mammouth-models-dashboard/update_models.py

import os
import requests
import json
import time
from dotenv import load_dotenv

load_dotenv("../.env.global")

MAMMOUTH_APIKEY = os.getenv("MAMMOUTH_APIKEY")
AIANALASYS_APIKEY = os.getenv("AIANALASYS_APIKEY")

def get_mammouth_models():
    # Retour au point d'accès OpenRouter compatible qui est plus stable
    url = "https://openrouter.ai/api/v1/models"
    headers = {"Authorization": f"Bearer {MAMMOUTH_APIKEY}"}
    try:
        response = requests.get(url, headers=headers)
        response.raise_for_status()
        return response.json().get('data', [])
    except Exception as e:
        print(f"Error fetching Mammouth models: {e}")
        return []

def get_aa_data():
    url = "https://artificialanalysis.ai/api/v2/data/llms/models"
    headers = {"x-api-key": AIANALASYS_APIKEY}
    try:
        response = requests.get(url, headers=headers)
        response.raise_for_status()
        return response.json().get('data', [])
    except Exception as e:
        print(f"Error fetching Artificial Analysis data: {e}")
        return []

def generate_markdown(models_data):
    categories = {}
    for m in models_data:
        cat = m.get('category', 'General')
        if cat not in categories:
            categories[cat] = []
        categories[cat].append(m)

    md = "# Table des Modèles Mammouth.ai\n\n"
    md += "*Mise à jour automatique via Artificial Analysis & Mammouth API*\n\n"
    md += f"Dernière mise à jour : {time.strftime('%Y-%m-%d %H:%M:%S')}\n\n"

    order = ['Coding', 'Agents', 'General']
    sorted_cats = sorted(categories.keys(), key=lambda x: order.index(x) if x in order else 99)

    for cat in sorted_cats:
        models = categories[cat]
        # Tri par score (desc) puis prix (asc)
        models.sort(key=lambda x: (x.get('score') or 0, -(x.get('price_in') or 999)), reverse=True)
        
        md += f"## {cat}\n\n"
        md += "| Modèle | Prix (In / Out / 1M) | Intelligence Index | Vitesse (TPS) |\n"
        md += "| :--- | :--- | :--- | :--- |\n"
        for m in models:
            p_in = f"${m['price_in']:.2f}"
            p_out = f"${m['price_out']:.2f}"
            score = f"**{m['score']:.1f}**" if m['score'] else "N/A"
            speed = f"{m['speed']:.1f}" if m['speed'] else "N/A"
            md += f"| {m['name']} | {p_in} / {p_out} | {score} | {speed} |\n"
        md += "\n"
    return md

def main():
    print("Fetching models...")
    mammouth_models = get_mammouth_models()
    aa_raw = get_aa_data()
    
    # Mapping AA : Priorité aux noms exacts
    aa_map_exact = {m.get('model_name', '').lower(): m for m in aa_raw}
    aa_map_id = {m.get('model_id', '').lower(): m for m in aa_raw}

    enriched_models = []
    for m in mammouth_models:
        m_id = m.get('id', '')
        m_name = m.get('name', '').lower()
        short_id = m_id.split('/')[-1].lower()
        
        # Mapping plus strict pour éviter les scores identiques
        aa_info = aa_map_id.get(m_id.lower()) or aa_map_exact.get(m_name) or aa_map_id.get(short_id)
        
        # Si toujours pas de match, on ne fait PAS de recherche par sous-chaîne floue
        # pour éviter de polluer les données. On ne match que si le nom est très proche.
        if not aa_info:
            for name, info in aa_map_exact.items():
                if name in m_name and len(name) > 0.8 * len(m_name):
                    aa_info = info
                    break

        pricing = m.get('pricing', {})
        try:
            price_in = float(pricing.get('prompt', 0)) * 1000000
            price_out = float(pricing.get('completion', 0)) * 1000000
        except:
            price_in, price_out = 0, 0
            
        score = None
        speed = None
        if aa_info:
            evals = aa_info.get('evaluations', {})
            score = evals.get('artificial_analysis_intelligence_index')
            speed = aa_info.get('median_output_tokens_per_second')
            
        category = "General"
        m_lower = (m_id + m_name).lower()
        if any(x in m_lower for x in ['coding', 'code', 'starcoder', 'coder']):
            category = "Coding"
        elif any(x in m_lower for x in ['agent', 'hermes', 'tool', 'function']):
            category = "Agents"

        enriched_models.append({
            'name': m.get('name', m_id),
            'price_in': price_in,
            'price_out': price_out,
            'score': score,
            'speed': speed,
            'category': category
        })

    # Filtrer les modèles inutiles (prix nul et pas de score)
    final_list = [m for m in enriched_models if m['price_in'] > 0 or m['score'] is not None]
    
    with open("README.md", "w", encoding="utf-8") as f:
        f.write(generate_markdown(final_list))
    
    print(f"Done! {len(final_list)} models processed.")

if __name__ == "__main__":
    main()