mirror of
https://github.com/countbot-ai/CountBot.git
synced 2026-09-14 20:46:47 +08:00
RIC
28b635177a
BM25Index already supports jieba and falls back to per-character tokenization when it is missing. Because the dependency was commented out, a default 'pip install -r requirements.txt' silently ran the degraded path, so the wiki search quality users actually get is well below what the code (and our published eval numbers in #107) assume. Measured impact on the 60-question eval (with vs without jieba): - direct-question recall@5: 0.80 -> 0.53 - paraphrased-question recall@5: 0.20 -> 0.10 (no top-1 hit at all) - exact-fact recall@10: 1.00 -> 0.80 One-line change, no code changes: the graceful fallback stays for anyone who deliberately removes it. Refs: #107
67 lines
2.2 KiB
Plaintext
67 lines
2.2 KiB
Plaintext
# ============================================
|
||
# CountBot 依赖配置
|
||
# ============================================
|
||
|
||
# ── 核心框架 ──────────────────────────────
|
||
fastapi>=0.115.0
|
||
uvicorn==0.32.0
|
||
sqlalchemy>=2.0.36
|
||
pydantic>=2.10.0
|
||
pydantic-settings>=2.7.0
|
||
loguru>=0.7.3
|
||
python-dotenv>=1.0.1
|
||
|
||
# ── HTTP 客户端 ───────────────────────────
|
||
httpx>=0.28.0
|
||
requests>=2.32.0
|
||
|
||
# ── LLM 相关 ─────────────────────────────
|
||
anthropic>=0.40.0
|
||
openai>=1.58.0
|
||
|
||
# ── 数据库 ───────────────────────────────
|
||
aiosqlite>=0.20.0
|
||
|
||
# ── 定时任务 ─────────────────────────────
|
||
croniter>=5.0.0
|
||
|
||
# ── 桌面客户端 ────────────────────────────
|
||
pywebview>=5.3
|
||
|
||
# ── 网页内容提取 ──────────────────────────
|
||
beautifulsoup4>=4.14.2
|
||
|
||
# ── Wiki 知识库 ────────────────────────────
|
||
python-frontmatter>=1.0.0
|
||
PyYAML>=6.0
|
||
|
||
# ── 渠道 SDK ─────────────────────────────
|
||
qq-botpy>=1.2.1
|
||
dingtalk-stream>=0.24.3
|
||
lark-oapi>=1.5.3
|
||
python-telegram-bot>=20.7
|
||
websockets>=12.0
|
||
|
||
# ============================================
|
||
# 可选依赖（按需安装）
|
||
# ============================================
|
||
|
||
# ── 增强网页抓取（反爬虫、JS渲染）─────────
|
||
# 注意：scrapling 包体积较大（~100MB），包含 playwright 等依赖
|
||
# 不安装不影响运行，会自动回退到 httpx（基础 HTTP 请求）
|
||
# scrapling>=0.3.6
|
||
|
||
# ── 极强反爬虫（仅在 scrapling 安装后需要）──
|
||
# camoufox>=0.4.11
|
||
|
||
# ── MCP 客户端（默认关闭，按需安装）──
|
||
# 不安装不影响运行，MCP 功能默认关闭
|
||
mcp>=1.0.0
|
||
|
||
# ── Wiki 知识库中文分词（可选，提升搜索质量）──
|
||
# 不安装会自动回退为单字分词，不影响基本功能
|
||
jieba>=0.42
|
||
|
||
|
||
|