done

2026-04-21 19:33:24 +08:00
parent 9a72566753
commit 3268080401
26 changed files with 1816 additions and 318 deletions
@@ -0,0 +1,72 @@
+"""
+Zero-token backfill:
+  - Mark old posts with prefilter_reason (rt_only / url_only / empty) where applicable
+  - Fill ai_reasoning for posts that have a signal but empty reasoning
+  - Stamp analysis_version='legacy' on anything still missing it
+
+Run once after the schema migration. Safe to re-run.
+"""
+from __future__ import annotations
+
+import asyncio
+import os
+import sys
+from typing import Optional
+
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+from sqlalchemy import select
+
+from app.database import AsyncSessionLocal
+from app.models import Post
+
+
+def classify(text: str) -> Optional[str]:
+    stripped = (text or "").strip()
+    if not stripped:
+        return "empty"
+    if stripped.startswith("RT: https://") and len(stripped) < 60:
+        return "rt_only"
+    if stripped.startswith("https://") and " " not in stripped:
+        return "url_only"
+    return None
+
+
+PREFILTER_MSG = {
+    "rt_only": "Pre-filtered: retweet with no added commentary.",
+    "url_only": "Pre-filtered: bare URL with no text content.",
+    "empty": "Pre-filtered: empty post body.",
+}
+
+
+async def main():
+    async with AsyncSessionLocal() as db:
+        posts = (await db.execute(select(Post))).scalars().all()
+
+        pre = ai_fix = ver = 0
+        for p in posts:
+            reason = classify(p.text)
+            if reason and not p.prefilter_reason:
+                p.prefilter_reason = reason
+                if not p.ai_reasoning:
+                    p.ai_reasoning = PREFILTER_MSG[reason]
+                    ai_fix += 1
+                pre += 1
+
+            # Any remaining signal'd post with empty reasoning: fill a minimal note
+            if p.signal and not p.ai_reasoning:
+                p.ai_reasoning = "(legacy) reasoning not recorded for this post."
+                ai_fix += 1
+
+            if not p.analysis_version:
+                p.analysis_version = "legacy"
+                ver += 1
+
+        await db.commit()
+        print(f"✅ prefilter_reason set on {pre} posts")
+        print(f"✅ ai_reasoning filled on {ai_fix} posts")
+        print(f"✅ analysis_version stamped 'legacy' on {ver} posts")
+
+
+if __name__ == "__main__":
+    asyncio.run(main())
@@ -0,0 +1,86 @@
+"""
+把 AI 信号写回数据库。
+跳过纯RT/URL帖子和已有signal的帖子。
+并发执行提速，自动限速避免API超限。
+用法: python scripts/backfill_signals.py [--limit 500] [--overwrite]
+"""
+import asyncio
+import argparse
+import sys
+import os
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+from app.services.analysis import analyze_post
+from app.database import AsyncSessionLocal
+from app.models import Post
+from sqlalchemy import select
+
+CONCURRENCY = 5   # 并发数，避免触发API限速
+
+
+async def process_one(post: Post, semaphore: asyncio.Semaphore, overwrite: bool) -> str:
+    """分析单条帖子并写回DB，返回状态字符串"""
+    if not overwrite and post.signal is not None:
+        return "skip"
+
+    async with semaphore:
+        analysis = await analyze_post(post.text)
+        await asyncio.sleep(0.1)  # 轻度限速
+
+    async with AsyncSessionLocal() as db:
+        result = await db.execute(select(Post).where(Post.id == post.id))
+        p = result.scalar_one_or_none()
+        if p:
+            p.sentiment         = analysis["sentiment"]
+            p.signal            = analysis["signal"]
+            p.ai_confidence     = analysis["confidence"]
+            p.ai_reasoning      = analysis["reasoning"]
+            p.relevant          = analysis["relevant"]
+            p.prefilter_reason  = analysis.get("prefilter_reason")
+            p.analysis_version  = analysis.get("analysis_version")
+            if analysis["relevant"] and analysis["asset"] and not p.price_impact_asset:
+                p.price_impact_asset = analysis["asset"]
+            await db.commit()
+
+    sig = analysis["signal"]
+    return f"{sig}:{analysis['confidence']}%"
+
+
+async def main(limit: int, overwrite: bool):
+    async with AsyncSessionLocal() as db:
+        result = await db.execute(
