stock-analysis-plugin 0.1.10

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.
Files changed (84) hide show
  1. package/README.md +498 -0
  2. package/package.json +88 -0
  3. package/pi/index.ts +1142 -0
  4. package/schemas/market_review_schema.json +176 -0
  5. package/schemas/report_schema.json +131 -0
  6. package/scripts/setup-python.mjs +57 -0
  7. package/scripts/sync-version.sh +56 -0
  8. package/scripts/venv-python.mjs +20 -0
  9. package/skills/bottom-volume/SKILL.md +63 -0
  10. package/skills/bottom-volume/scripts/gather.py +22 -0
  11. package/skills/box-oscillation/SKILL.md +72 -0
  12. package/skills/box-oscillation/scripts/gather.py +22 -0
  13. package/skills/bull-trend/SKILL.md +60 -0
  14. package/skills/bull-trend/scripts/gather.py +22 -0
  15. package/skills/chan-theory/SKILL.md +71 -0
  16. package/skills/chan-theory/scripts/gather.py +22 -0
  17. package/skills/dragon-head/SKILL.md +69 -0
  18. package/skills/dragon-head/scripts/gather.py +22 -0
  19. package/skills/emotion-cycle/SKILL.md +65 -0
  20. package/skills/emotion-cycle/scripts/gather.py +22 -0
  21. package/skills/event-driven/SKILL.md +73 -0
  22. package/skills/event-driven/scripts/gather.py +22 -0
  23. package/skills/expectation-repricing/SKILL.md +72 -0
  24. package/skills/expectation-repricing/scripts/gather.py +22 -0
  25. package/skills/growth-quality/SKILL.md +74 -0
  26. package/skills/growth-quality/scripts/gather.py +22 -0
  27. package/skills/hot-theme/SKILL.md +74 -0
  28. package/skills/hot-theme/scripts/gather.py +22 -0
  29. package/skills/ma-crossover/SKILL.md +63 -0
  30. package/skills/ma-crossover/scripts/gather.py +22 -0
  31. package/skills/market-review/SKILL.md +70 -0
  32. package/skills/market-review/scripts/gather.py +23 -0
  33. package/skills/one-yang-three-yin/SKILL.md +74 -0
  34. package/skills/one-yang-three-yin/scripts/gather.py +22 -0
  35. package/skills/shrink-pullback/SKILL.md +59 -0
  36. package/skills/shrink-pullback/scripts/gather.py +22 -0
  37. package/skills/stock-analysis/SKILL.md +134 -0
  38. package/skills/stock-analysis/scripts/gather.py +22 -0
  39. package/skills/stock-screener/SKILL.md +68 -0
  40. package/skills/stock-screener/scripts/gather.py +22 -0
  41. package/skills/strategy-backtest/SKILL.md +129 -0
  42. package/skills/strategy-backtest/scripts/gather.py +23 -0
  43. package/skills/volume-breakout/SKILL.md +58 -0
  44. package/skills/volume-breakout/scripts/gather.py +22 -0
  45. package/skills/wave-theory/SKILL.md +67 -0
  46. package/skills/wave-theory/scripts/gather.py +22 -0
  47. package/skills/wisburg-research/SKILL.md +119 -0
  48. package/strategies/examples/bottom_volume.yaml +47 -0
  49. package/strategies/examples/box_oscillation.yaml +41 -0
  50. package/strategies/examples/bull_trend.yaml +42 -0
  51. package/strategies/examples/dragon_head.yaml +45 -0
  52. package/strategies/examples/emotion_cycle.yaml +43 -0
  53. package/strategies/examples/ma_crossover.yaml +34 -0
  54. package/strategies/examples/rsi_oversold.yaml +35 -0
  55. package/strategies/examples/shrink_pullback.yaml +45 -0
  56. package/strategies/examples/volume_breakout.yaml +42 -0
  57. package/templates/_macros.j2 +90 -0
  58. package/templates/market_review/full.md.j2 +103 -0
  59. package/templates/stock_analysis/brief.md.j2 +28 -0
  60. package/templates/stock_analysis/full.md.j2 +108 -0
  61. package/tools/__init__.py +0 -0
  62. package/tools/alert_rules.py +243 -0
  63. package/tools/anomaly_detect.py +575 -0
  64. package/tools/backtest.py +764 -0
  65. package/tools/capabilities.py +108 -0
  66. package/tools/diagnostics.py +146 -0
  67. package/tools/gather.py +157 -0
  68. package/tools/import_parser.py +279 -0
  69. package/tools/market_regime.py +252 -0
  70. package/tools/market_review.py +155 -0
  71. package/tools/name_resolver.py +136 -0
  72. package/tools/pattern.py +515 -0
  73. package/tools/position_context.py +161 -0
  74. package/tools/report_renderer.py +116 -0
  75. package/tools/requirements.txt +15 -0
  76. package/tools/risk_screening.py +323 -0
  77. package/tools/screener.py +219 -0
  78. package/tools/search_intel.py +504 -0
  79. package/tools/stock_data.py +1721 -0
  80. package/tools/technical.py +642 -0
  81. package/tools/trading_calendar.py +226 -0
  82. package/tools/volume_analysis.py +178 -0
  83. package/tools/watchlist.py +66 -0
  84. package/tools/watchlist_context.py +185 -0
@@ -0,0 +1,108 @@
1
+ #!/usr/bin/env python3
2
+ """Market capability boundaries — tell agent what a given market supports.
