conll2cgel.py: incorporate subtokens

convertor/conll2cgel.py

@@ -55,6 +55,8 @@ def infer_cgel_pos(dtree: DependencyGraph) -> None:
                     cpos = 'V_aux'
                 case 'JJ'|'JJR'|'JJS':
                     cpos = 'Adj'
+                case 'RB' if node['word'].lower()=="n't":  # negative clitic
+                    cpos = ':subt'
                 case 'RB'|'RBR'|'RBS'|'WRB':
                     cpos = 'Adv'    # TODO: some actually P
                 case 'TO':
@@ -81,7 +83,7 @@ def attach_subtokens(dtree: DependencyGraph) -> None:
         if i==0: continue
         if node['cpos']==':subt':    # 's clitic
             prevnode = dtree.nodes[i-1]
-            prevnode['cfeats'] = [(':subt', prevnode['word']), (':subt', node['word'])]
+            prevnode['subtoks'] = [(':subt', prevnode['word']), (':subt', node['word'])]
             prevnode.setdefault('extra',[]).append('poss')
             prevnode['word'] += node['word']
 
@@ -391,6 +393,8 @@ def build_cgel_tree(targettree: CGELTree, subtree: T, parent: int, dtree: Depend
                 if dnode['lemma'] != dnode['word']:
                     targettree.tokens[i].lemma = dnode['lemma']
                 targettree.tokens[i].xpos = dnode['tag']
+                if 'subtoks' in dnode:
+                    targettree.tokens[i].substrings = dnode['subtoks']
 
 def attach_punct(tree: CGELTree, sent: list[tuple[str,int]], dtree: dict) -> None:
     """Add punctuation terminals to the CGEL tree given the CGEL-tokenized

convertor/wsj-dev-sample-conll2008.conll

@@ -65,3 +65,25 @@
 12	market	market	NN	NN	market	market	NN	13	NMOD	_	_	_	_	A0
 13	jitters	jitters	NNS	NNS	jitters	jitters	NNS	11	OBJ	jitters.01	_	_	A1	_
 14	.	.	.	.	.	.	.	4	P	_	_	_	_	_
+
+1	``	``	``	``	``	``	``	3	P	_	_	_	_	_
+2	It	it	PRP	PRP	It	it	PRP	3	SBJ	_	_	_	_	_
+3	's	be	VBZ	VBZ	's	be	VBZ	0	ROOT	_	_	_	_	_
+4	the	the	DT	DT	the	the	DT	7	NMOD	_	_	_	_	_
+5	most	most	RBS	RBS	most	most	RBS	6	AMOD	_	AM-EXT	_	_	_
+6	distracting	distracting	VBG	JJ	distracting	distract	VBG	7	NMOD	distract.01	_	_	_	_
+7	thing	thing	NN	NN	thing	thing	NN	3	PRD	_	A0	_	_	_
+8	in	in	IN	IN	in	in	IN	7	LOC	_	AM-LOC	_	_	_
+9	my	my	PRP$	PRP$	my	my	PRP$	10	NMOD	_	_	A0	_	_
+10	life	life	NN	NN	life	life	NN	8	PMOD	life.01	_	_	_	_
+11	--	--	:	:	--	--	:	3	P	_	_	_	_	_
+12	I	i	PRP	PRP	I	i	PRP	13	SBJ	_	_	_	A0	_
+13	ca	ca	MD	MD	ca	ca	MD	3	COORD	_	_	_	AM-MOD	_
+14	n't	not	RB	RB	n't	not	RB	13	ADV	_	_	_	AM-NEG	_
+15	even	even	RB	RB	even	even	RB	13	ADV	_	_	_	AM-DIS	_
+16	attend	attend	VB	VB	attend	attend	VB	13	VC	attend.01	_	_	_	_
+17	to	to	TO	TO	to	to	TO	16	ADV	_	_	_	A1	_
+18	my	my	PRP$	PRP$	my	my	PRP$	19	NMOD	_	_	_	_	A0
+19	business	business	NN	NN	business	business	NN	17	PMOD	business.01	_	_	_	A1
+20	.	.	.	.	.	.	.	3	P	_	_	_	_	_
+21	''	''	''	''	''	''	''	3	P	_	_	_	_	_

convertor/wsj-dev-sample-conll2008.conll10

@@ -65,3 +65,25 @@
 12	market	market	NN	NN	_	13	NMOD	_	_
 13	jitters	jitters	NNS	NNS	_	11	OBJ	_	_
 14	.	.	.	.	_	4	P	_	_
+
+1	``	``	``	``	_	3	P	_	_
+2	It	it	PRP	PRP	_	3	SBJ	_	_
+3	's	be	VBZ	VBZ	_	0	ROOT	_	_
+4	the	the	DT	DT	_	7	NMOD	_	_
+5	most	most	RBS	RBS	_	6	AMOD	_	_
+6	distracting	distracting	VBG	JJ	_	7	NMOD	_	_
+7	thing	thing	NN	NN	_	3	PRD	_	_
+8	in	in	IN	IN	_	7	LOC	_	_
+9	my	my	PRP$	PRP$	_	10	NMOD	_	_
+10	life	life	NN	NN	_	8	PMOD	_	_
+11	--	--	:	:	_	3	P	_	_
+12	I	i	PRP	PRP	_	13	SBJ	_	_
+13	ca	ca	MD	MD	_	3	COORD	_	_
+14	n't	not	RB	RB	_	13	ADV	_	_
+15	even	even	RB	RB	_	13	ADV	_	_
+16	attend	attend	VB	VB	_	13	VC	_	_
+17	to	to	TO	TO	_	16	ADV	_	_
+18	my	my	PRP$	PRP$	_	19	NMOD	_	_
+19	business	business	NN	NN	_	17	PMOD	_	_
+20	.	.	.	.	_	3	P	_	_
+21	''	''	''	''	_	3	P	_	_