yuclaw 1.0.0__tar.gz

yuclaw-1.0.0/PKG-INFO

@@ -0,0 +1,74 @@
+Metadata-Version: 2.4
+Name: yuclaw
+Version: 1.0.0
+Summary: YUCLAW — Open Financial Intelligence Platform
+Home-page: https://github.com/YuClawLab/yuclaw-brain
+Author: YuClawLab
+License: MIT
+Project-URL: Homepage, https://github.com/YuClawLab
+Project-URL: Dashboard, https://yuclawlab.github.io/yuclaw-brain
+Keywords: finance,trading,quant,zkp,backtesting,ai
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Financial and Insurance Industry
+Classifier: License :: OSI Approved :: MIT License
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+Requires-Dist: yfinance>=0.2.0
+Requires-Dist: pandas>=2.0.0
+Requires-Dist: numpy>=1.24.0
+Requires-Dist: requests>=2.28.0
+Requires-Dist: fastapi>=0.100.0
+Requires-Dist: uvicorn>=0.20.0
+Requires-Dist: rich>=13.0.0
+Dynamic: author
+Dynamic: home-page
+Dynamic: requires-python
+
+# YUCLAW — Open Financial Intelligence Platform
+
+The open-source financial intelligence OS. Real backtests. ZKP audit. Local AI.
+
+## Install in one command
+```bash
+pip install yuclaw
+yuclaw start
+```
+
+## What YUCLAW gives you
+
+- **Real signals** — factor-scored buy/sell for 167+ tickers
+- **Real backtests** — Calmar 3.055 from 15 years of actual prices
+- **ZKP audit trail** — every decision on Ethereum blockchain
+- **Local AI** — Nemotron 120B, zero cloud dependency
+- **Market regime** — CRISIS/RISK_OFF/RISK_ON detection
+- **30-day track record** — verifiable signal performance
+
+## Quick start
+```bash
+pip install yuclaw
+yuclaw start          # Start all engines
+yuclaw signals        # Show current signals
+yuclaw regime         # Show market regime
+yuclaw brief          # Show institutional brief
+yuclaw track          # Show track record
+yuclaw dashboard      # Open live dashboard
+```
+
+## How it works
+```
+Market Data -> Factor Library -> Signal Aggregator
+     |              |                |
+SEC XBRL    Macro Regime      Risk Engine
+     |              |                |
+Evidence Graph -> Nemotron 120B -> ZKP Proof
+     |
+Institutional Brief -> Dashboard -> Track Record
+```
+
+## Community
+
+- GitHub: https://github.com/YuClawLab
+- Dashboard: https://yuclawlab.github.io/yuclaw-brain
+- Paper: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6461418
+
+MIT License — free for everyone.

