CoolFace
Apppublic

mbarnig/ACE-Step

sourceHugging Faceapache-2.0updated 10mo agoView on Hugging Face
0likes
lyric_tokenizer.py883 linesDownload Raw Back to lyrics_utils
1import os2import re3import textwrap4from functools import cached_property5 6import pypinyin7import torch8from hangul_romanize import Transliter9from hangul_romanize.rule import academic10from num2words import num2words11from spacy.lang.ar import Arabic12from spacy.lang.en import English13from spacy.lang.es import Spanish14from spacy.lang.ja import Japanese15from spacy.lang.zh import Chinese16from tokenizers import Tokenizer17 18from .zh_num2words import TextNorm as zh_num2words19from typing import Dict, List, Optional, Set, Union20 21 22#copy from https://github.com/coqui-ai/TTS/blob/dbf1a08a0d4e47fdad6172e433eeb34bc6b13b4e/TTS/tts/layers/xtts/tokenizer.py23def get_spacy_lang(lang):24    if lang == "zh":25        return Chinese()26    elif lang == "ja":27        return Japanese()28    elif lang == "ar":29        return Arabic()30    elif lang == "es":31        return Spanish()32    else:33        # For most languages, Enlish does the job34        return English()35 36 37def split_sentence(text, lang, text_split_length=250):38    """Preprocess the input text"""39    text_splits = []40    if text_split_length is not None and len(text) >= text_split_length:41        text_splits.append("")42        nlp = get_spacy_lang(lang)43        nlp.add_pipe("sentencizer")44        doc = nlp(text)45        for sentence in doc.sents:46            if len(text_splits[-1]) + len(str(sentence)) <= text_split_length:47                # if the last sentence + the current sentence is less than the text_split_length48                # then add the current sentence to the last sentence49                text_splits[-1] += " " + str(sentence)50                text_splits[-1] = text_splits[-1].lstrip()51            elif len(str(sentence)) > text_split_length:52                # if the current sentence is greater than the text_split_length53                for line in textwrap.wrap(54                    str(sentence),55                    width=text_split_length,56                    drop_whitespace=True,57                    break_on_hyphens=False,58                    tabsize=1,59                ):60                    text_splits.append(str(line))61            else:62                text_splits.append(str(sentence))63 64        if len(text_splits) > 1:65            if text_splits[0] == "":66                del text_splits[0]67    else:68        text_splits = [text.lstrip()]69 70    return text_splits71 72 73_whitespace_re = re.compile(r"\s+")74 75# List of (regular expression, replacement) pairs for abbreviations:76_abbreviations = {77    "en": [78        (re.compile("\\b%s\\." % x[0], re.IGNORECASE), x[1])79        for x in [80            ("mrs", "misess"),81            ("mr", "mister"),82            ("dr", "doctor"),83            ("st", "saint"),84            ("co", "company"),85            ("jr", "junior"),86            ("maj", "major"),87            ("gen", "general"),88            ("drs", "doctors"),89            ("rev", "reverend"),90            ("lt", "lieutenant"),91            ("hon", "honorable"),92            ("sgt", "sergeant"),93            ("capt", "captain"),94            ("esq", "esquire"),95            ("ltd", "limited"),96            ("col", "colonel"),97            ("ft", "fort"),98        ]99    ],100    "es": [101        (re.compile("\\b%s\\." % x[0], re.IGNORECASE), x[1])102        for x in [103            ("sra", "señora"),104            ("sr", "señor"),105            ("dr", "doctor"),106            ("dra", "doctora"),107            ("st", "santo"),108            ("co", "compañía"),109            ("jr", "junior"),110            ("ltd", "limitada"),111        ]112    ],113    "fr": [114        (re.compile("\\b%s\\." % x[0], re.IGNORECASE), x[1])115        for x in [116            ("mme", "madame"),117            ("mr", "monsieur"),118            ("dr", "docteur"),119            ("st", "saint"),120            ("co", "compagnie"),121            ("jr", "junior"),122            ("ltd", "limitée"),123        ]124    ],125    "de": [126        (re.compile("\\b%s\\." % x[0], re.IGNORECASE), x[1])127        for x in [128            ("fr", "frau"),129            ("dr", "doktor"),130            ("st", "sankt"),131            ("co", "firma"),132            ("jr", "junior"),133        ]134    ],135    "pt": [136        (re.compile("\\b%s\\." % x[0], re.IGNORECASE), x[1])137        for x in [138            ("sra", "senhora"),139            ("sr", "senhor"),140            ("dr", "doutor"),141            ("dra", "doutora"),142            ("st", "santo"),143            ("co", "companhia"),144            ("jr", "júnior"),145            ("ltd", "limitada"),146        ]147    ],148    "it": [149        (re.compile("\\b%s\\." % x[0], re.IGNORECASE), x[1])150        for x in [151            # ("sig.ra", "signora"),152            ("sig", "signore"),153            ("dr", "dottore"),154            ("st", "santo"),155            ("co", "compagnia"),156            ("jr", "junior"),157            ("ltd", "limitata"),158        ]159    ],160    "pl": [161        (re.compile("\\b%s\\." % x[0], re.IGNORECASE), x[1])162        for x in [163            ("p", "pani"),164            ("m", "pan"),165            ("dr", "doktor"),166            ("sw", "święty"),167            ("jr", "junior"),168        ]169    ],170    "ar": [171        (re.compile("\\b%s\\." % x[0], re.IGNORECASE), x[1])172        for x in [173            # There are not many common abbreviations in Arabic as in English.174        ]175    ],176    "zh": [177        (re.compile("\\b%s\\." % x[0], re.IGNORECASE), x[1])178        for x in [179            # Chinese doesn't typically use abbreviations in the same way as Latin-based scripts.180        ]181    ],182    "cs": [183        (re.compile("\\b%s\\." % x[0], re.IGNORECASE), x[1])184        for x in [185            ("dr", "doktor"),  # doctor186            ("ing", "inženýr"),  # engineer187            ("p", "pan"),  # Could also map to pani for woman but no easy way to do it188            # Other abbreviations would be specialized and not as common.189        ]190    ],191    "ru": [192        (re.compile("\\b%s\\b" % x[0], re.IGNORECASE), x[1])193        for x in [194            ("г-жа", "госпожа"),  # Mrs.195            ("г-н", "господин"),  # Mr.196            ("д-р", "доктор"),  # doctor197            # Other abbreviations are less common or specialized.198        ]199    ],200    "nl": [201        (re.compile("\\b%s\\." % x[0], re.IGNORECASE), x[1])202        for x in [203            ("dhr", "de heer"),  # Mr.204            ("mevr", "mevrouw"),  # Mrs.205            ("dr", "dokter"),  # doctor206            ("jhr", "jonkheer"),  # young lord or nobleman207            # Dutch uses more abbreviations, but these are the most common ones.208        ]209    ],210    "tr": [211        (re.compile("\\b%s\\." % x[0], re.IGNORECASE), x[1])212        for x in [213            ("b", "bay"),  # Mr.214            ("byk", "büyük"),  # büyük215            ("dr", "doktor"),  # doctor216            # Add other Turkish abbreviations here if needed.217        ]218    ],219    "hu": [220        (re.compile("\\b%s\\." % x[0], re.IGNORECASE), x[1])221        for x in [222            ("dr", "doktor"),  # doctor223            ("b", "bácsi"),  # Mr.224            ("nőv", "nővér"),  # nurse225            # Add other Hungarian abbreviations here if needed.226        ]227    ],228    "ko": [229        (re.compile("\\b%s\\." % x[0], re.IGNORECASE), x[1])230        for x in [231            # Korean doesn't typically use abbreviations in the same way as Latin-based scripts.232        ]233    ],234}235 236 237def expand_abbreviations_multilingual(text, lang="en"):238    for regex, replacement in _abbreviations[lang]:239        text = re.sub(regex, replacement, text)240    return text241 242 243_symbols_multilingual = {244    "en": [245        (re.compile(r"%s" % re.escape(x[0]), re.IGNORECASE), x[1])246        for x in [247            ("&", " and "),248            ("@", " at "),249            ("%", " percent "),250            ("#", " hash "),251            ("$", " dollar "),252            ("£", " pound "),253            ("°", " degree "),254        ]255    ],256    "es": [257        (re.compile(r"%s" % re.escape(x[0]), re.IGNORECASE), x[1])258        for x in [259            ("&", " y "),260            ("@", " arroba "),261            ("%", " por ciento "),262            ("#", " numeral "),263            ("$", " dolar "),264            ("£", " libra "),265            ("°", " grados "),266        ]267    ],268    "fr": [269        (re.compile(r"%s" % re.escape(x[0]), re.IGNORECASE), x[1])270        for x in [271            ("&", " et "),272            ("@", " arobase "),273            ("%", " pour cent "),274            ("#", " dièse "),275            ("$", " dollar "),276            ("£", " livre "),277            ("°", " degrés "),278        ]279    ],280    "de": [281        (re.compile(r"%s" % re.escape(x[0]), re.IGNORECASE), x[1])282        for x in [283            ("&", " und "),284            ("@", " at "),285            ("%", " prozent "),286            ("#", " raute "),287            ("$", " dollar "),288            ("£", " pfund "),289            ("°", " grad "),290        ]291    ],292    "pt": [293        (re.compile(r"%s" % re.escape(x[0]), re.IGNORECASE), x[1])294        for x in [295            ("&", " e "),296            ("@", " arroba "),297            ("%", " por cento "),298            ("#", " cardinal "),299            ("$", " dólar "),300            ("£", " libra "),301            ("°", " graus "),302        ]303    ],304    "it": [305        (re.compile(r"%s" % re.escape(x[0]), re.IGNORECASE), x[1])306        for x in [307            ("&", " e "),308            ("@", " chiocciola "),309            ("%", " per cento "),310            ("#", " cancelletto "),311            ("$", " dollaro "),312            ("£", " sterlina "),313            ("°", " gradi "),314        ]315    ],316    "pl": [317        (re.compile(r"%s" % re.escape(x[0]), re.IGNORECASE), x[1])318        for x in [319            ("&", " i "),320            ("@", " małpa "),321            ("%", " procent "),322            ("#", " krzyżyk "),323            ("$", " dolar "),324            ("£", " funt "),325            ("°", " stopnie "),326        ]327    ],328    "ar": [329        # Arabic330        (re.compile(r"%s" % re.escape(x[0]), re.IGNORECASE), x[1])331        for x in [332            ("&", " و "),333            ("@", " على "),334            ("%", " في المئة "),335            ("#", " رقم "),336            ("$", " دولار "),337            ("£", " جنيه "),338            ("°", " درجة "),339        ]340    ],341    "zh": [342        # Chinese343        (re.compile(r"%s" % re.escape(x[0]), re.IGNORECASE), x[1])344        for x in [345            ("&", " 和 "),346            ("@", " 在 "),347            ("%", " 百分之 "),348            ("#", " 号 "),349            ("$", " 美元 "),350            ("£", " 英镑 "),351            ("°", " 度 "),352        ]353    ],354    "cs": [355        # Czech356        (re.compile(r"%s" % re.escape(x[0]), re.IGNORECASE), x[1])357        for x in [358            ("&", " a "),359            ("@", " na "),360            ("%", " procento "),361            ("#", " křížek "),362            ("$", " dolar "),363            ("£", " libra "),364            ("°", " stupně "),365        ]366    ],367    "ru": [368        # Russian369        (re.compile(r"%s" % re.escape(x[0]), re.IGNORECASE), x[1])370        for x in [371            ("&", " и "),372            ("@", " собака "),373            ("%", " процентов "),374            ("#", " номер "),375            ("$", " доллар "),376            ("£", " фунт "),377            ("°", " градус "),378        ]379    ],380    "nl": [381        # Dutch382        (re.compile(r"%s" % re.escape(x[0]), re.IGNORECASE), x[1])383        for x in [384            ("&", " en "),385            ("@", " bij "),386            ("%", " procent "),387            ("#", " hekje "),388            ("$", " dollar "),389            ("£", " pond "),390            ("°", " graden "),391        ]392    ],393    "tr": [394        (re.compile(r"%s" % re.escape(x[0]), re.IGNORECASE), x[1])395        