Failed to run the Question Answering example about pyvespa HOT 14 CLOSED

Thanks for reporting! I will see if I can reproduce and update here.

from pyvespa.

I added

    @property
    def json(self):
        return self.json

to VespaQueryResponse in my local pyvespa install to be able to dump the full response object, I think we will add this to future releases

from pyvespa.

Awesome!

Yes, totally agree on the error check - will look into - and also add to troubleshooting (this is a common problem), keeping this issue open until done

from pyvespa.

We hope to release it this sprint, meaning in a week or two

from pyvespa.

Forgot to mention i'm using latest version from pypi (pyvespa==0.31.0). Didn't try using master branch directly

from pyvespa.

I was running this notebook earlier this week with no issues, so I suspect there is something with your environment. Often too little Docker memory is the culprit, you can review https://pyvespa.readthedocs.io/en/latest/troubleshooting.html. You will see in the Docker output if a process crashes.

I will try again with a 4G setting (I ran it using 12G) but this notebook takes some time to run so probably update here tomorrow

from pyvespa.

Thanks for the quick help. What was strange is that I didn't got any error message on the notebook and no error on the docker logs also. I'll try with manually setting more ram to the docker container. Will update tomorrow.

from pyvespa.

In step 18 I just got:

{'id': 'id:sentence:sentence::2',
 'relevance': 0.5540203635649571,
 'source': 'qa_content',
 'fields': {'sddocname': 'sentence',
  'documentid': 'id:sentence:sentence::2',
  'questions': [0],
  'dataset': 'squad',
  'context_id': 0,
  'text': 'It is a replica of the grotto at Lourdes, France where the Virgin Mary reputedly appeared to Saint Bernadette Soubirous in 1858.'}}

from pyvespa.

I run pyvespa from head (checked out local version), and 3.10, too

from pyvespa.

#465

from pyvespa.

Actually I just found the error. When feeding the data with app.feed_data_point(), I got a 'NO SPACE' error. Related to this: https://docs.vespa.ai/en/operations/feed-block.html. Will try to make space for docker and it should fix it.
Just as a note, maybe the notebook could be updated with a check for the returned code of feed_data_point() and throw an error with link to this feed-block page if NO SPACE error is found.

from pyvespa.

#466 and #467

from pyvespa.

pyvespa 0.32.0 will throw an exception for failed reads/writes, so easier to see the feed block. troubleshooting is updated

from pyvespa.

Hey, when are you planning on releasing a new version of pyvespa? (0.32). Cause currently, on the 0.31, package.Field doesn't have the argument 'match'.
Thanks

from pyvespa.