chore: auto-commit (95 archivos)

- cmd/fn/doctor.go
- cmd/fn/main.go
- cpp/apps/primitives_gallery/playground/tables/CMakeLists.txt
- cpp/apps/primitives_gallery/playground/tables/data_table.cpp
- cpp/apps/primitives_gallery/playground/tables/data_table_logic.cpp
- cpp/apps/primitives_gallery/playground/tables/data_table_logic.h
- cpp/apps/primitives_gallery/playground/tables/self_test.cpp
- cpp/apps/primitives_gallery/playground/tables/tql.cpp
- cpp/apps/primitives_gallery/playground/tables/viz.cpp
- cpp/apps/primitives_gallery/playground/tables/viz.h
- ...

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

python/functions/ml/diffusers_generate.py

@@ -0,0 +1,98 @@
+"""diffusers_generate — ejecuta inferencia con un pipeline diffusers y retorna ImageGenResult."""
+
+from __future__ import annotations
+
+import sys
+import os
+import time
+from typing import Any
+
+sys.path.insert(0, os.path.dirname(__file__))
+
+from generation_config import GenerationConfig
+from image_gen_result import ImageGenResult
+from genconfig_to_diffusers_kwargs import genconfig_to_diffusers_kwargs
+
+
+def diffusers_generate(pipe: Any, cfg: GenerationConfig) -> ImageGenResult:
+    """Ejecuta inferencia con un pipeline diffusers y retorna ImageGenResult.
+
+    Convierte el GenerationConfig a kwargs via genconfig_to_diffusers_kwargs,
+    crea un torch.Generator con la semilla configurada, mide duracion y pico
+    de VRAM (si CUDA disponible). El campo meta del resultado incluye backend,
+    modelo, sampler, seed y steps usados.
+
+    Args:
+        pipe:  Pipeline diffusers cargado (resultado de diffusers_load_pipeline).
+               Debe ser callable: pipe(prompt=..., ...) -> objeto con .images[0].
+        cfg:   Parametros de generacion. cfg.seed se usa para torch.Generator.
+               cfg.model.name se incluye en meta. cfg.sampler se incluye en meta
+               pero NO se aplica aqui — usar diffusers_set_scheduler antes si se
+               quiere cambiar el sampler.
+
+    Returns:
+        ImageGenResult con image=PIL.Image, meta con keys backend/model/sampler/
+        actual_steps/seed, duration_ms y vram_peak_mb (None si no hay CUDA).
+
+    Raises:
+        ImportError: Si torch o diffusers no estan instalados.
+        RuntimeError: Si la inferencia falla (OOM, modelo incompatible, etc.).
+    """
+    try:
+        import torch
+    except ImportError as exc:
+        raise ImportError(
+            "diffusers_generate requiere torch. "
+            "Instalar con: pip install torch"
+        ) from exc
+
+    # Determinar device del pipeline
+    device = "cpu"
+    if hasattr(pipe, "device"):
+        device = str(pipe.device)
+
+    # Medir VRAM solo en CUDA
+    cuda_available = torch.cuda.is_available()
+    if cuda_available:
+        torch.cuda.reset_peak_memory_stats()
+
+    # Construir kwargs desde GenerationConfig
+    kwargs = genconfig_to_diffusers_kwargs(cfg)
+
+    # Crear generator con semilla
+    seed = cfg.seed if cfg.seed >= 0 else int(time.time()) % (2**32)
+    generator = torch.Generator(device=device).manual_seed(seed)
+    kwargs["generator"] = generator
+
+    # Inferencia
+    t0 = time.perf_counter()
+    result = pipe(**kwargs)
+    t1 = time.perf_counter()
+
+    duration_ms = int((t1 - t0) * 1000)
+
+    # VRAM peak
+    vram_peak_mb: int | None = None
+    if cuda_available:
+        vram_peak_mb = torch.cuda.max_memory_allocated() // 1024 // 1024
+
+    # Nombre del modelo
+    model_name = cfg.model.name if cfg.model and hasattr(cfg.model, "name") else "unknown"
+
+    meta: dict[str, Any] = {
+        "backend": "diffusers",
+        "model": model_name,
+        "sampler": cfg.sampler,
+        "actual_steps": cfg.steps,
+        "seed": seed,
+        "width": cfg.width,
+        "height": cfg.height,
+        "cfg_scale": cfg.cfg_scale,
+    }
+
+    return ImageGenResult(
+        image=result.images[0],
+        meta=meta,
+        duration_ms=duration_ms,
+        vram_peak_mb=vram_peak_mb,
+    )