How to extract sentences from a doc given sentence does not contain token with a specific dep_ ?

[fatihbozdag]

Greetings all,

I am struggling to use != / not in syntax over Spacy. I would like get sentences from a doc if the dep_ of any token included in the sentence is not 'dobj'.

For instance:

import spacy
nlp = spacy.load('en_core_web_trf')
doc = nlp('The worker painted the wall white. He pushed the book off the shelf. The river froze solid. A leaf fell off the tree.')
for token in doc:
    print(token.text, token.dep_)
The det
worker nsubj
painted ROOT
the det
**wall dobj**
white oprd
. punct
He nsubj
pushed ROOT
the det
**book dobj**
off prep
the det
shelf pobj
. punct
The det
river nsubj
froze ROOT
solid advcl
. punct
A det
leaf nsubj
fell ROOT
off prep
the det
tree pobj
. punct

sentences 1 and 2 include 'dobj', However, I would like extract sentences only if the sentence does not contain any token whose dep_ is 'dobj'. The following code return all sentences including the ones with token.dep_ == 'dobj'.

for token in doc:
    if token.dep_ != 'dobj':
        print(token.sent)

What is the correct way to negate token.dep_ here?

[fatihbozdag]

Well after many attempts, I found the solution as follows;

for a in doc:
    if a.dep_ == "prep" and a.pos_ == "ADP" and a.head.pos_ == "VERB":
        for b in a.head.children:
            if b.dep_ == "nsubj":
                sents = [t.sent for t in a.sent]
                for n in sents:
                    for c in n:
                        if c.dep_ == 'dobj':
                            pattern2_sents = [c.sent]
                        if c.dep_ != 'pobj':
                            pattern4_sents = [c.sent]

It may be of the use for someone else. However I am not sure why simple iterating if token.dep_ != 'dobj' would not work in the original question.