Init

README.md

@@ -0,0 +1,3 @@
+# tasbot
+
+You can do it!

bot/index.ts

@@ -0,0 +1,181 @@
+import { 
+    MatrixClient,
+    SimpleFsStorageProvider,
+    AutojoinRoomsMixin,
+} from "matrix-bot-sdk";
+
+const http = require("http");
+
+// probably the worst code you have seen in your life
+
+const bannedKeywords = [
+        ['Sus', 'thigh high'],
+	['Amogus!!', 'amogus', 'among us'],
+
+	['LGBTQQIAAP++ propaganda',
+
+		'transgender', 'trans rights',
+
+		[
+			'thigh', 'sock', 'computer'
+		],
+		
+		'transhumanist evolution', 'pride parade', 'pride flag', 'gay pride'
+
+	],
+
+	['Chinese cartoons', 'anime', 'manga', 'loli', 'azur lane', 'mihoyo', 'kancolle', 'neko', 'touhou', 'foxgirl', 'wolf girl', 'ayanamikodon', 'hatsune miku', 'vocaloid', 'mikudayo', 'kyoto animation'],
+
+	['Furries', 'fursona', 'fursuit', 'e621', 'furry fandom', 'furry art'],
+
+	['Gore', ' gore ', ' guro ', ' gory '],
+];
+
+// This will be the URL where clients can reach your homeserver. Note that this might be different
+// from where the web/chat interface is hosted. The server must support password registration without
+// captcha or terms of service (public servers typically won't work).
+const homeserverUrl = "https://trygve.me";
+
+// Use the access token you got from login or registration above.
+const accessToken = "your token";
+
+// In order to make sure the bot doesn't lose its state between restarts, we'll give it a place to cache
+// any information it needs to. You can implement your own storage provider if you like, but a JSON file
+// will work fine for this example.
+const storage = new SimpleFsStorageProvider("hello-bot.json");
+
+// Finally, let's create the client and set it to autojoin rooms. Autojoining is typical of bots to ensure
+// they can be easily added to any room.
+const client = new MatrixClient(homeserverUrl, accessToken, storage);
+AutojoinRoomsMixin.setupOnClient(client);
+
+// Before we start the bot, register our command handler
+client.on("room.event", handleCommand);
+
+// Now that everything is set up, start the bot. This will start the sync loop and run until killed.
+client.start().then(() => console.log("Bot started!"));
+
+// This is the command handler we registered a few lines up
+async function handleCommand(roomId: string, event: any) {
+    let expression = '';
+    let image_url = '';
+    if (
+        event['content']?.['msgtype'] == 'm.image'
+    ) {
+        image_url = homeserverUrl + "/_matrix/media/r0/download/" + (event['content']?.['url']).slice(6);
+        expression = 'image';
+    }
+    if (
+        event["content"]["avatar_url"]
+    ) {
+        image_url = homeserverUrl + "/_matrix/media/r0/download/" + (event['content']?.['avatar_url']).slice(6);
+        expression = 'pfp';
+    }
+
+    if (event['sender'] === await client.getUserId()) return;
+    if (image_url == '') return;
+
+    console.log('New image received, analyzing...');
+    analyze_image(image_url, function(degenerate, degenerateKeywords, fullDegenerateKeywords, response) {
+	if (degenerate) {
+            console.log('====Degeneracy Detected!====');
+            console.log('User: ' + event['sender'] + ' sent image with url: ' + image_url);
+	    console.log("Full keywords: " + fullDegenerateKeywords);
+	    console.log("Sent keywords: " + degenerateKeywords);
+	    console.log(response);
+	    console.log('https://matrix.to/#/' + roomId + '/' + event['event_id'] + '?via=willy.club&via=trygve.me&via=matrix.org');
+	    console.log('============================');
+            // determine what type of event it was
+            if (expression == 'image') {
+                client.redactEvent(roomId, event['event_id']);
+                client.sendText(roomId, 'Woah! Please refrain from posting images containing: "' + degenerateKeywords + '".');
+            } else {
+                client.kickUser(event['sender'], roomId, 'Please change your profile picture');
+            }
+ 
+        } else {
