Custom NLU Components Failed to find component class

Rasa version: 0.14

Rasa X version (if used & relevant):

Python version: 3.6

Operating system: Ubuntu

Issue: I’m trying to make own Custom NLU Components, but rasa_nlu can’t find my component, probably I miss something

rasa_nlu : 
├── classifiers
│   ├── ....
├── emulators
│   ├── ...
├── extractors
│   ├── ...
├── featurizers
│   ├── ...
├── schemas
│   └── ...
├── tokenizers
│   ├── ...
├── training_data
│   ├── ...
├── utils
│   ├── ...
│
│ 
├── customcomponents
│   ├── __init__.py
│   └── sentiment.py
│
│
├── __init__.py
├── registry.py
├── config.json
│
│
├── components.py
├── config.py
├── convert.py
├── data_router.py
├── evaluate.py
├── model.py
├── persistor.py
├── project.py
├── run.py
├── server.py
├── train.py
└── version.py

My Custom component sentiment.py :

from rasa_nlu.components import Component
# all import needed

class MySentimentAnalyzer(Component):
    """A pre-trained sentiment component"""

    name = "MySentimentAnalyzer"
    provides = ["entities"]
    requires = ["tokens"]
    defaults = {}
    language_list = ["en"]

    def __init__(self, component_config=None):
        super(MySentimentAnalyzer, self).__init__(component_config)

  ## THE REST OF CODE ...

registry.py :

from rasa_nlu.customcomponents.sentiment import MySentimentAnalyzer
# all import needed

component_classes = [
#
# all rasa components added no change
#
MySentimentAnalyzer
]

## THE REST OF COD

Error (including full traceback):

{
  "error": "Failed to find component class for 'customcomponents.sentiment.MySentimentAnalyzer'. Unknown component name. Check your configured pipeline and make sure the mentioned component is not misspelled. If you are creating your own component, make sure it is either listed as part of the `component_classes` in `rasa_nlu.registry.py` or is a proper name of a class in a module."
}

Command or request that led to error: (Python)

requests.post(
                url='http://0.0.0.0:5000/train',
                params = (('project', PROJECT_NAME), ('model', MODEL_NAME)),
                headers = {'content-type': 'application/x-yml'},
                data = data
            )

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

{
   "language":"pl",
   "pipeline":[
      {
         "name":"tokenizer_whitespace"
      },
      {
         "name":"customcomponents.sentiment.MySentimentAnalyzer"
      },
      {
         "name":"ner_duckling_http",
         "locale":"pl_Pl",
         "url":"http://127.0.0.1:8000",
         "dimensions":[
            "time",
            "number",
            "money"
         ]
      }
   ]
}

NEED HELP :

in config file try different combinations, still rasa can’t find my component 😕 any ide wht’s wrong ??

Content of configuration file (config.json) which I try :

• name:“customcomponents.sentiment.MySentimentAnalyzer”
• name:“sentiment.MySentimentAnalyzer”
• name:“.MySentimentAnalyzer”

All of them don’t work for me 😕

[elyase]

The error message can be misleading. You will see the same error if your component fails to import for any reason i.e. missing dependencies, has any syntax or other error you might have in your custom component or component not in your PYTHONPATH. I had to fix all these three issues to get it working:

To debug try:

❯ python                                                                                                                     
>>> import importlib;importlib.import_module('your_custom_component')

Also check that you can import any dependencies you declared in required_packages . For example if your required_packages looks like:

    @classmethod
    def required_packages(cls) -> List[Text]:
        return ["nltk"]

check that you can import nltk.

Finally make sure that your component is in the PYTHONPATH (ex: export PYTHONPATH=.). It is not enough to have it on the root folder of the project. For some reason rasa doesn’t look there at least if you use the new cli (rasa train).

Related issues: https://github.com/RasaHQ/rasa/issues/3552 https://github.com/RasaHQ/rasa/issues/1513

[mjspeck]

I’ve had a similar problem. The path to my sentiment analyzer is this: nlu.custom_components.sentiment.SentimentAnalyzer, but when I run rasa train I get this error:

Exception: Failed to find component class for 'nlu.custom_components.sentiment.SentimentAnalyzer'.
Unknown component name. Check your configured pipeline and make sure the mentioned component
is not misspelled. If you are creating your own component, make sure it is either listed as part of the
`component_classes` in `rasa.nlu.registry.py` or is a proper name of a class in a module.

I haven’t added the directory to my Python path, but I shouldn’t need to do that, right?