Fix errors seen during longer runs: selenium exceptions, insecure certificates, recusrsion limit exceeded, BM25 document corpus too small

2026-03-26 12:22:13 +00:00
parent fbc688b8f9
commit 5e374a8bd6
3 changed files with 104 additions and 30 deletions
@@ -1,32 +1,92 @@
 import { Builder, Browser } from "selenium-webdriver";
 import firefox from "selenium-webdriver/firefox";
+import { backOff } from "exponential-backoff";
+import { logger } from "../utils/logger";

-export async function extractWebpageContent(url: string) : Promise<string[]>{
-    const options = new firefox.Options();
-    options.addArguments("--headless");
-
-    let driver = await new Builder().forBrowser(Browser.FIREFOX).setFirefoxOptions(options).build()
-    try {
-        await driver.get(url)
-        await driver.wait(async () => {
-            return await driver.executeScript(
-                "return document.readyState === 'complete'"
-            );
-        }, 5000);
-
-        const readableText = await driver.executeScript(
-            "return document.body.innerText;"
-        ) as string;
-
-        const filteredLines = readableText
-            .split(/\r?\n/)
-            .map(line => line.trim())
-            .filter(line => line.split(/\s+/).length > 1); 
-        
-        return filteredLines;
-    } finally {
-        await driver.quit()
-    }
+export async function extractWebpageContent(url: string): Promise<string[]> {
+  try {
+    const response = await backOff(async () => {
+      return await extractWebpageContentWorker(url);
+    }, {
+      numOfAttempts: 10,
+      startingDelay: 500,
+      timeMultiple: 2,
+      jitter: "full",
+      maxDelay: 50000,
+    });
+    return response;
+  } catch (err: any) {
+    logger.error(`Failed out of retry loop for URL "${url}", returning placeholder to pipeline`);
+    return ["API EXCEPTION"];
+  }
 }

-//console.log(await extractWebpageContent("https://www.bbc.co.uk/news/live/c74wd01egvyt"))
+async function extractWebpageContentWorker(url: string): Promise<string[]> {
+  let driver;
+  try {
+    const options = new firefox.Options();
+    options.addArguments("--headless");
+    driver = await new Builder()
+      .forBrowser(Browser.FIREFOX)
+      .setFirefoxOptions(options)
+      .build();
+  } catch (err: any) {
+    const desc = `Failed to launch Firefox driver: ${err.message}`;
+    logger.error(desc);
+    throw new Error(desc);
+  }
+
+  try {
+    try {
+      await driver.get(url);
+    } catch (err: any) {
+      const desc = `Failed to navigate to URL "${url}": ${err.message}`;
+      logger.error(desc);
+      throw new Error(desc);
+    }
+
+    try {
+      await driver.wait(async () => {
+        return await driver.executeScript(
+          "return document.readyState === 'complete'"
+        );
+      }, 5000);
+    } catch (err: any) {
+      logger.error(`Page load timed out for "${url}", attempting to read partial content: ${err.message}`);
+      // do not throw, attempt to read
+    }
+
+    let readableText: string;
+    try {
+      readableText = await driver.executeScript(
+        "return document.body.innerText;"
+      ) as string;
+    } catch (err: any) {
+      const desc = `Failed to extract page text from "${url}": ${err.message}`;
+      logger.error(desc);
+      throw new Error(desc);
+    }
+
+    const filteredLines = readableText
+      .split(/\r?\n/)
+      .map(line => line.trim())
+      .filter(line => line.split(/\s+/).length > 1);
+
+    if (filteredLines.length === 0) {
+      const desc = `No content extracted from "${url}"`;
+      logger.error(desc);
+      throw new Error(desc);
+    }
+
+    return filteredLines;
+  } finally {
+    try {
+      await driver.quit();
+    } catch (err: any) {
+      logger.error(`Failed to quit Firefox driver cleanly: ${err.message}`);
+    }
+  }
+}
+
+// console.log(await extractWebpageContent("https://www.bbc.co.uk/news/live/c74wd01egvyt"))
+// console.log(await extractWebpageContent("https://badcertificate.int.jeynes.uk/"))