Python build_mitton_datasets示例

示例#1

文件： build_datasets.py 项目： ndronen/spelling

raise ValueError("make this a spelling.job")

from spelling.mitton import CORPORA, build_mitton_datasets
import cPickle
import spelling.dictionary

constructors = [
    spelling.dictionary.Aspell,
    spelling.dictionary.AspellWithNorvigLanguageModel,
    spelling.dictionary.NorvigWithAspellVocabWithoutLanguageModel,
    spelling.dictionary.NorvigWithAspellVocabAndGoogleLanguageModel,
    spelling.dictionary.NorvigWithAspellVocabGoogleLanguageModelPhoneticCandidates,
]

corpus_names = [c.replace("data/", "") for c in CORPORA]
corpus_names = [c.replace(".dat", "") for c in corpus_names]

for i, corpus in enumerate(CORPORA):
    corpus_name = corpus_names[i]
    print(corpus_name)
    dfs = build_mitton_datasets(corpus, constructors=constructors)
    cPickle.dump(dfs, open("data/" + corpus_name + ".pkl", "w"))

示例#2

文件： compare_aspell_and_norvig.py 项目： ndronen/spelling

def build_datasets(constructors=CONSTRUCTORS):
    dfs = build_mitton_datasets('data/birbeck.dat',
        constructors=CONSTRUCTORS)