[feat] eval: add EvalResult aggregate type; ignore _assets/

Two small additions paired with the in-progress eval orchestration:

- types.py: EvalResult dataclass with summary/per_video fields plus
  from_raw / save / print helpers. Used by scripts and any future
  callers that aggregate per-sample MetricResults into a corpus-level
  summary (e.g. scripts/eval/run_vbench_e2e.py).

- .gitignore: exclude fastvideo/eval/_assets/ which holds calibration
  data, downloaded videos, and run dumps used by the synthetic-flow
  evaluator that we don't want to track.

Authored by dataloader.

.gitignore

@@ -85,6 +85,9 @@ docs/distillation/examples/
 dmd_t2v_output/
 preprocess_output_text/
 
+# Synthetic-flow eval assets (calibration data, downloaded videos, run dumps)
+fastvideo/eval/_assets/
+
 # Next.js / Node artifacts under ui/: see ui/.gitignore
 
 .claude/

fastvideo/eval/types.py

@@ -1,6 +1,8 @@
 from __future__ import annotations
 
+import json
 from dataclasses import dataclass, field
+from pathlib import Path
 from typing import Any
 
 
@@ -14,3 +16,36 @@ class MetricResult:
     name: str
     score: float | None
     details: dict[str, Any] = field(default_factory=dict)
+
+
+@dataclass
+class EvalResult:
+    """Aggregate output of :meth:`Evaluator.evaluate_dataset`.
+
+    ``summary`` maps metric name → mean score and includes a synthetic
+    ``"overall_average"`` key. ``per_video`` is one row per scored video
+    with at least ``prompt``, ``video``, and ``scores`` fields.
+    """
+    summary: dict[str, float]
+    per_video: list[dict[str, Any]]
+
+    @classmethod
+    def from_raw(cls, by_metric: dict[str, list[float]],
+                 per_video: list[dict[str, Any]]) -> EvalResult:
+        summary = {m: sum(v) / len(v) for m, v in by_metric.items() if v}
+        if summary:
+            summary["overall_average"] = sum(summary.values()) / len(summary)
+        return cls(summary=summary, per_video=per_video)
+
+    def save(self, path: str | Path) -> None:
+        path = Path(path)
+        path.parent.mkdir(parents=True, exist_ok=True)
+        path.write_text(json.dumps(
+            {"summary": self.summary, "per_video": self.per_video},
+            indent=2))
+
+    def print(self) -> None:
+        print()
+        print("== summary ==")
+        for k, v in sorted(self.summary.items()):
+            print(f"  {k:42s} {v:.4f}")