+            select(Post).order_by(Post.published_at.desc()).limit(limit)
+        )
+        posts = result.scalars().all()
+
+    total = len(posts)
+    print(f"共 {total} 条帖子，并发={CONCURRENCY}，overwrite={overwrite}\n")
+
+    semaphore = asyncio.Semaphore(CONCURRENCY)
+    tasks = [process_one(p, semaphore, overwrite) for p in posts]
+
+    done = 0
+    buy = short = sell = hold = skipped = 0
+
+    for coro in asyncio.as_completed(tasks):
+        status = await coro
+        done += 1
+        if status == "skip":
+            skipped += 1
+        elif status.startswith("buy"):   buy += 1
+        elif status.startswith("short"): short += 1
+        elif status.startswith("sell"):  sell += 1
+        else:                            hold += 1
+
+        if done % 20 == 0 or done == total:
+            print(f"  进度 {done}/{total} | BUY={buy} SHORT={short} SELL={sell} HOLD={hold} SKIP={skipped}")
+
+    print(f"\n✅ 完成: BUY={buy} SHORT={short} SELL={sell} HOLD={hold} SKIP={skipped}")
+
+
+if __name__ == "__main__":
+    ap = argparse.ArgumentParser()
+    ap.add_argument("--limit", type=int, default=400)
+    ap.add_argument("--overwrite", action="store_true", help="重新分析已有signal的帖子")
+    args = ap.parse_args()
+    asyncio.run(main(args.limit, args.overwrite))
@@ -0,0 +1,84 @@
+"""
+批量回跑语义分析 — 验证新prompt效果
+用法: python scripts/reanalyze_sample.py
+"""
+import asyncio
+import sys
+import os
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+from app.services.analysis import analyze_post
+from app.database import AsyncSessionLocal
+from app.models import Post
+from sqlalchemy import select
+
+
+async def main():
+    async with AsyncSessionLocal() as db:
+        result = await db.execute(
+            select(Post)
+            .order_by(Post.published_at.desc())
+            .limit(60)
+        )
+        posts = result.scalars().all()
+
+    print(f"分析 {len(posts)} 条帖子 (model: claude-sonnet-4-6)\n")
+
+    signals = {"buy": 0, "sell": 0, "short": 0, "hold": 0}
+    relevant_count = 0
+
+    correct_m5, checked_m5 = 0, 0
+    correct_m15, checked_m15 = 0, 0
+    correct_m1h, checked_m1h = 0, 0
+
+    for i, post in enumerate(posts):
+        analysis = await analyze_post(post.text)
+        signals[analysis["signal"]] += 1
+        if analysis["relevant"]:
+            relevant_count += 1
+
+        if analysis["relevant"]:
+            sig = analysis["signal"].upper()
+            conf = analysis["confidence"]
+
+            actual_m5  = post.price_impact_m5
+            actual_m15 = post.price_impact_m15
+            actual_m1h = post.price_impact_m1h
+
+            def dir_ok(actual, signal):
+                if actual is None: return None
+                if signal == "buy":   return actual > 0
+                if signal == "short": return actual < 0
+                return None
+
+            m5_ok  = dir_ok(actual_m5,  analysis["signal"])
+            m15_ok = dir_ok(actual_m15, analysis["signal"])
+            m1h_ok = dir_ok(actual_m1h, analysis["signal"])
+
+            if m5_ok  is not None: checked_m5  += 1; correct_m5  += int(m5_ok)
+            if m15_ok is not None: checked_m15 += 1; correct_m15 += int(m15_ok)
+            if m1h_ok is not None: checked_m1h += 1; correct_m1h += int(m1h_ok)
+
+            # accuracy marks
+            marks = ""
+            if m5_ok  is not None: marks += f" 5m:{'✅' if m5_ok  else '❌'}"
+            if m15_ok is not None: marks += f" 15m:{'✅' if m15_ok else '❌'}"
+            if m1h_ok is not None: marks += f" 1h:{'✅' if m1h_ok else '❌'}"
+
+            print(f"[{i+1:03d}] {post.published_at.strftime('%m-%d %H:%M')} | {sig:5s} {conf:3d}% |{marks}")
+            print(f"      {post.text[:110]}")
+            print(f"      → {analysis['reasoning'][:200]}")
+            print()
+
+        await asyncio.sleep(0.2)
+
+    print("=" * 70)
+    print(f"总计: {len(posts)} 帖 | Relevant: {relevant_count} ({relevant_count/len(posts)*100:.0f}%)")
+    print(f"信号: BUY={signals['buy']} SHORT={signals['short']} SELL={signals['sell']} HOLD={signals['hold']}")
+    print()
+    if checked_m5:  print(f"准确率  5m: {correct_m5}/{checked_m5}  = {correct_m5/checked_m5*100:.0f}%")