yuclaw-1.0.0/README.md

@@ -0,0 +1,137 @@
+# YUCLAW — Open Financial Intelligence Platform
+
+[![Stars](https://img.shields.io/github/stars/YuClawLab/yuclaw-brain?style=flat&color=ffd700)](https://github.com/YuClawLab/yuclaw-brain)
+[![License](https://img.shields.io/badge/license-MIT-green)](LICENSE)
+[![Python](https://img.shields.io/badge/python-3.10+-blue)](https://python.org)
+[![Model](https://img.shields.io/badge/model-Nemotron%20120B-brightgreen)](https://nvidia.com)
+[![Hardware](https://img.shields.io/badge/hardware-DGX%20Spark%20GB10-76b900)](https://nvidia.com)
+[![ZKP](https://img.shields.io/badge/ZKP-Ethereum%20Sepolia-purple)](https://sepolia.etherscan.io)
+[![SSRN](https://img.shields.io/badge/paper-SSRN%206461418-red)](https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6461418)
+
+> The open-source financial intelligence OS. Real backtests. ZKP audit trail. Local 120B AI. Like OpenClaw but built for finance.
+
+## Install in one command
+```bash
+pip install yuclaw
+yuclaw start
+```
+
+## Quick commands
+```bash
+yuclaw signals    # Show current buy/sell signals
+yuclaw regime     # Show market regime (CRISIS/RISK_OFF/RISK_ON)
+yuclaw brief      # Nemotron 120B institutional brief
+yuclaw track      # 30-day verifiable track record
+yuclaw zkp        # ZKP cryptographic proofs on-chain
+yuclaw dashboard  # Open live dashboard
+```
+
+## Live Dashboard
+
+**[yuclawlab.github.io/yuclaw-brain](https://yuclawlab.github.io/yuclaw-brain)**
+
+Updates every 30 minutes. Real signals. Real risk. Real model.
+
+## Why YUCLAW
+
+| Feature | OpenClaw | Claude | YUCLAW |
+|---|---|---|---|
+| Real backtests | No | No | Calmar 3.055 |
+| ZKP audit trail | No | No | On-chain |
+| Real VaR/Kelly | No | No | Historical sim |
+| Local 120B model | No | No | Nemotron DGX |
+| Verified track record | No | No | 30-day building |
+| Adversarial Red Team | No | No | Kills bad strategies |
+| Academic paper | No | No | SSRN #6461418 |
+
+## Track Record (Day 3 of 30)
+
+| Signal | Result | Change | ZKP Block |
+|---|---|---|---|
+| LUNR STRONG_BUY | CORRECT | +14.68% | 10515603 |
+| ASTS STRONG_BUY | CORRECT | +10.44% | 10515603 |
+| MRVL STRONG_BUY | CORRECT | +6.59% | 10515603 |
+| DELL STRONG_BUY | CORRECT | +4.01% | 10515603 |
+
+**Day 3 accuracy: 60% — every signal on Ethereum blockchain**
+
+## How it works
+
+```
+Market Data (167+ tickers)
+    |
+Factor Library (RSI, MACD, Bollinger, Momentum)
+    |
+Signal Aggregator -> Adversarial Red Team
+    |
+Nemotron 3 Super 120B (local, 18.9 tok/s)
+    |
+ZKP Proof -> Ethereum Sepolia Blockchain
+    |
+Institutional Brief -> Dashboard -> Track Record
+```
+
+## What runs 24/7 on DGX Spark GB10
+
+- **Signal loop** — 167 tickers every 30 min
+- **Factor engine** — RSI/MACD/Bollinger every 30 min
+- **Risk engine** — VaR/CVaR/Kelly every 1 hr
+- **Macro regime** — CRISIS detection every 1 hr
+- **Nemotron brief** — Institutional analysis every 15 min
+- **ZKP vault** — Cryptographic proofs every 1 hr
+- **FIX gateway** — Paper trading every 30 min
+- **Track record** — Verified performance every 1 hr
+
+## Architecture
+
+```
+yuclaw/
+  modules/    Signal aggregator, macro regime
+  factors/    Factor library (RSI, MACD, Bollinger)
+  risk/       VaR, CVaR, Kelly criterion
+  finclaw/    Institutional brief, competitive intel
+  trust/      ZKP vault, on-chain proofs
+  edge/       FIX gateway, execution levels 0-4
+  memory/     Portfolio memory, track record
+  brain/      Evidence graph, financial NER
+  plugins/    Plugin system (like ClawHub)
+  api/        REST API for OpenClaw integration
+```
+
+## Plugin System
+```python
+from yuclaw.plugins import register
+
+@register('my_strategy', 'My custom strategy', 'quant')
+class MyStrategy:
+    def run(self):
+        # Your financial intelligence here
+        pass
+```
+
+## Academic Paper
+
+**CRT Lock-Free Concurrent Scheduler for Financial Systems**
+SSRN Abstract #6461418 | DGX Spark GB10 | 1.37ms latency
+
+[Read on SSRN](https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6461418)
+
+## Community
+
+- Dashboard: [yuclawlab.github.io/yuclaw-brain](https://yuclawlab.github.io/yuclaw-brain)
+- Twitter: [@Vincenzhang2026](https://twitter.com/Vincenzhang2026)
+- Paper: [SSRN #6461418](https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6461418)
+- GitHub: [YuClawLab](https://github.com/YuClawLab)
+
+## Contributing
+
+YUCLAW improves every day — like OpenClaw but for finance.
+PRs welcome. Build plugins. Add data sources. Improve factors.
+
+## License
+
+MIT — free for everyone.
+
+---
+
+*Built on NVIDIA DGX Spark GB10 | Nemotron 3 Super 120B | Zero cloud dependency*

