o XεiS/@sJdZddlZddlmZGdddejZddZedkr#edSdS) zx Test sentence segmentation against the "Golden Rules": https://github.com/diasks2/pragmatic_segmenter#the-golden-rules N sentencesc@seZdZdZddZddZddZdd Zd d Zd d Z ddZ ddZ ddZ ddZ ddZddZddZddZddZd d!Zd"d#Zd$d%Zd&S)'GoldenRulesTestCasez*Test golden rules of sentence segmentationcC|tdddgdS)zSimple period to end sentencezHello World. My name is Jonas.z Hello World.My name is Jonas.N assertEqual_get_sentencesselfr K/home/ubuntu/.local/lib/python3.10/site-packages/tests/test_golden_rules.py test_rule_1zGoldenRulesTestCase.test_rule_1cCr)zQuestion mark to end sentencez$What is your name? My name is Jonas.zWhat is your name?rNrr r r r test_rule_2rzGoldenRulesTestCase.test_rule_2cCr)z!Exclamation point to end sentencezThere it is! I found it.z There it is!z I found it.Nrr r r r test_rule_3 zGoldenRulesTestCase.test_rule_3cC|tddgdS)z#One letter upper case abbreviationszMy name is Jonas E. Smith.zMy name is Jonas E Smith.Nrr r r r test_rule_4" zGoldenRulesTestCase.test_rule_4cCr)z#One letter lower case abbreviationszPlease turn to p. 55.zPlease turn to p 55.Nrr r r r test_rule_5)rzGoldenRulesTestCase.test_rule_5cCr)z?Two letter lower case abbreviations in the middle of a sentencezWere Jane and co. at the party?z#Were Jane and company at the party?Nrr r r r test_rule_60zGoldenRulesTestCase.test_rule_6cCr)z?Two letter upper case abbreviations in the middle of a sentencez5They closed the deal with Pitt, Briggs & Co. at noon.z;They closed the deal with Pitt, Briggs and Company at noon.Nrr r r r test_rule_78rzGoldenRulesTestCase.test_rule_7cCr)z&Two letter (prepositive) abbreviationszI can see Mt. Fuji from here.zI can see Mount Fuji from here.Nrr r r r test_rule_10Nrz GoldenRulesTestCase.test_rule_10cCr)z"Possesive two letter abbreviationszThat is JFK Jr.'s book.zThat is J F K Junior's book.Nrr r r r test_rule_12]rz GoldenRulesTestCase.test_rule_12cCr)z6Multi-period abbreviations in the middle of a sentencezI visited the U.S.A. last year.zI visited the U S A last year.Nrr r r r test_rule_13drz GoldenRulesTestCase.test_rule_13cCr)zU.S. as non sentence boundaryz&I have lived in the U.S. for 20 years.z%I have lived in the U S for 20 years.Nrr r r r test_rule_17xrz GoldenRulesTestCase.test_rule_17cCs|tddgdS)zNumber as non sentence boundaryzShe has $100.00 in her bag.Nrr r r r test_rule_19z GoldenRulesTestCase.test_rule_19cCr)z&Double punctuation (exclamation point)zHello!! Long time no see.zHello!!zLong time no see.Nrr r r r test_rule_27rz GoldenRulesTestCase.test_rule_27cCr)z"Double punctuation (question mark)zHello?? Who is there?zHello??z Who is there?Nrr r r r test_rule_28rz GoldenRulesTestCase.test_rule_28cCr)z6Double punctuation (exclamation point / question mark)zHello!? Is that you?zHello!? Is that you?Nrr r r r test_rule_29rz GoldenRulesTestCase.test_rule_29cCr)z6Double punctuation (question mark / exclamation point)zHello?! Is that you?zHello?!r"Nrr r r r test_rule_30rz GoldenRulesTestCase.test_rule_30cCr)z0Errant newlines in the middle of sentences (PDF)z5This is a sentence cut off in the middle because pdf.z5This is a sentence cut off in the middle because pdf.Nrr r r r test_rule_40rz GoldenRulesTestCase.test_rule_40cCr)z*Errant newlines in the middle of sentencesz!It was a cold night in the city.z It was a cold night in the city.Nrr r r r test_rule_41rz GoldenRulesTestCase.test_rule_41N)__name__ __module__ __qualname____doc__rrrrrrrrrrrrr r!r#r$r%r&r r r r r s( * rcCsddt|dddDS)NcSsg|]}|jqSr text).0sr r r sz"_get_sentences..F)verbalize_numbersverbalize_currencyrr+r r r r s r __main__) r*unittestgruutrTestCaserr r'mainr r r r s