Merge pull request #653 from cygaar/playwright_improvements

feat: improve browser service
elizaOS · Nov 28, 2024 · dbec9fd · dbec9fd
2 parents bc80654 + d274ef7
commit dbec9fd
Show file tree

Hide file tree

Showing 5 changed files with 63 additions and 82 deletions.
diff --git a/packages/client-discord/src/messages.ts b/packages/client-discord/src/messages.ts
@@ -29,11 +29,7 @@ import {
     discordShouldRespondTemplate,
     discordMessageHandlerTemplate,
 } from "./templates.ts";
-import {
-    generateSummary,
-    sendMessageInChunks,
-    canSendMessage,
-} from "./utils.ts";
+import { sendMessageInChunks, canSendMessage } from "./utils.ts";
 
 export type InterestChannels = {
     [key: string]: {
@@ -410,20 +406,16 @@ export class MessageManager {
                     throw new Error("Browser service not found");
                 }
 
-                const { title, bodyContent } =
+                const { title, description: summary } =
                     await browserService.getPageContent(url, this.runtime);
 
-                const { title: newTitle, description } = await generateSummary(
-                    this.runtime,
-                    title + "\n" + bodyContent
-                );
                 attachments.push({
                     id: `webpage-${Date.now()}`,
                     url: url,
-                    title: newTitle || "Web Page",
+                    title: title || "Web Page",
                     source: "Web",
-                    description,
-                    text: bodyContent,
+                    description: summary,
+                    text: summary,
                 });
             }
         }

diff --git a/packages/core/src/runtime.ts b/packages/core/src/runtime.ts
@@ -953,13 +953,9 @@ Text: ${attachment.text}
                 .join(" ");
         }
 
-
         const knowledegeData = await knowledge.get(this, message);
 
-        const formattedKnowledge = formatKnowledge(
-            knowledegeData
-        );
-
+        const formattedKnowledge = formatKnowledge(knowledegeData);
 
         const initialState = {
             agentId: this.agentId,
@@ -1242,5 +1238,7 @@ Text: ${attachment.text}
 }
 
 const formatKnowledge = (knowledge: KnowledgeItem[]) => {
-    return knowledge.map((knowledge) => `- ${knowledge.content.text}`).join("\n");
+    return knowledge
+        .map((knowledge) => `- ${knowledge.content.text}`)
+        .join("\n");
 };
diff --git a/packages/plugin-node/package.json b/packages/plugin-node/package.json
@@ -60,6 +60,7 @@
         "youtube-dl-exec": "3.0.10"
     },
     "devDependencies": {
+        "@types/node": "22.8.4",
         "eslint": "9.13.0",
         "eslint-config-prettier": "9.1.0",
         "eslint-plugin-prettier": "5.2.1",

diff --git a/packages/plugin-node/src/services/browser.ts b/packages/plugin-node/src/services/browser.ts
@@ -63,9 +63,6 @@ export class BrowserService extends Service implements IBrowserService {
     private captchaSolver: CaptchaSolver;
     private cacheKey = "content/browser";
 
-    private queue: string[] = [];
-    private processing: boolean = false;
-
     static serviceType: ServiceType = ServiceType.BROWSER;
 
     static register(runtime: IAgentRuntime): IAgentRuntime {
@@ -87,17 +84,43 @@ export class BrowserService extends Service implements IBrowserService {
         );
     }
 
-    async initialize() { }
+    async initialize() {}
 
     async initializeBrowser() {
         if (!this.browser) {
             this.browser = await chromium.launch({
-                args: ["--no-sandbox", "--disable-setuid-sandbox"],
+                headless: true,
+                args: [
+                    "--disable-dev-shm-usage", // Uses /tmp instead of /dev/shm. Prevents memory issues on low-memory systems
+                    "--block-new-web-contents", // Prevents creation of new windows/tabs
+                ],
             });
 
+            const platform = process.platform;
+            let userAgent = "";
+
+            // Change the user agent to match the platform to reduce bot detection
+            switch (platform) {
+                case "darwin":
+                    userAgent =
+                        "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36";
+                    break;
+                case "win32":
+                    userAgent =
+                        "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36";
+                    break;
+                case "linux":
+                    userAgent =
+                        "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36";
+                    break;
+                default:
+                    userAgent =
+                        "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36";
+            }
+
             this.context = await this.browser.newContext({
-                userAgent:
-                    "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36",
+                userAgent,
+                acceptDownloads: false,
             });
 
             this.blocker =
@@ -121,49 +144,13 @@ export class BrowserService extends Service implements IBrowserService {
         runtime: IAgentRuntime
     ): Promise<PageContent> {
         await this.initializeBrowser();
-        this.queue.push(url);
-        this.processQueue(runtime);
-
-        return new Promise((resolve, reject) => {
-            const checkQueue = async () => {
-                const index = this.queue.indexOf(url);
-                if (index !== -1) {
-                    setTimeout(checkQueue, 100);
-                } else {
-                    try {
-                        const result = await this.fetchPageContent(
-                            url,
-                            runtime
-                        );
-                        resolve(result);
-                    } catch (error) {
-                        reject(error);
-                    }
-                }
-            };
-            checkQueue();
-        });
+        return await this.fetchPageContent(url, runtime);
     }
 
     private getCacheKey(url: string): string {
         return stringToUuid(url);
     }
 
-    private async processQueue(runtime: IAgentRuntime): Promise<void> {
-        if (this.processing || this.queue.length === 0) {
-            return;
-        }
-
-        this.processing = true;
-
-        while (this.queue.length > 0) {
-            const url = this.queue.shift();
-            await this.fetchPageContent(url, runtime);
-        }
-
-        this.processing = false;
-    }
-
     private async fetchPageContent(
         url: string,
         runtime: IAgentRuntime
@@ -214,15 +201,15 @@ export class BrowserService extends Service implements IBrowserService {
             if (captchaDetected) {
                 await this.solveCaptcha(page, url);
             }
-            const title = await page.evaluate(() => document.title);
+            const documentTitle = await page.evaluate(() => document.title);
             const bodyContent = await page.evaluate(
                 () => document.body.innerText
             );
-            const { description } = await generateSummary(
+            const { title: parsedTitle, description } = await generateSummary(
                 runtime,
-                title + "\n" + bodyContent
+                documentTitle + "\n" + bodyContent
             );
-            const content = { title, description, bodyContent };
+            const content = { title: parsedTitle, description, bodyContent };
             await runtime.cacheManager.set(`${this.cacheKey}/${cacheKey}`, {
                 url,
                 content,

diff --git a/pnpm-lock.yaml b/pnpm-lock.yaml