Spaces:
Sleeping
Sleeping
| import re | |
| from transformers import pipeline | |
| from bs4 import BeautifulSoup | |
| import requests | |
| import openai | |
| # Initialize summarizer and NER pipeline | |
| summarizer = pipeline("summarization", model="facebook/bart-large-cnn") | |
| ner_pipeline = pipeline("ner", model="dslim/bert-base-NER") | |
| # Set up OpenAI API for nuanced text processing (GPT-4) | |
| openai.api_key = "YOUR_OPENAI_API_KEY" | |
| def fetch_company_pages(base_url): | |
| # Define additional pages to search for mission or company info | |
| pages = [base_url, f"{base_url}/about", f"{base_url}/company", f"{base_url}/about-us"] | |
| content = "" | |
| for url in pages: | |
| try: | |
| response = requests.get(url, timeout=5) | |
| soup = BeautifulSoup(response.text, 'html.parser') | |
| # Get main descriptive content from meta or visible text | |
| meta_desc = soup.find('meta', {'name': 'description'}) | |
| main_text = " ".join(p.get_text() for p in soup.find_all('p')) if not meta_desc else meta_desc['content'] | |
| content += main_text | |
| # Stop if we find substantial content | |
| if len(content) > 300: | |
| break | |
| except Exception as e: | |
| print(f"Error fetching from {url}: {e}") | |
| return content or "Company information not found." | |
| def classify_industry(description): | |
| # Use OpenAI to classify industry based on description | |
| prompt = f"Classify the following company description into an industry category (e.g., Finance, Retail, Technology, Healthcare, Automotive):\n\n{description}\n\nIndustry:" | |
| response = openai.Completion.create(model="gpt-4", prompt=prompt, max_tokens=10) | |
| return response.choices[0].text.strip() | |
| def detect_focus_areas(text): | |
| # Identify focus areas based on recognized entities | |
| entities = ner_pipeline(text) | |
| focus_areas = set() | |
| # Define common focus area patterns | |
| area_patterns = { | |
| "customer experience": ["customer", "experience", "support", "satisfaction"], | |
| "operations": ["operations", "supply chain", "efficiency", "logistics", "processes"], | |
| "product innovation": ["innovation", "product", "development", "R&D"], | |
| "sustainability": ["sustainability", "environment", "eco", "carbon"] | |
| } | |
| # Check for focus area matches | |
| for entity in entities: | |
| for area, keywords in area_patterns.items(): | |
| if any(keyword in entity['word'].lower() for keyword in keywords): | |
| focus_areas.add(area) | |
| return list(focus_areas) | |
| def summarize_mission(text): | |
| # Generate a summary for mission content | |
| mission_summary = summarizer(text, max_length=30, min_length=10, do_sample=False)[0]['summary_text'] | |
| return mission_summary | |
| def get_company_info(url): | |
| try: | |
| # Step 1: Fetch and process company page content | |
| content = fetch_company_pages(url) | |
| # Step 2: Summarize mission content | |
| mission = summarize_mission(content) | |
| # Step 3: Classify the company industry using GPT | |
| industry = classify_industry(content) | |
| # Step 4: Detect specific focus areas | |
| focus_areas = detect_focus_areas(content) | |
| # Final dictionary with company information | |
| return { | |
| "mission": mission, | |
| "industry": industry, | |
| "focus_areas": focus_areas or ["No specific focus areas identified"] | |
| } | |
| except Exception as e: | |
| print(f"Error fetching company info: {e}") | |
| return { | |
| "mission": "Mission not found", | |
| "industry": "Unknown", | |
| "focus_areas": [] | |
| } | |