A little regular expression story Part 1

prologue

Once upon a time, there were Yura, Moya-san, and Yura's entourage in a place called Twitter. Whenever Yura does something, the entourage says:

• "Yura-chan is cute" * One day, Ms. Moya said:
• "Let's discipline those who say'Yura-chan is cute' Yura is mine. * Then he tried to discipline his entourage ... (I'm not saying he's doing it)

Main subject

I will do the first half of the last word of Mr. Moya with Java regular expressions. Reference: http://www.javadrive.jp/regex/ I won't touch regular expressions in other languages.

Introduction

First, let's divide "Yura-chan cute" (A) into two parts. Let's say "Yura-chan" (hereinafter B) and "cute" (hereinafter C).

Complex Japanese

Japanese is complicated. There are some kana characters such as "a" and "a" that are the same. Let's make B and C correspond to this, assuming that it does not correspond to the same reading kanji. You can do this by using this. B = [Yuyu] [Lalala] [Chichichi] [Yayaya] [Nun] C = [Kakaka] [Wow Wow] [Iii] [Iii] For C, "i" is always two consecutive, so use this to shorten it. C = [Kakaka] [Wow Wow] [Iii] {2} This is fine.

Inversion of words

Regular expressions can be synthesized by concatenating the context as a character string unless the beginning and end are specified. For example, if it is B + C, 100% of the" Erachan Kawaii "will be caught. But what if you say this? "** Cute ** Yo ** Yura-chan **" It will be difficult if it is inverted. Here, use this and this Let's define a new D. D=.* D will be caught in a string of 0 or more characters. (That is, it will get caught even if it is not) If you use this, you can handle the previous inversion with C + D + B. Even if "Yura-chan is cute", you can handle it with B + D + C.

Regular expression composition

Finally, let's combine multiple regular expressions into one. When I thought about it, a person who said this time appeared.

• "Swaying cute" * Let's deal with this as well. Since "Yura" itself has already created a regular expression ([Yuyu] [Lalala]), combine this with B. So what do you do? Make sure that the "chan" part of B is also caught by "yura". Here, use Candidate to synthesize. B = [Yuyu] [Lalala]([Chi-Chi] [Yayaya] [Nun] | [Yuyu] [Lalala]) It's done.

Let's get back to the main subject. Combine using the above method so that the regular expression can search for B + D + C and C + D + B at the same time. It will be long, but it will be like this. (Hereafter E) E=([Yuyu][Lala]([ChiChiChi][Yayaya][Hmmmm]|[Yuyu][Lala]).*[Kakaka][WowWow][Iiii]{2}|[Kakaka][WowWow][Iiii]{2}.*[Yuyu][Lala]([ChiChiChi][Yayaya][Hmmmm]|[Yuyu][Lala])) looks strong You can probably find most of "Yura-chan cute" with this. Perhaps.

People who cooperated with the name

Moya: Twitter Yura: Twitter Thank you very much.

Finally

Yura is cute ~~ I'm sorry I'm not good at writing ~~ For the second part, go to here