tagtrees
if parent is None:
self._top = Tk()
self._top.title('NLTK')
self._top.bind('<Control-p>', lambda e: self.print_to_file())
self._top.bind('<Control-x>', self.destroy)
self._top.bind('<Control-q>', self.destroy)
self._top.geometry("1400x800")
else:
self._top = parent[0]
frame = Frame(self._top)
v = StringVar()
self.w = Label(frame, text='Regexp:')
self._e = Entry(frame, textvariable=v)
self._e.bind("<Return>", self.return_pressed)
self._show_fs = True
self._show_fs_button = Button(frame, text="Hide Features",
command=self.show_fs)
self._sfs_button = Button(frame, text="Add Start Features",
command=self.start_feat)
self.highlight_button = Button(frame, text="Highlight",
command=self.highlight)
self.remove_button = Button(frame, text="Remove",
command=self.remove)
self.keep_button = Button(frame, text="Keep",
command=self.keep)
self._sfs_button.pack(side=LEFT)
self._show_fs_button.pack(side=LEFT)
self.w.pack(side=LEFT)
self._e.pack(expand=1, fill='both', side = LEFT)
self.highlight_button.pack(side=RIGHT)
self.keep_button.pack(side=RIGHT)
self.remove_button.pack(side=RIGHT)
statframe = Frame(self._top)
self.notfl = Label(statframe, text='Tree Framilies:')
self.notf = StringVar()
self.nott = StringVar()
self.tfcl = Label(statframe, textvariable=self.notf)
self.nottl = Label(statframe, text='Trees:')
self.tcl = Label(statframe, textvariable=self.nott)
self.notf.set(str(self._trees.tree_family_count()))
self.nott.set(str(self._trees.tree_count()))
statframe.pack(side = BOTTOM, fill='both')
self.notfl.pack(side = LEFT)
self.tfcl.pack(side = LEFT)
self.nottl.pack(side = LEFT)
self.tcl.pack(side = LEFT)
frame.pack(side = BOTTOM, fill='both')
self._frame = Frame(self._top)
self._frame.pack(fill='both', side = LEFT)
self.cols = ('fullpath', 'type')
self._add_fs = True
self._tagview = ttk.Treeview(self._frame, columns=self.cols,
displaycolumns='',
yscrollcommand=lambda f, l:autoscroll(vsb, f, l),
xscrollcommand=lambda f, l:autoscroll(hsb, f, l))
ysb = ttk.Scrollbar(self._frame, orient=VERTICAL,
command=self._tagview.yview)
xsb = ttk.Scrollbar(self._frame, orient=HORIZONTAL,
command=self._tagview.xview)
self._tagview['yscroll'] = ysb.set
self._tagview['xscroll'] = xsb.set
if parent:
self._tagview.bind('<<TreeviewSelect>>', parent[1])
else:
self._tagview.bind('<<TreeviewSelect>>', self.display)
self.populate_tree('', self._trees)
self._tagview.configure(xscrollcommand=xsb.set,
yscrollcommand=ysb.set)
ysb.pack(fill='y', side='right')
xsb.pack(fill='x', side='bottom')
self._tagview.heading('#0', text='Trees', anchor=W)
self._tagview.column('#0', stretch=1, width=220)
self._tagview.pack(expand=1, fill='both')
self._tw = TAGTreeView(None, parent=self._top)
self._tw.pack(expand=1, fill='both', side = LEFT)
self.sfs_tree ={}
def return_pressed(self, event):
"""
Short-cut for pressing return to show feature structures
"""
words = self._e.get().split()
if len(words) == 0:
self._show_fs = False
self.show_fs()
return
def start_feat(self):
"""
Add or Remove start feature structure of TAG tree set
"""
if self._trees.start_fs is None:
raise TypeError("Should set start feature for TAG Trees First")
node = self._tagview.focus()
path = self._tagview.set(node, "fullpath").split('/')
tree = self._trees
for subpath in path[1:]:
if subpath in tree:
tree = tree[subpath]
else:
raise TypeError("%s: tree does not match"
% type(self).__name__)
if isinstance(tree, type(self._trees)):
return
if not tree.start_feat:
self._sfs_button['text'] = 'Delete Start Features'
self.add_start_fs(tree, self._trees.start_fs)
tree.start_feat = True
else:
self._sfs_button['text'] = 'Add Start Features'
self.del_start_fs(tree)
tree.start_feat = False
self._add_fs = not self._add_fs
def add_start_fs(self, tree, start_fs):
"""
Add start feature structure.
