dsh-stock-analysis 0.2.5 → 0.2.6
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/dist/index.js +101 -9
- package/package.json +1 -1
- package/schemas/report_schema.json +3 -3
- package/scripts/setup-python.mjs +6 -6
- package/skills/bull-trend/SKILL.md +34 -10
- package/skills/bull-trend/references/fat-tails.md +91 -0
- package/skills/bull-trend/scripts/tail_analysis.py +132 -0
- package/skills/event-driven/SKILL.md +2 -0
- package/skills/growth-quality/SKILL.md +2 -0
- package/skills/stock-analysis/SKILL.md +9 -3
- package/skills/stock-screener/SKILL.md +2 -0
- package/skills/strategy-backtest/SKILL.md +4 -0
- package/skills/wisburg-research/SKILL.md +2 -0
- package/templates/stock_analysis/brief.md.j2 +4 -1
- package/templates/stock_analysis/full.md.j2 +6 -3
- package/tools/alert_rules.py +32 -4
- package/tools/backtest.py +22 -4
- package/tools/capabilities.py +3 -0
- package/tools/gather.py +20 -10
- package/tools/lixinger.py +100 -0
- package/tools/market_regime.py +7 -3
- package/tools/market_review.py +7 -6
- package/tools/position_context.py +20 -2
- package/tools/report_renderer.py +2 -0
- package/tools/risk_screening.py +45 -17
- package/tools/screener.py +17 -9
- package/tools/sec_edgar.py +346 -0
- package/tools/sec_holders.py +263 -0
- package/tools/signal_tracker.py +142 -19
- package/tools/stock_data.py +80 -17
- package/tools/technical.py +16 -2
- package/tools/value_data.py +159 -0
- package/tools/watchlist.py +24 -1
- package/tools/watchlist_context.py +7 -3
|
@@ -7,6 +7,8 @@ description: 策略回测:定义或读取 AlphaEvo YAML 策略,回测收益
|
|
|
7
7
|
|
|
8
8
|
开始前读取[短线交易决策框架](../stock-analysis/references/short-term-trading.md)的风险单位和回测约束。回测用于反证策略,不把样本内收益包装成预测。
|
|
9
9
|
|
|
10
|
+
解读回测时,`period` 是实际取得的数据期间,`requested_period` 保留用户请求,`warnings` 提示两者差异。年化收益按实际期间计算;覆盖不完整时明确说明,不能声称完成了完整请求期间的验证。买入预算包含佣金,`position.size` 为 0..1 的资金比例。
|
|
11
|
+
|
|
10
12
|
## 工作流程
|
|
11
13
|
|
|
12
14
|
### 第一步:策略定义
|
|
@@ -77,6 +79,8 @@ python3 tools/backtest.py run <path.yaml> <symbol> --start <date> --end <date>
|
|
|
77
79
|
- 使用用户指定的互斥日期区间分别运行样本内和样本外回测。
|
|
78
80
|
- 成本假设引用结果中的 `costs` 字段(实际生效的佣金/滑点/印花税费率与实付费用合计),未模拟的费用与成交约束列入限制。
|
|
79
81
|
|
|
82
|
+
趋势策略或用户要求肥尾收益诊断时,读取 [肥尾度量与趋势收益诊断](../bull-trend/references/fat-tails.md)的盈利集中度部分,基于平仓 `pnl` 计算大赢家贡献和删除敏感性。只有用户询问市场收益分布时才进入 κ 分支。
|
|
83
|
+
|
|
80
84
|
未模拟的结算、涨跌停、复权或特定市场费用列入限制,结论限定为该模拟条件下的历史表现。
|
|
81
85
|
|
|
82
86
|
### 第四步:诊断
|
|
@@ -7,6 +7,8 @@ description: 智堡投研检索:查找机构研报、财报公告、电话会
|
|
|
7
7
|
|
|
8
8
|
智堡 MCP 涵盖海外投行研报、国内券商研究、资管报告、央行/政府文献、电话会纪要和财报公告。首先检查这些 MCP 工具是否可用;不可用时直接改用 `search_comprehensive_intel`。
|
|
9
9
|
|
|
10
|
+
港股估值功课调用 `get_valuation_history(symbol, years=10)`;配置 `LIXINGER_TOKEN` 时优先理杏仁非金融公司每日 PE-TTM/PB,失败回退百度。以返回的实际窗口和正值样本计算分位及 Band,不混接同一指标的不同来源,不将缺失解释为低估。
|
|
11
|
+
|
|
10
12
|
## 可用工具
|
|
11
13
|
|
|
12
14
|
智堡 MCP 提供以下工具(通过 `mcp__wisburg-mcp-server__` 前缀调用):
|
|
@@ -7,11 +7,14 @@
|
|
|
7
7
|
{% if risk_screening and risk_screening.veto_buy %}
|
|
8
8
|
⚠️ **一票否决** — 风险筛查触发,建议谨慎。
|
|
9
9
|
{% endif %}
|
|
10
|
+
{% if risk_screening and (risk_screening.data_complete is sameas false or risk_screening.risk_level == 'unknown') %}
|
|
11
|
+
⚠️ **数据不完整** — 当前风险筛查不能判定为低风险。
|
|
12
|
+
{% endif %}
|
|
10
13
|
|
|
11
14
|
{% set dp = data_perspective or {} -%}
|
|
12
15
|
{% set pp = dp.price_position or {} -%}
|
|
13
16
|
{% set ts = dp.trend_status or {} -%}
|
|
14
|
-
- 📈 **趋势:** {{ ts.ma_alignment or '-' }}{% if ts.is_bullish is
|
|
17
|
+
- 📈 **趋势:** {{ ts.ma_alignment or '-' }}{% if ts.is_bullish is boolean %}({{ '看多' if ts.is_bullish else '看空' }}){% endif %}
|
|
15
18
|
- 💰 **价位:** 现价 {{ m.fmt_num(pp.current_price) }} | 支撑 {{ m.fmt_num(pp.support_level) }} | 压力 {{ m.fmt_num(pp.resistance_level) }}
|
|
16
19
|
|
|
17
20
|
{% set bp = battle_plan or {} -%}
|
|
@@ -26,12 +26,15 @@
|
|
|
26
26
|
## 风险筛查
|
|
27
27
|
|
|
28
28
|
{% if risk_screening -%}
|
|
29
|
-
- **风险等级:** {{ m.risk_emoji(risk_screening.risk_level) }} {{ risk_screening.risk_level or '-' }} | **风险分:** {{ risk_screening.risk_score }}/100