+	    console.log('=======Nothing found========');
+            console.log('User: ' + event['sender'] + ' sent image with url: ' + image_url);
+	    console.log(response);
+	    console.log('https://matrix.to/#/' + roomId + '/' + event['event_id'] + '?via=willy.club&via=trygve.me&via=matrix.org');
+	    console.log('============================');
+        } 
+    });
+}
+
+function analyze_image(image_url, callback) {
+    http.get(`http://localhost:5000/?image_url=` + image_url, resp => {
+            let data = "";
+
+            // A chunk of data has been recieved.
+            resp.on("data", chunk => {
+            	data += chunk;
+            });
+
+            // The whole response has been received. Print out the result.
+            resp.on("end", () => {
+            	let response = JSON.parse(data).message;
+		let responseWithoutComma = response.replace(',', ' ');
+
+		let degenerate = false;
+		let degenerateKeywords = '';
+		let fullDegenerateKeywords = '';
+		bannedKeywords.forEach(keyword => {
+		    if (Array.isArray(keyword)) {
+		        let alreadyAddedKeywordForThisGroup = false;
+			let skippedFirstIterationBecauseItsTheTitle = false;
+			keyword.forEach(subKeyword => {
+			    if (!skippedFirstIterationBecauseItsTheTitle) {
+			    	skippedFirstIterationBecauseItsTheTitle = true;
+				return;
+			    }
+			    if (Array.isArray(subKeyword)) {
+				let allOfThemMatch = true;
+                                subKeyword.forEach(subsubKeyword => {
+				    if(responseWithoutComma.includes(subsubKeyword)) {
+				        fullDegenerateKeywords += subsubKeyword + ',';
+				    } else {
+                                        allOfThemMatch = false;
+				    }
+				});
+				if (allOfThemMatch) {
+				    degenerate = true;
+				    if (!alreadyAddedKeywordForThisGroup) {
+				        degenerateKeywords += keyword[0] + ',';
+				        alreadyAddedKeywordForThisGroup = true;
+				    }
+				}
+			    } else if (responseWithoutComma.includes(subKeyword)) {
+				degenerate = true;
+				if (!alreadyAddedKeywordForThisGroup) {
+				    degenerateKeywords += keyword[0] + ',';
+				    fullDegenerateKeywords += subKeyword + ',';
+				    alreadyAddedKeywordForThisGroup = true;
+				}
+			    }
+			});
+		    } else if (responseWithoutComma.includes(keyword)) {
+                        degenerate = true;	
+			degenerateKeywords += keyword + ',';
+			fullDegenerateKeywords += keyword + ',';
+		    }
+                });
+
+		degenerateKeywords = degenerateKeywords.substr(0, degenerateKeywords.length -1);
+		fullDegenerateKeywords = fullDegenerateKeywords.substr(0, fullDegenerateKeywords.length -1);
+
+                if (degenerate) {
+                    callback(true, degenerateKeywords, fullDegenerateKeywords, response);
+                } else {
+                    callback(false, '', '', response);
+		}
+            });
+        });
+}
+
+// sometimes we dont have permission to redact messages, lets just ignore that
+process.on('uncaughtException', function (exception) {
+    console.log('An error occured but it was probably not that important');
+    //console.log(exception);
+});
+

bot/package.json

@@ -0,0 +1,6 @@
+{
+  "dependencies": {
+    "matrix-bot-sdk": "^0.6.3",
+    "typescript": "^4.9.4"
+  }
+}

interrogator/app.py

@@ -0,0 +1,26 @@
+from flask import Flask, request
+
+app = Flask(__name__)
+
+from PIL import Image
+from clip_interrogator import Config, Interrogator
+
+ci = Interrogator(Config(clip_model_name="ViT-L-14/openai"))
+#ci = Interrogator(Config(clip_model_name="ViT-H-14/laion2b_s32b_b79k"))
+
+image_path = "/dev/shm/image_data"
+
+import requests
+
+@app.route("/", methods=['GET'])
+def hello_world():
+    args = request.args
+
+    img_data = requests.get(args['image_url']).content
+    with open(image_path, 'wb') as handler:
+        handler.write(img_data)
+
+    image = Image.open(image_path).convert('RGB')
+
+    #return {'message': ci.interrogate(image)}
+    return {'message': ci.interrogate_fast(image)}