3
+
4
+ Prevents agents from calling A-share-only tools on HK/US/JP/KR/TW and vice versa.
5
+ """
6
+
7
+ import argparse
8
+ import json
9
+ import sys
10
+ from datetime import datetime, timezone
11
+ from pathlib import Path
12
+
13
+ sys.path.insert(0, str(Path(__file__).resolve().parent))
14
+ from stock_data import detect_market
15
+
16
+ ALL_MARKETS = ["A", "HK", "US", "JP", "KR", "TW"]
17
+
18
+ # tool_name -> {markets: [supported markets], reason: str for unsupported}
19
+ TOOL_MATRIX = {
20
+ "get_kline": {"markets": ALL_MARKETS, "reason": None},
21
+ "get_quote": {"markets": ALL_MARKETS, "reason": None},
22
+ "get_news": {"markets": ALL_MARKETS, "reason": None},
23
+ "get_financials": {"markets": ALL_MARKETS, "reason": None},
24
+ "get_technical_analysis": {"markets": ALL_MARKETS, "reason": None},
25
+ "analyze_pattern": {"markets": ALL_MARKETS, "reason": None},
26
+ "get_market_indices": {"markets": ALL_MARKETS, "reason": None},
27
+ "calculate_ma": {"markets": ALL_MARKETS, "reason": None},
28
+ "get_volume_analysis": {"markets": ALL_MARKETS, "reason": None},
29
+ "detect_anomaly": {"markets": ALL_MARKETS, "reason": None},
30
+ "get_stock_info": {"markets": ALL_MARKETS, "reason": None},
31
+ "detect_market_regime": {"markets": ["A", "HK", "US"], "reason": "index coverage is CN/HK/US only"},
32
+ "get_market_review": {"markets": ["A", "HK", "US"], "reason": "review data coverage is CN/HK/US only"},
33
+ # A-share only
34
+ "get_capital_flow": {"markets": ["A"], "reason": "A-share only (akshare data)"},
35
+ "get_chip_distribution": {"markets": ["A"], "reason": "A-share only"},
36
+ "get_sector_rankings": {"markets": ["A"], "reason": "A-share only"},
37
+ "get_market_stats": {"markets": ["A"], "reason": "A-share only"},
38
+ "get_fundamental_context": {"markets": ["A"], "reason": "A-share only"},
39
+ "resolve_stock_name": {"markets": ["A"], "reason": "A-share name/pinyin index only"},
40
+ # US-leaning
41
+ "get_social_sentiment": {
42
+ "markets": ["US"],
43
+ "reason": "sentiment providers are US-centric; requires SENTIMENT_API_KEY",
44
+ },
45
+ "get_trending_sentiment": {
46
+ "markets": ALL_MARKETS,
47
+ "reason": None,
48
+ },
49
+ # search-based, market-agnostic
50
+ "search_stock_news": {"markets": ALL_MARKETS, "reason": None},
51
+ "search_comprehensive_intel": {"markets": ALL_MARKETS, "reason": None},
52
+ "extract_article": {"markets": ALL_MARKETS, "reason": None},
53
+ "screen_risk": {"markets": ALL_MARKETS, "reason": None},
54
+ }
55
+
56
+
57
+ def get_capabilities(market: str) -> dict:
58
+ market = market.upper()
59
+ supported = []
60
+ unsupported = []
61
+ for tool, spec in TOOL_MATRIX.items():
62
+ if market in spec["markets"]:
63
+ supported.append(tool)
64
+ else:
65
+ unsupported.append({"tool": tool, "reason": spec["reason"] or "not supported"})
66
+
67
+ return {
68
+ "meta": {
69
+ "provider": "capabilities",
70
+ "fetched_at": datetime.now(timezone.utc).isoformat(),