|
|
29
|
+
- **风险等级:** {{ m.risk_emoji(risk_screening.risk_level) }} {{ risk_screening.risk_level or '-' }} | **风险分:** {{ m.fmt_num(risk_screening.risk_score, 0) }}/100
|
|
30
|
+
{% if risk_screening.data_complete is sameas false or risk_screening.risk_level == 'unknown' -%}
|
|
31
|
+
- **数据不完整:** {{ risk_screening.warnings | join(';') }},当前结果不能判定为低风险。
|
|
32
|
+
{% endif %}
|
|
30
33
|
{% if risk_screening.flags -%}
|
|
31
34
|
- **风险标记:**
|
|
32
35
|
{% for f in risk_screening.flags %} - [{{ f.severity }}] {{ f.category }}:{{ f.description }}
|
|
33
36
|
{% endfor %}
|
|
34
|
-
{%-
|
|
37
|
+
{%- elif risk_screening.data_complete is not sameas false and risk_screening.risk_level != 'unknown' -%}
|
|
35
38
|
- **风险标记:** 无显著风险标记
|
|
36
39
|
{%- endif %}
|
|
37
40
|
{%- else -%}
|
|
@@ -45,7 +48,7 @@
|
|
|
45
48
|
{% set pp = dp.price_position or {} -%}
|
|
46
49
|
{% set vol = dp.volume_analysis or {} -%}
|
|
47
50
|
{% set chip = dp.chip_structure or {} -%}
|
|
48
|
-
- **趋势:** {{ ts.ma_alignment or '-' }}{% if ts.is_bullish is
|
|
51
|
+
- **趋势:** {{ ts.ma_alignment or '-' }}{% if ts.is_bullish is boolean %}({{ '看多' if ts.is_bullish else '看空' }}){% endif %}{% if ts.trend_score is not none %} 趋势分 {{ m.fmt_num(ts.trend_score) }}{% endif %}
|
|
49
52
|
- **价格结构:** 现价 {{ m.fmt_num(pp.current_price) }} | 偏离 MA5 {{ m.change_pct(pp.bias_ma5) }} | 支撑 {{ m.fmt_num(pp.support_level) }} | 压力 {{ m.fmt_num(pp.resistance_level) }}
|
|
50
53
|
- **量价:** 量比 {{ m.fmt_num(vol.volume_ratio) }}{% if vol.volume_status %}({{ vol.volume_status }}){% endif %}{% if vol.turnover_rate is not none %} | 换手率 {{ m.change_pct(vol.turnover_rate) }}{% endif %}
|
|
51
54
|
{% if chip and (chip.profit_ratio is not none or chip.avg_cost is not none) -%}
|
package/tools/alert_rules.py
CHANGED
|
@@ -6,6 +6,7 @@ Does NOT store history, does NOT push, does NOT schedule. Host agent decides whe
|
|
|
6
6
|
|
|
7
7
|
import argparse
|
|
8
8
|
import json
|
|
9
|
+
import math
|
|
9
10
|
import sys
|
|
10
11
|
from concurrent.futures import ThreadPoolExecutor
|
|
11
12
|
from datetime import datetime, timezone
|
|
@@ -31,8 +32,14 @@ RULE_TYPES = {
|
|
|
31
32
|
def _check_rule(rule: dict, quote: dict, tech: dict, anom: dict, risk: dict) -> dict | None:
|
|
32
33
|
rtype = rule.get("type")
|
|
33
34
|
value = rule.get("value")
|
|
34
|
-
if rtype not in RULE_TYPES:
|
|
35
|
+
if not isinstance(rtype, str) or rtype not in RULE_TYPES:
|
|
35
36
|
return {"rule": rtype, "triggered": False, "error": f"unknown rule type: {rtype}"}
|
|
37
|
+
if rtype in ("price_below", "price_above", "change_pct_above", "change_pct_below", "volume_ratio_above"):
|
|
38
|
+
try:
|
|
39
|
+
if isinstance(value, bool) or not math.isfinite(float(value)):
|
|
40
|
+
raise ValueError
|
|
41
|
+
except (ValueError, TypeError):
|
|
42
|
+
return {"rule": rtype, "triggered": False, "error": "Rule value must be a finite number"}
|
|
36
43
|
|
|
37
44
|
price = quote.get("price")
|
|
38
45
|
change_pct = quote.get("change_pct")
|
|
@@ -90,6 +97,8 @@ def _check_rule(rule: dict, quote: dict, tech: dict, anom: dict, risk: dict) ->
|
|
|
90
97
|
"message": f"量比 {ratio} 超过 {value}",
|
|
91
98
|
}
|
|
92
99
|
elif rtype == "anomaly":
|
|
100
|
+
if not isinstance(value, str) or not value:
|
|
101
|
+
return {"rule": rtype, "triggered": False, "error": "Anomaly type must be a nonempty string"}
|
|
93
102
|
# value is an anomaly type string, e.g. "macd_golden_cross"
|
|
94
103
|
anomalies = (anom or {}).get("anomalies", [])
|
|
95
104
|
for a in anomalies:
|
|
@@ -106,6 +115,8 @@ def _check_rule(rule: dict, quote: dict, tech: dict, anom: dict, risk: dict) ->
|
|
|
106
115
|
return {"rule": rtype, "triggered": True, "severity": "high", "message": "风险筛查触发一票否决"}
|
|
107
116
|
elif rtype == "risk_level_at_least":
|
|
108
117
|
order = {"low": 0, "medium": 1, "high": 2}
|
|
118
|
+
if not isinstance(value, str) or value not in order:
|
|
119
|
+
