config.py

import os
from typing import Dict, Any
try:
    from dotenv import load_dotenv  # type: ignore
    load_dotenv()
except Exception:
    # dotenv is optional; if not installed, skip silently
    pass


# Local LLM settings
class ModelConfig:
    LOCAL_WRITER_MODEL = os.getenv("LOCAL_WRITER_MODEL", "llama3.1:8b")
    LOCAL_RESEARCHER_MODEL = os.getenv("LOCAL_RESEARCHER_MODEL", "llama3.1:8b")
    OLLAMA_BASE_URL = os.getenv("OLLAMA_BASE_URL", "http://localhost:11434")

    # Fallback to Groq if local not available (optional)
    GROQ_WRITER = "llama3-70b-8192"
    GROQ_RESEARCHER = "mixtral-8x7b-32768"

# Model parameters
def get_model_params(role: str) -> dict:
    """Get model parameters for a specific role"""
    return {
        "temperature": float(os.getenv(f"{role.upper()}_TEMPERATURE", 
                                      "0.7" if role == "writer" else "0.3")),
        "max_tokens": int(os.getenv(f"{role.upper()}_MAX_TOKENS", 
                                   "4000" if role == "writer" else "2000")),
        "top_k": int(os.getenv(f"{role.upper()}_TOP_K", "40")),
        "top_p": float(os.getenv(f"{role.upper()}_TOP_P", "0.9"))
    }


# - Add summarization settings
class SummarizationConfig:
    MAX_CONTENT_LENGTH = 30000
    CHUNK_SIZE = 4000
    HF_MODEL = "facebook/bart-large-cnn"  # Fallback model
    ENABLE_HF_FALLBACK = True  # Set to False if you want pure local LLM only
    FAST_THRESHOLD = 800
    LLM_THRESHOLD = 1500


# ADD VALIDATION
def validate_environment() -> Dict[str, Any]:
    """Validate all required environment variables and dependencies"""
    issues = []
    warnings = []

    # Check local LLM availability
    try:
        import requests
        response = requests.get("http://localhost:11434/api/tags", timeout=5)
        if response.status_code != 200:
            issues.append("Ollama not running on localhost:11434")
    except:
        issues.append("Cannot connect to Ollama. Please ensure it's installed and running.")

    # Check required APIs
    if not os.getenv("TAVILY_API_KEY") and not os.getenv("PERPLEXITY_API_KEY"):
        warnings.append("No web search API configured. Research capabilities will be limited.")

    # Check plagiarism threshold
    plagiarism_threshold = int(os.getenv("PLAGIARISM_THRESHOLD", 15))
    if plagiarism_threshold < 5 or plagiarism_threshold > 50:
        issues.append(f"PLAGIARISM_THRESHOLD should be between 5 and 50, got {plagiarism_threshold}")

    # Check EmergentMind API (optional but recommended for Arxiv)
    if not os.getenv("EMERGENTMIND_API_KEY"):
        warnings.append("No EmergentMind API key configured. Arxiv research will use basic search.")
    
    # Check Google Trends API (optional but recommended for SEO)
    if not os.getenv("GOOGLE_TRENDS_API_KEY"):
        warnings.append("No Google Trends API key configured. SEO keyword research will use estimation.")

    return {
        "valid": len(issues) == 0,
        "issues": issues,
        "warnings": warnings
    }


# Add to config.py
VALIDATION_RESULT = validate_environment()

# Other configurations remain the same
PLAGIARISM_THRESHOLD = int(os.getenv("PLAGIARISM_THRESHOLD", 15))

# Research configuration
RESEARCH_QUERY_COUNT = int(os.getenv("RESEARCH_QUERY_COUNT", 3))
RESEARCH_MAX_TOKENS = int(os.getenv("RESEARCH_MAX_TOKENS", 900000))
FREE_TIER_CREDITS = int(os.getenv("FREE_TIER_CREDITS", 100))

# EmergentMind API configuration
EMERGENTMIND_API_KEY = os.getenv("EMERGENTMIND_API_KEY", "")
EMERGENTMIND_DAILY_LIMIT = int(os.getenv("EMERGENTMIND_DAILY_LIMIT", 25))

# Google Trends API configuration
GOOGLE_TRENDS_API_KEY = os.getenv("GOOGLE_TRENDS_API_KEY", "")