chore: script para listar modelos que respondem para chat

scripts/list_chat_models.py

@@ -0,0 +1,111 @@
+#!/usr/bin/env python3
+"""Lista todos os modelos disponíveis que efetivamente respondem para chat."""
+
+import time
+import json
+from pathlib import Path
+
+import httpx
+
+from thc_cli.core.config import load_config
+from thc_cli.core.providers import get_provider
+
+PROVIDERS_TO_TEST = ["mistral", "gemini", "openrouter", "huggingface", "github_models"]
+
+
+def _get_hf_key():
+    painel_path = Path.home() / '.painel_keys'
+    painel_data = json.loads(painel_path.read_text(encoding='utf-8'))
+    return next((v.get('chave') if isinstance(v, dict) else v for k, v in painel_data.items() if 'hf' in k.lower() or 'hugging' in k.lower()), None)
+
+
+def _format_error(e: Exception) -> str:
+    text = str(e)
+    if "Rate limited" in text or "429" in text:
+        return "rate_limited"
+    if "413" in text:
+        return "413_payload_too_large"
+    if "403" in text:
+        return "403_forbidden"
+    if "402" in text:
+        return "402_payment_required"
+    if "Name or service not known" in text or "ConnectError" in text:
+        return "network_error"
+    return text[:80]
+
+
+def main():
+    cfg = load_config()
+    results = []
+
+    for provider_name in PROVIDERS_TO_TEST:
+        try:
+            provider = get_provider(provider_name, cfg)
+        except Exception as e:
+            print(f"[{provider_name}] erro ao obter provider: {e}")
+            continue
+
+        try:
+            models = provider.list_models()
+        except Exception as e:
+            print(f"[{provider_name}] erro ao listar modelos: {e}")
+            continue
+
+        print(f"\n[{provider_name}] testando {len(models)} modelos...")
+        for m in models:
+            model_key = m.get("key", "")
+            start = time.time()
+            try:
+                if provider_name == "huggingface":
+                    url = "https://router.huggingface.co/v1/chat/completions"
+                    headers = {
+                        "Authorization": f"Bearer {provider.api_key}",
+                        "Content-Type": "application/json",
+                    }
+                    payload = {
+                        "model": model_key,
+                        "messages": [{"role": "user", "content": "diga oi em uma palavra"}],
+                        "max_tokens": 20,
+                    }
+                    with httpx.Client(timeout=30) as client:
+                        r = client.post(url, headers=headers, json=payload)
+                    data = r.json()
+                    text = data.get("choices", [{}])[0].get("message", {}).get("content", "")
+                    elapsed = time.time() - start
+                    if r.status_code == 200 and text:
+                        results.append((provider_name, model_key, "ok", elapsed, text[:60]))
+                        print(f"  ok {model_key} ({elapsed:.1f}s) -> {text[:40]}")
+                    else:
+                        results.append((provider_name, model_key, f"http_{r.status_code}", elapsed, ""))
+                        print(f"  http_{r.status_code} {model_key} ({elapsed:.1f}s)")
+                else:
+                    reply = provider.chat_completion(
+                        messages=[{"role": "user", "content": "diga oi em uma palavra"}],
+                        model=model_key,
+                        max_tokens=20,
+                    )
+                    text = reply.get("choices", [{}])[0].get("message", {}).get("content", "")
+                    elapsed = time.time() - start
+                    if text:
+                        results.append((provider_name, model_key, "ok", elapsed, text[:60]))
+                        print(f"  ok {model_key} ({elapsed:.1f}s) -> {text[:40]}")
+                    else:
+                        results.append((provider_name, model_key, "sem_texto", elapsed, ""))
+                        print(f"  sem_texto {model_key} ({elapsed:.1f}s)")
+            except Exception as e:
+                elapsed = time.time() - start
+                err = _format_error(e)
+                results.append((provider_name, model_key, err, elapsed, ""))
+                print(f"  {err} {model_key} ({elapsed:.1f}s)")
+
+    print("\n=== MODELOS QUE RESPONDEM PARA CHAT ===")
+    ok = [r for r in results if r[2] == "ok"]
+    for provider_name, model_key, status, elapsed, text in ok:
+        print(f"{provider_name} | {model_key} | {elapsed:.1f}s | {text}")
+
+    print(f"\nTotal respondendo: {len(ok)}/{len(results)}")
+    return ok
+
+
+if __name__ == "__main__":
+    main()