Getting KeyError when I convert json training samples to .md format

Rasa version: rasa 1.4.1

Python version: Python 3.6.7

Operating system (Ubuntu):

Issue: I want to convert my json training data into .md format. My data doesn’t contain any intent, because I only want to extract entities.

Error (including full traceback):

---------------------------------------------------------------------------
KeyError                                  Traceback (most recent call last)
<ipython-input-7-fe8a5bb0d89a> in <module>
     19 
     20 with open(output_md_file,'w') as f:
---> 21     f.write(load_data(input_training_file).as_markdown())

~/Documents/job_description_ner_api/env/lib/python3.6/site-packages/rasa/nlu/training_data/training_data.py in as_markdown(self)
    246         )
    247 
--> 248         return self.nlu_as_markdown()
    249 
    250     def persist_nlu(self, filename: Text = DEFAULT_TRAINING_DATA_OUTPUT_PATH):

~/Documents/job_description_ner_api/env/lib/python3.6/site-packages/rasa/nlu/training_data/training_data.py in nlu_as_markdown(self)
    237         )
    238 
--> 239         return MarkdownWriter().dumps(self)
    240 
    241     def as_markdown(self) -> Text:

~/Documents/job_description_ner_api/env/lib/python3.6/site-packages/rasa/nlu/training_data/formats/markdown.py in dumps(self, training_data)
    201         """Transforms a TrainingData object into a markdown string."""
    202         md = ""
--> 203         md += self._generate_training_examples_md(training_data)
    204         md += self._generate_synonyms_md(training_data)
    205         md += self._generate_regex_features_md(training_data)

~/Documents/job_description_ner_api/env/lib/python3.6/site-packages/rasa/nlu/training_data/formats/markdown.py in _generate_training_examples_md(self, training_data)
    212         training_examples = sorted(
    213             [e.as_dict_nlu() for e in training_data.training_examples],
--> 214             key=lambda k: k[MESSAGE_INTENT_ATTRIBUTE],
    215         )
    216         md = ""

~/Documents/job_description_ner_api/env/lib/python3.6/site-packages/rasa/nlu/training_data/formats/markdown.py in <lambda>(k)
    212         training_examples = sorted(
    213             [e.as_dict_nlu() for e in training_data.training_examples],
--> 214             key=lambda k: k[MESSAGE_INTENT_ATTRIBUTE],
    215         )
    216         md = ""

KeyError: 'intent'

Command or request that led to error:

from rasa.nlu.training_data import load_data
input_training_file = 'data/train_data/training_data.json'
output_md_file = 'data/train_data/training_data.md'
with open(output_md_file,'w') as f:
    f.write(load_data(input_training_file).as_markdown())

Content of configuration file (config.yml) (if relevant):

language: "en"

pipeline:
- name: "SpacyNLP"
- name: "SpacyFeaturizer"
- name: "SpacyTokenizer"
- name: "CRFEntityExtractor"
- name: "EntitySynonymMapper"
- name: "RegexFeaturizer"
  BILOU_flag: true
  features: [
     # Features to be extracted from token before
     ['low', 'title', 'upper', 'pos', 'pos2'],
     # Features to be extracted from current token
     ['bias', 'low', 'word3', 'word2', 'upper','title', 'digit','pos', 'pos2','pattern'],
     # Features to be extracted from token after
     ['low', 'title', 'upper', 'pos', 'pos2']]
  max_iterations: 50
  L1_c: 1
  L2_c: 0.001

[erohmensing]

Sure, it’s just the json equivalent. Wouldn’t say it’s a hack, as it’s rather the official rasa format to have intent as part of your example 😃

[stale[bot]]

This issue has been automatically closed due to inactivity. Please create a new issue if you need more help.