arg1 and arg2 repeating same text

[mishumausam]

http://trusty:9020/sentences?arg1=There+are+strong++indications+that+Abu+Sayyaf%2C+Jemaah+Islamiyah%2C+and+other++terrorist+groups+are+really+collaborating+to+have+a+common+type+of+IED&rel=said&arg2=There+are+strong++indications+that+Abu+Sayyaf%2C+Jemaah+Islamiyah%2C+and+other++terrorist+groups+are+really+collaborating+to+have+a+common+type+of+IED%2C++Ferro

[schmmd]

The arg1 is contained as one of the arguments in the arg2 on the sentence page.

[schmmd]

This is actually indexed as such. The SRL demo does not have the same output.

http://nlpweb.cs.washington.edu/log/1945

<doc>
<str name="id">827699</str>
<str name="arg1">
There are strong indications that Abu Sayyaf, Jemaah Islamiyah, and other terrorist groups are really collaborating to have a common type of IED
</str>
<str name="arg1_exact">
There are strong indications that Abu Sayyaf, Jemaah Islamiyah, and other terrorist groups are really collaborating to have a common type of IED
</str>
<str name="rel">said</str>
<str name="rel_exact">said</str>
<arr name="arg2">
<str>
There are strong indications that Abu Sayyaf, Jemaah Islamiyah, and other terrorist groups are really collaborating to have a common type of IED, Ferro
</str>
</arr>
<arr name="arg2_exact">
<str>
There are strong indications that Abu Sayyaf, Jemaah Islamiyah, and other terrorist groups are really collaborating to have a common type of IED, Ferro
</str>
</arr>
<str name="arg1_postag">
EX VBP JJ NNS IN NNP NNP , NNP NNP , CC JJ JJ NNS VBP RB VBG TO VB DT JJ NN IN NNP
</str>
<str name="rel_postag">.</str>
<str name="arg2_postag">NNP VBD</str>
<arr name="arg1_types">
<str>ArabicName</str>
<str>StanfordPERSON</str>
</arr>
<arr name="arg2_types">
<str>StanfordPERSON</str>
</arr>
<str name="context"/>
<str name="context_exact"/>
<double name="confidence">0.9429314535406047</double>
<str name="sentence">
There are strong indications that Abu Sayyaf, Jemaah Islamiyah, and other terrorist groups are really collaborating to have a common type of IED, Ferro said.
</str>
<str name="sentence_exact">
There are strong indications that Abu Sayyaf, Jemaah Islamiyah, and other terrorist groups are really collaborating to have a common type of IED, Ferro said.
</str>
<str name="extractor">Srl</str>
<str name="url">
/scratch/dovetail/corpus/kdd3-sentences-parsed/tosir283.pdf.txt.sentences
</str>
<long name="_version_">1435956443506278401</long>
</doc>.

[schmmd]

This is a problem with the SRL extractor. It's hard to troubleshoot because it stems from the unicode issues. Here is the unicode sentence and the Clear dependencies.

There are “strong indications that Abu Sayyaf, Jemaah Islamiyah, and other terrorist groups are really collaborating to have a common type of IED,” Ferro said.

expl(are_VBP_1_6, There_EX_0_0); attr(are_VBP_1_6, indications_NNS_3_18); amod(indications_NNS_3_18, strong_JJ_2_10); ccomp(indications_NNS_3_18, collaborating_VBG_17_104); nn(Sayyaf_NNP_6_39, Abu_NNP_5_35); punct(Sayyaf_NNP_6_39, ,_,_7_45); conj(Sayyaf_NNP_6_39, Islamiyah_NNP_9_54); nn(Islamiyah_NNP_9_54, Jemaah_NNP_8_47); punct(Islamiyah_NNP_9_54, ,_,_10_63); cc(Islamiyah_NNP_9_54, and_CC_11_65); conj(Islamiyah_NNP_9_54, groups_NNS_14_86); amod(groups_NNS_14_86, other_JJ_12_69); amod(groups_NNS_14_86, terrorist_JJ_13_76); complm(collaborating_VBG_17_104, that_IN_4_30); nsubj(collaborating_VBG_17_104, Sayyaf_NNP_6_39); aux(collaborating_VBG_17_104, are_VBP_15_93); advmod(collaborating_VBG_17_104, really_RB_16_97); xcomp(collaborating_VBG_17_104, have_VB_19_121); aux(have_VB_19_121, to_TO_18_118); dobj(have_VB_19_121, type_NN_22_135); det(type_NN_22_135, a_DT_20_126); amod(type_NN_22_135, common_JJ_21_128); prep(type_NN_22_135, of_IN_23_140); pobj(of_IN_23_140, IED_NN_24_143); nn(Ferro_NNP_27_149, _NNP_26_147); ccomp(said_VBD_28_155, are_VBP_1_6); punct(said_VBD_28_155, ,_,_25_146); nsubj(said_VBD_28_155, Ferro_NNP_27_149); punct(said_VBD_28_155, ._._29_159)