3 d@sdZddlZddlmZmZddlmZddlmZddlm Z ddl m Z ddl m Z Gd d d ed Zd d ZddZdddZGddded ZdS)zLanguage Model Interface.N)ABCMetaabstractmethod)bisect) accumulate) NgramCounter) log_base2) Vocabularyc@s0eZdZdZddZeddZeddZdS) SmoothingzNgram Smoothing Interface Implements Chen & Goodman 1995's idea that all smoothing algorithms have certain features in common. This should ideally allow smoothing algorithms to work both with Backoff and Interpolation. cCs||_||_dS)z :param vocabulary: The Ngram vocabulary object. :type vocabulary: nltk.lm.vocab.Vocabulary :param counter: The counts of the vocabulary items. :type counter: nltk.lm.counter.NgramCounter N)vocabcounts)self vocabularycounterr+/tmp/pip-build-v9q4h5k9/nltk/nltk/lm/api.py__init__szSmoothing.__init__cCs tdS)N)NotImplementedError)r wordrrr unigram_score%szSmoothing.unigram_scorecCs tdS)N)r)r rcontextrrr alpha_gamma)szSmoothing.alpha_gammaN)__name__ __module__ __qualname____doc__rrrrrrrrr s r ) metaclasscCst|t|S)z0Return average (aka mean) for sequence of items.)sumlen)itemsrrr_mean.srcCst|tjr|Stj|S)N) isinstancerandomRandom)Zseed_or_generatorrrr_random_generator3s r#cCsR|s tdt|t|kr$tdtt|}|d}|j}|t|||S)z`Like random.choice, but with weights. Heavily inspired by python 3.6 `random.choices`. z"Can't choose from empty populationz3The number of weights does not match the population) ValueErrorrlistrr!r) populationweightsrandom_generator cum_weightstotal thresholdrrr_weighted_choice9s r.c@sheZdZdZdddZdddZdddZedd d Zdd d Z d dZ ddZ ddZ dddZ dS) LanguageModelzKABC for Language Models. Cannot be directly instantiated itself. NcCs2||_|dkrtn||_|dkr(tn||_dS)asCreates new LanguageModel. :param vocabulary: If provided, this vocabulary will be used instead of creating a new one when training. :type vocabulary: `nltk.lm.Vocabulary` or None :param counter: If provided, use this object to count ngrams. :type vocabulary: `nltk.lm.NgramCounter` or None :param ngrams_fn: If given, defines how sentences in training text are turned to ngram sequences. :type ngrams_fn: function or None :param pad_fn: If given, defines how sentences in training text are padded. :type pad_fn: function or None N)orderrr rr )r r0r rrrrrOszLanguageModel.__init__cs@js"|dkrtdjj|jjfdd|DdS)zeTrains the model on a text. :param text: Training text as a sequence of sentences. Nz:Cannot fit without a vocabulary or text to create it from.c3s|]}jj|VqdS)N)r lookup).0sent)r rr msz$LanguageModel.fit..)r r&updater )r textZvocabulary_textr)r rfitas  zLanguageModel.fitcCs$|j|jj||r|jj|ndS)zMasks out of vocab (OOV) words and computes their model score. For model-specific logic of calculating scores, see the `unmasked_score` method. N)unmasked_scorer r1)r rrrrrscoreoszLanguageModel.scorecCs tdS)aScore a word given some optional context. Concrete models are expected to provide an implementation. Note that this method does not mask its arguments with the OOV label. Use the `score` method for that. :param str word: Word for which we want the score :param tuple(str) context: Context the word is in. If `None`, compute unigram score. :param context: tuple(str) or None :rtype: float N)r)r rrrrrr8yszLanguageModel.unmasked_scorecCst|j||S)zEvaluate the log score of this word in this context. The arguments are the same as for `score` and `unmasked_score`. )rr9)r rrrrrlogscoreszLanguageModel.logscorecCs"|r|jt|d|S|jjS)zHelper method for retrieving counts for a given context. Assumes context has been checked and oov words in it masked. :type context: tuple(str) or None r$)r rZunigrams)r rrrrcontext_countsszLanguageModel.context_countscsdtfdd|DS)zCalculate cross-entropy of model for given evaluation text. :param Iterable(tuple(str)) text_ngrams: A sequence of ngram tuples. :rtype: float r$cs$g|]}j|d|ddqS)r$Nr%r%)r:)r2Zngram)r rr sz)LanguageModel.entropy..r%)r)r text_ngramsr)r rentropyszLanguageModel.entropycCstd|j|S)zCalculates the perplexity of the given text. This is simply 2 ** cross-entropy for the text, so the arguments are the same. g@)powr>)r r=rrr perplexityszLanguageModel.perplexityr$cs|dkr gnt|}t|}|dkrt|jkrF|j ddn|jjj}x<r| rtdkrddngjjj}q^Wt|}t|t fdd|D|Sg}x*t |D]}|j j d|||dqW|S)aGenerate words from the model. :param int num_words: How many words to generate. By default 1. :param text_seed: Generation can be conditioned on preceding context. :param random_seed: A random seed or an instance of `random.Random`. If provided, makes the random sampling part of generation reproducible. :return: One (str) word or a list of words generated from model. Examples: >>> from nltk.lm import MLE >>> lm = MLE(2) >>> lm.fit([[("a", "b"), ("b", "c")]], vocabulary_text=['a', 'b', 'c']) >>> lm.fit([[("a",), ("b",), ("c",)]]) >>> lm.generate(random_seed=3) 'a' >>> lm.generate(text_seed=['a']) 'b' Nr$c3s|]}j|VqdS)N)r9)r2w)rr rrr4sz)LanguageModel.generate..) num_words text_seed random_seed) r'r#rr0r;r r1sortedr.tuplerangeappendgenerate)r rBrCrDr*Zsamples generated_r)rr rrIs," zLanguageModel.generate)NN)N)N)N)N)r$NN)rrrrrr7r9rr8r:r;r>r@rIrrrrr/Hs       r/)N)rr!abcrrr itertoolsrZnltk.lm.counterrZ nltk.lm.utilrZnltk.lm.vocabularyrr rr#r.r/rrrrs