From 2fb29ee46ee4b1bd194d6eb575e63cff049e3f43 Mon Sep 17 00:00:00 2001 From: Aparup Ganguly Date: Thu, 20 Mar 2025 15:26:02 +0530 Subject: [PATCH] Add examples/ mistra- small-3.1-crawler --- .../mistral-small-3.1-crawler.py | 272 ++++++++++++++++++ 1 file changed, 272 insertions(+) create mode 100644 examples/mistral-small-3.1-crawler/mistral-small-3.1-crawler.py diff --git a/examples/mistral-small-3.1-crawler/mistral-small-3.1-crawler.py b/examples/mistral-small-3.1-crawler/mistral-small-3.1-crawler.py new file mode 100644 index 00000000..9c8c3173 --- /dev/null +++ b/examples/mistral-small-3.1-crawler/mistral-small-3.1-crawler.py @@ -0,0 +1,272 @@ +import os +from firecrawl import FirecrawlApp +import json +from dotenv import load_dotenv +from mistralai import Mistral + +# 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") +mistral_api_key = os.getenv("MISTRAL_API_KEY") + +# Initialize the FirecrawlApp and Mistral client +app = FirecrawlApp(api_key=firecrawl_api_key) +client = Mistral(api_key=mistral_api_key) + +# Find the page that most likely contains the objective +def find_relevant_page_via_map(objective, url, app, client): + 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}") + completion = client.chat.complete( + model="mistral-small-latest", + messages=[ + { + "role": "user", + "content": map_prompt + } + ] + ) + + map_search_parameter = completion.choices[0].message.content + 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}") + map_website = app.map_url(url, params={"search": map_search_parameter}) + + # Debug print to see the response structure + print(f"{Colors.MAGENTA}Debug - Map response structure: {json.dumps(map_website, indent=2)}{Colors.RESET}") + + print(f"{Colors.GREEN}Website mapping completed successfully.{Colors.RESET}") + + # Handle the response based on its structure + if isinstance(map_website, dict): + # Assuming the links are in a 'urls' or similar key + links = map_website.get('urls', []) or map_website.get('links', []) + elif isinstance(map_website, str): + try: + parsed = json.loads(map_website) + links = parsed.get('urls', []) or parsed.get('links', []) + except json.JSONDecodeError: + links = [] + else: + links = map_website if isinstance(map_website, list) else [] + + if not links: + print(f"{Colors.RED}No links found in map response.{Colors.RESET}") + return None + + rank_prompt = f""" + Given this list of URLs and the objective: {objective} + Analyze each URL and rank the top 3 most relevant ones that are most likely to contain the information we need. + Return your response as a JSON array with exactly 3 objects, each containing: + - "url": the full URL + - "relevance_score": number between 0-100 indicating relevance to objective + - "reason": brief explanation of why this URL is relevant + + Example output: + [ + {{ + "url": "https://example.com/about", + "relevance_score": 95, + "reason": "Main about page containing company information" + }}, + {{ + "url": "https://example.com/team", + "relevance_score": 80, + "reason": "Team page with leadership details" + }}, + {{ + "url": "https://example.com/contact", + "relevance_score": 70, + "reason": "Contact page with location information" + }} + ] + + URLs to analyze: + {json.dumps(links, indent=2)} + """ + + print(f"{Colors.YELLOW}Ranking URLs by relevance to objective...{Colors.RESET}") + completion = client.chat.complete( + model="mistral-small-latest", + messages=[ + { + "role": "user", + "content": rank_prompt + } + ] + ) + + # Debug print to see Mistral's raw response + print(f"{Colors.MAGENTA}Debug - Mistral's raw response:{Colors.RESET}") + print(f"{Colors.MAGENTA}{completion.choices[0].message.content}{Colors.RESET}") + + try: + # Try to clean the response by stripping any potential markdown or extra whitespace + cleaned_response = completion.choices[0].message.content.strip() + print(f"{Colors.YELLOW}Attempting to extract JSON from response...{Colors.RESET}") + + # Extract JSON block if it exists between triple backticks + if "```json" in cleaned_response and "```" in cleaned_response.split("```json", 1)[1]: + print(f"{Colors.GREEN}Found JSON code block, extracting...