@Blog{zip358.com}

#前処理 mecab-python3バージョンは1.0.8です
sudo apt-get install mecab libmecab-dev mecab-ipadic-utf8
sudo pip install mecab-python3
pip install unidic-lite

<?php
class sharpPost
{
    /**
     * mecab.pyを使って文章を解析（名詞と形容詞を取り出す）
     * @param $posstData
     * @return array|null
     */
    public function analysis($postData)
    {
        if(!$postData)return null;
        $word = null;
        exec('python py/mecab.py "'.strip_tags(htmlentities($postData)).'"',$output);// 2>&1
        if(is_array($output)){
            foreach($output as $val){
                $analysisWord = explode("\t",$val);
                if(isset($analysisWord[1]) && preg_match('/(名詞|形容詞)/',$analysisWord[1])){
                    $word[] = $analysisWord[0];
                    $word = array_unique($word);
                }
            }
        }
        return $word;
    }
    
    /**
     * 文字列を置き換える処理
     * @param $postData
     * @param $word|null
     * @return string
     */
    public function replacePostData($postData='',$word=null)
    {
        if(is_array($word)){
            foreach($word as $val){
                $postData = preg_replace("/({$val})/u"," #{$val} ",$postData);
            }
        }
        return $postData;
    }
}
$textData = '単なる自分が使いたい機能です、無かったので作ってみただけです。';
$sharpPost = new sharpPost();
$word = $sharpPost->analysis($textData);
print(($sharpPost->replacePostData($textData,$word)).PHP_EOL);

import MeCab
import sys
args = sys.argv
if(args[1]):
    tagger = MeCab.Tagger()
    print(tagger.parse(args[1]))

#!/usr/local/bin/python3
# coding:utf-8
import os
import sys
import MeCab
import gensim
import markovify
import unicodedata
model = gensim.models.KeyedVectors.load_word2vec_format('/var/www/html/model.vec', binary=False)
f = open('merosu.txt')
tagger = MeCab.Tagger("-Owakati")
tagger.parse('')
text0 = tagger.parse(f.read())
text1 = text0
text0 = text0.replace('\n','')
text0 = text0.replace('\r','')
text1x = text0.split(" ")
text2 = []
try:
    for item in text1x:
        if item.strip():
            results  = model.most_similar(positive=[item],topn=2)
            #"print(results)
            for val1 in results:
                text2.append(val1[0] + "\n")
#
    # print (text1)
    # print (" ".join(text2))
    model_a = markovify.Text(text1 + "\n")
    print(str(model_a.make_sentence()).replace(' ',''))
    model_b = markovify.Text(" ".join(text2))
    print(str(model_b.make_sentence()).replace(' ',''))
    model_combo = markovify.combine([model_a, model_b], [1, 1])
    print(str(model_combo.make_sentence()).replace(' ',''))
except Exception as e:
    print("動作エラー", e.args)
    pass