return {"rule": rtype, "triggered": False, "error": "Risk threshold must be low/medium/high"}
|
|
109
120
|
cur = order.get((risk or {}).get("risk_level"))
|
|
110
121
|
want = order.get(value)
|
|
111
122
|
if cur is not None and want is not None and cur >= want:
|
|
@@ -122,7 +133,7 @@ def _check_rule(rule: dict, quote: dict, tech: dict, anom: dict, risk: dict) ->
|
|
|
122
133
|
|
|
123
134
|
|
|
124
135
|
def check_rules(symbol: str, rules: list[dict]) -> dict:
|
|
125
|
-
if not rules:
|
|
136
|
+
if not isinstance(rules, list) or not rules or any(not isinstance(rule, dict) for rule in rules):
|
|
126
137
|
return {"symbol": symbol, "error": "no rules provided"}
|
|
127
138
|
|
|
128
139
|
needs_tech = any(r.get("type") == "volume_ratio_above" for r in rules)
|
|
@@ -141,6 +152,9 @@ def check_rules(symbol: str, rules: list[dict]) -> dict:
|
|
|
141
152
|
with ThreadPoolExecutor(max_workers=len(jobs)) as pool:
|
|
142
153
|
futures = {key: pool.submit(_run_json, script, args) for key, (script, args) in jobs.items()}
|
|
143
154
|
fetched = {key: future.result() for key, future in futures.items()}
|
|
155
|
+
fetched = {
|
|
156
|
+
key: value if isinstance(value, dict) and "error" not in value else None for key, value in fetched.items()
|
|
157
|
+
}
|
|
144
158
|
quote = fetched["quote"] or {}
|
|
145
159
|
tech = fetched.get("tech")
|
|
146
160
|
anom = fetched.get("anom")
|
|
@@ -153,13 +167,25 @@ def check_rules(symbol: str, rules: list[dict]) -> dict:
|
|
|
153
167
|
warnings.append("technical analysis fetch failed — volume_ratio rules cannot fire")
|
|
154
168
|
if needs_anom and not anom:
|
|
155
169
|
warnings.append("anomaly fetch failed — anomaly rules cannot fire")
|
|
156
|
-
if needs_risk and not risk:
|
|
170
|
+
if needs_risk and (not risk or risk.get("data_complete") is False or risk.get("risk_level") == "unknown"):
|
|
157
171
|
warnings.append("risk screening fetch failed — risk rules cannot fire")
|
|
158
172
|
|
|
159
173
|
hits = []
|
|
160
174
|
evaluated = []
|
|
161
175
|
for rule in rules:
|
|
162
|
-
|
|
176
|
+
rtype = rule.get("type")
|
|
177
|
+
missing = (
|
|
178
|
+
(rtype in ("price_below", "price_above") and quote.get("price") is None)
|
|
179
|
+
or (rtype in ("change_pct_below", "change_pct_above") and quote.get("change_pct") is None)
|
|
180
|
+
or (rtype == "volume_ratio_above" and (tech or {}).get("volume", {}).get("volume_ratio") is None)
|
|
181
|
+
or (rtype == "anomaly" and not isinstance((anom or {}).get("anomalies"), list))
|
|
182
|
+
or (rtype == "risk_veto" and (risk or {}).get("veto_buy") is None)
|
|
183
|
+
or (rtype == "risk_level_at_least" and (risk or {}).get("risk_level") not in ("low", "medium", "high"))
|
|
184
|
+
)
|
|
185
|
+
if missing:
|
|
186
|
+
result = {"rule": rtype, "triggered": False, "error": "Required source data unavailable"}
|
|
187
|
+
else:
|
|
188
|
+
result = _check_rule(rule, quote, tech, anom, risk)
|
|
163
189
|
if result is None:
|
|
164
190
|
evaluated.append({"rule": rule.get("type"), "triggered": False})
|
|
165
191
|
else:
|
|
@@ -214,6 +240,8 @@ def main():
|
|
|
214
240
|
result = check_rules(args.symbol, rules)
|
|
215
241
|
json.dump(result, sys.stdout, ensure_ascii=False, indent=2, default=str)
|
|
216
242
|
print()
|
|
243
|
+
if "error" in result:
|
|
244
|
+
sys.exit(1)
|
|
217
245
|
|
|
218
246
|
|
|
219
247
|
if __name__ == "__main__":
|
package/tools/backtest.py
CHANGED
|
@@ -16,6 +16,7 @@ import yaml
|
|
|
16
16
|
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
|
|
17
17
|
|
|
18
18
|
from _subproc import json_safe, run_tool, utf8_stdio
|
|
19
|
+
from stock_data import detect_market
|
|
19
20
|
from technical import bollinger_series, ema_series, ma_series, macd_series, rsi_series
|
|
20
21
|
|
|
21
22
|
# --------------- Strategy Loading ---------------
|
|
@@ -67,7 +68,7 @@ def fetch_kline(symbol: str, start: str | None, end: str | None) -> pd.DataFrame
|
|
|
67
68
|
if start and end:
|
|
68
69
|
d0 = datetime.strptime(start, "%Y-%m-%d")
|
|
69
