1
2026-07-19 228d838fdb7f7dde7edc4993fdbb9654c9c31df7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
from __future__ import annotations
 
import hashlib
import json
import os
import re
from pathlib import Path
 
import pandas as pd
import pymysql
from PIL import Image, ImageDraw, ImageFont
 
 
RUN_ID = "RUN-ANA-WUJI-BASELINE-PILOT-20260607-001"
ROOT = Path(__file__).resolve().parents[1]
LOCAL_DB_INDEX = Path(
    r"D:\strategy_project\s-system-doc\observer\天下模型沉淀\数据库索引数据.md"
)
 
 
def read_password() -> str:
    env = os.environ.get("TIANXIA_MYSQL_PASSWORD") or os.environ.get("MYSQL_PWD")
    if env:
        return env
    text = LOCAL_DB_INDEX.read_text(encoding="utf-8")
    match = re.search(r"^\s*-\s*密码:`([^`]+)`", text, re.MULTILINE)
    if not match:
        raise RuntimeError("Unable to read local MySQL credential from approved local index.")
    return match.group(1)
 
 
def get_conn():
    return pymysql.connect(
        host="127.0.0.1",
        port=3306,
        user="root",
        password=read_password(),
        database="tianxia",
        charset="utf8mb4",
        connect_timeout=5,
        read_timeout=120,
    )
 
 
def sha256_file(path: Path) -> str:
    h = hashlib.sha256()
    with path.open("rb") as f:
        for chunk in iter(lambda: f.read(1024 * 1024), b""):
            h.update(chunk)
    return h.hexdigest()
 
 
def font(size: int):
    for p in [
        Path("C:/Windows/Fonts/msyh.ttc"),
        Path("C:/Windows/Fonts/simhei.ttf"),
        Path("C:/Windows/Fonts/simsun.ttc"),
    ]:
        if p.exists():
            return ImageFont.truetype(str(p), size)
    return ImageFont.load_default()
 
 
FONT_TITLE = font(26)
FONT_MID = font(17)
FONT_SMALL = font(13)
 
 
def short_id(value: object) -> str:
    return str(value).replace(f"LOT-{RUN_ID}-", "LOT-")
 
 
def signal_label(signal_type: str) -> str:
    return {
        "STOP5_SELL_SIGNAL_CANDIDATE": "-5%止损候选",
        "TREND_TAKE_PROFIT_SIGNAL_CANDIDATE": "趋势止盈候选",
        "THREE_HIGH_NOT_RISING_SIGNAL_CANDIDATE": "三高不升候选",
        "HOLD_REVIEW_CANDIDATE": "继续持仓复核",
    }.get(signal_type, signal_type)
 
 
def wrap_text(text: object, max_chars: int = 22) -> list[str]:
    lines: list[str] = []
    for raw in str(text).splitlines():
        item = raw.strip()
        if not item:
            lines.append("")
            continue
        while len(item) > max_chars:
            lines.append(item[:max_chars])
            item = item[max_chars:]
        lines.append(item)
    return lines
 
 
def normalize_time(value) -> str:
    text = str(value)
    if "days" in text:
        text = text.split()[-1]
    if "." in text:
        text = text.split(".")[0]
    parts = text.split(":")
    if len(parts) >= 3:
        return f"{int(parts[0]):02d}:{int(parts[1]):02d}:{int(float(parts[2])):02d}"
    return text
 
 
def y_price(value: float, low: float, high: float, top: int, bottom: int) -> int:
    if high <= low:
        return (top + bottom) // 2
    return bottom - int((value - low) / (high - low) * (bottom - top))
 
