import unittest from src.great_ai.utilities import clean class TestClean(unittest.TestCase): def test_xml_handling(self) -> None: xml = 'Hi, my name
is András! <3 <> < ><> <> <|' clean_xml = "Hi, my name is András! <3 <> <|" clean_xml_ascii = "Hi, my name is Andras! <3 <> <|" assert clean(xml) == clean_xml assert clean(xml, ignore_xml=True, ignore_latex=True) == xml assert clean(xml, convert_to_ascii=True) == clean_xml_ascii def test_simple_latex_handling(self) -> None: latex = 'Bj\\"{o}rn is \\textit{happy} 🙂' clean_latex = "Björn is happy 🙂" clean_latex_ascii = "Bjorn is happy" assert clean(latex) == clean_latex assert clean(latex, ignore_latex=True) == latex assert clean(latex, convert_to_ascii=True) == clean_latex_ascii def test_cursed(self) -> None: cursed = """ t̴̨̢̝̻͚͓͉̀̄̃́͜o̸͉̰̼͖͖̅̀̓̿̇̚͝ ̶̢͕̄͊̾͑̂̀̉͑ṱ̸̨̛͈̣̠̤͍̰̂̅͋̀́h̸̹̐͗̅̉͐͂͝e̶̜̳̞̍͘ ̴̟̗̻̤̤̮̒̂̋̓́͐͘͜m̵̺̫̥̙̣̥͐̌͜o̴̖͙̙̎̒͂o̷̬̤͑̆̂̉̊̂n̸̥͒̊ """ cleaned = "to the moon" assert clean(cursed) == cursed.strip() assert clean(cursed, convert_to_ascii=True) == cleaned def test_whitespace(self) -> None: text = """ word1 word2 \n\n\n\t wo\t\f rd3 """ # noqa: W293 cleaned = "word1 word2 wo rd3" assert clean(text, convert_to_ascii=True) == cleaned assert clean(text, ignore_xml=True, ignore_latex=True) == cleaned assert clean(text) == cleaned def test_hyphens(self) -> None: text = """ break - word break- word break -word break -word break-\tword """ cleaned = "break - word break-word break-word break-word break-word" assert clean(text) == cleaned def test_T_I_T_L_E_case(self) -> None: text = "an a r t i c l e is to F I N D the purpose of a tree" cleaned = "an article is to FIND the purpose of a tree" assert clean(text) == cleaned def test_bracket_removal(self) -> None: text = "something [0], and [frefe, ferf]" cleaned = "something, and" assert clean(text, remove_brackets=True) == cleaned self.assertEqual( clean(text, ignore_xml=True, ignore_latex=True, remove_brackets=True), cleaned, ) self.assertEqual( clean(text, convert_to_ascii=True, remove_brackets=True), cleaned ) assert clean(text) == text def test_latex_math_handling(self) -> None: latex = "An increase of 3% was achieved with $q_1 = \\frac{3}{5}$." bad_latex = "An increase of 3% was achieved with $q_1 = \\frac{3}/5$." clean_latex = "An increase of 3% was achieved with q_1 = 3/5." clean_bad_latex = "An increase of 3% was achieved with q_1 = 3//5." assert clean(latex) == clean_latex assert clean(bad_latex) == clean_bad_latex assert clean(latex, ignore_latex=True) == latex assert clean(bad_latex, ignore_latex=True) == bad_latex def test_everything(self) -> None: text = """ Hi % 3 >2

my paper

there! cost- effective cost - effective cost -effective \\frac{1/2} hi \\frac{1}{2} \\textit{italic} \\`acc\\^ented text 北亰 fi æ IJ ij ff András öô 2132 rgrv \n\\   fd [32] [Bei et al., 2003]\n 🤡 🙄 😶‍🌫️ 🇲🇲 """ cleaned = "Hi% 3 >2 my paper there! cost-effective cost - effective cost-effective 1/2/hi 1/2 italic accented text Bei Jing fi ae IJ ij ff Andras oo 2132 rgrv fd" self.assertEqual( clean(text, convert_to_ascii=True, remove_brackets=True), cleaned ) def test_empty(self) -> None: assert clean("", convert_to_ascii=True) == "" def test_punctuation_fixing(self) -> None: text = " Dear reader ( or listener ) , welcome ! " fixed = "Dear reader (or listener), welcome!" assert clean(text) == fixed