{Colors.RESET}") + cleaned_response = cleaned_response.split("```json", 1)[1].split("```", 1)[0].strip() + # If no code blocks but contains square brackets, try to extract just the JSON array + elif "[" in cleaned_response and "]" in cleaned_response: + print(f"{Colors.GREEN}Found JSON array markers, extracting...{Colors.RESET}") + start_idx = cleaned_response.find("[") + end_idx = cleaned_response.rfind("]") + 1 + cleaned_response = cleaned_response[start_idx:end_idx].strip() + + print(f"{Colors.YELLOW}Parsing extracted content: {cleaned_response[:100]}...{Colors.RESET}") + ranked_results = json.loads(cleaned_response) + print(f"{Colors.GREEN}Successfully parsed JSON response{Colors.RESET}") + + # Validate the structure of the results + if not isinstance(ranked_results, list): + raise ValueError("Response is not a list") + + for result in ranked_results: + if not all(key in result for key in ["url", "relevance_score", "reason"]): + raise ValueError("Response items missing required fields") + + links = [result["url"] for result in ranked_results] + + # Print detailed ranking info + print(f"{Colors.CYAN}Top 3 ranked URLs:{Colors.RESET}") + for result in ranked_results: + print(f"{Colors.GREEN}URL: {result['url']}{Colors.RESET}") + print(f"{Colors.YELLOW}Relevance Score: {result['relevance_score']}{Colors.RESET}") + print(f"{Colors.BLUE}Reason: {result['reason']}{Colors.RESET}") + print("---") + + if not links: + print(f"{Colors.RED}No relevant links identified.{Colors.RESET}") + return None + + except (json.JSONDecodeError, KeyError) as e: + print(f"{Colors.RED}Error parsing ranked results: {str(e)}{Colors.RESET}") + return None + + print(f"{Colors.GREEN}Located {len(links)} relevant links.{Colors.RESET}") + return 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, client): + try: + # Get top 3 links from the map result + if not map_website: + print(f"{Colors.RED}No links found to analyze.{Colors.RESET}") + return None + + top_links = map_website[:3] + 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_result = app.scrape_url(link, params={'formats': ['markdown']}) + print(f"{Colors.GREEN}Page scraping completed successfully.{Colors.RESET}") + + 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 exactly '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. If returning JSON, ensure it's valid JSON format without any markdown formatting. + 4. If objective is not met, respond only with 'Objective not met'. + """ + + completion = client.chat.complete( + model="mistral-small-latest", + messages=[{"role": "user", "content": check_prompt}] + ) + + result = completion.choices[0].message.content.strip() + + # Clean up the response if it contains markdown formatting + # Extract JSON block if it exists between triple backticks + if "```json" in result and "```" in result.split("```json", 1)[1]: + print(f"{Colors.GREEN}Found JSON code block, extracting...{Colors.RESET}") + result = result.split("```json", 1)[1].split("```", 1)[0].strip() + # If no code blocks but contains curly braces, try to extract just the JSON object + elif result != "Objective not met" and "{" in result and "}" in result: + print(f"{Colors.GREEN}Found JSON object markers, extracting...{Colors.RESET}") + start_idx = result.find("{") + end_idx = result.rfind("}") + 1 + result = result[start_idx:end_idx].strip() + + if result == "Objective not met": + print(f"{Colors.YELLOW}Objective not met on this page. Proceeding to next link...{Colors.RESET}") + continue + + try: + print(f"{Colors.YELLOW}Parsing extracted content: {result[:100]}...{Colors.RESET}") + json_result = json.loads(result) + print(f"{Colors.GREEN}Successfully parsed JSON response{Colors.RESET}") + print(f"{Colors.GREEN}Objective fulfilled. Relevant information found.{Colors.RESET}") + return json_result + except json.JSONDecodeError as e: + print(f"{Colors.RED}Error parsing JSON response: {str(e)}{Colors.RESET}") + print(f"{Colors.MAGENTA}Raw response: {result}{Colors.RESET}") + continue + + 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, client) + + if map_website: + print(f"{Colors.GREEN}Relevant pages identified. Proceeding with detailed analysis using Mistral Small 3.1...{Colors.RESET}") + # Find objective in top pages + result = find_objective_in_top_pages(map_website, objective, app, client) + + 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() \ No newline at end of file