Rewrite how tensorflow works

Xiao.js

@@ -235,14 +235,6 @@ client.on('ready', async () => {
 		client.logger.error(`[TIMEZONES] Failed to set timezones\n${err.stack}`);
 	}
 
-	// Set up parse-domain
-	try {
-		await client.loadParseDomain();
-		client.logger.info('[PARSE DOMAIN] parse-domain loaded.');
-	} catch (err) {
-		client.logger.error(`[PARSE DOMAIN] Failed to load parse-domain\n${err.stack}`);
-	}
-
 	// Fetch adult site list
 	try {
 		await client.fetchAdultSiteList();
@@ -253,7 +245,7 @@ client.on('ready', async () => {
 
 	// Fetch NSFW model
 	try {
-		await client.loadNSFWModel();
+		await client.tensorflow.loadNSFWModel();
 		client.logger.info('[NSFW MODEL] Loaded NSFW model.');
 	} catch (err) {
 		client.logger.error(`[NSFW MODEL] Failed to load NSFW model\n${err.stack}`);
@@ -261,7 +253,7 @@ client.on('ready', async () => {
 
 	// Set up face detection
 	try {
-		await client.loadFaceDetector();
+		await client.tensorflow.loadFaceDetector();
 		client.logger.info('[FACE DETECTOR] Loaded face detector.');
 	} catch (err) {
 		client.logger.error(`[FACE DETECTOR] Failed to load face detector\n${err.stack}`);

commands/analyze/faces.js

@@ -25,7 +25,7 @@ module.exports = class FacesCommand extends Command {
 
 	async run(msg, { image }) {
 		const imgData = await request.get(image);
-		const faces = await this.client.detectFaces(imgData.body);
+		const faces = await this.client.tensorflow.detectFaces(imgData.body);
 		if (!faces) return msg.reply('There are no faces in this image.');
 		if (faces === 'size') return msg.reply('This image is too large.');
 		const base = await loadImage(imgData.body);

commands/analyze/is-it-down.js

@@ -1,5 +1,11 @@
 const Command = require('../../framework/Command');
 const request = require('node-superfetch');
+let parseDomain;
+let ParseResultType;
+import('parse-domain').then(loadedModule => {
+	parseDomain = loadedModule.parseDomain;
+	ParseResultType = loadedModule.ParseResultType;
+});
 
 module.exports = class IsItDownCommand extends Command {
 	constructor(client) {
@@ -26,8 +32,8 @@ module.exports = class IsItDownCommand extends Command {
 	}
 
 	async run(msg, { url }) {
-		const { type, domain, topLevelDomains } = this.client.parseDomain(url.hostname);
+		const { type, domain, topLevelDomains } = parseDomain(url.hostname);
-		if (type !== this.client.ParseResultType.Listed) return msg.reply('This domain is not supported.');
+		if (type !== ParseResultType.Listed) return msg.reply('This domain is not supported.');
 		const { text } = await request
 			.post('https://www.isitdownrightnow.com/check.php')
 			.query({ domain: `${domain}.${topLevelDomains.join('.')}` });

commands/analyze/nsfw-image.js

@@ -1,7 +1,6 @@
 const Command = require('../../framework/Command');
 const request = require('node-superfetch');
 const { stripIndents } = require('common-tags');
-const { isImageNSFW } = require('../../util/Util');
 const displayNames = {
 	Drawing: 'SFW (Drawing)',
 	Neutral: 'SFW',
@@ -33,7 +32,7 @@ module.exports = class NsfwImageCommand extends Command {
 
 	async run(msg, { image }) {
 		const { body } = await request.get(image);
-		const predictions = await isImageNSFW(this.client.nsfwModel, body, false);
+		const predictions = await this.client.tensorflow.isImageNSFW(body, false);
 		const formatted = predictions.map(result => {
 			const percentage = Math.round(result.probability * 100);
 			return `${percentage}% ${displayNames[result.className]}`;

commands/analyze/screenshot.js

@@ -1,7 +1,7 @@
 const Command = require('../../framework/Command');
 const { PermissionFlagsBits } = require('discord.js');
 const request = require('node-superfetch');
-const { isImageNSFW, isUrlNSFW } = require('../../util/Util');
+const { isUrlNSFW } = require('../../util/Util');
 
 module.exports = class ScreenshotCommand extends Command {
 	constructor(client) {
@@ -41,7 +41,7 @@ module.exports = class ScreenshotCommand extends Command {
 			}
 			const { body } = await request.get(`https://image.thum.io/get/width/1920/crop/675/noanimate/${url.href}`);
 			if (!msg.channel.nsfw) {
-				const aiDetect = await isImageNSFW(this.client.nsfwModel, body);
+				const aiDetect = await this.client.tensorflow.isImageNSFW(body);
 				if (aiDetect) return msg.reply('This site isn\'t NSFW, but the resulting image was.');
 			}
 			return msg.say({ files: [{ attachment: body, name: 'screenshot.png' }] });