:param tree: display tree
:type: TAGTree
:param start_fs: start feature structure
:type: FeatStruct
"""
root = tree.get_node_name() + '.t'
all_fs = tree.get_all_fs()
self._old_sfs = copy.deepcopy(all_fs[root])
for i in start_fs:
all_fs[root][i] = start_fs[i]
tree.set_all_fs(all_fs)
self._tw.redraw(self._show_fs, tree)
def del_start_fs(self, tree):
"""
Remove start feature structure.
:param tree: display tree
:type: TAGTree
"""
root = tree.get_node_name() + '.t'
all_fs = tree.get_all_fs()
all_fs[root] = self._old_sfs
tree.set_all_fs(all_fs)
self._old_sfs = None
self._tw.redraw(self._show_fs, tree)
def pack(self):
"""
Pack the canvas frame of ``TAGTreeView``.
"""
self._tagview.pack(expand=1, fill='both')
self._frame.pack(fill='both', side = LEFT)
self._tw.pack(expand=1, fill='both', side = LEFT)
def focus(self):
"""
Get selected TAGTree
:return: selected treeview
:rtype: TAGTree
"""
node = self._tagview.focus()
path = self._tagview.set(node, "fullpath").split('/')
tree = self._trees
for subpath in path[1:]:
if subpath in tree:
if not isinstance(tree, type(self._trees)):
if tree._lex:
tree[subpath] = tree[subpath].copy(True)
tree = tree[subpath]
else:
raise TypeError("%s: tree does not match"
% type(self).__name__)
if not isinstance(tree, type(self._trees)):
if tree._lex:
tree.lexicalize()
tree._lex = False
return (tree, subpath)
def display(self, event=None):
"""
Display the tag tree on the canvas when the tree
is selected.
"""
node = self._tagview.focus()
path = self._tagview.set(node, "fullpath").split('/')
tree = self._trees
for subpath in path[1:]:
if subpath in tree:
if not isinstance(tree, type(self._trees)):
if tree._lex:
tree[subpath] = tree[subpath].copy(True)
tree = tree[subpath]
else:
raise TypeError("%s: tree does not match"
% type(self).__name__)
if not isinstance(tree, type(self._trees)):
if tree.start_feat:
self._sfs_button['text'] = 'Delete Start Features'
else:
self._sfs_button['text'] = 'Add Start Features'
if tree._lex:
tree.lexicalize()
tree._lex = False
self._tw.redraw(self._show_fs, tree)
else:
self._tw.redraw(self._show_fs, tree)
def populate_tree(self, parent, trees):
"""
Popluate the trees on the treeview.
"""
if not trees:
return
for t in sorted(trees.keys()):
node = parent
parent_path = self._tagview.set(parent, "fullpath")
path = parent_path + '/' + t
if ord(t[0]) < 10:
f_chr = self.greek(t[0])
else:
f_chr = t[0]
if isinstance(trees[t], type(trees)):
node = self._tagview.insert(parent, END, text=f_chr+t[1:],
values=[path, 'directory'])
self.populate_tree(node, trees[t])
else:
self._tagview.insert(parent, END, text=f_chr+t[1:],
values=[path, 'file'])
def clear(self):
"""
Empty the treeview and TAG tree set.
"""
self._sfs_button['text'] = 'Add Start Features'
self._add_fs = True
x = self._tagview.get_children()
for item in x:
self._tagview.delete(item)
self._trees = TAGTreeSet()
def update(self, trees):
"""
Update the window when the change the TAG tree set.
"""
self._tw.clear()
self._trees = trees
self.populate_tree('', trees)
self.notf.set(str(self._trees.tree_family_count()))
self.nott.set(str(self._trees.tree_count()))
def greek(self, ascii):
"""
Translate ASCII to greek letter
"""
i = ord(u'\u03af') + ord(ascii)
return chr(i)
def destroy(self, *e):
if self._top is None: return
self._top.destroy()
self._top = None
def mainloop(self, *args, **kwargs):
"""
Enter the Tkinter mainloop. This function must be called if
this demo is created from a non-interactive program (e.g.
from a secript); otherwise, the demo will close as soon as
the script completes.
"""
if in_idle(): return
self._top.mainloop(*args, **kwargs)
def show_fs(self):
"""
Display or hide the feature structure on the canvas.
"""
if self._show_fs:
self._show_fs_button['text'] = 'Show Feature'
else:
self._show_fs_button['text'] = 'Hide Feature'
self._e.delete(0, END)
self._show_fs = not self._show_fs
node = self._tagview.focus()
path = self._tagview.set(node, "fullpath").split('/')
tree = self._trees
for subpath in path[1:]:
if subpath in tree:
tree = tree[subpath]
else:
raise TypeError("%s: tree does not match"
% type(self).__name__)
if not isinstance(tree, type(self._trees)):
self._tw.redraw(self._show_fs, tree)
def keep(self):
"""
Display the feature structures which match the
input regular expression.