yuclaw-1.0.0/README_PACKAGE.md

@@ -0,0 +1,48 @@
+# YUCLAW — Open Financial Intelligence Platform
+
+The open-source financial intelligence OS. Real backtests. ZKP audit. Local AI.
+
+## Install in one command
+```bash
+pip install yuclaw
+yuclaw start
+```
+
+## What YUCLAW gives you
+
+- **Real signals** — factor-scored buy/sell for 167+ tickers
+- **Real backtests** — Calmar 3.055 from 15 years of actual prices
+- **ZKP audit trail** — every decision on Ethereum blockchain
+- **Local AI** — Nemotron 120B, zero cloud dependency
+- **Market regime** — CRISIS/RISK_OFF/RISK_ON detection
+- **30-day track record** — verifiable signal performance
+
+## Quick start
+```bash
+pip install yuclaw
+yuclaw start          # Start all engines
+yuclaw signals        # Show current signals
+yuclaw regime         # Show market regime
+yuclaw brief          # Show institutional brief
+yuclaw track          # Show track record
+yuclaw dashboard      # Open live dashboard
+```
+
+## How it works
+```
+Market Data -> Factor Library -> Signal Aggregator
+     |              |                |
+SEC XBRL    Macro Regime      Risk Engine
+     |              |                |
+Evidence Graph -> Nemotron 120B -> ZKP Proof
+     |
+Institutional Brief -> Dashboard -> Track Record
+```
+
+## Community
+
+- GitHub: https://github.com/YuClawLab
+- Dashboard: https://yuclawlab.github.io/yuclaw-brain
+- Paper: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6461418
+
+MIT License — free for everyone.

yuclaw-1.0.0/pyproject.toml

@@ -0,0 +1,33 @@
+[build-system]
+requires = ["setuptools>=61.0", "wheel"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "yuclaw"
+version = "1.0.0"
+description = "YUCLAW — Open Financial Intelligence Platform"
+readme = "README_PACKAGE.md"
+license = {text = "MIT"}
+requires-python = ">=3.10"
+keywords = ["finance", "trading", "quant", "zkp", "backtesting", "ai"]
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Intended Audience :: Financial and Insurance Industry",
+    "License :: OSI Approved :: MIT License",
+]
+dependencies = [
+    "yfinance>=0.2.0",
+    "pandas>=2.0.0",
+    "numpy>=1.24.0",
+    "requests>=2.28.0",
+    "fastapi>=0.100.0",
+    "uvicorn>=0.20.0",
+    "rich>=13.0.0",
+]
+
+[project.scripts]
+yuclaw = "yuclaw.cli:main"
+
+[project.urls]
+Homepage = "https://github.com/YuClawLab"
+Dashboard = "https://yuclawlab.github.io/yuclaw-brain"

yuclaw-1.0.0/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

yuclaw-1.0.0/setup.py

@@ -0,0 +1,35 @@
+from setuptools import setup, find_packages
+import os
+
+setup(
+    name='yuclaw',
+    version='1.0.0',
+    description='YUCLAW — Open Financial Intelligence Platform. Real backtests. ZKP audit. Local AI.',
+    long_description=open('README_PACKAGE.md').read() if os.path.exists('README_PACKAGE.md') else '',
+    long_description_content_type='text/markdown',
+    author='YuClawLab',
+    url='https://github.com/YuClawLab/yuclaw-brain',
+    packages=find_packages(),
+    install_requires=[
+        'yfinance>=0.2.0',
+        'pandas>=2.0.0',
+        'numpy>=1.24.0',
+        'requests>=2.28.0',
+        'fastapi>=0.100.0',
+        'uvicorn>=0.20.0',
+        'rich>=13.0.0',
+    ],
+    entry_points={
+        'console_scripts': [
+            'yuclaw=yuclaw.cli:main',
+        ],
+    },
+    classifiers=[
+        'Development Status :: 4 - Beta',
+        'Intended Audience :: Financial and Insurance Industry',
+        'Topic :: Office/Business :: Financial :: Investment',
+        'License :: OSI Approved :: MIT License',
+        'Programming Language :: Python :: 3.10',
+    ],
+    python_requires='>=3.10',
+)

