zhaolei
6f402ffcee
- 后端: coworker 智能体框架, WS API, 文件上传, 附件处理 - 前端: Open WebUI, 文件全量走 upload API (含 MD/TXT/JSON 等文本类) - 技能: md-to-office (pandoc + wkhtmltopdf) - 修复: 上传文件路径丢失, Agent 搜索浪费, 输出文件跑到 uploads/ - 打包: PyInstaller one-dir, 预打包 pandoc/wkhtmltopdf/chromium
26 lines
881 B
Python
26 lines
881 B
Python
"""One-off corpus coverage stats for the permission-mode analysis (not shipped)."""
|
|
import json
|
|
import collections
|
|
|
|
for name in ("benign", "dangerous", "injection"):
|
|
tags = collections.Counter()
|
|
tools = collections.Counter()
|
|
correct = collections.Counter()
|
|
holdout = 0
|
|
n = 0
|
|
with open(f"tests/corpora/{name}.jsonl", encoding="utf-8") as fh:
|
|
for line in fh:
|
|
line = line.strip()
|
|
if not line:
|
|
continue
|
|
d = json.loads(line)
|
|
n += 1
|
|
correct[d["correct"]] += 1
|
|
tools[d["action"]["tool"]] += 1
|
|
holdout += bool(d.get("holdout"))
|
|
for t in d.get("tags", []):
|
|
tags[t] += 1
|
|
print(f"== {name}: {n} rows, holdout={holdout}, correct={dict(correct)}")
|
|
print(" tools:", dict(tools))
|
|
print(" tags:", dict(tags.most_common()))
|