70
|
d1 = datetime.strptime(end, "%Y-%m-%d")
|
|
70
|
-
count = max((d1 - d0).days + 60, 300)
|
|
71
|
+
count = max((datetime.now() - d0).days + 60, (d1 - d0).days + 60, 300)
|
|
71
72
|
|
|
72
73
|
# run_tool gives venv-python + a bounded timeout + None on any failure
|
|
73
74
|
# (non-zero exit, empty/invalid stdout); callers turn that into clean JSON.
|
|
@@ -251,7 +252,11 @@ def evaluate_conditions(df: pd.DataFrame, conditions: list, logic: str, idx: int
|
|
|
251
252
|
|
|
252
253
|
|
|
253
254
|
def simulate(df: pd.DataFrame, strategy: dict, capital: float, symbol: str) -> dict:
|
|
254
|
-
|
|
255
|
+
if not math.isfinite(capital) or capital <= 0:
|
|
256
|
+
raise ValueError("capital must be finite and positive")
|
|
257
|
+
if not np.isfinite(df["close"]).all() or (df["close"] <= 0).any():
|
|
258
|
+
raise ValueError("backtest prices must be finite and positive")
|
|
259
|
+
market = detect_market(symbol)
|
|
255
260
|
lot_size = 100 if market == "A" else 1
|
|
256
261
|
sim_cfg = strategy.get("simulation", {})
|
|
257
262
|
slippage = float(sim_cfg.get("slippage", 0.001))
|
|
@@ -268,6 +273,8 @@ def simulate(df: pd.DataFrame, strategy: dict, capital: float, symbol: str) -> d
|
|
|
268
273
|
exit_ = strategy.get("exit", {})
|
|
269
274
|
position_cfg = strategy.get("position", {})
|
|
270
275
|
size_frac = float(position_cfg.get("size", 1.0))
|
|
276
|
+
if not 0 <= size_frac <= 1:
|
|
277
|
+
raise ValueError("position.size must be a fraction in [0, 1]")
|
|
271
278
|
stop_loss = float(exit_.get("stop_loss", -0.10))
|
|
272
279
|
take_profit = float(exit_.get("take_profit", 0.50))
|
|
273
280
|
ts_val = exit_.get("trailing_stop")
|
|
@@ -297,7 +304,7 @@ def simulate(df: pd.DataFrame, strategy: dict, capital: float, symbol: str) -> d
|
|
|
297
304
|
if hit:
|
|
298
305
|
buy_price = price * (1 + slippage)
|
|
299
306
|
buy_amount = cash * size_frac
|
|
300
|
-
shares = int(buy_amount / buy_price / lot_size) * lot_size
|
|
307
|
+
shares = int(buy_amount / (buy_price * (1 + commission)) / lot_size) * lot_size
|
|
301
308
|
if shares <= 0:
|
|
302
309
|
continue
|
|
303
310
|
cost = shares * buy_price * (1 + commission)
|
|
@@ -417,7 +424,7 @@ def compute_metrics(trades: list, equity_curve: list, initial: float, final: flo
|
|
|
417
424
|
years = 1
|
|
418
425
|
annual_return = (1 + total_return) ** (1 / max(years, 0.01)) - 1 if total_return > -1 else total_return
|
|
419
426
|
|
|
420
|
-
equities = pd.Series([e["equity"] for e in equity_curve], dtype=float)
|
|
427
|
+
equities = pd.Series([initial, *[e["equity"] for e in equity_curve]], dtype=float)
|
|
421
428
|
peak = equities.cummax()
|
|
422
429
|
dd = (equities - peak) / peak
|
|
423
430
|
max_drawdown = float(dd.min()) if len(dd) > 0 else 0
|
|
@@ -653,6 +660,13 @@ def run_backtest(strategy_path: str, symbol: str, start: str | None, end: str |
|
|
|
653
660
|
if len(df) < 30:
|
|
654
661
|
return {"error": f"Insufficient data: {len(df)} rows (need at least 30)"}
|
|
655
662
|
|
|
663
|
+
requested_period = f"{start} to {end}"
|
|
664
|
+
actual_start = df["date"].iloc[0].strftime("%Y-%m-%d")
|
|
665
|
+
actual_end = df["date"].iloc[-1].strftime("%Y-%m-%d")
|
|
666
|
+
warnings = []
|
|
667
|
+
if actual_start != start or actual_end != end:
|
|
668
|
+
warnings.append(f"Actual data covers {actual_start} to {actual_end}; requested {requested_period}")
|
|
669
|
+
start, end = actual_start, actual_end
|
|
656
670
|
sim = simulate(df, strategy, capital, symbol)
|
|
657
671
|
final = sim["final_equity"]
|
|
658
672
|
metrics = compute_metrics(sim["trades"], sim["equity_curve"], capital, final, start, end)
|
|
@@ -661,6 +675,8 @@ def run_backtest(strategy_path: str, symbol: str, start: str | None, end: str |
|
|
|
661
675
|
"strategy": strategy.get("name", "unnamed"),
|
|
662
676
|
"symbol": symbol,
|
|
663
677
|
"period": f"{start} to {end}",
|
|
678
|
+
"requested_period": requested_period,
|
|
679
|
+
"warnings": warnings,
|
|
664
680
|
"initial_capital": capital,
