o 5ti#@sddlmZmZddlmZmZmZmZmZm Z ddddej dfdee deee d efd d Z ej d fdee deee d eed efddZdddej dfde dee de d eededed efddZejejejddfdee deee dededededed efddZejejejddfde dee dededededed efddZ    d(dee deee d ed!ed"ed#ed e fd$d%Z    d(de dee d ed!ed"ed#ed e fd&d'ZdS)))SequenceOptional)BLEUCHRFTER BLEUScore CHRFScoreTERScoreexpNF hypotheses referencesreturnc Cs t||||||d}|||S)aJComputes BLEU for a corpus against a single (or multiple) reference(s). This is the main CLI entry point for computing BLEU between a system output and a reference sentence. :param hypotheses: A sequence of hypothesis strings. :param references: A sequence of reference documents with document being defined as a sequence of reference strings. :param smooth_method: The smoothing method to use ('floor', 'add-k', 'exp' or 'none') :param smooth_value: The smoothing value for `floor` and `add-k` methods. `None` falls back to default value. :param force: Ignore data that looks already tokenized :param lowercase: Lowercase the data :param tokenize: The tokenizer to use :param use_effective_order: Don't take into account n-gram orders without any match. :return: a `BLEUScore` object ) lowercaseforcetokenize smooth_method smooth_valueeffective_order)r corpus_score) r r rrrrruse_effective_ordermetricrD/home/ubuntu/.local/lib/python3.10/site-packages/sacrebleu/compat.py corpus_bleu s  rfloorrc Cst||d|ddddS)aComputes BLEU for a corpus against a single (or multiple) reference(s). This convenience function assumes a particular set of arguments i.e. it disables tokenization and applies a `floor` smoothing with value `0.1`. This convenience call does not apply any tokenization at all, neither to the system output nor the reference. It just computes BLEU on the "raw corpus" (hence the name). :param hypotheses: A sequence of hypothesis strings. :param references: A sequence of reference documents with document being defined as a sequence of reference strings. :param smooth_value: The smoothing value for `floor`. If not given, the default of 0.1 is used. :return: Returns a `BLEUScore` object. rTnone)rrrrr)r)r r rrrrraw_corpus_bleu(s rT hypothesisrrrcCs t||d|||d}|||S)a Computes BLEU for a single sentence against a single (or multiple) reference(s). Disclaimer: Computing BLEU at the sentence level is not its intended use as BLEU is a corpus-level metric. :param hypothesis: A single hypothesis string. :param references: A sequence of reference strings. :param smooth_method: The smoothing method to use ('floor', 'add-k', 'exp' or 'none') :param smooth_value: The smoothing value for `floor` and `add-k` methods. `None` falls back to default value. :param lowercase: Lowercase the data :param tokenize: The tokenizer to use :param use_effective_order: Don't take into account n-gram orders without any match. :return: Returns a `BLEUScore` object. F)rrrrrr)rsentence_score)rr rrrrrrrrr sentence_bleu@s  r char_order word_orderbetaremove_whitespace eps_smoothingcC t|||| |d}|||S)a Computes chrF for a corpus against a single (or multiple) reference(s). If `word_order` equals to 2, the metric is referred to as chrF++. :param hypotheses: A sequence of hypothesis strings. :param references: A sequence of reference documents with document being defined as a sequence of reference strings. :param char_order: Character n-gram order. :param word_order: Word n-gram order. If equals to 2, the metric is referred to as chrF++. :param beta: Determine the importance of recall w.r.t precision. :param eps_smoothing: If `True`, applies epsilon smoothing similar to reference chrF++.py, NLTK and Moses implementations. Otherwise, it takes into account effective match order similar to sacreBLEU < 2.0.0. :param remove_whitespace: If `True`, removes whitespaces prior to character n-gram extraction. :return: A `CHRFScore` object. r!r"r# whitespacer%)rr)r r r!r"r#r$r%rrrr corpus_chrf^s r)cCr&)aA Computes chrF for a single sentence against a single (or multiple) reference(s). If `word_order` equals to 2, the metric is referred to as chrF++. :param hypothesis: A single hypothesis string. :param references: A sequence of reference strings. :param char_order: Character n-gram order. :param word_order: Word n-gram order. If equals to 2, the metric is referred to as chrF++. :param beta: Determine the importance of recall w.r.t precision. :param eps_smoothing: If `True`, applies epsilon smoothing similar to reference chrF++.py, NLTK and Moses implementations. Otherwise, it takes into account effective match order similar to sacreBLEU < 2.0.0. :param remove_whitespace: If `True`, removes whitespaces prior to character n-gram extraction. :return: A `CHRFScore` object. r')rr)rr r!r"r#r$r%rrrr sentence_chrf~s r* normalizedno_punct asian_supportcase_sensitivecCt||||d}|||S)a  Computes TER for a corpus against a single (or multiple) reference(s). :param hypotheses: A sequence of hypothesis strings. :param references: A sequence of reference documents with document being defined as a sequence of reference strings. :param normalized: Enable character normalization. :param no_punct: Remove punctuation. :param asian_support: Enable special treatment of Asian characters. :param case_sensitive: Enables case-sensitivity. :return: A `TERScore` object. r+r,r-r.)rr)r r r+r,r-r.rrrr corpus_ters r1cCr/)a Computes TER for a single hypothesis against a single (or multiple) reference(s). :param hypothesis: A single hypothesis string. :param references: A sequence of reference strings. :param normalized: Enable character normalization. :param no_punct: Remove punctuation. :param asian_support: Enable special treatment of Asian characters. :param case_sensitive: Enable case-sensitivity. :return: A `TERScore` object. r0)rr)rr r+r,r-r.rrrr sentence_ters r2)FFFF)typingrrmetricsrrrrr r TOKENIZER_DEFAULTstrrSMOOTH_DEFAULTSfloatrboolr CHAR_ORDER WORD_ORDERBETAintr)r*r1r2rrrrs   !      " !