71
+ "freshness": "static",
72
+ "fallback_used": False,
73
+ "warnings": [],
74
+ },
75
+ "market": market,
76
+ "supported": supported,
77
+ "unsupported": unsupported,
78
+ }
79
+
80
+
81
+ def main():
82
+ parser = argparse.ArgumentParser(description="Market capability boundaries")
83
+ sub = parser.add_subparsers(dest="command")
84
+ p = sub.add_parser("get")
85
+ grp = p.add_mutually_exclusive_group(required=True)
86
+ grp.add_argument("--market", type=str.upper, choices=ALL_MARKETS, help="Market code")
87
+ grp.add_argument("--symbol", help="Stock symbol (auto-detects market)")
88
+ args = parser.parse_args()
89
+
90
+ if not args.command:
91
+ parser.print_help()
92
+ sys.exit(1)
93
+
94
+ market = args.market or detect_market(args.symbol)
95
+ result = get_capabilities(market)
96
+ if args.symbol:
97
+ result["symbol"] = args.symbol
98
+ json.dump(result, sys.stdout, ensure_ascii=False, indent=2, default=str)
99
+ print()
100
+
101
+
102
+ if __name__ == "__main__":
103
+ # Windows defaults stdio to a legacy code page (cp1252) that cannot encode the
104
+ # Chinese text these tools emit — force UTF-8 so stdout never crashes there.
105
+ for _s in (sys.stdout, sys.stderr):
106
+ if hasattr(_s, "reconfigure"):
107
+ _s.reconfigure(encoding="utf-8")
108
+ main()
@@ -0,0 +1,146 @@
1
+ #!/usr/bin/env python3
2
+ """Data source diagnostics — report which providers are usable per market.
3
+
4
+ Checks: (1) python package importable, (2) required env vars present.
5
+ Does NOT hit the network — this is a cheap capability probe, not a health check.
6
+ """
7
+
8
+ import argparse
9
+ import importlib
10
+ import importlib.util
11
+ import json
12
+ import os
13
+ import sys
14
+ from datetime import datetime, timezone
15
+ from pathlib import Path
16
+
17
+ sys.path.insert(0, str(Path(__file__).resolve().parent))
18
+ from capabilities import ALL_MARKETS # noqa: E402
19
+
20
+ # provider -> (import_name, required_env_vars, markets, note)
21
+ PROVIDERS = {
22
+ "akshare": ("akshare", [], ["A"], None),
23
+ "tushare": ("tushare", ["TUSHARE_TOKEN"], ["A"], None),
24
+ "efinance": ("efinance", [], ["A"], None),
25
+ "pytdx": ("pytdx", [], ["A"], "connects to public Tencent servers"),
26
+ "yfinance": ("yfinance", [], ["HK", "US", "JP", "KR", "TW"], "quote may be delayed 15-20min"),
27
+ "finnhub": ("finnhub", ["FINNHUB_API_KEY"], ["HK", "US"], None),
28
+ "longbridge": (
29
+ "longport.openapi",
30
+ ["LONGBRIDGE_APP_KEY", "LONGBRIDGE_APP_SECRET", "LONGBRIDGE_ACCESS_TOKEN"],
31
+ ["HK", "US"],
32
+ None,
33
+ ),
34
+ "alphavantage": ("requests", ["ALPHAVANTAGE_API_KEY"], ["US"], None),
35
+ }
36
+
37
+
38
+ def _check_provider(name: str) -> dict:
39
+ import_name, env_vars, markets, note = PROVIDERS[name]
40
+
41
+ # find_spec avoids executing the package's top-level code — cheaper and
42
+ # side-effect-free vs import_module. This is a capability probe, not a call.