|
|
665
681
|
"final_capital": final,
|
|
666
682
|
"costs": sim["costs"],
|
|
@@ -837,6 +853,8 @@ def main():
|
|
|
837
853
|
sys.exit(1)
|
|
838
854
|
|
|
839
855
|
print(json.dumps(json_safe(result), ensure_ascii=False, default=str))
|
|
856
|
+
if "error" in result:
|
|
857
|
+
sys.exit(1)
|
|
840
858
|
|
|
841
859
|
|
|
842
860
|
if __name__ == "__main__":
|
package/tools/capabilities.py
CHANGED
|
@@ -18,6 +18,9 @@ ALL_MARKETS = ["A", "HK", "US", "JP", "KR", "TW"]
|
|
|
18
18
|
|
|
19
19
|
# tool_name -> {markets: [supported markets], reason: str for unsupported}
|
|
20
20
|
TOOL_MATRIX = {
|
|
21
|
+
"get_announcements": {"markets": ["A"], "reason": "cninfo A-share disclosures only"},
|
|
22
|
+
"get_us_holdings": {"markets": ["US"], "reason": "SEC manager holdings or verified issuer-CUSIP reverse lookup"},
|
|
23
|
+
"get_insider_trades": {"markets": ["US"], "reason": "SEC issuer Form 4 only"},
|
|
21
24
|
"get_valuation_history": {"markets": ["A", "HK"], "reason": "A/HK historical valuation providers only"},
|
|
22
25
|
"get_index_valuation": {"markets": ["A"], "reason": "Chinese index valuation only"},
|
|
23
26
|
"get_earnings_calendar": {
|
package/tools/gather.py
CHANGED
|
@@ -10,6 +10,7 @@ from pathlib import Path
|
|
|
10
10
|
sys.path.insert(0, str(Path(__file__).resolve().parent))
|
|
11
11
|
from _subproc import json_safe, run_tool, utf8_stdio # noqa: E402
|
|
12
12
|
from stock_data import detect_market # noqa: E402
|
|
13
|
+
from technical import analyze_records # noqa: E402
|
|
13
14
|
|
|
14
15
|
|
|
15
16
|
# gather fans out the heavier CLIs, so the default timeout stays 60s; raw stdout
|
|
@@ -32,36 +33,41 @@ def _parse_json(raw: str | None):
|
|
|
32
33
|
return raw
|
|
33
34
|
|
|
34
35
|
|
|
35
|
-
def _run_all(tasks: dict) -> dict:
|
|
36
|
+
def _run_all(tasks: dict, symbol: str) -> dict:
|
|
36
37
|
with ThreadPoolExecutor(max_workers=len(tasks)) as pool:
|
|
37
38
|
futures = {pool.submit(_run, script, args): key for key, (script, args) in tasks.items()}
|
|
38
|
-
|
|
39
|
+
result = {key: _parse_json(f.result()) for f, key in futures.items()}
|
|
40
|
+
result["technical"] = analyze_records(symbol, result["kline"]) if result["kline"] is not None else None
|
|
41
|
+
return result
|
|
39
42
|
|
|
40
43
|
|
|
41
44
|
def gather_analysis(symbol: str) -> dict:
|
|
45
|
+
market = detect_market(symbol)
|
|
42
46
|
tasks = {
|
|
43
47
|
"quote": ("stock_data.py", ["quote", symbol]),
|
|
44
48
|
"kline": ("stock_data.py", ["kline", symbol, "--period", "daily", "--count", "120"]),
|
|
45
|
-
"technical": ("technical.py", ["analyze", symbol, "--period", "daily", "--count", "120"]),
|
|
46
49
|
"financials": ("stock_data.py", ["financials", symbol]),
|
|
47
50
|
"capital_flow": ("stock_data.py", ["capital_flow", symbol]),
|
|
48
51
|
"chip_distribution": ("stock_data.py", ["chip_distribution", symbol]),
|
|
49
52
|
"news": ("search_intel.py", ["search", _news_query(symbol)]),
|
|
50
53
|
"social_sentiment": ("search_intel.py", ["sentiment", symbol]),
|
|
51
54
|
"risk": ("risk_screening.py", ["screen", symbol]),
|
|
52
|
-
"regime": ("market_regime.py", ["detect"]),
|
|
53
55
|
}
|
|
54
|
-
|
|
56
|
+
if market in ("A", "HK", "US"):
|
|
57
|
+
tasks["regime"] = ("market_regime.py", ["detect", market])
|
|
58
|
+
result = _run_all(tasks, symbol)
|
|
59
|
+
if market not in ("A", "HK", "US"):
|
|
60
|
+
result["regime"] = {"error": f"Market regime unavailable for {market}"}
|
|
61
|
+
return result
|
|
55
62
|
|
|
56
63
|
|
|
57
64
|
def gather_technical(symbol: str, kline_count: int = 120, with_quote: bool = False) -> dict:
|
|
58
65
|
tasks = {
|
|
59
66
|
"kline": ("stock_data.py", ["kline", symbol, "--period", "daily", "--count", str(kline_count)]),
|
|
60
|
-
"technical": ("technical.py", ["analyze", symbol, "--period", "daily", "--count", str(kline_count)]),
|
|
61
67
|
}
|
|
62
68
|
if with_quote:
|
|
63
69
|
tasks["quote"] = ("stock_data.py", ["quote", symbol])
|
|
64
|
-
return _run_all(tasks)
