import requests import os from dotenv import load_dotenv # Ensure environment variables are loaded from root .env basedir = os.path.abspath(os.path.join(os.path.dirname(__file__), '../../')) load_dotenv(os.path.join(basedir, '.env'), override=True) from flask import Blueprint, request, jsonify, current_app from flask_jwt_extended import jwt_required, get_jwt_identity from app.models.chat import ChatMessage from app.models.user import User from app.models.learning import LearningMaterial from app import db chat_bp = Blueprint('chat', __name__) def call_groq_api(prompt): # Try app config first, then env vars print(f"DEBUG: GROQ_API_KEY in config: {'Yes' if current_app.config.get('GROQ_API_KEY') else 'No'}") print(f"DEBUG: GROQ_API_KEY in getenv: {'Yes' if os.getenv('GROQ_API_KEY') else 'No'}") print(f"DEBUG: GROQ_API_KEY in environ: {'Yes' if 'GROQ_API_KEY' in os.environ else 'No'}") api_key = current_app.config.get("GROQ_API_KEY") or os.getenv("GROQ_API_KEY") or os.getenv("GOOGLE_API_KEY") if not api_key: print("!!! ERROR: No API Key found in ANY source.") is_vercel = os.environ.get('VERCEL') == '1' msg = "I'm sorry, but my AI core is currently offline (API Key Missing)." if is_vercel: msg += " Please add 'GROQ_API_KEY' to your Vercel Environment Variables in the project settings." else: msg += " Please check your .env file or contact the administrator." return msg url = "https://api.groq.com/openai/v1/chat/completions" payload = { "model": "groq/compound-mini", "messages": [ { "role": "system", "content": "You are a professional educational tutor. Provide clear, accurate, and concise academic support." }, { "role": "user", "content": prompt } ], "temperature": 0.7, "max_completion_tokens": 1024, "top_p": 1, "stream": False } headers = { 'Content-Type': 'application/json', 'Authorization': f'Bearer {api_key}' } try: response = requests.post(url, json=payload, headers=headers, timeout=20) print(f" * Groq Response Status: {response.status_code}") response.raise_for_status() result = response.json() if 'choices' in result and len(result['choices']) > 0: content = result['choices'][0]['message']['content'] return content return "I understood your question, but I'm having trouble generating a response. Could you try asking in a different way?" except requests.exceptions.RequestException as e: print(f"Groq API Request Error: {e}") return "I'm having a brief technical moment connecting to my brain. Please try again in a few seconds!" except Exception as e: print(f"Groq Unexpected Error: {e}") return "I'm having a brief technical moment. Please try again in a few seconds!" @chat_bp.route('/send', methods=['POST']) @jwt_required() def send_message(): try: print(">>> CHAT SEND REQUEST START") user_id = int(get_jwt_identity()) data = request.get_json() print(f"User ID: {user_id}, Message: {data.get('message')}") if not data or not data.get('message'): return jsonify({"success": False, "message": "Missing message content"}), 400 user_msg = ChatMessage(user_id=user_id, message=data.get('message'), is_bot=False) db.session.add(user_msg) # Fetch user details for personalization user = db.session.get(User, user_id) user_lang = user.language if user else 'en' user_lvl = user.education_level if user else 'Secondary' # Fetch some context from learning materials relevant to user level materials = LearningMaterial.query.filter_by(language=user_lang, education_level=user_lvl).limit(5).all() if not materials: materials = LearningMaterial.query.filter_by(language='en', education_level=user_lvl).limit(5).all() context = "\n".join([f"{m.subject}: {m.content[:300]}" for m in materials]) # Map language codes to names for the AI lang_map = { 'en': 'English', 'ha': 'Hausa', 'yo': 'Yoruba', 'ig': 'Igbo', 'pi': 'Nigerian Pidgin' } lang_name = lang_map.get(user_lang, 'English') # Determine Persona based on user level if user_lvl == 'Vocational': persona_name = "Workshop Expert" persona_role = "a master craftsman and professional vocational mentor" persona_goal = "provide expert-level technical guidance, troubleshooting, and career advice for vocational skills" persona_context = "Your focus is purely on practical skills, industrial standards, and hands-on craftsmanship." else: persona_name = "QualiLearn AI" persona_role = "a professional academic mentor and expert tutor in Mathematics, English, Science, Economics, and Government" persona_goal = f"provide expert-level, accurate, and curriculum-aligned academic guidance for a {user_lvl} student" persona_context = "Your focus is on academic excellence, deep subject mastery, and precise exam preparation. You must provide the most accurate and up-to-date information available." # Build prompt safely prompt_template = f""" You are {{persona_name}}, {{persona_role}}. Your goal is to {{persona_goal}}. IMPORTANT: Respond ONLY in {{lang_name}}. STRICT GUIDELINES: - DIRECT ANSWER ONLY: Do NOT include greetings, introductions, or any conversational filler. Start the very first sentence with the answer. - NO META-COMMENTARY: Do NOT say "Sure, here is the answer" or "I can help with that". - Tone: Highly direct, professional, and authoritative. - {{persona_feature}} - CLEAN TEXT: Use ONLY plain text and LaTeX math. ABSOLUTELY NO markdown stars (*), bold markers (**), or headers (#). - MAXIMUM ACCURACY: Double-check all facts, dates, and formulas. Provide the definitive correct answer. - {{persona_context}} Context: {{context}} Student Inquiry: {{message}} """ persona_feature = "Technical Guidance: Provide step-by-step practical instructions where applicable." if user_lvl == 'Vocational' else "Math Formulas: Use professional LaTeX formatting for all mathematical equations. Use $ formula $ for inline and $$ formula $$ for large block equations." prompt = prompt_template.replace("{persona_name}", persona_name) \ .replace("{persona_role}", persona_role) \ .replace("{persona_goal}", persona_goal) \ .replace("{lang_name}", lang_name) \ .replace("{persona_feature}", persona_feature) \ .replace("{persona_context}", persona_context) \ .replace("{context}", context) \ .replace("{message}", data.get('message', '')) print("Calling Groq API...") bot_response = call_groq_api(prompt) print(f"Groq API returned: {bot_response[:100]}...") bot_msg = ChatMessage(user_id=user_id, message=bot_response, is_bot=True) db.session.add(bot_msg) # Reward progress if user: user.study_time = (user.study_time or 0) + 2 user.overall_progress = min(100, (user.overall_progress or 0) + 0.05) db.session.commit() print("<<< CHAT SEND REQUEST END SUCCESS") return jsonify({ "success": True, "data": bot_msg.to_dict() }) except Exception as e: print(f"!!! CRITICAL CHAT ERROR: {str(e)}") import traceback traceback.print_exc() return jsonify({ "success": False, "message": f"Server encountered an error processing your chat: {str(e)}" }), 500 @chat_bp.route('/history', methods=['GET']) @jwt_required() def get_history(): user_id = int(get_jwt_identity()) messages = ChatMessage.query.filter_by(user_id=user_id).order_by(ChatMessage.timestamp.asc()).all() return jsonify({ "success": True, "data": [m.to_dict() for m in messages] }) @chat_bp.route('/clear', methods=['DELETE']) @jwt_required() def clear_history(): user_id = int(get_jwt_identity()) ChatMessage.query.filter_by(user_id=user_id).delete() db.session.commit() return jsonify({ "success": True, "message": "Chat history cleared" })