zh-mistake-text-gen 0.2.1

中文錯誤類型文字增量

錯誤類型中文語料生成

安裝

pip install zh-mistake-text-gen

使用 (Pipeline)

from zh_mistake_text_gen import Pipeline
import random

random.seed(7)
pipeline = Pipeline()
augs = pipeline("中文語料生成",k=8)
for aug in augs:
    print(aug)

type='MissingWordMaker' correct='中文語料生成' incorrect='中文料生成' incorrect_start_at=2 incorrect_end_at=2 span='語'
type='MissingVocabMaker' correct='中文語料生成' incorrect='語料生成' incorrect_start_at=0 incorrect_end_at=2 span='中文'
type='PronounceSimilarWordMaker' correct='中文語料生成' incorrect='中文語尥生成' incorrect_start_at=3 incorrect_end_at=3 span='尥'
type='PronounceSameWordMaker' correct='中文語料生成' incorrect='諥文語料生成' incorrect_start_at=0 incorrect_end_at=0 span='諥'
type='PronounceSimilarVocabMaker' correct='中文語料生成' incorrect='鍾文語料生成' incorrect_start_at=0 incorrect_end_at=2 span='鍾文'
type='PronounceSameVocabMaker' correct='中文語料生成' incorrect='中文预料生成' incorrect_start_at=2 incorrect_end_at=4 span='预料'
type='RedundantWordMaker' correct='中文語料生成' incorrect='成中文語料生成' incorrect_start_at=0 incorrect_end_at=0 span='成'
type='MistakWordMaker' correct='中文語料生成' incorrect='谁文語料生成' incorrect_start_at=0 incorrect_end_at=0 span='谁'

文檔

Pipeline

• __init__

  • makers = None : maker實例，可選
  • maker_weight = None : maker被抽中的機率，可選

• __call__

  • x : 輸入句(str)，必需
  • k=1 : 期待最多返回多少結果，可選
  • verbose=True : debug 訊息，可選

可用方法

from zh_mistake_text_gen.data_maker import *

Data Maker	Description
MissingWordMaker	隨機缺字
MissingVocabMaker	隨機缺詞
PronounceSimilarWordMaker	隨機相似字替換
PronounceSimilarWordPlusMaker	編輯距離找發音相似並且用高頻字替換
PronounceSimilarVocabMaker	發音相似詞替換
PronounceSameWordMaker	發音相同字替換
PronounceSameVocabMaker	發音相同詞替換
RedundantWordMaker	隨機複製旁邊一個字作為沆於字
RandomInsertVacabMaker	隨機插入詞彙
MistakWordMaker	隨機替換字
MistakeWordHighFreqMaker	隨機替換高頻字
MissingWordHighFreqMaker	隨機刪除高頻字