rewriting rule 8: negation

[vcvpaiva]

According to stats (https://github.com/own-pt/rte-sick/blob/master/expanded/conllu/all.conllu.stats.xml) we have 436 neg relations. They do not cover all the negations, e.g. No person is hiking should be negation and has no trace.

1 No no DET DT _ 2 det _ DT|?|?
2 person person NOUN NN _ 4 nsubj _ NN|00007846-n|Human=
3 is be VERB VBZ _ 4 aux _ VBZ|02604760-v|Entity+
4 hiking hike VERB VBG _ 0 ROOT _ VBG|01920932-v|Walking+

in this case the rep gives us the opposite meaning of the sentence. Bad.

How many sentences in this situation?
it would 565 if the Turku query L=no was good enough. But of course there are restricted negations like

1 A a DET DT _ 2 det _ DT|?|?
2 man man NOUN NN _ 7 nsubj _ NN|10287213-n|Man=
3 with with ADP IN _ 2 prep _ IN|?|?
4 no no DET DT _ 5 det _ DT|?|?
5 shirt shirt NOUN NN _ 3 pobj _ NN|04197391-n|Shirt=
6 is be VERB VBZ _ 7 aux _ VBZ|02604760-v|Entity+
7 jumping jump VERB VBG _ 0 ROOT _ VBG|01963942-v|Ambulating=
8 over over ADP IN _ 7 prep _ RP|00226550-r|SubjectiveAssessmentAttribute+
9 a a DET DT _ 10 det _ DT|?|?
10 log log NOUN NN _ 8 pobj _ NN|06812631-n|LogFn=

Are there automatic ways of knowing which ones are clausal negation versus restricted negation in this corpus?

[vcvpaiva]

query _ <neg NOUN only 7 cases.
query _ <neg !VERB 8 cases
query _<neg VERB = 413 cases?

[vcvpaiva]

part of the issue here is that negation should introduce a context, of the preposition that is negated.