Error : Loading domain from 'domain.yml' failed. Using empty domain

Rasa Open Source
1

Rasa version:2.6.2

Python version:3.7.4

Operating system (windows, osx, …): Ubuntu

Issue: Hey! I am trying to validate my rasa project using rasa data validate command , I get an error stating UserWarning: Loading domain from ‘domain.yml’ failed. Using empty domain. Error: 'If you use the deprecated list syntax for forms, all form names have to be strings. Error (including full traceback):

/home/agsc_rdi_a_arcotlo/venv/lib/python3.7/site-packages/rasa/shared/utils/io.py:97: UserWarning: Loading domain from 'domain.yml' failed. Using empty domain. Error: 'If you use the deprecated list syntax for forms, all form names have to be strings. Please see https://rasa.com/docs/rasa/forms for more information.'
/home/agsc_rdi_a_arcotlo/venv/lib/python3.7/site-packages/rasa/shared/utils/io.py:97: UserWarning: Issue found in 'data/rules.yml':
Found intent 'greetings' in stories which is not part of the domain.
  More info at https://rasa.com/docs/rasa/stories
/home/agsc_rdi_a_arcotlo/venv/lib/python3.7/site-packages/rasa/shared/utils/io.py:97: UserWarning: Issue found in 'data/rules.yml':
Found intent 'how_are_you' in stories which is not part of the domain.
  More info at https://rasa.com/docs/rasa/stories
/home/agsc_rdi_a_arcotlo/venv/lib/python3.7/site-packages/rasa/shared/utils/io.py:97: UserWarning: Issue found in 'data/rules.yml':
Found intent 'feedback' in stories which is not part of the domain.
  More info at https://rasa.com/docs/rasa/stories
/home/agsc_rdi_a_arcotlo/venv/lib/python3.7/site-packages/rasa/shared/utils/io.py:97: UserWarning: Issue found in 'data/rules.yml':
Found intent 'attached_file' in stories which is not part of the domain.
  More info at https://rasa.com/docs/rasa/stories
/home/agsc_rdi_a_arcotlo/venv/lib/python3.7/site-packages/rasa/shared/utils/io.py:97: UserWarning: Issue found in 'data/rules.yml':
Found intent 'dialog_access_time' in stories which is not part of the domain.
  More info at https://rasa.com/docs/rasa/stories
/home/agsc_rdi_a_arcotlo/venv/lib/python3.7/site-packages/rasa/shared/utils/io.py:97: UserWarning: Issue found in 'data/rules.yml':
Found intent 'help' in stories which is not part of the domain.
  More info at https://rasa.com/docs/rasa/stories
/home/agsc_rdi_a_arcotlo/venv/lib/python3.7/site-packages/rasa/shared/utils/io.py:97: UserWarning: Issue found in 'data/rules.yml':
Found intent 'praise' in stories which is not part of the domain.
  More info at https://rasa.com/docs/rasa/stories
/home/agsc_rdi_a_arcotlo/venv/lib/python3.7/site-packages/rasa/shared/utils/io.py:97: UserWarning: Issue found in 'data/rules.yml':
Found intent 'creator' in stories which is not part of the domain.
  More info at https://rasa.com/docs/rasa/stories
/home/agsc_rdi_a_arcotlo/venv/lib/python3.7/site-packages/rasa/shared/utils/io.py:97: UserWarning: Issue found in 'data/rules.yml':
Found intent 'later' in stories which is not part of the domain.
  More info at https://rasa.com/docs/rasa/stories
/home/agsc_rdi_a_arcotlo/venv/lib/python3.7/site-packages/rasa/shared/utils/io.py:97: UserWarning: Issue found in 'data/rules.yml':
Found intent 'thanks_goodbye' in stories which is not part of the domain.
  More info at https://rasa.com/docs/rasa/stories
/home/agsc_rdi_a_arcotlo/venv/lib/python3.7/site-packages/rasa/shared/utils/io.py:97: UserWarning: Issue found in 'data/rules.yml':
Found intent 'affirm' in stories which is not part of the domain.
  More info at https://rasa.com/docs/rasa/stories
/home/agsc_rdi_a_arcotlo/venv/lib/python3.7/site-packages/rasa/shared/utils/io.py:97: UserWarning: Issue found in 'data/rules.yml':
Found intent 'deny' in stories which is not part of the domain.
  More info at https://rasa.com/docs/rasa/stories
/home/agsc_rdi_a_arcotlo/venv/lib/python3.7/site-packages/rasa/shared/utils/io.py:97: UserWarning: Issue found in 'data/rules.yml':
Found intent 'training_details' in stories which is not part of the domain.
  More info at https://rasa.com/docs/rasa/stories
/home/agsc_rdi_a_arcotlo/venv/lib/python3.7/site-packages/rasa/shared/utils/io.py:97: UserWarning: Issue found in 'data/rules.yml':
Project validation completed with errors.
3

