Rasa: empty parsed text

[ctrado18]

Rasa NLU version : development

Used backend / pipeline (mitie, spacy_sklearn, ...): spacy_sklearn

Operating system (windows, osx, ...): windows

Issue:

With this example of current version:

http://nlu.rasa.ai/master/python.html

`from rasa_nlu.training_data import load_data
from rasa_nlu.config import RasaNLUConfig
from rasa_nlu.model import Trainer

training_data = load_data('data/examples/rasa/germ_test.json')
trainer = Trainer(RasaNLUConfig("sample_configs/config_spacy.json"))
trainer.train(training_data)
model_directory = trainer.persist('./projects/')  # Returns the directory the model is stored in

from rasa_nlu.model import Metadata, Interpreter

# where `model_directory points to the folder the model is persisted in
interpreter = Interpreter.load(model_directory, RasaNLUConfig("sample_configs/config_spacy.json"))

interpreter.parse(u"Wie viel Uhr ist es?")`

I receive this:

C:\Users\XXXX\Anaconda3\lib\site-packages\sklearn\preprocessing\label.py:151: DeprecationWarning: The truth value of an empty array is ambiguous. Returning False, but in future this will result in an error. Use array.size > 0 to check that an array is not empty.
if diff:

And the result of interpreter.parse seems to be empty.

What does this mean?

[ctrado18]

I want to update

I traced back this error until the update function for the output in the parse function of model.Interpreter.

`    def parse(self, text, time=None, only_output_properties=True):
        # type: (Text) -> Dict[Text, Any]
        """Parse the input text, classify it and return pipeline result.

        The pipeline result usually contains intent and entities."""

        if not text:
            # Not all components are able to handle empty strings. So we need
            # to prevent that... This default return will not contain all
            # output attributes of all components, but in the end, no one should
            # pass an empty string in the first place.
            output = self.default_output_attributes()
            output["text"] = ""
            return output

        message = Message(text, self.default_output_attributes(), time=time)

        for component in self.pipeline:
            component.process(message, **self.context) #here

        output = self.default_output_attributes()


        output.update(message.as_dict(only_output_properties=only_output_properties))
        #output is not printed as_dict works fine!! So it is the update function
        return output`

The return value of as_dict gives the proper dict back. So it works but the output.update seems to malfunction.

I have yet not looked at this update function.

[leewonmoh]

I have the same issue:
OS: CentOS 7.0
pipeline: spacy_sklearn
Will share investigation findings if any soon.

[ohld]

Same Issue, macOS, spacy_sklearn

[tmbo]

So the DeprecationWarning is normal (sklearn needs to update its internals to fix this, I looked into it). I don't think the "empty response" issue is related to this.

[tmbo]

Can you provide more details on how to reproduce this?

[ctrado18]

I used the standard example, like I wrote and showed in the first post,

Is it because of some experimental changes due to developer version? I will try an older version soon.

And as I wrote in the 2nd post, the return value of as_dict is fine. So it seem to be the update function...

[MnHung]

Same Issue, maybe it is about sklearn's version?

[tmbo]

I don't think it is due to the update function, as this is just pythons dict update.

@MnHung which spacy version + model are you using?

[tmbo]

As said, the deprecation warning itself is not an issue.

[MnHung]

Hi @tmbo
I am using spacy-2.0.8
and thank you, I am sure that rasa_nlu works well.

[ctrado18]

I used now version 11.3 and still same. I cannot see output of entities and intents etc. if use it wih python. Just the depreciated warning...

Why is the dict not returned properly?

[ayashjorden]

I think that this is a dependacy update issue see scikit-learn/scikit-learn/pull/9816

OS: macOS 10.13.3
pipeline: spacy_sklearn

[giantstonex7]

OS: Ubuntu 16.04
pipeline: spacy_sklearn
spacy.__version__'2.0.9'
sklearn.__version__'0.19.1'

the same error
Use array.size > 0 to check that an array is not empty.
if diff:

scikit-learn/scikit-learn#9816 solved the issue.

[ayashjorden]

But how do I get scikit-learn/scikit-learn#9816 ?
I've tried upgrading both rasa_core and rasa_nlu, but it didn't update _sklearn_

[pandasMX]

same error

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[sebastien34250]

same error

[akelad]

it's not really an error so much as a warning, if this is really something you have a problem with then please create a new issue :)