SDP2CONLL

[moonstarsky]

How can I use this to convert SDP format to CONLL format?
Thank you

[danielhers]

$ python semstr/convert.py test_files/20001001.sdp -f conll
Converting: 100%|████████████| 1/1 [00:00<00:00, 87.65file/s, file=20001001.sdp]
$ cat 20001001.conll 
# format = sdp
# sent_id = 20001001
1	Pierre	Pierre	NNP	NNP	_	0	ROOT	_	_
2	Vinken	_generic_proper_ne_	NNP	NNP	_	1	compound_	_
2	Vinken	_generic_proper_ne_	NNP	NNP	_	6	ARG1	__
2	Vinken	_generic_proper_ne_	NNP	NNP	_	9	ARG1	__
3	,	_	,	,	_	0	ROOT	_	_
4	61	_generic_card_ne_	CD	CD	_	0	ROOT	__
5	years	year	NNS	NNS	_	4	ARG1	_	_
6	old	old	JJ	JJ	_	5	measure	_	_
7	,	_	,	,	_	0	ROOT	_	_
8	will	will	MD	MD	_	0	ROOT	_	_
9	join	join	VB	VB	_	12	ARG1	_	_
9	join	join	VB	VB	_	17	loc	_	_
10	the	the	DT	DT	_	0	ROOT	_	_
11	board	board	NN	NN	_	9	ARG2	_	_
11	board	board	NN	NN	_	10	BV	_	_
12	as	as	IN	IN	_	0	ROOT	_	_
13	a	a	DT	DT	_	0	ROOT	_	_
14	nonexecutive	_generic_jj_	JJ	JJ	_	0	ROOT	__
15	director	director	NN	NN	_	12	ARG2	__
15	director	director	NN	NN	_	13	BV	__
15	director	director	NN	NN	_	14	ARG1	__
16	Nov.	Nov.	NNP	NNP	_	0	ROOT	_	_
17	29	_generic_dom_card_ne_	CD	CD	_	16	of	__
18	.	_	.	.	_	0	ROOT	_	_

[danielhers]

Note that the CoNLL format does not allow multiple heads, so these are discarded in the example above.
If, however, you want CoNLL-U format, just say so:

$ python semstr/convert.py test_files/20001001.sdp -f conllu
Converting: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 76.49file/s, file=20001001.sdp]
$ cat 20001001.conllu
# format = sdp
# sent_id = 20001001
# text = Pierre Vinken , 61 years old , will join the board as a nonexecutive director Nov. 29 .
1	Pierre	Pierre	NNP	NNP	_	0	root	0:root	_
2	Vinken	_generic_proper_ne_	NNP	NNP	_	1	compound	1:compound|6:ARG1|9:ARG1	_
3	,	_	,	,	_	1	orphan	1:orphan	_
4	61	_generic_card_ne_	CD	CD	_	1	orphan	1:orphan	_
5	years	year	NNS	NNS	_	4	ARG1	4:ARG1	_
6	old	old	JJ	JJ	_	5	measure	5:measure	_
7	,	_	,	,	_	1	orphan	1:orphan	_
8	will	will	MD	MD	_	1	orphan	1:orphan	_
9	join	join	VB	VB	_	1	orphan	1:orphan	_
10	the	the	DT	DT	_	1	orphan	1:orphan	_
11	board	board	NN	NN	_	9	ARG2	9:ARG2|10:BV	_
12	as	as	IN	IN	_	1	orphan	1:orphan	_
13	a	a	DT	DT	_	1	orphan	1:orphan	_
14	nonexecutive	_generic_jj_	JJ	JJ	_	1	orphan	1:orphan	_
15	director	director	NN	NN	_	12	ARG2	12:ARG2|13:BV|14:ARG1	_
16	Nov.	Nov.	NNP	NNP	_	1	orphan	1:orphan	_
17	29	_generic_dom_card_ne_	CD	CD	_	16	of	16:of	_
18	.	_	.	.	_	1	orphan	1:orphan	_

This way you get all heads in the last column (the deps column).