Training from wit.ai data causes exception in extractor.py

[sciguy14]

Rasa version: 2.2.2

Rasa SDK version (if used & relevant): 2.2.0

Rasa X version (if used & relevant): None

Python version: 3.8.5

Operating system (windows, osx, ...): WSL (Linux-4.4.0-19041-Microsoft-x86_64-with-glibc2.29)

Issue:
Following the instructions to convert my wit.ai data to Rasa training data results in a Python exception. I followed these instructions: https://rasa.com/docs/rasa/migrate-from/facebook-wit-ai-to-rasa/

My utterances-1.json file is present in the data/ folder as noted in the migration instructions.

Error (including full traceback):

(venv) jeremy@PHOENIX:/mnt/c/wit.ai to Rasa$ rasa train nlu
2021-01-03 21:14:05.237561: W tensorflow/stream_executor/platform/default/dso_loader.cc:59] Could not load dynamic library 'libcudart.so.10.1'; dlerror: libcudart.so.10.1: cannot open shared object file: No such file or directory
2021-01-03 21:14:05.238107: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.
2021-01-03 21:14:07.937718: W tensorflow/stream_executor/platform/default/dso_loader.cc:59] Could not load dynamic library 'libcuda.so.1'; dlerror: libcuda.so.1: cannot open shared object file: No such file or directory
2021-01-03 21:14:07.938268: W tensorflow/stream_executor/cuda/cuda_driver.cc:312] failed call to cuInit: UNKNOWN ERROR (303)
2021-01-03 21:14:07.938525: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:156] kernel driver does not appear to be running on this host (PHOENIX): /proc/driver/nvidia/version does not exist
The configuration for pipeline was chosen automatically. It was written into the config file at 'config.yml'.
Training NLU model...
2021-01-03 21:14:08 INFO     rasa.shared.nlu.training_data.training_data  - Training data stats:
2021-01-03 21:14:08 INFO     rasa.shared.nlu.training_data.training_data  - Number of intent examples: 1066 (20 distinct intents)