43
+ try:
44
+ spec = importlib.util.find_spec(import_name)
45
+ pkg_ok = spec is not None
46
+ pkg_err = None if pkg_ok else f"package '{import_name}' not installed"
47
+ except Exception as e:
48
+ pkg_ok = False
49
+ pkg_err = f"package '{import_name}' not installed: {e}"
50
+
51
+ missing_env = [v for v in env_vars if not os.environ.get(v)]
52
+
53
+ available = pkg_ok and not missing_env
54
+ reasons = []
55
+ if not pkg_ok:
56
+ reasons.append(pkg_err)
57
+ if missing_env:
58
+ reasons.append(f"missing env: {', '.join(missing_env)}")
59
+
60
+ entry = {
61
+ "name": name,
62
+ "available": available,
63
+ "markets": markets,
64
+ "reason": "; ".join(reasons) if reasons else None,
65
+ }
66
+ if note:
67
+ entry["note"] = note
68
+ return entry
69
+
70
+
71
+ def diagnose(market: str = "all") -> dict:
72
+ market = (market or "all").upper()
73
+ all_providers = [_check_provider(p) for p in PROVIDERS]
74
+
75
+ markets_to_report = ALL_MARKETS if market == "ALL" else [market]
76
+
77
+ result = {
78
+ "meta": {
79
+ "provider": "diagnostics",
80
+ "fetched_at": datetime.now(timezone.utc).isoformat(),
81
+ "freshness": "instant",
82
+ "fallback_used": False,
83
+ "warnings": [],
84
+ },
85
+ "markets": [],
86
+ }
87
+
88
+ for m in markets_to_report:
89
+ providers = [p for p in all_providers if m in p["markets"]]
90
+ available = [p for p in providers if p["available"]]
91
+
92
+ warnings = []
93
+ if not available:
94
+ warnings.append(f"no working data source for {m}")
95
+ # market-specific caveats
96
+ if m == "A" and not any(p["name"] == "tushare" and p["available"] for p in providers):
97
+ warnings.append("TUSHARE_TOKEN not set — akshare-only, some deep data unavailable")
98
+ if (
99
+ m in ("HK", "US")
100
+ and not any(p["name"] == "longbridge" and p["available"] for p in providers)
101
+ and not any(p["name"] == "finnhub" and p["available"] for p in providers)
102
+ ):
103
+ warnings.append(f"{m} quotes will be delayed (yfinance only)")
104
+ if m in ("JP", "KR", "TW"):
105
+ warnings.append(f"{m} is served by yfinance only — quotes delayed 15-20min, no failover")
106
+
107
+ result["markets"].append(
108
+ {
109
+ "market": m,
110
+ "available": bool(available),
111
+ "providers": providers,
112
+ "warnings": warnings,
113
+ }
114
+ )
115
+
116
+ return result
117
+
118
+
119
+ def main():
120
+ parser = argparse.ArgumentParser(description="Data source diagnostics")
121
+ sub = parser.add_subparsers(dest="command")
122
+ p = sub.add_parser("check")
123
+ p.add_argument(
124
+ "--market",
125
+ default="ALL",
126
+ type=str.upper,
127
+ choices=ALL_MARKETS + ["ALL"],
128
+ help="Market to diagnose (default: ALL)",
129
+ )
130
+ args = parser.parse_args()
131
+ if not args.command:
132
+ parser.print_help()
133
+ sys.exit(1)
134
+
135
+ result = diagnose(args.market)
136
+ json.dump(result, sys.stdout, ensure_ascii=False, indent=2, default=str)
137
+ print()
138
+
139
+
140
+ if __name__ == "__main__":
141
+ # Windows defaults stdio to a legacy code page (cp1252) that cannot encode the
142
+ # Chinese text these tools emit — force UTF-8 so stdout never crashes there.
143
+ for _s in (sys.stdout, sys.stderr):
144
+ if hasattr(_s, "reconfigure"):
145
+ _s.reconfigure(encoding="utf-8")
146
+ main()
@@ -0,0 +1,157 @@
1
+ #!/usr/bin/env python3
2
+ """Unified data gathering — parallel subprocess calls to tools for skill scripts."""
