Merge pull request #199 from openchatai/add-standalone-selenium

Add headless crawling

llm-server/requirements.txt

@@ -127,3 +127,4 @@ Werkzeug==3.0.1
 wikipedia==1.4.0
 yarl==1.9.2
 zipp==3.16.2
+selenium==4.14.0

llm-server/utils/crawler.py

@@ -0,0 +1,78 @@
+from selenium import webdriver
+from urllib.parse import urlparse, urljoin
+from collections import deque
+
+from selenium.webdriver.common.by import By
+
+
+def recursive_scraper(root_url, max_pages):
+    visited_links = set()
+    links_to_visit = deque()
+
+    # Initialize the WebDriver (Firefox)
+    firefox_options = webdriver.FirefoxOptions()
+
+    firefox_options.add_argument('--headless')
+    driver = webdriver.Remote(
+        command_executor='http://selenium:4444/wd/hub',
+        options=firefox_options
+    )
+
+    # Remove trailing slashes from the root URL
+    root_url = root_url.rstrip('/')
+
+    # Add the root URL to the queue
+    links_to_visit.append(root_url)
+
+    while links_to_visit and len(visited_links) < max_pages:
+        current_url = links_to_visit.popleft()
+
+        # Check if the URL has already been visited
+        if current_url in visited_links:
+            continue
+
+        try:
+            driver.get(current_url)
+
+            # Print the current URL for observation
+            print("Visiting:", current_url)
+
+            # Extract information from the current page here
+            # For example, you can use driver.page_source and BeautifulSoup
+
+            # Add the current URL to the visited set
+            visited_links.add(current_url)
+
+            # Extract all links on the current page
+            sublinks = [a.get_attribute('href') for a in driver.find_elements(By.TAG_NAME, "a")]
+
+            # Filter and enqueue sublinks with the same domain, and without in-page anchors
+            for link in sublinks:
+                if not link or link.startswith("#"):
+                    continue  # Skip empty or in-page anchor links
+
+                # Remove trailing slashes from the link
+                link = link.rstrip('/')
+
+                parsed_link = urlparse(link)
+                parsed_root_url = urlparse(root_url)
+
+                if parsed_link.netloc == parsed_root_url.netloc:
+                    absolute_link = urljoin(root_url, link)
+
+                    # Check if the link is not already visited and is valid
+                    if absolute_link not in visited_links and absolute_link not in links_to_visit:
+                        print("Found sublink:", absolute_link)
+                        links_to_visit.append(absolute_link)
+
+        except Exception as e:
+            print(f"Error visiting {current_url}: {str(e)}")
+
+    # Clean up and close the WebDriver
+    driver.quit()
+
+
+# # Example usage
+# root_url = 'https://opencopilot.so'
+# max_pages = 10
+# recursive_scraper(root_url, max_pages)