All the intents are available in domain.yml file but still rasa is unable to read the intents from domain.yml file.

Kindly suggest how to resolve this issue. Thanks in advance

4

this is most likely the reason your domain can’t be read, can you post the forms section of your domain?

6

forms:

  • host_reg_form: comp_name: - type: text entity: comp_name project_domain: - type: categorical entity: project_domain system_type: - type: categorical entity: system_type
7

Any updates ?

8

@Richa Did this issue solved by you? If not what the current status of this error?

9

Actually this error is getting generated because of model is not getting trained in one go, like when i am using “rasa train” then nlu model is getting train but failed to trained core model. But when i did “rasa train core” so it is training core model but failed to fetch nlu model at the same time.

I am getting below error when i am trying “rasa data validate”

/home/agsc_rdi_a_arcotlo/venv/lib/python3.7/site-packages/rasa/shared/utils/io.py:97: UserWarning: You have provided an invalid key model_url for component ConveRTTokenizer in your pipeline. Valid options for ConveRTTokenizer are:

  • intent_tokenization_flag
  • intent_split_symbol
  • token_pattern RasaException: Parameter “model_url” was not specified in the configuration of “ConveRTFeaturizer”. It is mandatory to pass a value for this parameter. You can either use a community hosted URL of the model or if you have a local copy of the model, pass the path to the directory containing the model files.
10

@Richa Can you share your original config file or your project directory structure and rasa version including Rasa SDK or rasa core sdk.

11

Rasa version : 2.7.0

Config.yml

language: en

pipeline:

policies:

  • name: MemoizationPolicy
  • name: TEDPolicy max_history: 5 epochs: 100
  • name: RulePolicy core_fallback_threshold: 0.3 fallback_action_name: “action_default_fallback”
12

@Richa you have two models? DIETClassifier epochs: 600 (how much time you taking to train the model?) Are you using ResponseSelector?

13

Yes… It is taking about 6-7 hrs

14

@Richa Insane (6-7hrs) did you coded yourself whole project? or you only taking a pre-trained models?

15

I guess the conversation is going off the track. Can we please stick to the actual issue i.e. UserWarning: You have provided an invalid key model_url for component ConveRTTokenizer in your pipeline. Valid options for ConveRTTokenizer are:

  • intent_tokenization_flag
  • token_pattern
  • intent_split_symbol
17

I am using a local model, but still it is getting invalid url error. I tried checking the reachability through curl command from linux and direct access from browser, both worked fine.

But still the error is not getting resolved.

18

@Richa This of your url error is because the original ConveRT was taken offline by the original authors, the tokenizers and featurizers cannot download the model from the original URL of the model.Which you given. When you run the pipeline URL (Uniform Resource Locator) search for the path which you have mention and he was unable to fetch the information that’s why url error you facing. If you have downloaded the model’s pass without the url address. I hope my point is clear to you? Thanks

19

@Richa Tips: Try pass your local directory path of the model’s (only both model’s Size: 146.1 MB) or relative path as value to model_url, may be it will work for you. Hope now you got my point? or may be I can be wrong as I not run this code. Good Luck!

20

@Richa Hey! Are you able to solve this? Yes or No. As I am getting similar query. If got time reply.

21

Yes The issue is resolved. Kindly close the issue.

22

@Richa Please mark the solution from your side or which post you think solved your issue with the suggestion or solution. Thanks.

23

There no solution from the above comments, there was a redundancy in the stories which was a part of story n rule that is why the error happened.