BeWe11/rasa_composite_entities

Failed to find component class for 'rasa_composite_entities.CompositeEntityExtractor' on $ rasa train

Closed this issue · 4 comments

Hello, i did pip install rasa_composite_entitites, and next tried with the latest rasa version to train with this config :

language: "fr"

pipeline:
    - name: "WhitespaceTokenizer"
    - name: "RegexFeaturizer"
    - name: "CRFEntityExtractor"
    - name: "EntitySynonymMapper"
    - name: "CountVectorsFeaturizer"
    - name: "CountVectorsFeaturizer"
      analyzer: "char_wb"
      min_ngram: 1
      max_ngram: 4
    - name: "EmbeddingIntentClassifier"
      batch_strategy: "sequence"
    - BILOU_flag: true
      name: "CRFEntityExtractor"
      features:
          - [low, title, upper]
          - [low, bias, prefix5, prefix2, suffix5, suffix3, suffix2, upper, title, digit, pattern]
          - [low, title, upper]

Result :
image

Do you have any idea ?

To resume :

  • pip install rasa
  • pip install rasa_composite_entities
  • add to config this : - name: "rasa_composite_entities.CompositeEntityExtractor")
  • (shell cmd) rasa train
  • error (see screenshot)

hey @MaximeGendre7,

does this problem persist? If yes, can you post your complete config file here? I can train the component successfully using the latest rasa version (1.5.0).

Closing because of inactivity and because I cannot reproduce this error. Feel free to post here again with new insights, I'll reopen the issue then.

I am facing the same issue with rasa==1.6.2.
Here is the config file.

language: en
pipeline:

  • name: WhitespaceTokenizer
  • name: RegexFeaturizer
  • name: CRFEntityExtractor
    max_iterations: 150
  • name: EntitySynonymMapper
  • name: CountVectorsFeaturizer
    lowercase: true
  • name: EmbeddingIntentClassifier
    epochs: 20
    random_seed: 1
    hidden_layers_sizes_a:
    • 256
    • 128
      batch_size:
    • 128
    • 256
      embed_dim: 20
      evaluate_every_num_epochs: 30
      intent_tokenization_flag: true
      intent_split_symbol: +
  • name: DucklingHTTPExtractor
    url: http://duckling:7000
    dimensions:
    • sys_time
    • sys_day
    • sys_month
    • sys_year
    • phone-number
    • number
    • amount-of-money
    • distance
    • duration
    • volume
    • ordinal
    • temperature
    • email
    • url
    • time
  • name: "rasa_composite_entities.CompositeEntityExtractor"
    composite_patterns_path: "composite_entity_patterns.json"
    policies:
  • name: KerasPolicy
    epochs: 300
    batch_size: 20
    max_training_samples: 300
  • name: TwoStageFallbackPolicy
    nlu_threshold: 0.78
    core_threshold: 0.3
    fallback_core_action_name: action_default_fallback
    fallback_nlu_action_name: action_default_fallback
    deny_suggestion_intent_name: out_of_scope
  • name: MemoizationPolicy
  • name: FormPolicy
  • name: MappingPolicy

FYI. Now my rasa is at 1.6.2 and my composite entity packages was at 1.0.3 when i tried to train it it threw an error "ModuleNotFoundError: Failed to find module 'rasa_composite_entities'". But with combination rasa is at 1.10.0 and my composite entity package==1.0.3 it was working fine.