2021-01-03 21:14:08 INFO     rasa.shared.nlu.training_data.training_data  -   Found intents: 'introductions', 'date', 'snooze', 'play_news', 'prev_track', 'weather', 'play_music', 'out_of_scope', 'choose_music', 'curtains', 'stop_audio', 'goodnight', 'greetings', 'pause_audio', 'alarm', 'volume', 'lighting', 'time', 'next_track', 'wikipedia'
2021-01-03 21:14:08 INFO     rasa.shared.nlu.training_data.training_data  - Number of response examples: 0 (0 distinct responses)
2021-01-03 21:14:08 INFO     rasa.shared.nlu.training_data.training_data  - Number of entity examples: 706 (10 distinct entities)
2021-01-03 21:14:08 INFO     rasa.shared.nlu.training_data.training_data  -   Found entity types: 'relative_adjustment', 'wit$location', 'location_1', 'wit$wikipedia_search_query', 'wit$datetime', 'wit$number', 'artist', 'light_type_entity', 'mood', 'song'
2021-01-03 21:14:08 INFO     rasa.shared.nlu.training_data.training_data  -   Found entity roles: 'wikipedia_search_query', 'relative_adjustment', 'number', 'location', 'location_1', 'datetime', 'artist', 'light_type_entity', 'mood', 'song'
2021-01-03 21:14:08 WARNING  rasa.shared.utils.common  - The Entity Roles and Groups feature is currently experimental and might change or be removed in the future 🔬 Please share your feedback on it in the forum (https://forum.rasa.com) to help us make this feature ready for production.
/mnt/c/wit.ai to Rasa/venv/lib/python3.8/site-packages/rasa/shared/utils/io.py:93: UserWarning: Intent 'out_of_scope' has only 1 training examples! Minimum is 2, training may fail.
2021-01-03 21:14:08 INFO     rasa.nlu.model  - Starting to train component WhitespaceTokenizer
2021-01-03 21:14:08 INFO     rasa.nlu.model  - Finished training component.
2021-01-03 21:14:08 INFO     rasa.nlu.model  - Starting to train component RegexFeaturizer
2021-01-03 21:14:08 INFO     rasa.nlu.model  - Finished training component.
2021-01-03 21:14:08 INFO     rasa.nlu.model  - Starting to train component LexicalSyntacticFeaturizer
2021-01-03 21:14:09 INFO     rasa.nlu.model  - Finished training component.
2021-01-03 21:14:09 INFO     rasa.nlu.model  - Starting to train component CountVectorsFeaturizer
2021-01-03 21:14:09 INFO     rasa.nlu.featurizers.sparse_featurizer.count_vectors_featurizer  - 578 vocabulary slots consumed out of 1578 slots configured for text attribute.
2021-01-03 21:14:09 INFO     rasa.nlu.model  - Finished training component.
2021-01-03 21:14:09 INFO     rasa.nlu.model  - Starting to train component CountVectorsFeaturizer
2021-01-03 21:14:09 INFO     rasa.nlu.featurizers.sparse_featurizer.count_vectors_featurizer  - 3866 vocabulary slots consumed out of 5799 slots configured for text attribute.
2021-01-03 21:14:09 INFO     rasa.nlu.model  - Finished training component.
2021-01-03 21:14:09 INFO     rasa.nlu.model  - Starting to train component DIETClassifier
Traceback (most recent call last):
  File "/mnt/c/wit.ai to Rasa/venv/bin/rasa", line 10, in <module>
    sys.exit(main())
  File "/mnt/c/wit.ai to Rasa/venv/lib/python3.8/site-packages/rasa/__main__.py", line 116, in main
    cmdline_arguments.func(cmdline_arguments)
  File "/mnt/c/wit.ai to Rasa/venv/lib/python3.8/site-packages/rasa/cli/train.py", line 195, in train_nlu
    return train_nlu(
  File "/mnt/c/wit.ai to Rasa/venv/lib/python3.8/site-packages/rasa/train.py", line 700, in train_nlu
    return rasa.utils.common.run_in_loop(
  File "/mnt/c/wit.ai to Rasa/venv/lib/python3.8/site-packages/rasa/utils/common.py", line 308, in run_in_loop
    result = loop.run_until_complete(f)
  File "uvloop/loop.pyx", line 1456, in uvloop.loop.Loop.run_until_complete
  File "/mnt/c/wit.ai to Rasa/venv/lib/python3.8/site-packages/rasa/train.py", line 749, in _train_nlu_async
    return await _train_nlu_with_validated_data(
  File "/mnt/c/wit.ai to Rasa/venv/lib/python3.8/site-packages/rasa/train.py", line 811, in _train_nlu_with_validated_data
    await rasa.nlu.train(
  File "/mnt/c/wit.ai to Rasa/venv/lib/python3.8/site-packages/rasa/nlu/train.py", line 116, in train
    interpreter = trainer.train(training_data, **kwargs)
  File "/mnt/c/wit.ai to Rasa/venv/lib/python3.8/site-packages/rasa/nlu/model.py", line 209, in train
    updates = component.train(working_data, self.config, **context)
  File "/mnt/c/wit.ai to Rasa/venv/lib/python3.8/site-packages/rasa/nlu/classifiers/diet_classifier.py", line 803, in train
    self.check_correct_entity_annotations(training_data)
  File "/mnt/c/wit.ai to Rasa/venv/lib/python3.8/site-packages/rasa/nlu/extractors/extractor.py", line 418, in check_correct_entity_annotations
    entities_repr = [
  File "/mnt/c/wit.ai to Rasa/venv/lib/python3.8/site-packages/rasa/nlu/extractors/extractor.py", line 422, in <listcomp>
    entity[ENTITY_ATTRIBUTE_VALUE],
KeyError: 'value'

Command or request that led to error:

rasa train nlu

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

Content of domain file (domain.yml) (if relevant): N/A

[sara-tagger]

Thanks for raising this issue, @degiz will get back to you about it soon✨

Please also check out the docs and the forum in case your issue was raised there too 🤗

[degiz]

cc @RasaHQ/enable-squad

[wochinge]

@degiz Can you please add it our inbox in the future?

@sciguy14 Seems there is some issue with converting the entities correctly. Do you want to have a go solving it (I'd be happy to support you with the PR) or should we handle it?

[stale]

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.

[roots-ai]

@degiz Can you please add it our inbox in the future?

@sciguy14 Seems there is some issue with converting the entities correctly. Do you want to have a go solving it (I'd be happy to support you with the PR) or should we handle it?

Facing the same issue. Please help. @sara-tagger @wochinge

[wochinge]

@roots-ai Would you mind creating a PR with the necessary fixes? Unfortunately the team is currently very busy with 3.0 so I don't think we can tackle this one soon ourselves

[roots-ai]

@roots-ai Would you mind creating a PR with the necessary fixes? Unfortunately the team is currently very busy with 3.0 so I don't think we can tackle this one soon ourselves
@wochinge
Can you please guide me; any ideas or suggestions on what are the possible reasons and how can I go about finding a fix?

[roots-ai]

@wochinge ?

[wochinge]

It seems like we are missing to set the ENTITY_ATTRIBUTE_VALUE here. We'd probably just need to add the line e[ENTITY_ATTRIBUTE_ROLE] = e.pop("body")

[roots-ai]

@roots-ai Would you mind creating a PR with the necessary fixes? Unfortunately the team is currently very busy with 3.0 so I don't think we can tackle this one soon ourselves

Any expected timeline for when it be possible for the team to fix this?