"""
node = self._tagview.focus()
path = self._tagview.set(node, "fullpath").split('/')
tree = self._trees
for subpath in path[1:]:
if subpath in tree:
tree = tree[subpath]
else:
raise TypeError("%s: tree does not match"
% type(self).__name__)
if not isinstance(tree, type(self._trees)):
self._tw.redraw(self._show_fs, tree, keep=True, reg=self._e.get())
self._show_fs_button['text'] = 'Show ALL Features'
self._show_fs = False
#return
def highlight(self):
"""
Remove the feature structures which match the
input regular expression from the canvas.
"""
node = self._tagview.focus()
path = self._tagview.set(node, "fullpath").split('/')
tree = self._trees
for subpath in path[1:]:
if subpath in tree:
tree = tree[subpath]
else:
raise TypeError("%s: tree does not match"
% type(self).__name__)
if not isinstance(tree, type(self._trees)):
self._tw.redraw(self._show_fs, tree,
highlight=True, reg=self._e.get())
def remove(self):
"""
Highlight the feature structures which match
the input regular expression.
"""
node = self._tagview.focus()
path = self._tagview.set(node, "fullpath").split('/')
tree = self._trees
for subpath in path[1:]:
if subpath in tree:
tree = tree[subpath]
else:
raise TypeError("%s: tree does not match"
% type(self).__name__)
if not isinstance(tree, type(self._trees)):
self._tw.redraw(self._show_fs, tree,
remove=True, reg=self._e.get())
self._show_fs_button['text'] = 'Show ALL Features'
self._show_fs = False
class LexView(object):
"""
A window that lexicalize a TAG Tree set. When type in
word in the text entry, the TAG Tree Set will be lexicalized.
Only TAG Trees that can be lexicalized with that word
remain in the window. The word will be attached to the
substitution node of the tree, and feature structure
will be updated to make the tree consistent with the word.
"""
def __init__(self, grammar):
self._top = Tk()
self._top.title('NLTK')
self._top.bind('<Control-x>', self.destroy)
self._top.bind('<Control-q>', self.destroy)
self._top.geometry("1400x800")
self.frame = Frame(self._top)
v = StringVar()
self._e = Entry(self.frame, textvariable=v)
self._e.bind("<Return>", self.return_pressed)
all_button = Button(self.frame, text='Show All', command=self.show_all)
button = Button(self.frame, text="Search", command=self.lex_search)
self._e.pack(expand=1, fill='both', side=LEFT)
self.phrases = StringVar()
self.phrases.set("")
update_button = Button(self.frame, text="Select", command=self.update_lex_tree)
self._w = OptionMenu(self.frame, self.phrases, [])
update_button.pack(side=RIGHT)
self._w.pack(side=RIGHT)
wl = Label(self.frame, text='Anchors:')
wl.pack(side=RIGHT)
all_button.pack(side=RIGHT)
button.pack(side=RIGHT)
self.frame.pack(fill='both')
self.grammar = grammar
self._tagset = grammar.trees
self._treeview = TAGTreeSetView(self._tagset, (self._top, self.display))
self._treeview.pack()
self._dicword = {}
self._count = {}
def display(self, event=None):
"""
Display the tag tree on the canvas when the tree
is selected.