3
+
4
+ import argparse
5
+ import json
6
+ import subprocess
7
+ import sys
8
+ from concurrent.futures import ThreadPoolExecutor, as_completed
9
+ from pathlib import Path
10
+
11
+ TOOLS_DIR = Path(__file__).resolve().parent
12
+
13
+
14
+ def _find_python() -> str:
15
+ venv = TOOLS_DIR.parent / ".venv" / "bin" / "python3"
16
+ if venv.exists():
17
+ return str(venv)
18
+ return sys.executable
19
+
20
+
21
+ def _run(script: str, args: list[str], timeout: int = 60) -> str | None:
22
+ python = _find_python()
23
+ cmd = [python, str(TOOLS_DIR / script)] + args
24
+ try:
25
+ r = subprocess.run(cmd, capture_output=True, text=True, encoding="utf-8", timeout=timeout)
26
+ if r.returncode == 0 and r.stdout.strip():
27
+ return r.stdout.strip()
28
+ except subprocess.TimeoutExpired:
29
+ pass
30
+ except Exception:
31
+ pass
32
+ return None
33
+
34
+
35
+ def _parse_json(raw: str | None):
36
+ if raw is None:
37
+ return None
38
+ try:
39
+ return json.loads(raw)
40
+ except json.JSONDecodeError:
41
+ return raw
42
+
43
+
44
+ def gather_analysis(symbol: str) -> dict:
45
+ tasks = {
46
+ "quote": ("stock_data.py", ["quote", symbol]),
47
+ "kline": ("stock_data.py", ["kline", symbol, "--period", "daily", "--count", "120"]),
48
+ "technical": ("technical.py", ["analyze", symbol, "--period", "daily", "--count", "120"]),
49
+ "financials": ("stock_data.py", ["financials", symbol]),
50
+ "capital_flow": ("stock_data.py", ["capital_flow", symbol]),
51
+ "news": ("search_intel.py", ["search", f"{symbol} stock news"]),
52
+ "risk": ("risk_screening.py", ["screen", symbol]),
53
+ "regime": ("market_regime.py", ["detect"]),
54
+ }
55
+
56
+ results = {}
57
+ with ThreadPoolExecutor(max_workers=8) as pool:
58
+ futures = {pool.submit(_run, script, args): key for key, (script, args) in tasks.items()}
59
+ for future in as_completed(futures):
60
+ key = futures[future]
61
+ results[key] = _parse_json(future.result())
62
+
63
+ return results
64
+
65
+
66
+ def gather_technical(symbol: str, kline_count: int = 120, with_quote: bool = False) -> dict:
67
+ tasks = {
68
+ "kline": ("stock_data.py", ["kline", symbol, "--period", "daily", "--count", str(kline_count)]),
69
+ "technical": ("technical.py", ["analyze", symbol, "--period", "daily", "--count", str(kline_count)]),
70
+ }
71
+ if with_quote:
72
+ tasks["quote"] = ("stock_data.py", ["quote", symbol])
73
+
74
+ results = {}
75
+ with ThreadPoolExecutor(max_workers=4) as pool:
76
+ futures = {pool.submit(_run, script, args): key for key, (script, args) in tasks.items()}
77
+ for future in as_completed(futures):
78
+ key = futures[future]
79
+ results[key] = _parse_json(future.result())
80
+
81
+ return results
82
+
83
+
84
+ def gather_screen(market: str = "A", top: int = 20, config: str | None = None) -> dict:
85
+ args = ["screen", "--market", market, "--top", str(top)]
86
+ if config:
87
+ args += ["--config", config]
88
+ raw = _run("screener.py", args, timeout=120)
89
+ return {"screen_results": _parse_json(raw)}
90
+
91
+
92
+ def gather_fundamental(symbol: str) -> dict:
93
+ tasks = {
94
+ "quote": ("stock_data.py", ["quote", symbol]),
95
+ "kline": ("stock_data.py", ["kline", symbol, "--period", "daily", "--count", "60"]),
96
+ "technical": ("technical.py", ["analyze", symbol, "--period", "daily", "--count", "60"]),