for x in [396            ("&", " ve "),397            ("@", " at "),398            ("%", " yüzde "),399            ("#", " diyez "),400            ("$", " dolar "),401            ("£", " sterlin "),402            ("°", " derece "),403        ]404    ],405    "hu": [406        (re.compile(r"%s" % re.escape(x[0]), re.IGNORECASE), x[1])407        for x in [408            ("&", " és "),409            ("@", " kukac "),410            ("%", " százalék "),411            ("#", " kettőskereszt "),412            ("$", " dollár "),413            ("£", " font "),414            ("°", " fok "),415        ]416    ],417    "ko": [418        # Korean419        (re.compile(r"%s" % re.escape(x[0]), re.IGNORECASE), x[1])420        for x in [421            ("&", " 그리고 "),422            ("@", " 에 "),423            ("%", " 퍼센트 "),424            ("#", " 번호 "),425            ("$", " 달러 "),426            ("£", " 파운드 "),427            ("°", " 도 "),428        ]429    ],430}431 432 433def expand_symbols_multilingual(text, lang="en"):434    for regex, replacement in _symbols_multilingual[lang]:435        text = re.sub(regex, replacement, text)436        text = text.replace("  ", " ")  # Ensure there are no double spaces437    return text.strip()438 439 440_ordinal_re = {441    "en": re.compile(r"([0-9]+)(st|nd|rd|th)"),442    "es": re.compile(r"([0-9]+)(º|ª|er|o|a|os|as)"),443    "fr": re.compile(r"([0-9]+)(º|ª|er|re|e|ème)"),444    "de": re.compile(r"([0-9]+)(st|nd|rd|th|º|ª|\.(?=\s|$))"),445    "pt": re.compile(r"([0-9]+)(º|ª|o|a|os|as)"),446    "it": re.compile(r"([0-9]+)(º|°|ª|o|a|i|e)"),447    "pl": re.compile(r"([0-9]+)(º|ª|st|nd|rd|th)"),448    "ar": re.compile(r"([0-9]+)(ون|ين|ث|ر|ى)"),449    "cs": re.compile(r"([0-9]+)\.(?=\s|$)"),  # In Czech, a dot is often used after the number to indicate ordinals.450    "ru": re.compile(r"([0-9]+)(-й|-я|-е|-ое|-ье|-го)"),451    "nl": re.compile(r"([0-9]+)(de|ste|e)"),452    "tr": re.compile(r"([0-9]+)(\.|inci|nci|uncu|üncü|\.)"),453    "hu": re.compile(r"([0-9]+)(\.|adik|edik|odik|edik|ödik|ödike|ik)"),454    "ko": re.compile(r"([0-9]+)(번째|번|차|째)"),455}456_number_re = re.compile(r"[0-9]+")457_currency_re = {458    "USD": re.compile(r"((\$[0-9\.\,]*[0-9]+)|([0-9\.\,]*[0-9]+\$))"),459    "GBP": re.compile(r"((£[0-9\.\,]*[0-9]+)|([0-9\.\,]*[0-9]+£))"),460    "EUR": re.compile(r"(([0-9\.\,]*[0-9]+€)|((€[0-9\.\,]*[0-9]+)))"),461}462 463_comma_number_re = re.compile(r"\b\d{1,3}(,\d{3})*(\.\d+)?\b")464_dot_number_re = re.compile(r"\b\d{1,3}(.\d{3})*(\,\d+)?\b")465_decimal_number_re = re.compile(r"([0-9]+[.,][0-9]+)")466 467 468def _remove_commas(m):469    text = m.group(0)470    if "," in text:471        text = text.replace(",", "")472    return text473 474 475def _remove_dots(m):476    text = m.group(0)477    if "." in text:478        text = text.replace(".", "")479    return text480 481 482def _expand_decimal_point(m, lang="en"):483    amount = m.group(1).replace(",", ".")484    return num2words(float(amount), lang=lang if lang != "cs" else "cz")485 486 487def _expand_currency(m, lang="en", currency="USD"):488    amount = float((re.sub(r"[^\d.]", "", m.group(0).replace(",", "."))))489    full_amount = num2words(amount, to="currency", currency=currency, lang=lang if lang != "cs" else "cz")490 491    and_equivalents = {492        "en": ", ",493        "es": " con ",494        "fr": " et ",495        "de": " und ",496        "pt": " e ",497        "it": " e ",498        "pl": ", ",499        "cs": ", ",500        "ru": ", ",501        "nl": ", ",502        "ar": ", ",503        "tr": ", ",504        "hu": ", ",505        "ko": ", ",506    }507 508    if amount.is_integer():509        last_and = full_amount.rfind(and_equivalents[lang])510        if last_and != -1:511            full_amount = full_amount[:last_and]512 513    return full_amount514 515 516def _expand_ordinal(m, lang="en"):517    return num2words(int(m.group(1)), ordinal=True, lang=lang if lang != "cs" else "cz")518 519 520def _expand_number(m, lang="en"):521    return num2words(int(m.group(0)), lang=lang if lang != "cs" else "cz")522 523 524def