"""
node = self._treeview._tagview.focus()
path = self._treeview._tagview.set(node, "fullpath").split('/')
tree = self._treeview._trees
for subpath in path[1:]:
if subpath in tree:
if not isinstance(tree, type(self._treeview._trees)):
if tree._lex:
tree[subpath] = tree[subpath].copy(True)
tree = tree[subpath]
else:
raise TypeError("%s: tree does not match"
% type(self).__name__)
if not isinstance(tree, type(self._treeview._trees)):
# Words is supposed to be all words with associated with a tree
words = []
if len(words) > 0:
self.update_anchor(words)
if tree.start_feat:
self._treeview._sfs_button['text'] = 'Delete Start Features'
else:
self._treeview._sfs_button['text'] = 'Add Start Features'
if tree._lex:
tree.lexicalize()
tree._lex = False
self._treeview._tw.redraw(self._treeview._show_fs, tree)
else:
self._treeview._tw.redraw(self._treeview._show_fs, tree)
else:
for path in tree:
if isinstance(tree[path], type(self._treeview._trees)):
return
self.phrases.set("")
if isinstance(tree, type(self._treeview._trees)) and subpath[-6:] == '.trees':
tree_fam = tree[subpath]
treename = subpath[:-6]
else:
treename = subpath
def update_anchor(self, words):
"""
Update anchors with selected words
:param: selected words
:type: basestring
"""
self.clean_anchor()
for choice in words:
phrase = ''
for term in choice:
phrase = phrase + term[0] + ' '
if not phrase in self._dicword:
self._dicword[phrase] = choice
else:
continue
import tkinter as tk
self._w['menu'].add_command(label=phrase, command=tk._setit(self.phrases, phrase))
def clean_anchor(self):
"""
Clean anchors of this viewer
"""
self._dicword = {}
self._w['menu'].delete(0, 'end')
def update_lex_tree(self):
"""
Update the viewer with selected words
"""
words = self.phrases.get().split()
if len(words) == 0:
return
self._tagset = TAGTreeSet()
self._e.delete(0, END)
self._treeview.clear()
for word in words:
self._tagset[word] = TAGTreeSet()
fset = self._tagset[word]
self.lexicalize_tagset(lex_list, fset)
for morph in lex_list:
if len(morph[0]) > 1:
phrases = [v[0] for v in morph[0]]
if phrases == words:
index = ''
for i in phrases:
index = index + i + ' '
if not index in self._tagset:
self._tagset[index] = TAGTreeSet()
fset = self._tagset[index]
self.lexicalize_tagset([morph], fset)
self._tagset.set_start_fs(self._alltrees.start_fs)
self._treeview.update(self._tagset)
self._count = {}
def return_pressed(self, event):
"""
Short-cut for pressing return to show feature structures
"""
self.clean_anchor()
words = self._e.get().split()
if len(words) == 0:
self.show_all()
return
self.lex_search()
def resize(self, *e):
"""
Resize the window
"""
self._treeview(e)
def show_all(self):
"""
Show all the TAG Tree in window
"""
self.clean_anchor()
self._e.delete(0, END)
self._tagset = self._alltrees
self._treeview.clear()
self._treeview.update(self._tagset)
return
def lex_search(self):
"""
Update the viewer after lexicalization
"""
words = self._e.get().split()
if len(words) == 0:
return
self.clean_anchor()
self._tagset = TAGTreeSet()
self._treeview.clear()
for word in words:
self._tagset[word] = TAGTreeSet()
fset = self._tagset[word]
lex_list = word_to_features(word)
self.lexicalize_tagset(lex_list, fset)
for morph in lex_list:
if len(morph[0]) > 1:
phrases = [v[0] for v in morph[0]]
if words == phrases:
index = ''
for i in phrases:
index = index + i + ' '
if not isinstance(self._tagset[index], TAGTreeSet):
self._tagset[index] = TAGTreeSet()
fset = self._tagset[index]
self.lexicalize_tagset([morph], fset)
self._tagset.set_start_fs(self._alltrees.start_fs)
self._treeview.update(self._tagset)
self._count = {}
def lexicalize_tagset(self, lex_list, fset):
"""
Lexicalize all TAG tree set
:param lex_list: a list of lexicalized node
:type: list
:alltrees: TAG tree set to be lexicalized
:type: TAGTreeSet
"""
for morph in lex_list:
index = ''
for i in morph[0]:
index = index + i[0] + '.' + i[1] + ' '
for i in morph[5][1]:
index = index + i + '_'
index = index[:-1]
if index not in fset:
fset[index] = TAGTreeSet()
sset = fset[index]
if len(morph[2]) > 0:
for tf in morph[2]:
ckey = index+tf
tf = tf + '.trees'
if tf not in sset:
key = tf
self._count[ckey] = 0
else:
key = tf[:-5] + '_' + str(self._count[ckey]) + '.trees'
sset[key] = TAGTreeSet()
index = None
for sub in self._alltrees:
if tf in self._alltrees[sub]:
index = sub
if not index:
raise NameError('No tree family')
sset[key] += self._alltrees[index][tf].copy(True)
for t in sset[key]:
if sset[key][t]._lex:
sset[key][t]._lex_fs
sset[key][t].init_lex(morph[0], morph[3], morph[4])
self._count[ckey] += 1
else:
for t in morph[1]:
ckey = index+t
if t not in sset:
key = t
self._count[ckey] = 0
else:
key = t + '_' + str(self._count[ckey])
for sub in self._alltrees:
for tr in self._alltrees[sub]:
if t in self._alltrees[sub][tr]:
sset[key] = self._alltrees[sub][tr][t].copy(True)
self._count[ckey] += 1
if not isinstance(sset[key], TAGTree):
raise TypeError('Not TAGTree')
sset[key].init_lex(morph[0], morph[3], morph[4])
def destroy(self, *e):
if self._top is None: return
self._top.destroy()
self._top = None
def mainloop(self, *args, **kwargs):
if in_idle(): return
self._top.mainloop(*args, **kwargs)
