Trying to create a basic regex for spanish language without success.

[Idomingog]

I'm trying to create a basic zip code recognition for spanish language,
I modify the example given, but appears an error that I couldn't solve. Anyone could give me a hint?
Thanks in advance.

from presidio_analyzer import AnalyzerEngine, PatternRecognizer, EntityRecognizer, Pattern, RecognizerResult
from presidio_analyzer.recognizer_registry import RecognizerRegistry
from presidio_analyzer.nlp_engine import NlpEngine, SpacyNlpEngine, NlpArtifacts, NlpEngineProvider

from presidio_anonymizer import AnonymizerEngine

Create configuration containing engine name and models

configuration = {
"nlp_engine_name": "spacy",
"models": [{"lang_code": "es", "model_name": "es_core_news_lg"},
{"lang_code": "en", "model_name": "en_core_web_lg"}],
}

Create NLP engine based on configuration

provider = NlpEngineProvider(nlp_configuration=configuration)
nlp_engine_with_spanish = provider.create_engine()

Pass the created NLP engine and supported_languages to the AnalyzerEngine

analyzer = AnalyzerEngine(
nlp_engine=nlp_engine_with_spanish,
supported_languages=["en", "es"]
)

Define the regex pattern

regex = r"(\b\d{5}(?:-\d{4})?\b)" # very weak regex pattern
cp_pattern = Pattern(name="cp_ 5d", regex=regex, score=0.01)

Define the recognizer with the defined pattern

cp_recognizer = PatternRecognizer(supported_entity="ES_ZIP_CODE",
patterns = [cp_pattern],
context= ["Código postal", "código Postal", "CP", "C.P."],
supported_language='es')

registry = RecognizerRegistry()
registry.add_recognizer(cp_recognizer)
analyzer = AnalyzerEngine(registry=registry)

Test

cp_result = analyzer.analyze(text="Mi CP es 28001.", entities=["ES_ZIP_CODE"], language='es')
print(f"Result:\n {cp_result}")

---

KeyError Traceback (most recent call last)
~\AppData\Local\Temp/ipykernel_25648/3379088721.py in
23
24 # Test
---> 25 cp_result = analyzer.analyze(text="Mi CP es 28770.",entities=["ES_ZIP_CODE"], language='es')
26 print(f"Result:\n {cp_result}")

~\anaconda3\envs\Presidio\lib\site-packages\presidio_analyzer\analyzer_engine.py in analyze(self, text, language, entities, correlation_id, score_threshold, return_decision_process, ad_hoc_recognizers)
161 # run the nlp pipeline over the given text, store the results in
162 # a NlpArtifacts instance
--> 163 nlp_artifacts = self.nlp_engine.process_text(text, language)
164
165 if self.log_decision_process:

~\anaconda3\envs\Presidio\lib\site-packages\presidio_analyzer\nlp_engine\spacy_nlp_engine.py in process_text(self, text, language)
41 def process_text(self, text: str, language: str) -> NlpArtifacts:
42 """Execute the SpaCy NLP pipeline on the given text and language."""
---> 43 doc = self.nlplanguage
44 return self._doc_to_nlp_artifact(doc, language)
45

KeyError: 'es'

[omri374]

Hi @Idomingog, from a first glance it appears like you create the AnalyzerEngine object twice. Once with the nlp engine and then with the registry.

Try this :

registry = RecognizerRegistry()
registry.add_recognizer(cp_recognizer)

analyzer = AnalyzerEngine(
nlp_engine=nlp_engine_with_spanish, 
registry=registry, 
supported_languages=["en", "es"]
)

[Idomingog]

Thanks a lot for your help.