I installed RASA from the source and made the above changes and now it gives this new error

/root/miniconda3/lib/python3.8/site-packages/rasa/shared/utils/io.py:97: UserWarning: Entity role 'vent_adjective' has only 1 training examples! The minimum is 2, because of this the training may fail. /root/miniconda3/lib/python3.8/site-packages/rasa/shared/utils/io.py:97: UserWarning: Entity entity 'which_person' has only 1 training examples! The minimum is 2, because of this the training may fail. /root/miniconda3/lib/python3.8/site-packages/rasa/shared/utils/io.py:97: UserWarning: Entity role 'which_person' has only 1 training examples! The minimum is 2, because of this the training may fail. 2021-10-07 01:53:03 INFO rasa.nlu.model - Starting to train component WhitespaceTokenizer 2021-10-07 01:53:03 INFO rasa.nlu.model - Finished training component. 2021-10-07 01:53:03 INFO rasa.nlu.model - Starting to train component RegexFeaturizer 2021-10-07 01:53:03 INFO rasa.nlu.model - Finished training component. 2021-10-07 01:53:03 INFO rasa.nlu.model - Starting to train component LexicalSyntacticFeaturizer 2021-10-07 01:53:03 INFO rasa.nlu.model - Finished training component. 2021-10-07 01:53:03 INFO rasa.nlu.model - Starting to train component CountVectorsFeaturizer 2021-10-07 01:53:03 INFO rasa.nlu.featurizers.sparse_featurizer.count_vectors_featurizer - 2014 vocabulary items were created for text attribute. 2021-10-07 01:53:04 INFO rasa.nlu.model - Finished training component. 2021-10-07 01:53:04 INFO rasa.nlu.model - Starting to train component CountVectorsFeaturizer 2021-10-07 01:53:04 INFO rasa.nlu.featurizers.sparse_featurizer.count_vectors_featurizer - 8363 vocabulary items were created for text attribute. 2021-10-07 01:53:05 INFO rasa.nlu.model - Finished training component. 2021-10-07 01:53:05 INFO rasa.nlu.model - Starting to train component DIETClassifier /root/miniconda3/lib/python3.8/site-packages/rasa/utils/tensorflow/model_data_utils.py:395: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray np.array([v[0] for v in values]), number_of_dimensions=3 Traceback (most recent call last): File "/root/miniconda3/bin/rasa", line 8, in <module> sys.exit(main()) File "/root/miniconda3/lib/python3.8/site-packages/rasa/__main__.py", line 117, in main cmdline_arguments.func(cmdline_arguments) File "/root/miniconda3/lib/python3.8/site-packages/rasa/cli/train.py", line 196, in run_nlu_training return train_nlu( File "/root/miniconda3/lib/python3.8/site-packages/rasa/model_training.py", line 646, in train_nlu return rasa.utils.common.run_in_loop( File "/root/miniconda3/lib/python3.8/site-packages/rasa/utils/common.py", line 296, in run_in_loop result = loop.run_until_complete(f) File "uvloop/loop.pyx", line 1456, in uvloop.loop.Loop.run_until_complete File "/root/miniconda3/lib/python3.8/site-packages/rasa/model_training.py", line 696, in train_nlu_async return await _train_nlu_with_validated_data( File "/root/miniconda3/lib/python3.8/site-packages/rasa/model_training.py", line 758, in _train_nlu_with_validated_data await rasa.nlu.train.train( File "/root/miniconda3/lib/python3.8/site-packages/rasa/nlu/train.py", line 111, in train interpreter = trainer.train(training_data, **kwargs) File "/root/miniconda3/lib/python3.8/site-packages/rasa/nlu/model.py", line 221, in train component.train(working_data, self.config, **context) File "/root/miniconda3/lib/python3.8/site-packages/rasa/nlu/classifiers/diet_classifier.py", line 846, in train self.check_correct_entity_annotations(training_data) File "/root/miniconda3/lib/python3.8/site-packages/rasa/nlu/extractors/extractor.py", line 459, in check_correct_entity_annotations entities_repr = [ File "/root/miniconda3/lib/python3.8/site-packages/rasa/nlu/extractors/extractor.py", line 463, in <listcomp> entity[ENTITY_ATTRIBUTE_VALUE], KeyError: 'value'

[wochinge]

@roots-ai Sorry, as I said in my previous comment it's currently very busy and I can't provide a reliable timeline

[roots-ai]

@wochinge Would be good to put this as a disclaimer on the migration from Wit.ai page so that others don't end up wasting their time with RASA.

[wochinge]

Since you've asked so nicely I've created a PR with the necessary changes: #9864

[wochinge]

Cutting a release today/tomorrow for this