object has no attribute '_formatter_parser'

[buddharajuramaraju]

Rasa NLU version :"version": "0.10.1"

Used backend / pipeline spacy_sklearn,

Operating system RHEL 7

Issue:
while executing query following error is comming

{"error": "'NoneType' object has no attribute '_formatter_parser'"}

Example:
curl -XPOST localhost:5000/parse -d '{"q":"I am looking for Chinese food"}'
{"error": "'NoneType' object has no attribute '_formatter_parser'"}

Content of configuration file (if used & relevant):

[wrathagom]

Can you provide more information on how you're running Rasa: command line, python, docker, etc?

Have you already trained a model? What is in the /projects folder? Can you post the metadata.json file of the model that is being loaded.

[derde]

I have the same on a brand new installation on Ubuntu from github: command line, running as regular user, following the tutorial without modification.
Running as such:

python -m rasa_nlu.train -c sample_configs/config_spacy.json
python -m rasa_nlu.server -c sample_configs/config_spacy.json &
# interactive:
curl -XPOST localhost:5000/parse -d '{"q":"I want some italian", "model": "model_20171020-125436" }' 

metadata.json as below:

{
    "entity_synonyms": "entity_synonyms.json", 
    "pipeline": [
        "nlp_spacy", 
        "tokenizer_spacy", 
        "intent_featurizer_spacy", 
        "intent_entity_featurizer_regex", 
        "ner_crf", 
        "ner_synonyms", 
        "intent_classifier_sklearn"
    ], 
    "spacy_model_name": "en", 
    "language": "en", 
    "intent_classifier_sklearn": "intent_classifier.pkl", 
    "trained_at": "20171020-125436", 
    "training_data": "training_data.json", 
    "entity_extractor_crf": {
        "BILOU_flag": true, 
        "model_file": "crf_model.pkl", 
        "version": 1, 
        "crf_features": [
            [
                "low", 
                "title", 
                "upper", 
                "pos", 
                "pos2"
            ], 
            [
                "bias", 
                "low", 
                "word3", 
                "word2", 
                "upper", 
                "title", 
                "digit", 
                "pos", 
                "pos2", 
                "pattern"
            ], 
            [
                "low", 
                "title", 
                "upper", 
                "pos", 
                "pos2"
            ]
        ]
    }, 
    "rasa_nlu_version": "0.10.1", 
    "regex_featurizer": "regex_featurizer.json"
}

[tmbo]

which python version do you use?

[tmbo]

And can you please also add the stacktrace from the console

[Jonathan-Wei]

Hi,@wrathagom,I have the same problem.
This is my metadata.json.
{ "entity_synonyms": "entity_synonyms.json", "pipeline": [ "nlp_mitie", "tokenizer_jieba", "ner_mitie", "ner_synonyms", "intent_entity_featurizer_regex", "intent_featurizer_mitie", "intent_classifier_sklearn" ], "mitie_feature_extractor_fingerprint": 1810187658478185215, "regex_featurizer": null, "language": "zh", "mitie_file": "./data/total_word_feature_extractor_zh.dat", "intent_classifier_sklearn": "intent_classifier.pkl", "trained_at": "20171030-153135", "training_data": "training_data.json", "entity_extractor_mitie": "entity_extractor.dat", "rasa_nlu_version": "0.10.1" }

[Jonathan-Wei]

@wrathagom @tmbo python version is 2.7.0

[derde]

I pushed a fix/work-around (it's a log statement that doesn't have a format string), so if you get the current version of rasa_nlu you should find the visible part of this problem has gone away.

[tmbo]

@Jonathan-Wei we also just recently cut a release, so the issue is also fixed in 0.10.4