97
+ "financials": ("stock_data.py", ["financials", symbol]),
98
+ "news": ("search_intel.py", ["search", f"{symbol} stock news"]),
99
+ "stock_info": ("stock_data.py", ["stock_info", symbol]),
100
+ "sector_rankings": ("stock_data.py", ["sector_rankings", "--top", "5", "--direction", "both"]),
101
+ }
102
+
103
+ results = {}
104
+ with ThreadPoolExecutor(max_workers=7) as pool:
105
+ futures = {pool.submit(_run, script, args): key for key, (script, args) in tasks.items()}
106
+ for future in as_completed(futures):
107
+ key = futures[future]
108
+ results[key] = _parse_json(future.result())
109
+
110
+ return results
111
+
112
+
113
+ def main():
114
+ parser = argparse.ArgumentParser(description="Unified data gathering for skills")
115
+ sub = parser.add_subparsers(dest="command")
116
+
117
+ p_analysis = sub.add_parser("analysis")
118
+ p_analysis.add_argument("symbol")
119
+
120
+ p_technical = sub.add_parser("technical")
121
+ p_technical.add_argument("symbol")
122
+ p_technical.add_argument("--kline-count", type=int, default=120)
123
+ p_technical.add_argument("--with-quote", action="store_true")
124
+
125
+ p_screen = sub.add_parser("screen")
126
+ p_screen.add_argument("--market", default="A")
127
+ p_screen.add_argument("--top", type=int, default=20)
128
+ p_screen.add_argument("--config")
129
+
130
+ p_fundamental = sub.add_parser("fundamental")
131
+ p_fundamental.add_argument("symbol")
132
+
133
+ args = parser.parse_args()
134
+
135
+ if args.command == "analysis":
136
+ result = gather_analysis(args.symbol)
137
+ elif args.command == "technical":
138
+ result = gather_technical(args.symbol, args.kline_count, args.with_quote)
139
+ elif args.command == "screen":
140
+ result = gather_screen(args.market, args.top, args.config)
141
+ elif args.command == "fundamental":
142
+ result = gather_fundamental(args.symbol)
143
+ else:
144
+ parser.print_help()
145
+ sys.exit(1)
146
+
147
+ json.dump(result, sys.stdout, ensure_ascii=False, indent=2)
148
+ print()
149
+
150
+
151
+ if __name__ == "__main__":
152
+ # Windows defaults stdio to a legacy code page (cp1252) that cannot encode the
153
+ # Chinese text these tools emit — force UTF-8 so stdout never crashes there.
154
+ for _s in (sys.stdout, sys.stderr):
155
+ if hasattr(_s, "reconfigure"):
156
+ _s.reconfigure(encoding="utf-8")
157
+ main()
@@ -0,0 +1,279 @@
1
+ #!/usr/bin/env python3
2
+ """Parse stock list from freeform text / CSV / markdown.
3
+
4
+ Extracts symbols using regex heuristics + name_resolver for A-share Chinese names.
5
+ """
6
+
7
+ import argparse
8
+ import json
9
+ import re
10
+ import sys
11
+ from datetime import datetime, timezone
12
+ from pathlib import Path
13
+
14
+ sys.path.insert(0, str(Path(__file__).resolve().parent))
15
+ from name_resolver import resolve # noqa: E402
16
+ from stock_data import detect_market # noqa: E402
17
+
18
+ # regex patterns
19
+ RE_A_CODE = re.compile(r"(?<!\d)(\d{6})(?!\d)")
20
+ RE_HK_CODE = re.compile(r"(?<![A-Za-z0-9])(\d{4,5}\.HK)(?![A-Za-z0-9])", re.IGNORECASE)
21
+ RE_ASIA_SUFFIX_CODE = re.compile(
22
+ r"(?<![A-Za-z0-9])(\d{4}\.T|\d{6}\.(?:KS|KQ)|\d{4}\.(?:TW|TWO))(?![A-Za-z0-9])", re.IGNORECASE
23
+ )
24
+ RE_US_TICKER = re.compile(r"(?<![A-Za-z0-9])([A-Z]{1,5})(?![A-Za-z0-9\.])")
