wip

src/tools/ai/categorizer.ts

@@ -1,93 +0,0 @@
-import * as fs from 'fs';
-
-const bayes = require('./naive-bayes.js');
-const MeCab = require('mecab-async');
-import * as msgpack from 'msgpack-lite';
-
-import Post from '../../api/models/post';
-import config from '../../conf';
-
-/**
- * 投稿を学習したり与えられた投稿のカテゴリを予測します
- */
-export default class Categorizer {
-	private classifier: any;
-	private categorizerDbFilePath: string;
-	private mecab: any;
-
-	constructor() {
-		this.categorizerDbFilePath = `${__dirname}/../../../data/category`;
-
-		this.mecab = new MeCab();
-		if (config.categorizer.mecab_command) this.mecab.command = config.categorizer.mecab_command;
-
-		// BIND -----------------------------------
-		this.tokenizer = this.tokenizer.bind(this);
-	}
-
-	private tokenizer(text: string) {
-		return this.mecab.wakachiSync(text);
-	}
-
-	public async init() {
-		try {
-			const buffer = fs.readFileSync(this.categorizerDbFilePath);
-			const db = msgpack.decode(buffer);
-
-			this.classifier = bayes.import(db);
-			this.classifier.tokenizer = this.tokenizer;
-		} catch (e) {
-			this.classifier = bayes({
-				tokenizer: this.tokenizer
-			});
-
-			// 訓練データ
-			const verifiedPosts = await Post.find({
-				is_category_verified: true
-			});
-
-			// 学習
-			verifiedPosts.forEach(post => {
-				this.classifier.learn(post.text, post.category);
-			});
-
-			this.save();
-		}
-	}
-
-	public async learn(id, category) {
-		const post = await Post.findOne({ _id: id });
-
-		Post.update({ _id: id }, {
-			$set: {
-				category: category,
-				is_category_verified: true
-			}
-		});
-
-		this.classifier.learn(post.text, category);
-
-		this.save();
-	}
-
-	public async categorize(id) {
-		const post = await Post.findOne({ _id: id });
-
-		const category = this.classifier.categorize(post.text);
-
-		Post.update({ _id: id }, {
-			$set: {
-				category: category
-			}
-		});
-	}
-
-	public async test(text) {
-		return this.classifier.categorize(text);
-	}
-
-	private save() {
-		const buffer = msgpack.encode(this.classifier.export());
-		fs.writeFileSync(this.categorizerDbFilePath, buffer);
-	}
-}

src/tools/ai/predict-all-post-category.ts

@@ -0,0 +1,57 @@
+const bayes = require('./naive-bayes.js');
+const MeCab = require('mecab-async');
+
+import Post from '../../api/models/post';
+import config from '../../conf';
+
+const classifier = bayes({
+	tokenizer: this.tokenizer
+});
+
+const mecab = new MeCab();
+if (config.categorizer.mecab_command) mecab.command = config.categorizer.mecab_command;
+
+// 訓練データ取得
+Post.find({
+	is_category_verified: true
+}, {
+	fields: {
+		_id: false,
+		text: true,
+		category: true
+	}
+}).then(verifiedPosts => {
+	// 学習
+	verifiedPosts.forEach(post => {
+		classifier.learn(post.text, post.category);
+	});
+
+	// 全ての(人間によって証明されていない)投稿を取得
+	Post.find({
+		text: {
+			$exists: true
+		},
+		is_category_verified: {
+			$ne: true
+		}
+	}, {
+		sort: {
+			_id: -1
+		},
+		fields: {
+			_id: true,
+			text: true
+		}
+	}).then(posts => {
+		posts.forEach(post => {
+			console.log(`predicting... ${post._id}`);
+			const category = classifier.categorize(post.text);
+
+			Post.update({ _id: post._id }, {
+				$set: {
+					category: category
+				}
+			});
+		});
+	});
+});

src/tools/ai/predict-user-interst.ts

@@ -0,0 +1,45 @@
+import Post from '../../api/models/post';
+import User from '../../api/models/user';
+
+export async function predictOne(id) {
+	console.log(`predict interest of ${id} ...`);
+
+	// TODO: repostなども含める
+	const recentPosts = await Post.find({
+		user_id: id,
+		category: {
+			$exists: true
+		}
+	}, {
+		sort: {
+			_id: -1
+		},
+		limit: 1000,
+		fields: {
+			_id: false,
+			category: true
+		}
+	});
+
+	const categories = {};
+
+	recentPosts.forEach(post => {
+		if (categories[post.category]) {
+			categories[post.category]++;
+		} else {
+			categories[post.category] = 1;
+		}
+	});
+}
+
+export async function predictAll() {
+	const allUsers = await User.find({}, {
+		fields: {
+			_id: true
+		}
+	});
+
+	allUsers.forEach(user => {
+		predictOne(user._id);
+	});
+}