diff --git a/examples/grok_web_crawler/grok_web_crawler.py b/examples/grok_web_crawler/grok_web_crawler.py new file mode 100644 index 00000000..8d598e86 --- /dev/null +++ b/examples/grok_web_crawler/grok_web_crawler.py @@ -0,0 +1,150 @@ +import os +from firecrawl import FirecrawlApp +import json +from dotenv import load_dotenv +import requests + +# ANSI color codes +class Colors: + CYAN = '\033[96m' + YELLOW = '\033[93m' + GREEN = '\033[92m' + RED = '\033[91m' + MAGENTA = '\033[95m' + BLUE = '\033[94m' + RESET = '\033[0m' + +# Load environment variables +load_dotenv() + +# Retrieve API keys from environment variables +firecrawl_api_key = os.getenv("FIRECRAWL_API_KEY") +grok_api_key = os.getenv("GROK_API_KEY") + +# Initialize the FirecrawlApp +app = FirecrawlApp(api_key=firecrawl_api_key) + +# Function to make Grok API calls +def grok_completion(prompt): + url = "https://api.x.ai/v1/chat/completions" + headers = { + "Content-Type": "application/json", + "Authorization": f"Bearer {grok_api_key}" + } + data = { + "messages": [ + { + "role": "system", + "content": "You are a helpful assistant." + }, + { + "role": "user", + "content": prompt + } + ], + "model": "grok-beta", + "stream": False, + "temperature": 0 + } + response = requests.post(url, headers=headers, json=data) + return response.json()['choices'][0]['message']['content'] + +# Find the page that most likely contains the objective +def find_relevant_page_via_map(objective, url, app): + try: + print(f"{Colors.CYAN}Understood. The objective is: {objective}{Colors.RESET}") + print(f"{Colors.CYAN}Initiating search on the website: {url}{Colors.RESET}") + + map_prompt = f""" + The map function generates a list of URLs from a website and it accepts a search parameter. Based on the objective of: {objective}, come up with a 1-2 word search parameter that will help us find the information we need. Only respond with 1-2 words nothing else. + """ + + print(f"{Colors.YELLOW}Analyzing objective to determine optimal search parameter...{Colors.RESET}") + map_search_parameter = grok_completion(map_prompt) + print(f"{Colors.GREEN}Optimal search parameter identified: {map_search_parameter}{Colors.RESET}") + + print(f"{Colors.YELLOW}Mapping website using the identified search parameter...{Colors.RESET}") + print(f"{Colors.MAGENTA}{map_search_parameter}{Colors.RESET}") + map_website = app.map_url(url, params={"search": map_search_parameter}) + print(f"{Colors.GREEN}Website mapping completed successfully.{Colors.RESET}") + print(f"{Colors.GREEN}Located {len(map_website['links'])} relevant links.{Colors.RESET}") + print(f"{Colors.MAGENTA}{map_website}{Colors.RESET}") + return map_website["links"] + except Exception as e: + print(f"{Colors.RED}Error encountered during relevant page identification: {str(e)}{Colors.RESET}") + return None + +# Scrape the top 3 pages and see if the objective is met, if so return in json format else return None +def find_objective_in_top_pages(map_website, objective, app): + try: + print(f"{Colors.MAGENTA}{map_website}{Colors.RESET}") + # Get top 3 links from the map result + top_links = map_website[:3] if isinstance(map_website, list) else [] + print(f"{Colors.CYAN}Proceeding to analyze top {len(top_links)} links: {top_links}{Colors.RESET}") + + for link in top_links: + print(f"{Colors.YELLOW}Initiating scrape of page: {link}{Colors.RESET}") + # Scrape the page + scrape_result = app.scrape_url(link, params={'formats': ['markdown']}) + print(f"{Colors.GREEN}Page scraping completed successfully.{Colors.RESET}") + + + # Check if objective is met + check_prompt = f""" + Given the following scraped content and objective, determine if the objective is met. + If it is, extract the relevant information in a simple and concise JSON format. Use only the necessary fields and avoid nested structures if possible. + If the objective is not met with confidence, respond with 'Objective not met'. + + Objective: {objective} + Scraped content: {scrape_result['markdown']} + + Remember: + 1. Only return JSON if you are confident the objective is fully met. + 2. Keep the JSON structure as simple and flat as possible. + 3. Do not include any explanations or markdown formatting in your response. + """ + + result = grok_completion(check_prompt) + print(f"{Colors.MAGENTA}{result}{Colors.RESET}") + if result != "Objective not met": + print(f"{Colors.GREEN}Objective potentially fulfilled. Relevant information identified.{Colors.RESET}") + try: + result = result.replace("```json", "").replace("```", "") + return json.loads(result) + except json.JSONDecodeError: + print(f"{Colors.RED}Error in parsing response. Proceeding to next page...{Colors.RESET}") + else: + print(f"{Colors.YELLOW}Objective not met on this page. Proceeding to next link...{Colors.RESET}") + + print(f"{Colors.RED}All available pages analyzed. Objective not fulfilled in examined content.{Colors.RESET}") + return None + + except Exception as e: + print(f"{Colors.RED}Error encountered during page analysis: {str(e)}{Colors.RESET}") + return None + +# Main function to execute the process +def main(): + # Get user input + url = input(f"{Colors.BLUE}Enter the website to crawl: {Colors.RESET}") + objective = input(f"{Colors.BLUE}Enter your objective: {Colors.RESET}") + + print(f"{Colors.YELLOW}Initiating web crawling process...{Colors.RESET}") + # Find the relevant page + map_website = find_relevant_page_via_map(objective, url, app) + + if map_website: + print(f"{Colors.GREEN}Relevant pages identified. Proceeding with detailed analysis...{Colors.RESET}") + # Find objective in top pages + result = find_objective_in_top_pages(map_website, objective, app) + + if result: + print(f"{Colors.GREEN}Objective successfully fulfilled. Extracted information:{Colors.RESET}") + print(f"{Colors.MAGENTA}{json.dumps(result, indent=2)}{Colors.RESET}") + else: + print(f"{Colors.RED}Unable to fulfill the objective with the available content.{Colors.RESET}") + else: + print(f"{Colors.RED}No relevant pages identified. Consider refining the search parameters or trying a different website.{Colors.RESET}") + +if __name__ == "__main__": + main()