# SentimentTermNormalizer **`class SentimentTermNormalizer(config)`** Bases: [`Normalizer`](squirro.lib.nlp.steps.normalizers.Normalizer.md#squirro.lib.nlp.steps.normalizers.Normalizer) Extracts positive and negative terms/phrases from given text. It tries to detect negations and phrases based on detected [Vader-Valence-Terms](https://github.com/cjhutto/vaderSentiment). Extracted phrases are ranked according to their sentiment-polarity (valence score). **Input** - all input fields need to be of type [`str`](https://docs.python.org/3.11/library/stdtypes.html#str). **Output** - the output fields positive_terms and negative_terms are filled with data of type [`list`](https://docs.python.org/3.11/library/stdtypes.html#list) [ [`str`](https://docs.python.org/3.11/library/stdtypes.html#str) ]. Parameters - `step` ([`str`](https://docs.python.org/3.11/library/stdtypes.html#str)`, ``"normalizer"`) – The step - `type` ([`str`](https://docs.python.org/3.11/library/stdtypes.html#str)`, ``"sentiment_terms"`) – Custom extractor - `fields` ([`list`](https://docs.python.org/3.11/library/stdtypes.html#list)`, ``[``"body"``]`) – Input text - `take_top_n_terms` ([`int`](https://docs.python.org/3.11/library/functions.html#int)`, ``10`) – Extract top-N positive / negative ranked phrases (by polarity) **Example** ```json { "step": "normalizer", "type": "sentiment_terms", "fields": ["body"] } ``` Methods SummaryMethods Documentation **`process_doc(doc)`** Process a document Parameters `doc` ([`Document`](../technical/libnlp/base.md#squirro.lib.nlp.document.Document)) – Document Returns Processed document Return type [Document](../technical/libnlp/base.md#squirro.lib.nlp.document.Document)