25
+ RE_CN_NAME = re.compile(r"[一-鿿][一-鿿\w]{1,7}")
26
+
27
+ # US ticker false-positive filter (common ALL-CAPS words that aren't tickers)
28
+ US_STOPWORDS = {
29
+ "A",
30
+ "I",
31
+ "AM",
32
+ "PM",
33
+ "AN",
34
+ "AS",
35
+ "AT",
36
+ "BE",
37
+ "BY",
38
+ "DO",
39
+ "GO",
40
+ "IF",
41
+ "IN",
42
+ "IS",
43
+ "IT",
44
+ "ME",
45
+ "MY",
46
+ "NO",
47
+ "OF",
48
+ "ON",
49
+ "OR",
50
+ "SO",
51
+ "TO",
52
+ "UP",
53
+ "US",
54
+ "WE",
55
+ "THE",
56
+ "AND",
57
+ "FOR",
58
+ "BUT",
59
+ "NOT",
60
+ "YOU",
61
+ "ALL",
62
+ "CAN",
63
+ "HAD",
64
+ "HAS",
65
+ "HER",
66
+ "WAS",
67
+ "ONE",
68
+ "OUR",
69
+ "OUT",
70
+ "DAY",
71
+ "GET",
72
+ "HIM",
73
+ "HIS",
74
+ "HOW",
75
+ "MAN",
76
+ "NEW",
77
+ "NOW",
78
+ "OLD",
79
+ "SEE",
80
+ "TWO",
81
+ "WAY",
82
+ "WHO",
83
+ "BOY",
84
+ "DID",
85
+ "ITS",
86
+ "LET",
87
+ "PUT",
88
+ "SAY",
89
+ "SHE",
90
+ "TOO",
91
+ "USE",
92
+ "PE",
93
+ "PB",
94
+ "PS",
95
+ "ROE",
96
+ "MA",
97
+ "RSI",
98
+ "MACD",
99
+ "KDJ",
100
+ "BOLL",
101
+ "ETF",
102
+ "IPO",
103
+ "AI",
104
+ "ML",
105
+ "GDP",
106
+ "CPI",
107
+ "PPI",
108
+ "CEO",
109
+ "CFO",
110
+ "CTO",
111
+ "COO",
112
+ "USA",
113
+ "PDF",
114
+ "CSV",
115
+ "API",
116
+ "URL",
117
+ "HTTP",
118
+ "SQL",
119
+ "JSON",
120
+ "XML",
121
+ "HTML",
122
+ "CSS",
123
+ "OK",
124
+ "TV",
125
+ "PC",
126
+ "OS",
127
+ "USD",
128
+ "CNY",
129
+ "HKD",
130
+ "EUR",
131
+ "JPY",
132
+ "GBP",
133
+ }
134
+
135
+
136
+ def _dedupe(items: list[dict]) -> list[dict]:
137
+ seen = set()
138
+ result = []
139
+ for it in items:
140
+ key = it.get("symbol")
141
+ if key and key not in seen:
142
+ seen.add(key)
143
+ result.append(it)
144
+ return result
145
+
146
+
147
+ def parse(text: str) -> dict:
148
+ text = text.strip()
149
+ items: list[dict] = []
150
+ unresolved: list[str] = []
151
+ seen_symbols: set[str] = set()
152
+
153
+ # 1. HK codes (do first — they contain digits that A-share regex would also match)
154
+ hk_matches: list[str] = []
155
+ for m in RE_HK_CODE.finditer(text):
156
+ code = m.group(1).upper()
157
+ # normalize to 5-digit
158
+ num, suf = code.split(".")