 
def draw_daily_exit_chart(df: pd.DataFrame, lot: pd.Series, signal: dict, out_path: Path) -> None:
    w, h = 1500, 860
    img = Image.new("RGB", (w, h), "#fbfbf7")
    d = ImageDraw.Draw(img)
    d.rectangle([0, 0, w - 1, h - 1], outline="#cbd5e1")
    title = f"持仓/卖点日K复核:{lot.symbol}  入场 {lot.entry_trade_date}"
    d.text((32, 24), title, fill="#111827", font=FONT_TITLE)
    d.text((32, 58), "当前为卖点信号材料,不直接生成 SELL;真实卖出需 AI/人工账本裁决。", fill="#334155", font=FONT_MID)
    left, top, right, bottom = 80, 105, 1060, 590
    vol_top, vol_bottom = 640, 790
    note_left, note_top = 1090, 110
    d.rectangle([left, top, right, bottom], outline="#94a3b8")
    d.rectangle([left, vol_top, right, vol_bottom], outline="#94a3b8")
    if df.empty:
        d.text((left + 200, top + 180), "无日线数据", fill="#b91c1c", font=FONT_TITLE)
        price_low, price_high = 0, 1
    else:
        price_low = min(float(df.low_price.min()), float(lot.entry_price) * 0.95) * 0.98
        price_high = max(float(df.high_price.max()), float(lot.entry_price) * 1.05) * 1.02
        max_vol = max(float(df.volume.max()), 1.0)
        n = len(df)
        gap = (right - left) / max(n, 1)
        body_w = max(5, int(gap * 0.55))
        for i, row in df.reset_index(drop=True).iterrows():
            cx = int(left + gap * i + gap / 2)
            op, hi, lo, cl = [float(row[c]) for c in ["open_price", "high_price", "low_price", "close_price"]]
            color = "#dc2626" if cl >= op else "#16a34a"
            d.line([cx, y_price(lo, price_low, price_high, top, bottom), cx, y_price(hi, price_low, price_high, top, bottom)], fill=color, width=2)
            y1, y2 = y_price(op, price_low, price_high, top, bottom), y_price(cl, price_low, price_high, top, bottom)
            d.rectangle([cx - body_w // 2, min(y1, y2), cx + body_w // 2, max(y1, y2)], fill=color, outline=color)
            vh = int(float(row.volume) / max_vol * (vol_bottom - vol_top))
            d.rectangle([cx - body_w // 2, vol_bottom - vh, cx + body_w // 2, vol_bottom], fill=color, outline=color)
            date_s = str(row.trade_date)
            if date_s == str(lot.entry_trade_date):
                d.line([cx, top, cx, vol_bottom], fill="#2563eb", width=2)
                d.text((cx + 5, top + 8), "买入", fill="#2563eb", font=FONT_SMALL)
            if date_s == signal.get("signal_trade_date", ""):
                d.line([cx, top, cx, vol_bottom], fill="#b91c1c", width=2)
                d.text((cx + 5, top + 30), "信号", fill="#b91c1c", font=FONT_SMALL)
            if i % max(1, n // 6) == 0:
                d.text((cx - 28, vol_bottom + 8), date_s[5:], fill="#64748b", font=FONT_SMALL)
        for ref, label, color in [
            (float(lot.entry_price), "买入价", "#2563eb"),
            (float(lot.entry_price) * 0.95, "-5%止损", "#b91c1c"),
        ]:
            yy = y_price(ref, price_low, price_high, top, bottom)
            d.line([left, yy, right, yy], fill=color, width=2)
            d.text((right - 98, yy - 8), f"{label} {ref:.2f}", fill=color, font=FONT_SMALL)
    d.rounded_rectangle([note_left, note_top, 1460, 790], radius=8, outline="#334155", fill="#ffffff")
    d.text((note_left + 18, note_top + 18), "卖点信号", fill="#111827", font=FONT_TITLE)
    notes = [
        f"lot:{short_id(lot.trade_lot_id)}",
        f"入场:{lot.entry_trade_date} {lot.entry_time}",
        f"买入价:{float(lot.entry_price):.2f}",
        f"可卖日:{lot.sellable_from_trade_date}",
        f"信号:{signal_label(signal['signal_type'])}",
        f"信号日:{signal.get('signal_trade_date', '')}",
        "说明:",
        *wrap_text(signal["signal_note_cn"], 20),
        "",
        "本图只准备卖点材料,",
        "不写真实 SELL。",
    ]
    yy = note_top + 58
    for line in notes:
        d.text((note_left + 18, yy), line, fill="#334155", font=FONT_SMALL)
        yy += 24 if line else 12
    img.save(out_path)
 
 
def main() -> None:
    lots = pd.read_csv(ROOT / "position_lot_ledger.csv", encoding="utf-8-sig")
    if lots.empty:
        raise RuntimeError("No open lots found.")
    lots["entry_trade_date"] = pd.to_datetime(lots["entry_trade_date"])
    lots["sellable_from_trade_date"] = pd.to_datetime(lots["sellable_from_trade_date"])
    lots["entry_price"] = pd.to_numeric(lots["entry_price"], errors="coerce")
    symbols = sorted(lots.symbol.unique().tolist())
    min_date = (lots.entry_trade_date.min() - pd.Timedelta(days=90)).strftime("%Y-%m-%d")
    max_date = (lots.entry_trade_date.max() + pd.Timedelta(days=25)).strftime("%Y-%m-%d")
    sym_ph = ",".join(["%s"] * len(symbols))
    with get_conn() as conn:
        daily = pd.read_sql(
            f"""
            SELECT trade_date, symbol, open_price, high_price, low_price, close_price, volume, amount
            FROM a_share_daily_price
            WHERE symbol IN ({sym_ph}) AND trade_date BETWEEN %s AND %s
            ORDER BY symbol, trade_date
            """,
            conn,
            params=[*symbols, min_date, max_date],
        )
        calendar = pd.read_sql(
            """
            SELECT calendar_date
            FROM a_share_trading_calendar
            WHERE is_trading_day=1 AND calendar_date BETWEEN %s AND %s
            ORDER BY calendar_date
            """,
            conn,
            params=[min_date, max_date],
        )
    daily.trade_date = pd.to_datetime(daily.trade_date)
    for col in ["open_price", "high_price", "low_price", "close_price", "volume", "amount"]:
        daily[col] = pd.to_numeric(daily[col], errors="coerce")
    calendar.calendar_date = pd.to_datetime(calendar.calendar_date)
    trade_dates = list(calendar.calendar_date.dt.strftime("%Y-%m-%d"))
 
