Example of using OpenIE from Stanza CoreNLP client

[jeromemassot]

Dear Stanza team,

I try to use OpenIE from Python code and as much as possible I want to stay with Stanford pieces of code.
I see that we can use the CoreNLP java functionality after installing it from Stanza. This is great, especially with automatic install available in v.1.1.

However, despite you introduced several examples of usage in the documentation, it seems to me that the OpenIE pipeline example is not available. I have read the CoreNLP document related to OpenIE, using Java, but I cannot find anything related to a direct usage of OpenIE from Stanza lib.

Is it possible to include such example in the documentation?

Thanks a lot

Best regards

Jerome

[AngledLuffa]

Sorry for dropping this on the floor. Actually, it's pretty straightforward, so there's not really a good excuse for making you wait.

from stanza.server import CoreNLPClient

text = "I miss Mox Opal"

#with CoreNLPClient(annotators=["tokenize","ssplit","pos","lemma","depparse","natlog","openie"], be_quiet=False) as client:
with CoreNLPClient(annotators=["openie"], be_quiet=False) as client:
    ann = client.annotate(text)
    #print(ann)
    for sentence in ann.sentence:
        for triple in sentence.openieTriple:
            print(triple)

The exact layout of openieTriple can be found by reading doc/CoreNLP.proto

In general corenlp will add the needed annotators for "openie" so you don't need to worry about providing the whole list.

Which documentation do you recommend us upgrading? Something here?

https://stanfordnlp.github.io/stanza/corenlp_client.html

[jeromemassot]

Sorry for the late feedback John. And thanks a lot for your reply. I will test it ASAP. Best regards Jerome Le ven. 28 août 2020 à 14:27, John Bauer <notifications@github.com> a écrit :

…

Sorry for dropping this on the floor. Actually, it's pretty straightforward, so there's not really a good excuse for making you wait. from stanza.server import CoreNLPClient text = "I miss Mox Opal" #with CoreNLPClient(annotators=["tokenize","ssplit","pos","lemma","depparse","natlog","openie"], be_quiet=False) as client: with CoreNLPClient(annotators=["openie"], be_quiet=False) as client: ann = client.annotate(text) #print(ann) for sentence in ann.sentence: for triple in sentence.openieTriple: print(triple) The exact layout of openieTriple can be found by reading doc/CoreNLP.proto <https://github.com/stanfordnlp/stanza/blob/master/doc/CoreNLP.proto> In general corenlp will add the needed annotators for "openie" so you don't need to worry about providing the whole list. Which documentation do you recommend us upgrading? Something here? https://stanfordnlp.github.io/stanza/corenlp_client.html — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#441 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AE2Q4ZVDQWVM7TJ2NJFP47LSDAOLRANCNFSM4QERW7BA> .

[Johnnee]

Sorry for dropping this on the floor. Actually, it's pretty straightforward, so there's not really a good excuse for making you wait.

from stanza.server import CoreNLPClient

text = "I miss Mox Opal"

#with CoreNLPClient(annotators=["tokenize","ssplit","pos","lemma","depparse","natlog","openie"], be_quiet=False) as client:
with CoreNLPClient(annotators=["openie"], be_quiet=False) as client:
    ann = client.annotate(text)
    #print(ann)
    for sentence in ann.sentence:
        for triple in sentence.openieTriple:
            print(triple)

The exact layout of openieTriple can be found by reading doc/CoreNLP.proto

In general corenlp will add the needed annotators for "openie" so you don't need to worry about providing the whole list.

Which documentation do you recommend us upgrading? Something here?

https://stanfordnlp.github.io/stanza/corenlp_client.html

Hi, does this methods support Chinese information extraction ? If it do, pls tell me. thanks.

[AngledLuffa]

I do not believe there is an openie model for Chinese: https://stanfordnlp.github.io/CoreNLP/human-languages.html

[Johnnee]

Fine, thank you.