expand_numbers_multilingual(text, lang="en"):525    if lang == "zh":526        text = zh_num2words()(text)527    else:528        if lang in ["en", "ru"]:529            text = re.sub(_comma_number_re, _remove_commas, text)530        else:531            text = re.sub(_dot_number_re, _remove_dots, text)532        try:533            text = re.sub(_currency_re["GBP"], lambda m: _expand_currency(m, lang, "GBP"), text)534            text = re.sub(_currency_re["USD"], lambda m: _expand_currency(m, lang, "USD"), text)535            text = re.sub(_currency_re["EUR"], lambda m: _expand_currency(m, lang, "EUR"), text)536        except:537            pass538        if lang != "tr":539            text = re.sub(_decimal_number_re, lambda m: _expand_decimal_point(m, lang), text)540        text = re.sub(_ordinal_re[lang], lambda m: _expand_ordinal(m, lang), text)541        text = re.sub(_number_re, lambda m: _expand_number(m, lang), text)542    return text543 544 545def lowercase(text):546    return text.lower()547 548 549def collapse_whitespace(text):550    return re.sub(_whitespace_re, " ", text)551 552 553def multilingual_cleaners(text, lang):554    text = text.replace('"', "")555    if lang == "tr":556        text = text.replace("İ", "i")557        text = text.replace("Ö", "ö")558        text = text.replace("Ü", "ü")559    text = lowercase(text)560    try:561        text = expand_numbers_multilingual(text, lang)562    except:563        pass564    try:565        text = expand_abbreviations_multilingual(text, lang)566    except:567        pass568    try:569        text = expand_symbols_multilingual(text, lang=lang)570    except:571        pass572    text = collapse_whitespace(text)573    return text574 575 576def basic_cleaners(text):577    """Basic pipeline that lowercases and collapses whitespace without transliteration."""578    text = lowercase(text)579    text = collapse_whitespace(text)580    return text581 582 583def chinese_transliterate(text):584    return "".join(585        [p[0] for p in pypinyin.pinyin(text, style=pypinyin.Style.TONE3, heteronym=False, neutral_tone_with_five=True)]586    )587 588 589def japanese_cleaners(text, katsu):590    text = katsu.romaji(text)591    text = lowercase(text)592    return text593 594 595def korean_transliterate(text):596    r = Transliter(academic)597    return r.translit(text)598 599 600DEFAULT_VOCAB_FILE = os.path.join(os.path.dirname(os.path.realpath(__file__)), "vocab.json")601 602 603class VoiceBpeTokenizer:604    def __init__(self, vocab_file=DEFAULT_VOCAB_FILE):605        self.tokenizer = None606        if vocab_file is not None:607            self.tokenizer = Tokenizer.from_file(vocab_file)608        self.char_limits = {609            "en": 10000,610            "de": 253,611            "fr": 273,612            "es": 239,613            "it": 213,614            "pt": 203,615            "pl": 224,616            "zh": 82,617            "ar": 166,618            "cs": 186,619            "ru": 182,620            "nl": 251,621            "tr": 226,622            "ja": 71,623            "hu": 224,624            "ko": 95,625        }626 627    @cached_property628    def katsu(self):629        import cutlet630 631        return cutlet.Cutlet()632 633    def check_input_length(self, txt, lang):634        lang = lang.split("-")[0]  # remove the region635        limit = self.char_limits.get(lang, 250)636        # if len(txt) > limit:637        #     print(638        #         f"[!] Warning: The text length exceeds the character limit of {limit} for language '{lang}', this might cause truncated audio."639        #     )640 641    def preprocess_text(self, txt, lang):642        if lang in {"ar", "cs", "de", "en", "es", "fr", "hu", "it", "nl", "pl", "pt", "ru", "tr", "zh", "ko"}:643            txt = multilingual_cleaners(txt, lang)644            if lang == "zh":645                txt = chinese_transliterate(txt)646            if lang == "ko":647                txt = korean_transliterate(txt)648        elif lang == "ja":649            txt = japanese_cleaners(txt, self.katsu)650        elif lang == "hi":651            # @manmay