159
+ code = f"{num.zfill(5)}.HK"
160
+ if code not in seen_symbols:
161
+ items.append({"symbol": code, "market": "HK"})
162
+ seen_symbols.add(code)
163
+ hk_matches.append(m.group(1))
164
+
165
+ # 1b. JP/KR/TW suffixed codes (7203.T, 005930.KS, 2330.TW — digits would hit A-share regex too)
166
+ asia_matches: list[str] = []
167
+ for m in RE_ASIA_SUFFIX_CODE.finditer(text):
168
+ code = m.group(1).upper()
169
+ if code not in seen_symbols:
170
+ items.append({"symbol": code, "market": detect_market(code)})
171
+ seen_symbols.add(code)
172
+ asia_matches.append(m.group(1))
173
+
174
+ # strip suffixed codes so their digit part doesn't get picked up as A-share
175
+ # and their suffix doesn't get picked up as a US ticker
176
+ text_without_suffixed = text
177
+ for m in hk_matches + asia_matches:
178
+ text_without_suffixed = text_without_suffixed.replace(m, " ")
179
+
180
+ # 2. A-share 6-digit codes
181
+ for m in RE_A_CODE.finditer(text_without_suffixed):
182
+ code = m.group(1)
183
+ if code not in seen_symbols:
184
+ items.append({"symbol": code, "market": "A"})
185
+ seen_symbols.add(code)
186
+
187
+ # 3. US tickers (1-5 uppercase letters), filtered against stopwords
188
+ # runs on stripped text so HK/JP/KR/TW suffixes (HK, TW, KS...) aren't misread as tickers
189
+ for m in RE_US_TICKER.finditer(text_without_suffixed):
190
+ tick = m.group(1)
191
+ if tick in US_STOPWORDS:
192
+ continue
193
+ if len(tick) == 1:
194
+ continue # too noisy
195
+ if tick not in seen_symbols:
196
+ items.append({"symbol": tick, "market": "US"})
197
+ seen_symbols.add(tick)
198
+
199
+ # 4. Chinese names → resolve via name_resolver
200
+ cn_names = set(RE_CN_NAME.findall(text))
201
+ # filter obviously non-stock words
202
+ noise = {"成本", "持仓", "买入", "卖出", "止损", "止盈", "股票", "代码", "行情", "分析", "策略"}
203
+ cn_names = {n for n in cn_names if n not in noise and len(n) >= 2}
204
+
205
+ for name in cn_names:
206
+ try:
207
+ hits = resolve(name, top=1)
208
+ except Exception:
209
+ hits = []
210
+ if hits and isinstance(hits, list) and hits[0].get("code"):
211
+ top_hit = hits[0]
212
+ # ponytail: name_resolver returns nearest fuzzy match even for names not in A-share
213
+ # (e.g. "腾讯" → "拓新药业" via pinyin). Reject low-confidence hits.
214
+ score = top_hit.get("score")
215
+ hit_name = top_hit.get("name", "")
216
+ is_exact = (name in hit_name) or (hit_name in name)
217
+ if not is_exact and score is not None and score < 0.85:
218
+ unresolved.append(name)
219
+ continue
220
+ code = top_hit["code"]
221
+ if code not in seen_symbols:
222
+ items.append(
223
+ {
224
+ "symbol": code,
225
+ "market": "A",
226
+ "name": hit_name,
227
+ "resolved_from": name,
228
+ }
229
+ )
230
+ seen_symbols.add(code)
231
+ else:
232
+ unresolved.append(name)
233
+
234
+ return {
235
+ "meta": {
236
+ "provider": "import_parser",
237
+ "fetched_at": datetime.now(timezone.utc).isoformat(),
238
+ "freshness": "instant",
239
+ "fallback_used": False,
240
+ "warnings": [],
241
+ },
242
+ "items": _dedupe(items),
243
+ "unresolved": unresolved,
244
+ "count": len(items),
245
+ }
246
+
247
+
248
+ def main():
249
+ parser = argparse.ArgumentParser(description="Parse stock list from text")
250
+ sub = parser.add_subparsers(dest="command")
251
+ p = sub.add_parser("parse")
252
+ p.add_argument("--text", help='Text to parse (or "-" for stdin)')
253
+ p.add_argument("--text-b64", help="Base64-encoded text")
254
+
255
+ args = parser.parse_args()
256
+ if not args.command:
257
+ parser.print_help()
258
+ sys.exit(1)
259
+
260
+ if args.text_b64:
261
+ import base64
262
+
263
+ text = base64.b64decode(args.text_b64).decode("utf-8")
264
+ elif not args.text or args.text == "-":
265
+ text = sys.stdin.read()
266
+ else:
267
+ text = args.text
268
+ result = parse(text)
269
+ json.dump(result, sys.stdout, ensure_ascii=False, indent=2, default=str)
270
+ print()
271
+
272
+
273
+ if __name__ == "__main__":
274
+ # Windows defaults stdio to a legacy code page (cp1252) that cannot encode the
275
+ # Chinese text these tools emit — force UTF-8 so stdout never crashes there.
276
+ for _s in (sys.stdout, sys.stderr):
277
+ if hasattr(_s, "reconfigure"):
278
+ _s.reconfigure(encoding="utf-8")
279
+ main()