+    if checked_m15: print(f"准确率 15m: {correct_m15}/{checked_m15} = {correct_m15/checked_m15*100:.0f}%")
+    if checked_m1h: print(f"准确率  1h: {correct_m1h}/{checked_m1h} = {correct_m1h/checked_m1h*100:.0f}%")
+
+if __name__ == "__main__":
+    asyncio.run(main())
@@ -0,0 +1,128 @@
+"""
+Hyperliquid 接入测试脚本
+
+用法:
+  source venv/bin/activate
+  HL_API_PRIVATE_KEY="0x你的API钱包私钥" \
+  HL_ACCOUNT_ADDRESS="0x你的MetaMask地址" \
+  python scripts/test_hyperliquid.py
+
+会做以下验证（只读，不下单）:
+  1. 连通性检查
+  2. 账户 USDC 余额
+  3. BTC-PERP 当前价格
+  4. 当前持仓
+  5. 模拟开仓参数（不实际下单）
+
+加 --trade 参数才会真正下一个 0.001 BTC 的小仓位并立即平掉（mainnet 慎用）
+"""
+
+import asyncio
+import argparse
+import os
+import sys
+
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+from app.services.hyperliquid import HyperliquidTrader
+
+
+async def main(do_trade: bool):
+    api_key = os.getenv("HL_API_PRIVATE_KEY") or os.getenv("HL_API_KEY", "")
+    account = os.getenv("HL_ACCOUNT_ADDRESS", "")
+
+    if not api_key:
+        print("❌ 缺少 HL_API_PRIVATE_KEY 环境变量")
+        print("   export HL_API_PRIVATE_KEY='0x你的API钱包私钥'")
+        sys.exit(1)
+    if not account:
+        print("❌ 缺少 HL_ACCOUNT_ADDRESS 环境变量")
+        print("   export HL_ACCOUNT_ADDRESS='0x你的MetaMask地址'")
+        sys.exit(1)
+
+    mainnet = os.getenv("HL_MAINNET", "true").lower() != "false"
+    net_label = "MAINNET" if mainnet else "TESTNET"
+    print(f"\n=== Hyperliquid 接入测试 ({net_label}) ===\n")
+
+    trader = HyperliquidTrader(
+        api_private_key=api_key,
+        account_address=account,
+        leverage=3,
+        mainnet=mainnet,
+    )
+
+    # 1. USDC 余额
+    print("① 查询 USDC 余额...")
+    balance = await trader.get_balance()
+    print(f"   可提取余额: ${balance:.2f} USDC")
+    if balance < 10:
+        print("   ⚠️  余额不足 $10，交易可能失败")
+    else:
+        print("   ✅ 余额充足")
+
+    # 2. BTC mid price
+    print("\n② 查询 BTC-PERP 价格...")
+    try:
+        from hyperliquid.info import Info
+        base_url = "https://api.hyperliquid.xyz" if mainnet else "https://api.hyperliquid-testnet.xyz"
+        info = Info(base_url, skip_ws=True)
+        mids = info.all_mids()
+        btc_price = float(mids.get("BTC", 0))
+        print(f"   BTC mid price: ${btc_price:,.2f}")
+        print("   ✅ 价格获取正常")
+    except Exception as e:
+        print(f"   ❌ 价格获取失败: {e}")
+
+    # 3. 当前持仓
+    print("\n③ 查询当前持仓...")
+    positions = await trader.get_open_positions()
+    if positions:
+        for p in positions:
+            side = "多" if p["szi"] > 0 else "空"
+            print(f"   {p['coin']} {side}仓: {abs(p['szi'])} 张 @ 均价 ${p['entry_px']:,.2f}  未实现盈亏: ${p['unrealized_pnl']:.2f}")
+        print(f"   共 {len(positions)} 个持仓")
+    else:
+        print("   无持仓 ✅")
+
+    # 4. 模拟开仓参数
+    print("\n④ 模拟开仓参数（不实际下单）...")
+    sz_dec = await trader._get_sz_decimals("BTC")
+    mid = await trader._mid_price("BTC")
+    size_usd = 100
+    size_coins = round(size_usd / mid, sz_dec)
+    limit_buy = round(mid * 1.01, 1)
+    print(f"   开多 $100 notional:")
+    print(f"   → 数量: {size_coins} BTC")
+    print(f"   → 限价(1%滑点): ${limit_buy:,.1f}")
+    print(f"   → 杠杆: 3x 隔离保证金")
+    print(f"   → 所需保证金: ${size_usd / 3:.2f} USDC")
+
+    # 5. 真实测试下单（需要 --trade 参数）
+    if do_trade:
+        print(f"\n⑤ 真实下单测试 (0.001 BTC long → 立即平仓)...")
+        if not mainnet:
+            print("   使用 testnet，无需担心资金损失")
+        else:
+            print("   ⚠️  MAINNET 真实资金！继续中...")
+
+        try:
+            result = await trader.open_position("BTC", "long", size_usd=50)
+            print(f"   ✅ 开仓成功: fill_price=${result['fill_price']:,.2f}, size={result['size_coins']} BTC")
+
+            await asyncio.sleep(2)
+
+            close = await trader.close_position("BTC")
+            print(f"   ✅ 平仓成功: fill_price=${close['fill_price']:,.2f}")
+        except Exception as e:
+            print(f"   ❌ 下单失败: {e}")
+    else:
+        print("\n提示: 加 --trade 参数进行真实下单+平仓测试")
+
+    print("\n=== 测试完成 ===")
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--trade", action="store_true", help="执行真实下单+平仓测试")
+    args = parser.parse_args()
+    asyncio.run(main(args.trade))