will implement this652            txt = basic_cleaners(txt)653        else:654            raise NotImplementedError(f"Language '{lang}' is not supported.")655        return txt656    657    def encode(self, txt, lang):658        lang = lang.split("-")[0]  # remove the region659        self.check_input_length(txt, lang)660        txt = self.preprocess_text(txt, lang)661        lang = "zh-cn" if lang == "zh" else lang662        txt = f"[{lang}]{txt}"663        txt = txt.replace(" ", "[SPACE]")664        return self.tokenizer.encode(txt).ids665 666    def decode(self, seq, skip_special_tokens=False):667        if isinstance(seq, torch.Tensor):668            seq = seq.cpu().numpy()669        txt = self.tokenizer.decode(seq, skip_special_tokens=False).replace(" ", "")670        txt = txt.replace("[SPACE]", " ")671        txt = txt.replace("[STOP]", "")672        # txt = txt.replace("[UNK]", "")673        return txt674    675 676    #copy from https://github.com/huggingface/transformers/blob/main/src/transformers/tokenization_utils_base.py#L3936677    def batch_decode(678        self,679        sequences: Union[List[int], List[List[int]], "np.ndarray", "torch.Tensor", "tf.Tensor"],680        skip_special_tokens: bool = False,681    ) -> List[str]:682        """683        Convert a list of lists of token ids into a list of strings by calling decode.684 685        Args:686            sequences (`Union[List[int], List[List[int]], np.ndarray, torch.Tensor, tf.Tensor]`):687                List of tokenized input ids. Can be obtained using the `__call__` method.688            skip_special_tokens (`bool`, *optional*, defaults to `False`):689                Whether or not to remove special tokens in the decoding.690            kwargs (additional keyword arguments, *optional*):691                Will be passed to the underlying model specific decode method.692 693        Returns:694            `List[str]`: The list of decoded sentences.695        """696        return [697            self.decode(seq)698            for seq in sequences699        ]700    701    #https://github.com/coqui-ai/TTS/blob/dev/TTS/tts/layers/xtts/trainer/dataset.py#L202702    # def pad(self): 703 704    def __len__(self):705        return self.tokenizer.get_vocab_size()706 707    def get_number_tokens(self):708        return max(self.tokenizer.get_vocab().values()) + 1709 710 711def test_expand_numbers_multilingual():712    test_cases = [713        # English714        ("In 12.5 seconds.", "In twelve point five seconds.", "en"),715        ("There were 50 soldiers.", "There were fifty soldiers.", "en"),716        ("This is a 1st test", "This is a first test", "en"),717        ("That will be $20 sir.", "That will be twenty dollars sir.", "en"),718        ("That will be 20€ sir.", "That will be twenty euro sir.", "en"),719        ("That will be 20.15€ sir.", "That will be twenty euro, fifteen cents sir.", "en"),720        ("That's 100,000.5.", "That's one hundred thousand point five.", "en"),721        # French722        ("En 12,5 secondes.", "En douze virgule cinq secondes.", "fr"),723        ("Il y avait 50 soldats.", "Il y avait cinquante soldats.", "fr"),724        ("Ceci est un 1er test", "Ceci est un premier test", "fr"),725        ("Cela vous fera $20 monsieur.", "Cela vous fera vingt dollars monsieur.", "fr"),726        ("Cela vous fera 20€ monsieur.", "Cela vous fera vingt euros monsieur.", "fr"),727        ("Cela vous fera 20,15€ monsieur.", "Cela vous fera vingt euros et quinze centimes monsieur.", "fr"),728        ("Ce sera 100.000,5.", "Ce sera cent mille virgule cinq.", "fr"),729        # German730        ("In 12,5 Sekunden.", "In zwölf Komma fünf Sekunden.", "de"),731        ("Es gab 50 Soldaten.", "Es gab fünfzig Soldaten.", "de"),732        ("Dies ist ein 1. Test", "Dies ist ein erste Test", "de"),  # Issue with gender733        ("Das macht $20 Herr.", "Das macht zwanzig Dollar Herr.", "de"),734        ("Das macht 20€ Herr.", "Das macht zwanzig Euro Herr.", "de"),735        ("Das macht 20,15€ Herr.", "Das macht zwanzig Euro und fünfzehn Cent Herr.", "de"),736        # Spanish737        ("En 12,5 segundos.", "En doce punto cinco segundos.", "es"),738        ("Había 50 soldados.", "Había cincuenta soldados.", "es"),739        ("Este es un 1er test", "Este es un primero test", "es"),740        ("Eso le costará $20 señor.", "Eso le costará veinte dólares señor.", "es"),741        ("Eso le costará 20€ señor.", "Eso le costará veinte euros señor.", "es"),742        ("Eso le costará 20,15€ señor.", "Eso le costará veinte euros con quince céntimos señor.", "es"),743        # Italian744        ("In 12,5 secondi.", "In dodici virgola cinque secondi.", "it"),745        ("C'erano 50 soldati.", "C'erano cinquanta soldati.", "it"),746        ("Questo è un 1° test", "Questo è un primo test", "it"),747        ("Ti costerà $20 signore.", "Ti costerà venti dollari signore.", "it"),748        ("Ti costerà 20€ signore.", "Ti costerà venti euro signore.", "it"),749        ("Ti costerà 20,15€ signore.", "Ti costerà venti euro e quindici centesimi signore.", "it"),750        # Portuguese751        ("Em 12,5 segundos.", "Em doze vírgula cinco segundos.", "pt"),752        ("Havia 50 soldados.", "Havia cinquenta soldados.", "pt"),753        ("Este é um 1º teste", "Este é um primeiro teste", "pt"),754        ("Isso custará $20 senhor.", "Isso custará vinte dólares senhor.", "pt"),755        ("Isso custará 20€ senhor.", "Isso custará vinte euros senhor.", "pt"),756        (757            "Isso custará 20,15€ senhor.",758            "Isso custará vinte euros e quinze cêntimos senhor.",759            "pt",760        ),  # "cêntimos" should be "centavos" num2words issue761        # Polish762        ("W 12,5 sekundy.", "W dwanaście przecinek pięć sekundy.", "pl"),763        ("Było 50 żołnierzy.", "Było pięćdziesiąt żołnierzy.", "pl"),764        ("To będzie kosztować 20€ panie.", "To będzie kosztować dwadzieścia euro panie.", "pl"),765        ("To będzie kosztować 20,15€ panie.", "To będzie kosztować dwadzieścia euro, piętnaście centów panie.", "pl"),766        # Arabic767        ("في الـ 12,5 ثانية.", "في الـ اثنا عشر  , خمسون ثانية.", "ar"),768        ("كان هناك 50 جنديًا.", "كان هناك خمسون جنديًا.", "ar"),769        # ("ستكون النتيجة $20 يا سيد.", 'ستكون النتيجة عشرون دولار يا سيد.', 'ar'), # $ and € are mising from num2words770        # ("ستكون النتيجة 20€ يا سيد.", 'ستكون النتيجة عشرون يورو يا سيد.', 'ar'),771        # Czech772        ("Za 12,5 vteřiny.", "Za dvanáct celá pět vteřiny.", "cs"),773        ("Bylo tam 50 vojáků.", "Bylo tam padesát vojáků.", "cs"),774        ("To bude stát 20€ pane.", "To bude stát dvacet euro pane.", "cs"),775        ("To bude 20.15€ pane.", "To bude dvacet euro, patnáct centů pane.", "cs"),776        # Russian777        ("Через 12.5 секунды.", "Через двенадцать запятая пять секунды.", "ru"),778        ("Там было 50 солдат.", "Там было пятьдесят солдат.", "ru"),779        ("Это будет 20.15€ сэр.", "Это будет двадцать евро, пятнадцать центов сэр.", "ru"),780        ("Это будет стоить 20€ господин.", "Это будет стоить двадцать евро господин.", "ru"),781        # Dutch782        ("In 12,5 seconden.", "In twaalf komma vijf seconden.", "nl"),783        ("Er waren 50 soldaten.", "Er waren vijftig soldaten.", "nl"),784        ("Dat wordt dan $20 meneer.", "Dat wordt dan twintig dollar meneer.", "nl"),785        ("Dat wordt dan 20€ meneer.", "Dat wordt dan twintig euro meneer.", "nl"),786        # Chinese (Simplified)787        ("在12.5秒内", "在十二点五秒内", "zh"),788        ("有50名士兵", "有五十名士兵", "zh"),789        # ("那将是$20先生", '那将是二十美元先生', 'zh'), currency doesn't work790        # ("那将是20€先生", '那将是二十欧元先生', 'zh'),791        # Turkish792        # ("12,5 saniye içinde.", 'On iki virgül beş saniye içinde.', 'tr'), # decimal doesn't work for TR793        ("50 asker vardı.", "elli asker vardı.", "tr"),794        ("Bu 1. test", "Bu birinci test", "tr"),795        # ("Bu 100.000,5.", 'Bu yüz bin virgül beş.', 'tr'),796        # Hungarian797        ("12,5 másodperc alatt.", "tizenkettő egész öt tized másodperc alatt.", "hu"),798        ("50 katona volt.", "ötven katona volt.", "hu"),799        ("Ez az 1. teszt", "Ez az első