    signal_rows = []
    manifest_rows = []
    for _, lot in lots.iterrows():
        entry = lot.entry_trade_date.strftime("%Y-%m-%d")
        sellable = lot.sellable_from_trade_date.strftime("%Y-%m-%d")
        idx = trade_dates.index(entry)
        obs_dates = trade_dates[idx : idx + 11]
        obs = daily[(daily.symbol == lot.symbol) & (daily.trade_date.dt.strftime("%Y-%m-%d").isin(obs_dates))].copy()
        sellable_obs = obs[obs.trade_date.dt.strftime("%Y-%m-%d") >= sellable].copy()
        signal = {
            "signal_type": "HOLD_REVIEW_CANDIDATE",
            "signal_trade_date": sellable if not sellable_obs.empty else "",
            "signal_note_cn": "观察窗口内未触发硬止损或趋势止盈代码信号,需人工继续复核卖点。",
            "proxy_or_substitute_used_flag": False,
        }
        if not sellable_obs.empty:
            stop = sellable_obs[sellable_obs.low_price.le(float(lot.entry_price) * 0.95)]
            trend = sellable_obs[
                sellable_obs.high_price.ge(float(lot.entry_price) * 1.05)
                & sellable_obs.close_price.ge(sellable_obs.open_price)
            ]
            three_high_signal = None
            if len(sellable_obs) >= 3:
                for i in range(2, len(sellable_obs)):
                    tri = sellable_obs.iloc[i - 2 : i + 1]
                    highs = list(tri.high_price)
                    if not (highs[0] < highs[1] < highs[2]):
                        three_high_signal = sellable_obs.iloc[i]
                        break
            if not stop.empty:
                r = stop.iloc[0]
                signal = {
                    "signal_type": "STOP5_SELL_SIGNAL_CANDIDATE",
                    "signal_trade_date": r.trade_date.strftime("%Y-%m-%d"),
                    "signal_note_cn": "卖出候选:可卖日后触及单票-5%硬止损线,需AI/人工确认真实SELL。",
                    "proxy_or_substitute_used_flag": False,
                }
            elif not trend.empty:
                r = trend.iloc[0]
                signal = {
                    "signal_type": "TREND_TAKE_PROFIT_SIGNAL_CANDIDATE",
                    "signal_trade_date": r.trade_date.strftime("%Y-%m-%d"),
                    "signal_note_cn": "卖出候选:可卖日后出现相对买入价5%以上趋势性上涨,需AI/人工确认是否按原文止盈卖出。",
                    "proxy_or_substitute_used_flag": False,
                }
            elif three_high_signal is not None:
                signal = {
                    "signal_type": "THREE_HIGH_NOT_RISING_SIGNAL_CANDIDATE",
                    "signal_trade_date": three_high_signal.trade_date.strftime("%Y-%m-%d"),
                    "signal_note_cn": "卖出候选:三日高点未逐步抬高,需AI/人工确认是否卖出。",
                    "proxy_or_substitute_used_flag": False,
                }
        case_dir = ROOT / "cases" / lot.case_id
        img_dir = case_dir / "img"
        img_dir.mkdir(parents=True, exist_ok=True)
        out_path = img_dir / f"05_exit_daily_signal_{lot.symbol.replace('.', '_')}_{entry}.png"
        window = daily[(daily.symbol == lot.symbol) & (daily.trade_date.dt.strftime("%Y-%m-%d") <= obs_dates[-1])].tail(80).copy()
        window["trade_date"] = window.trade_date.dt.strftime("%Y-%m-%d")
        lot_for_chart = lot.copy()
        lot_for_chart.entry_trade_date = entry
        lot_for_chart.sellable_from_trade_date = sellable
        draw_daily_exit_chart(window, lot_for_chart, signal, out_path)
        rel = out_path.relative_to(ROOT).as_posix()
        signal_rows.append(
            {
                "trade_lot_id": lot.trade_lot_id,
                "order_id": lot.order_id,
                "case_id": lot.case_id,
                "symbol": lot.symbol,
                "entry_trade_date": entry,
                "entry_time": lot.entry_time,
                "entry_price": f"{float(lot.entry_price):.4f}",
                "sellable_from_trade_date": sellable,
                "signal_type": signal["signal_type"],
                "signal_trade_date": signal.get("signal_trade_date", ""),
                "signal_note_cn": signal["signal_note_cn"],
                "manual_sell_decision": "PENDING_AI_MANUAL_DECISION",
                "decision_chart_path": rel,
                "t1_guard_passed_flag": signal.get("signal_trade_date", "") >= sellable if signal.get("signal_trade_date", "") else "",
                "proxy_or_substitute_used_flag": signal["proxy_or_substitute_used_flag"],
            }
        )
        manifest_rows.append(
            {
                "case_id": lot.case_id,
                "symbol": lot.symbol,
                "trade_date": signal.get("signal_trade_date", ""),
                "event_id": f"{lot.trade_lot_id}_exit_signal",
                "chart_role": "exit_daily_signal_review_view",
                "decision_time": f"{signal.get('signal_trade_date', '')} close",
                "path": rel,
                "sha256": sha256_file(out_path),
                "status": "PASS",
                "note": "卖点/持仓日K复核图;不自动生成SELL。",
            }
        )
 
