Files
Youssef faf7d56233 Eval harness: direct-API bench matrix, screenshots, sloplint + vision-judge scoring, compare gallery
Arms: Anthropic baseline, GLM on Together, Kimi on OpenRouter (DeepSeek
and Qwen staged, disabled). Six briefs incl. metric-temptation,
component-scope, and vision-probe probes. Floor-only vs full skill packs
to measure prompt bloat. run-matrix orchestrates gen, shots, score,
judge, report; the repair loop feeds sloplint FAILs back to the model.
2026-07-23 13:08:45 +01:00

24 lines
771 B
Python

#!/usr/bin/env python3
"""serve.py: local static server for the eval gallery.
Serves the eval/ directory on http://127.0.0.1:4201 (loopback only).
PORT env var overrides the port.
"""
import functools
import http.server
import os
import socketserver
from pathlib import Path
port = int(os.environ.get("PORT", "4201"))
root = str(Path(__file__).resolve().parent)
Handler = functools.partial(http.server.SimpleHTTPRequestHandler, directory=root)
socketserver.ThreadingTCPServer.allow_reuse_address = True
socketserver.ThreadingTCPServer.daemon_threads = True
if __name__ == "__main__":
with socketserver.ThreadingTCPServer(("127.0.0.1", port), Handler) as httpd:
print(f"serving {root} on http://127.0.0.1:{port}", flush=True)
httpd.serve_forever()