commands/edit-face/anime-eyes.js

@@ -28,7 +28,7 @@ module.exports = class AnimeEyesCommand extends Command {
 		const leftEye = await loadImage(path.join(__dirname, '..', '..', 'assets', 'images', 'anime-eyes', 'left.png'));
 		const rightEye = await loadImage(path.join(__dirname, '..', '..', 'assets', 'images', 'anime-eyes', 'right.png'));
 		const imgData = await request.get(image);
-		const faces = await this.client.detectFaces(imgData.body);
+		const faces = await this.client.tensorflow.detectFaces(imgData.body);
 		if (!faces) return msg.reply('There are no faces in this image.');
 		if (faces === 'size') return msg.reply('This image is too large.');
 		const base = await loadImage(imgData.body);

commands/edit-face/danny-devito.js

@@ -34,7 +34,7 @@ module.exports = class DannyDevitoCommand extends Command {
 	async run(msg, { image }) {
 		const danny = await loadImage(path.join(__dirname, '..', '..', 'assets', 'images', 'danny-devito.png'));
 		const imgData = await request.get(image);
-		const faces = await this.client.detectFaces(imgData.body);
+		const faces = await this.client.tensorflow.detectFaces(imgData.body);
 		if (!faces) return msg.reply('There are no faces in this image.');
 		if (faces === 'size') return msg.reply('This image is too large.');
 		const base = await loadImage(imgData.body);

commands/edit-face/emoji-face.js

@@ -41,7 +41,7 @@ module.exports = class EmojiFaceCommand extends Command {
 		const emojiData = await request.get(emojiURL);
 		const emojiImg = await loadImage(emojiData.body);
 		const imgData = await request.get(image);
-		const faces = await this.client.detectFaces(imgData.body);
+		const faces = await this.client.tensorflow.detectFaces(imgData.body);
 		if (!faces) return msg.reply('There are no faces in this image.');
 		if (faces === 'size') return msg.reply('This image is too large.');
 		const base = await loadImage(imgData.body);

commands/edit-face/eyes.js

@@ -26,7 +26,7 @@ module.exports = class EyesCommand extends Command {
 	async run(msg, { image }) {
 		const eyes = await loadImage(path.join(__dirname, '..', '..', 'assets', 'images', 'eyes.png'));
 		const imgData = await request.get(image);
-		const faces = await this.client.detectFaces(imgData.body);
+		const faces = await this.client.tensorflow.detectFaces(imgData.body);
 		if (!faces) return msg.reply('There are no faces in this image.');
 		if (faces === 'size') return msg.reply('This image is too large.');
 		const base = await loadImage(imgData.body);

commands/edit-face/shrek.js

@@ -34,7 +34,7 @@ module.exports = class ShrekCommand extends Command {
 	async run(msg, { image }) {
 		const shrek = await loadImage(path.join(__dirname, '..', '..', 'assets', 'images', 'shrek.png'));
 		const imgData = await request.get(image);
-		const faces = await this.client.detectFaces(imgData.body);
+		const faces = await this.client.tensorflow.detectFaces(imgData.body);
 		if (!faces) return msg.reply('There are no faces in this image.');
 		if (faces === 'size') return msg.reply('This image is too large.');
 		const base = await loadImage(imgData.body);

package.json

@@ -1,6 +1,6 @@
 {
 	"name": "xiao",
-	"version": "147.0.1",
+	"version": "148.0.0",
 	"description": "Your personal server companion.",
 	"main": "Xiao.js",
 	"private": true,

structures/Client.js

@@ -3,20 +3,16 @@ const request = require('node-superfetch');
 const { Collection } = require('@discordjs/collection');
 const winston = require('winston');
 const fontFinder = require('font-finder');
-const nsfw = require('nsfwjs');
-const tfnode = require('@tensorflow/tfjs-node');
-const faceDetection = require('@tensorflow-models/face-detection');
-const model = faceDetection.SupportedModels.MediaPipeFaceDetector;
 const moment = require('moment-timezone');
 const fs = require('fs');
-const url = require('url');
 const path = require('path');
 const Redis = require('./Redis');
+const Tensorflow = require('./Tensorflow');
 const Font = require('./Font');
 const PhoneManager = require('./phone/PhoneManager');
 const TimerManager = require('./remind/TimerManager');
 const PokemonStore = require('./pokemon/PokemonStore');
-const activities = require('./activity');
+const activities = require('./Activity');
 const { REPORT_CHANNEL_ID, JOIN_LEAVE_CHANNEL_ID } = process.env;
 
 module.exports = class XiaoClient extends CommandClient {
@@ -37,17 +33,9 @@ module.exports = class XiaoClient extends CommandClient {
 		this.dispatchers = new Map();
 		this.cleverbots = new Map();
 		this.phone = new PhoneManager(this);
+		this.tensorflow = new Tensorflow(this);
 		this.activities = activities;
 		this.adultSiteList = null;
-		this.nsfwModel = null;
-		this.faceDetector = null;
-	}
-
-	async loadParseDomain() {
-		const parseDomainModule = await import('parse-domain');
-		this.parseDomain = parseDomainModule.parseDomain;
-		this.ParseResultType = parseDomainModule.ParseResultType;
-		return parseDomainModule;
 	}
 
 	async registerFontsIn(filepath) {
@@ -76,31 +64,6 @@ module.exports = class XiaoClient extends CommandClient {
 		return this.adultSiteList;
 	}
 