    signals = pd.DataFrame(signal_rows)
    signals.to_csv(ROOT / "sell_signal_candidates.csv", index=False, encoding="utf-8-sig")
    root_manifest = pd.read_csv(ROOT / "image_manifest.csv", encoding="utf-8-sig")
    combined = pd.concat([root_manifest, pd.DataFrame(manifest_rows)], ignore_index=True)
    combined = combined.drop_duplicates(subset=["case_id", "symbol", "event_id", "chart_role"], keep="last")
    combined.to_csv(ROOT / "image_manifest.csv", index=False, encoding="utf-8-sig")
 
    for case_id, group in pd.DataFrame(manifest_rows).groupby("case_id"):
        case_dir = ROOT / "cases" / case_id
        board_path = case_dir / "case_image_board.md"
        existing = board_path.read_text(encoding="utf-8") if board_path.exists() else f"# {case_id} 图片审核板\n"
        existing = re.sub(r"\n## 5\. 卖点 / 持仓日K信号图\n[\s\S]*$", "", existing.rstrip())
        lines = [existing.rstrip(), "", "## 5. 卖点 / 持仓日K信号图", ""]
        for _, row in group.iterrows():
            rel = Path(row.path).relative_to(f"cases/{case_id}").as_posix()
            lines.extend([f"### {row.symbol}", "", f"![{row.symbol}]({rel})", "", "- 本图只提示卖点/持仓信号,不写真实 SELL。", ""])
        board_path.write_text("\n".join(lines) + "\n", encoding="utf-8")
        combined[combined.case_id == case_id].to_csv(case_dir / "image_manifest.csv", index=False, encoding="utf-8-sig")
 
    summary = {
        "schema_version": "1.0",
        "run_id": RUN_ID,
        "generated_at": "2026-06-08T01:45:00+08:00",
        "stage": "SELL_SIGNAL_MATERIAL_READY",
        "signal_counts": signals.signal_type.value_counts().to_dict(),
        "signal_rows": len(signals),
        "exit_signal_images": len(manifest_rows),
        "artifacts": {
            "sell_signal_candidates.csv": {
                "size": (ROOT / "sell_signal_candidates.csv").stat().st_size,
                "sha256": sha256_file(ROOT / "sell_signal_candidates.csv"),
            },
            "image_manifest.csv": {
                "size": (ROOT / "image_manifest.csv").stat().st_size,
                "sha256": sha256_file(ROOT / "image_manifest.csv"),
            },
        },
        "boundary": "Signal material only; no SELL orders and no return statistics.",
        "next_step": "AI/manual sell decision from sell_signal_candidates.csv and images.",
    }
    (ROOT / "exit_review_generation_summary.json").write_text(
        json.dumps(summary, ensure_ascii=False, indent=2) + "\n",
        encoding="utf-8",
    )
    (ROOT / "exit_review_generation_summary.md").write_text(
        "\n".join(
            [
                "# exit_review_generation_summary",
                "",
                f"run_id:`{RUN_ID}`",
                "阶段:`SELL_SIGNAL_MATERIAL_READY`",
                "",
                f"- 卖点/持仓信号行:{len(signals)}",
                f"- 卖点/持仓日K图:{len(manifest_rows)}",
                "",
                "当前只生成卖点材料,不产生 SELL 或收益结论。",
                "",
            ]
        ),
        encoding="utf-8",
    )
 
 
if __name__ == "__main__":
    main()