|
|
70
|
+
return _run_all(tasks, symbol)
|
|
65
71
|
|
|
66
72
|
|
|
67
73
|
def gather_screen(market: str = "A", top: int = 20, config: str | None = None) -> dict:
|
|
@@ -73,16 +79,20 @@ def gather_screen(market: str = "A", top: int = 20, config: str | None = None) -
|
|
|
73
79
|
|
|
74
80
|
|
|
75
81
|
def gather_fundamental(symbol: str) -> dict:
|
|
82
|
+
market = detect_market(symbol)
|
|
76
83
|
tasks = {
|
|
77
84
|
"quote": ("stock_data.py", ["quote", symbol]),
|
|
78
85
|
"kline": ("stock_data.py", ["kline", symbol, "--period", "daily", "--count", "60"]),
|
|
79
|
-
"technical": ("technical.py", ["analyze", symbol, "--period", "daily", "--count", "60"]),
|
|
80
86
|
"financials": ("stock_data.py", ["financials", symbol]),
|
|
81
87
|
"news": ("search_intel.py", ["search", _news_query(symbol)]),
|
|
82
88
|
"stock_info": ("stock_data.py", ["stock_info", symbol]),
|
|
83
|
-
"sector_rankings": ("stock_data.py", ["sector_rankings", "--top", "5", "--direction", "both"]),
|
|
84
89
|
}
|
|
85
|
-
|
|
90
|
+
if market == "A":
|
|
91
|
+
tasks["sector_rankings"] = ("stock_data.py", ["sector_rankings", "--top", "5", "--direction", "both"])
|
|
92
|
+
result = _run_all(tasks, symbol)
|
|
93
|
+
if market != "A":
|
|
94
|
+
result["sector_rankings"] = {"error": f"Sector rankings unavailable for {market}"}
|
|
95
|
+
return result
|
|
86
96
|
|
|
87
97
|
|
|
88
98
|
def main():
|
|
@@ -0,0 +1,100 @@
|
|
|
1
|
+
"""Optional Lixinger HK non-financial datasets (same env token as index data)."""
|
|
2
|
+
|
|
3
|
+
import os
|
|
4
|
+
from datetime import date, timedelta
|
|
5
|
+
|
|
6
|
+
import requests
|
|
7
|
+
from _cache import DAY, cached
|
|
8
|
+
|
|
9
|
+
|
|
10
|
+
def hk_data(symbol, dataset, metrics):
|
|
11
|
+
code = symbol.split(".")[0].zfill(5)
|
|
12
|
+
end = date.today()
|
|
13
|
+
try:
|
|
14
|
+
start = end.replace(year=end.year - 10)
|
|
15
|
+
except ValueError:
|
|
16
|
+
start = end.replace(year=end.year - 10, day=28) + timedelta(days=1)
|
|
17
|
+
|
|
18
|
+
def fetch():
|
|
19
|
+
response = requests.post(
|
|
20
|
+
f"https://open.lixinger.com/api/hk/company/{dataset}/non_financial",
|
|
21
|
+
json={
|
|
22
|
+
"token": os.environ["LIXINGER_TOKEN"],
|
|
23
|
+
"stockCodes": [code],
|
|
24
|
+
"startDate": start.isoformat(),
|
|
25
|
+
"endDate": end.isoformat(),
|
|
26
|
+
"metricsList": metrics,
|
|
27
|
+
},
|
|
28
|
+
timeout=20,
|
|
29
|
+
)
|
|
30
|
+
response.raise_for_status()
|
|
31
|
+
payload = response.json()
|
|
32
|
+
if payload.get("code") != 1 or not isinstance(payload.get("data"), list) or not payload["data"]:
|
|
33
|
+
raise ValueError("Lixinger HK non-financial data unavailable; check token permissions/company type")
|
|
34
|
+
return payload["data"]
|
|
35
|
+
|
|
36
|
+
return cached(f"lixinger-hk:{dataset}:{code}:{end}:{','.join(metrics)}", DAY, fetch)
|
|
37
|
+
|
|
38
|
+
|
|
39
|
+
def financials_hk(symbol, periods, annual):
|
|
40
|
+
fields = {
|
|
41
|
+
"roe": "m.roe",
|
|
42
|
+
"gross_margin": "ps.gp_m",
|
|
43
|
+
"net_profit_margin": "ps.np_s_r",
|
|
44
|
+
"total_revenue": "ps.toi",
|
|
45
|
+
"net_income": "ps.np",
|
|
46
|
+
"operating_cash_flow": "cfs.ncffoa",
|
|
47
|
+
}
|
|
48
|
+
data = hk_data(symbol, "fs", [f"q.{field}.t" for field in fields.values()])
|
|
49
|
+
history = []
|
|
50
|
+
for row in sorted(data, key=lambda r: r.get("date", ""), reverse=True):
|
|
51
|
+
if annual and row.get("reportType") != "annual_report":
|
|
52
|
+
continue
|
|
53
|
+
if not row.get("date") or row.get("stockCode", symbol.split(".")[0].zfill(5)) != symbol.split(".")[0].zfill(5):
|
|
54
|
+
continue
|
|
55
|
+
entry = {
|
|
56
|
+
"report_date": row["date"][:10],
|
|
57
|
+
"disclosure_date": (row.get("reportDate") or "")[:10] or None,
|
|
58
|
+
"report_type": row.get("reportType"),
|
|
59
|
+
"currency": row.get("currency"),
|
|
60
|
+
"source": "lixinger",
|
|
61
|
+
}
|
|
62
|
+
for key, field in fields.items():
|
|
63
|
+
table, metric = field.split(".")
