Adding a custom preprocesor to MessageProcessor

Hello, I have been trying to implement a custom preprocessor that replaces certain words in the received message by fuzzy matching. I have tried to import a function that I wrote but when I set the message_preprocessor to the function it does not change and stays None.

Here is the code for the fuzzy matching function:

from fuzzywuzzy import process

FUZZY_SEARCH_DATA = ['krzysztof', 'laura', 'agnieszka', 'milena']

def fuzzy_match_preprocessor(message):
    words = message.split()
    new_string = []
    for word in words:
        fuzzy_matched_value = process.extractOne(word, FUZZY_SEARCH_DATA, score_cutoff= 65)
        if fuzzy_matched_value != None:
    message_text = ' '.join(new_string)
    return message_text

I import it from a file called with this line:

from rasa.core.fuzzy_matcher import fuzzy_match_preprocessor

And this is how it is assigned to the message_preprocessor:

message_preprocessor: Optional[LambdaType] = fuzzy_match_preprocessor

Any help would be appreciated.

Hi @pixpack! Have you thought about adding this as a custom component in the NLU pipeline? You can also find more info about it in the docs here

Hello @tyd. :slight_smile: Thank you very much. I’ve managed to fix my issue with the preprocessor but I’m now considering reimplementing it as a custom component.

1 Like

In case anyone is intrested in the component I’m posting the code here. This solution is optimised for fuzzy matching full names or pairs of words. You can change this behavior by changing the ngram_range.

 from rasa.nlu.components import Component
    import typing
    from typing import Any, Optional, Text, Dict

    if typing.TYPE_CHECKING:
        from rasa.nlu.model import Metadata

    class SpellChecker(Component):
        """A new component"""

        provides = ['text']

        requires = []

        defaults = {}

        language_list = None

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

        def train(self, training_data, cfg, **kwargs):

        def process(self, message, **kwargs):
            from fuzzywuzzy import process

            FULLNAMES = ['john kowalski', 'andy michaels']

            text = message.text

            ngram_range = 2
            words = text.split()
            ngrams = zip(*[words[i:] for i in range(ngram_range)])
            tokens = [' '.join(ngram) for ngram in ngrams]

            fuzzy_matched_values = []
            scores = []
            if tokens:
                for token in tokens:
                    fuzzy_match = process.extract(token, FULLNAMES, limit = 1)[0]

                cutoff_threshold = 75

                if max(scores) >= cutoff_threshold:
                    max_value_index = scores.index(max(scores))
                    text = text.replace(tokens[max_value_index],fuzzy_matched_values[max_value_index][0])

            message.text = text

        def persist(self, file_name: Text, model_dir: Text) -> Optional[Dict[Text, Any]]:
            """Persist this component to disk for future loading."""


        def load(
            meta: Dict[Text, Any],
            model_dir: Optional[Text] = None,
            model_metadata: Optional["Metadata"] = None,
            cached_component: Optional["Component"] = None,
            **kwargs: Any
        ) -> "Component":
            """Load this component from file."""

            if cached_component:
                return cached_component
                return cls(meta)

@pixpack where to add it in the pipeline and does it work for every chatbot or do we need change anything