teszt", "hu"),800        # Korean801        ("12.5 초 안에.", "십이 점 다섯 초 안에.", "ko"),802        ("50 명의 병사가 있었다.", "오십 명의 병사가 있었다.", "ko"),803        ("이것은 1 번째 테스트입니다", "이것은 첫 번째 테스트입니다", "ko"),804    ]805    for a, b, lang in test_cases:806        out = expand_numbers_multilingual(a, lang=lang)807        assert out == b, f"'{out}' vs '{b}'"808 809 810def test_abbreviations_multilingual():811    test_cases = [812        # English813        ("Hello Mr. Smith.", "Hello mister Smith.", "en"),814        ("Dr. Jones is here.", "doctor Jones is here.", "en"),815        # Spanish816        ("Hola Sr. Garcia.", "Hola señor Garcia.", "es"),817        ("La Dra. Martinez es muy buena.", "La doctora Martinez es muy buena.", "es"),818        # French819        ("Bonjour Mr. Dupond.", "Bonjour monsieur Dupond.", "fr"),820        ("Mme. Moreau est absente aujourd'hui.", "madame Moreau est absente aujourd'hui.", "fr"),821        # German822        ("Frau Dr. Müller ist sehr klug.", "Frau doktor Müller ist sehr klug.", "de"),823        # Portuguese824        ("Olá Sr. Silva.", "Olá senhor Silva.", "pt"),825        ("Dra. Costa, você está disponível?", "doutora Costa, você está disponível?", "pt"),826        # Italian827        ("Buongiorno, Sig. Rossi.", "Buongiorno, signore Rossi.", "it"),828        # ("Sig.ra Bianchi, posso aiutarti?", 'signora Bianchi, posso aiutarti?', 'it'), # Issue with matching that pattern829        # Polish830        ("Dzień dobry, P. Kowalski.", "Dzień dobry, pani Kowalski.", "pl"),831        ("M. Nowak, czy mogę zadać pytanie?", "pan Nowak, czy mogę zadać pytanie?", "pl"),832        # Czech833        ("P. Novák", "pan Novák", "cs"),834        ("Dr. Vojtěch", "doktor Vojtěch", "cs"),835        # Dutch836        ("Dhr. Jansen", "de heer Jansen", "nl"),837        ("Mevr. de Vries", "mevrouw de Vries", "nl"),838        # Russian839        ("Здравствуйте Г-н Иванов.", "Здравствуйте господин Иванов.", "ru"),840        ("Д-р Смирнов здесь, чтобы увидеть вас.", "доктор Смирнов здесь, чтобы увидеть вас.", "ru"),841        # Turkish842        ("Merhaba B. Yılmaz.", "Merhaba bay Yılmaz.", "tr"),843        ("Dr. Ayşe burada.", "doktor Ayşe burada.", "tr"),844        # Hungarian845        ("Dr. Szabó itt van.", "doktor Szabó itt van.", "hu"),846    ]847 848    for a, b, lang in test_cases:849        out = expand_abbreviations_multilingual(a, lang=lang)850        assert out == b, f"'{out}' vs '{b}'"851 852 853def test_symbols_multilingual():854    test_cases = [855        ("I have 14% battery", "I have 14 percent battery", "en"),856        ("Te veo @ la fiesta", "Te veo arroba la fiesta", "es"),857        ("J'ai 14° de fièvre", "J'ai 14 degrés de fièvre", "fr"),858        ("Die Rechnung beträgt £ 20", "Die Rechnung beträgt pfund 20", "de"),859        ("O meu email é ana&joao@gmail.com", "O meu email é ana e joao arroba gmail.com", "pt"),860        ("linguaggio di programmazione C#", "linguaggio di programmazione C cancelletto", "it"),861        ("Moja temperatura to 36.6°", "Moja temperatura to 36.6 stopnie", "pl"),862        ("Mám 14% baterie", "Mám 14 procento baterie", "cs"),863        ("Těším se na tebe @ party", "Těším se na tebe na party", "cs"),864        ("У меня 14% заряда", "У меня 14 процентов заряда", "ru"),865        ("Я буду @ дома", "Я буду собака дома", "ru"),866        ("Ik heb 14% batterij", "Ik heb 14 procent batterij", "nl"),867        ("Ik zie je @ het feest", "Ik zie je bij het feest", "nl"),868        ("لدي 14% في البطارية", "لدي 14 في المئة في البطارية", "ar"),869        ("我的电量为 14%", "我的电量为 14 百分之", "zh"),870        ("Pilim %14 dolu.", "Pilim yüzde 14 dolu.", "tr"),871        ("Az akkumulátorom töltöttsége 14%", "Az akkumulátorom töltöttsége 14 százalék", "hu"),872        ("배터리 잔량이 14%입니다.", "배터리 잔량이 14 퍼센트입니다.", "ko"),873    ]874 875    for a, b, lang in test_cases:876        out = expand_symbols_multilingual(a, lang=lang)877        assert out == b, f"'{out}' vs '{b}'"878 879 880if __name__ == "__main__":881    test_expand_numbers_multilingual()882    test_abbreviations_multilingual()883    test_symbols_multilingual()