KeyError: 'data'

[danielchang1985]

hi all I tried to execute the codes in documentation to create an NER component and I am getting this error.
Appreciate if there's someone who knows how to troubleshoot this.

import spacy

nlp = spacy.blank("en")
llm_ner = nlp.add_pipe("llm_ner")
llm_ner.add_label("PERSON")
llm_ner.add_label("LOCATION")
nlp.initialize()
doc = nlp("Jack and Jill rode up the hill in Les Deux Alpes")
print([(ent.text, ent.label_) for ent in doc.ents])

Error
KeyError Traceback (most recent call last)
Cell In[1], line 4
1 import spacy
3 nlp = spacy.blank("en")
----> 4 llm_ner = nlp.add_pipe("llm_ner")
5 llm_ner.add_label("PERSON")
6 llm_ner.add_label("LOCATION")

File c:\Users\danie\anaconda3\envs\lymphoma\Lib\site-packages\spacy\language.py:821, in Language.add_pipe(self, factory_name, name, before, after, first, last, source, config, raw_config, validate)
817 pipe_component, factory_name = self.create_pipe_from_source(
818 factory_name, source, name=name
819 )
820 else:
--> 821 pipe_component = self.create_pipe(
822 factory_name,
823 name=name,
824 config=config,
825 raw_config=raw_config,
826 validate=validate,
827 )
828 pipe_index = self._get_pipe_index(before, after, first, last)
829 self._pipe_meta[name] = self.get_factory_meta(factory_name)

File c:\Users\danie\anaconda3\envs\lymphoma\Lib\site-packages\spacy\language.py:709, in Language.create_pipe(self, factory_name, name, config, raw_config, validate)
...
---> 70 response = r.json()["data"]
71 models = [response[i]["id"] for i in range(len(response))]
72 if self._name not in models:

KeyError: 'data'

Packages
Spacy 3.7.2
Spacy-llm 0.6.4

[rmitsch]

Hi @danielchang1985, I can't reproduce this. Can you access the OpenAI API by any other means, e. g. using the openai library?

[sarafrr]

Hi @danielchang1985, I had the same issue, I've solved initialising the environment variable OPENAI_API_KEY by setting your API key that you can create here.

import os
os.environ["OPENAI_API_KEY"] = <yr_api_key>

You will see that the add of the labels will return 1, True, using the key.

import spacy
nlp = spacy.blank("en")
llm_ner = nlp.add_pipe("llm_ner")
print(llm_ner.add_label("PERSON"))
print(llm_ner.add_label("LOCATION"))
nlp.initialize()
doc = nlp("Jack and Jill rode up the hill in Les Deux Alpes")
print([(ent.text, ent.label_) for ent in doc.ents])

You can see by setting and unsetting (put an empty string) the environment variable, the two different behaviours.