|
|
64
|
+
value = row.get("q", {}).get(table, {}).get(metric, {}).get("t")
|
|
65
|
+
entry[key] = (
|
|
66
|
+
value * 100 if value is not None and key in ("roe", "gross_margin", "net_profit_margin") else value
|
|
67
|
+
)
|
|
68
|
+
if any(entry.get(key) is not None for key in fields):
|
|
69
|
+
history.append(entry)
|
|
70
|
+
if not history:
|
|
71
|
+
raise ValueError("Lixinger returned no matching financial periods")
|
|
72
|
+
history = history[:periods]
|
|
73
|
+
latest = history[0]
|
|
74
|
+
result = {
|
|
75
|
+
**dict.fromkeys(("name", "market_cap", "pe", "forward_pe", "pb", "debt_to_equity", "dividend_yield")),
|
|
76
|
+
"snapshot_available": False,
|
|
77
|
+
"symbol": symbol,
|
|
78
|
+
"source": "lixinger",
|
|
79
|
+
"report_date": latest["report_date"],
|
|
80
|
+
"currency": latest["currency"],
|
|
81
|
+
"total_revenue": latest["total_revenue"],
|
|
82
|
+
"net_income": latest["net_income"],
|
|
83
|
+
"roe": latest["roe"] / 100 if latest["roe"] is not None else None,
|
|
84
|
+
"profit_margin": latest["net_profit_margin"] / 100 if latest["net_profit_margin"] is not None else None,
|
|
85
|
+
"periods": history,
|
|
86
|
+
"requested_periods": periods,
|
|
87
|
+
"available_periods": len(history),
|
|
88
|
+
"history_units": {
|
|
89
|
+
"roe": "percent",
|
|
90
|
+
"gross_margin": "percent",
|
|
91
|
+
"net_profit_margin": "percent",
|
|
92
|
+
"total_revenue": "reported currency",
|
|
93
|
+
"net_income": "reported currency",
|
|
94
|
+
"operating_cash_flow": "reported currency",
|
|
95
|
+
},
|
|
96
|
+
"accounting_basis": "provider-reported HK financial statements; not normalized to A-share accounting",
|
|
97
|
+
}
|
|
98
|
+
if len(history) < periods:
|
|
99
|
+
result["warnings"] = ["Requested financial history incomplete; only actual available periods returned"]
|
|
100
|
+
return result
|
package/tools/market_regime.py
CHANGED
|
@@ -54,7 +54,9 @@ def _get_index_kline(market: str) -> list:
|
|
|
54
54
|
def _compute_indicators(klines: list) -> dict:
|
|
55
55
|
import numpy as np
|
|
56
56
|
|
|
57
|
-
|
|
57
|
+
df = pd.DataFrame(klines).reindex(columns=["close", "high", "low"]).apply(pd.to_numeric, errors="coerce")
|
|
58
|
+
df = df[np.isfinite(df).all(axis=1) & (df > 0).all(axis=1)]
|
|
59
|
+
closes = df["close"].to_numpy()
|
|
58
60
|
if len(closes) < 20:
|
|
59
61
|
return {}
|
|
60
62
|
|
|
@@ -65,8 +67,8 @@ def _compute_indicators(klines: list) -> dict:
|
|
|
65
67
|
ma60 = float(ma_series(close_s, 60).iloc[-1]) if len(closes) >= 60 else float(np.mean(closes))
|
|
66
68
|
close = float(closes[-1])
|
|
67
69
|
|
|
68
|
-
highs =
|
|
69
|
-
lows =
|
|
70
|
+
highs = df["high"].to_numpy()
|
|
71
|
+
lows = df["low"].to_numpy()
|
|
70
72
|
|
|
71
73
|
tr_values = []
|
|
72
74
|
for i in range(-min(14, len(closes) - 1), 0):
|
|
@@ -226,6 +228,8 @@ def main():
|
|
|
226
228
|
if args.command == "detect":
|
|
227
229
|
result = detect_regime(args.market)
|
|
228
230
|
print(json.dumps(result, ensure_ascii=False, default=str))
|
|
231
|
+
if "error" in result:
|
|
232
|
+
sys.exit(1)
|
|
229
233
|
|
|
230
234
|
|
|
231
235
|
if __name__ == "__main__":
|
package/tools/market_review.py
CHANGED
|
@@ -19,7 +19,7 @@ sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
|
|
|
19
19
|
|
|
20
20
|
from _subproc import run_tool as _run_tool
|
|
21
21
|
from _subproc import utf8_stdio
|
|
22
|
-
from signal_tracker import _load_records
|
|
22
|
+
from signal_tracker import _append_record, _load_records, _store_lock
|
|
23
23
|
|
|
24
24
|
|
|
25
25
|
def calc_temperature(stats: dict, indices: list) -> dict:
|
|
@@ -145,10 +145,9 @@ def save_review(report: dict, store: Path | None = None) -> dict:
|
|
|
145
145
|
"""Append one review report object to the JSONL archive (side effect of render --save)."""