-	async loadNSFWModel() {
-		const nsfwModel = await nsfw.load(
-			`${url.pathToFileURL(path.join(__dirname, '..', 'tf_models', 'nsfw', 'web_model')).href}/`,
-			{ type: 'graph' }
-		);
-		this.nsfwModel = nsfwModel;
-		return this.nsfwModel;
-	}
-
-	async loadFaceDetector() {
-		this.faceDetector = await faceDetection.createDetector(model, { runtime: 'tfjs', maxFaces: 10 });
-		return this.faceDetector;
-	}
-
-	async detectFaces(imgData) {
-		if (Buffer.byteLength(imgData) >= 4e+6) return 'size';
-		tfnode.setBackend('tensorflow');
-		const image = tfnode.node.decodeImage(imgData, 3);
-		tfnode.setBackend('cpu');
-		const faces = await this.faceDetector.estimateFaces(image);
-		tfnode.setBackend('tensorflow');
-		if (!faces || !faces.length) return null;
-		return faces;
-	}
-
 	fetchReportChannel() {
 		if (!REPORT_CHANNEL_ID) return null;
 		return this.channels.fetch(REPORT_CHANNEL_ID);

structures/Tensorflow.js

@@ -0,0 +1,58 @@
+const tfnode = require('@tensorflow/tfjs-node');
+const nsfw = require('nsfwjs');
+const faceDetection = require('@tensorflow-models/face-detection');
+const url = require('url');
+const path = require('path');
+
+module.exports = class Tensorflow {
+	constructor(client) {
+		Object.defineProperty(this, 'client', { value: client });
+
+		this.nsfwModel = null;
+		this.faceModel = faceDetection.SupportedModels.MediaPipeFaceDetector;
+		this.faceDetector = null;
+	}
+
+	async loadNSFWModel() {
+		const nsfwModel = await nsfw.load(
+			`${url.pathToFileURL(path.join(__dirname, '..', 'tf_models', 'nsfw', 'web_model')).href}/`,
+			{ type: 'graph' }
+		);
+		this.nsfwModel = nsfwModel;
+		return this.nsfwModel;
+	}
+
+	async loadFaceDetector() {
+		const faceDetector = await faceDetection.createDetector(this.faceModel, { runtime: 'tfjs', maxFaces: 10 });
+		this.faceDetector = faceDetector;
+		return this.faceDetector;
+	}
+
+	async detectFaces(imgData) {
+		if (Buffer.byteLength(imgData) >= 4e+6) return 'size';
+		tfnode.setBackend('tensorflow');
+		const image = tfnode.node.decodeImage(imgData, 3);
+		tfnode.setBackend('cpu');
+		const faces = await this.faceDetector.estimateFaces(image);
+		tfnode.setBackend('tensorflow');
+		image.dispose();
+		if (!faces || !faces.length) return null;
+		return faces;
+	}
+
+	async isImageNSFW(image, bool = true) {
+		const img = await tfnode.node.decodeImage(image, 3);
+		const predictions = await this.nsfwModel.classify(img);
+		img.dispose();
+		if (bool) {
+			const results = [];
+			results.push(predictions[0]);
+			for (const result of predictions) {
+				if (result.className === predictions[0].className) continue;
+				if (result.probability >= predictions[0].probability - 0.1) results.push(result);
+			}
+			return results.some(result => result.className !== 'Drawing' && result.className !== 'Neutral');
+		}
+		return predictions;
+	}
+};

util/Util.js

@@ -2,7 +2,6 @@ const { ActionRowBuilder, ButtonBuilder, PermissionFlagsBits, ButtonStyle, Compo
 const crypto = require('crypto');
 const request = require('node-superfetch');
 const fs = require('fs');
-const tf = require('@tensorflow/tfjs-node');
 let parseDomain;
 let ParseResultType;
 import('parse-domain').then(loadedModule => {
@@ -244,22 +243,6 @@ module.exports = class Util {
 		return str;
 	}
 
-	static async isImageNSFW(model, image, bool = true) {
-		const img = await tf.node.decodeImage(image, 3);
-		const predictions = await model.classify(img);
-		img.dispose();
-		if (bool) {
-			const results = [];
-			results.push(predictions[0]);
-			for (const result of predictions) {
-				if (result.className === predictions[0].className) continue;
-				if (result.probability >= predictions[0].probability - 0.1) results.push(result);
-			}
-			return results.some(result => result.className !== 'Drawing' && result.className !== 'Neutral');
-		}
-		return predictions;
-	}
-
 	static async reactIfAble(msg, user, emoji, fallbackEmoji) {
 		const dm = !msg.guild;
 		if (!emoji) emoji = fallbackEmoji;