|
|
146
146
|
store = store or _default_store()
|
|
147
147
|
try:
|
|
148
|
-
store
|
|
149
|
-
|
|
150
|
-
|
|
151
|
-
except OSError as e:
|
|
148
|
+
with _store_lock(store):
|
|
149
|
+
_append_record(store, report)
|
|
150
|
+
except (OSError, ValueError) as e:
|
|
152
151
|
return {"error": str(e)}
|
|
153
152
|
return {"saved": True, "date": report.get("review_date") or report.get("date")}
|
|
154
153
|
|
|
@@ -156,7 +155,9 @@ def save_review(report: dict, store: Path | None = None) -> dict:
|
|
|
156
155
|
def cmd_history(args) -> dict:
|
|
157
156
|
"""Recent archived reviews, newest first. Empty store -> clean empty list."""
|
|
158
157
|
try:
|
|
159
|
-
|
|
158
|
+
store = Path(args.store) if args.store else _default_store()
|
|
159
|
+
with _store_lock(store):
|
|
160
|
+
records = _load_records(store)
|
|
160
161
|
except (OSError, UnicodeDecodeError) as e:
|
|
161
162
|
# store is a directory / not UTF-8 — same clean-error contract as save_review
|
|
162
163
|
return {"error": str(e)}
|
|
@@ -6,6 +6,7 @@ Not a portfolio system. Stateless: user passes cost/quantity/stops each call.
|
|
|
6
6
|
|
|
7
7
|
import argparse
|
|
8
8
|
import json
|
|
9
|
+
import math
|
|
9
10
|
import sys
|
|
10
11
|
from datetime import datetime, timezone
|
|
11
12
|
from pathlib import Path
|
|
@@ -48,13 +49,28 @@ def analyze_position(
|
|
|
48
49
|
stop_loss: float | None = None,
|
|
49
50
|
take_profit: float | None = None,
|
|
50
51
|
) -> dict:
|
|
52
|
+
for field, value in (
|
|
53
|
+
("cost", cost),
|
|
54
|
+
("quantity", quantity),
|
|
55
|
+
("stop_loss", stop_loss),
|
|
56
|
+
("take_profit", take_profit),
|
|
57
|
+
):
|
|
58
|
+
if value is not None and (not isinstance(value, (int, float)) or not math.isfinite(value) or value <= 0):
|
|
59
|
+
return {"symbol": symbol, "error": f"{field} must be finite and positive"}
|
|
51
60
|
quote = _run_json("stock_data.py", ["quote", symbol]) or {}
|
|
52
61
|
tech = _run_json("technical.py", ["analyze", symbol, "--period", "daily", "--count", "60"]) or {}
|
|
53
62
|
|
|
54
|
-
price = quote.get("price")
|
|
63
|
+
price = quote.get("price") if isinstance(quote, dict) and "error" not in quote else None
|
|
55
64
|
if price is None:
|
|
56
65
|
return {"symbol": symbol, "error": "cannot fetch current price"}
|
|
57
|
-
|
|
66
|
+
try:
|
|
67
|
+
price = float(price)
|
|
68
|
+
except (TypeError, ValueError):
|
|
69
|
+
return {"symbol": symbol, "error": "current price must be finite and positive"}
|
|
70
|
+
if not math.isfinite(price) or price <= 0:
|
|
71
|
+
return {"symbol": symbol, "error": "current price must be finite and positive"}
|
|
72
|
+
if not isinstance(tech, dict):
|
|
73
|
+
tech = {}
|
|
58
74
|
|
|
59
75
|
pnl_per_share = price - cost
|
|
60
76
|
pnl_pct = (pnl_per_share / cost * 100) if cost else 0.0
|
|
@@ -133,6 +149,8 @@ def main():
|
|
|
133
149
|
)
|
|
134
150
|
json.dump(result, sys.stdout, ensure_ascii=False, indent=2, default=str)
|
|
135
151
|
print()
|
|
152
|
+
if "error" in result:
|
|
153
|
+
sys.exit(1)
|
|
136
154
|
|
|
137
155
|
|
|
138
156
|
if __name__ == "__main__":
|
package/tools/report_renderer.py
CHANGED