yuclaw-1.0.0/yuclaw/agents/agents.py

@@ -0,0 +1,247 @@
+"""yuclaw/agents/agents.py — All YUCLAW agents."""
+from __future__ import annotations
+import json
+from datetime import datetime, timezone
+from dataclasses import dataclass, field
+from ..core.router import get_router
+from ..core.ontology.models import EvidenceNode, ValidationStatus
+from ..memory.evidence_graph import EvidenceGraph
+from ..memory.portfolio_memory import PortfolioMemory
+
+
+# ─────────────────────────────────────────────────────────
+#  CIO AGENT
+# ─────────────────────────────────────────────────────────
+
+CIO_SYSTEM = """You are the CIO Agent of YUCLAW financial intelligence system.
+Decompose the research request into ordered tasks for specialist agents.
+Output JSON array only. Each task: {"task_id":str, "agent":str, "instruction":str, "priority":"high|medium|low"}
+agent options: research | macro | quant | risk | compliance
+No preamble. JSON only."""
+
+
+@dataclass
+class Task:
+    task_id: str
+    agent: str
+    instruction: str
+    priority: str = "medium"
+    result: str | None = None
+
+
+@dataclass  
+class WorkPlan:
+    request: str
+    tasks: list[Task]
+    capital_limit_usd: float = 0.0  # Always 0 until Execution Agent activated
+
+
+class CIOAgent:
+    def __init__(self):
+        self._router = get_router()
+
+    async def decompose(self, request: str) -> WorkPlan:
+        response = await self._router.complete(
+            prompt=f"Research request: {request}",
+            system=CIO_SYSTEM,
+            fast=True
+        )
+        try:
+            tasks = [Task(**t) for t in json.loads(response)]
+        except Exception:
+            tasks = [Task(task_id="t1", agent="research",
+                          instruction=request, priority="high")]
+        return WorkPlan(request=request, tasks=tasks, capital_limit_usd=0.0)
+
+
+# ─────────────────────────────────────────────────────────
+#  RESEARCH AGENT
+# ─────────────────────────────────────────────────────────
+
+RESEARCH_SYSTEM = """You are the YUCLAW Research Agent. Analyze financial documents.
+CRITICAL: Every factual claim MUST include a citation: [DOC:<id>|PAGE:<n>]
+Output JSON only:
+{
+  "summary": "comprehensive investment analysis",
+  "key_metrics": [{"name":str, "value":str, "page":int, "significance":str}],
+  "thesis": {"bull":str, "base":str, "bear":str},
+  "key_assumptions": [str],
+  "catalysts": [str],
+  "risks": [str],
+  "claims": [{"claim":str, "page":int, "confidence":float, "ontology_tags":[str]}]
+}
+JSON only. No preamble."""
+
+
+class ResearchAgent:
+    def __init__(self, evidence_graph: EvidenceGraph, portfolio_memory: PortfolioMemory):
+        self._router   = get_router()
+        self._graph    = evidence_graph
+        self._memory   = portfolio_memory
+        # Wire in real PDF parser for grounding with actual financial data
+        try:
+            from ..data.parsers.real_pdf_parser import RealPDFParser
+            self._real_parser = RealPDFParser()
+        except ImportError:
+            self._real_parser = None
+
+    async def analyze(self, doc_text: str, doc_id: str, ticker: str, query: str) -> dict:
+        # Pull prior thesis from memory
+        history = await self._memory.get_history(ticker)
+        history_context = ""
+        if history:
+            history_context = (
+                f"\n\nPRIOR THESIS HISTORY FOR {ticker} ({len(history)} records):\n"
+                + "\n".join(f"- [{h.created_at[:10]}] {h.thesis[:120]}" for h in history[:3])
+                + "\n\nCompare current analysis against prior thesis.\n"
+            )
+
+        # Ground with REAL financial data from yfinance
+        grounding_context = ""
+        if self._real_parser:
+            try:
+                parsed = self._real_parser.parse_yfinance(ticker)
+                if parsed.facts:
+                    grounding_context = "\n\n" + parsed.to_grounding_context() + "\n"
+            except Exception:
+                pass
+
+        prompt = (
+            f"DOCUMENT ID: {doc_id}\nTICKER: {ticker}\n"
+            f"{grounding_context}"
+            f"{history_context}"
+            f"\nDOCUMENT TEXT:\n{doc_text[:50000]}\n\n"
+            f"QUERY: {query}"
+        )
+
+        response = await self._router.complete(
+            prompt=prompt, system=RESEARCH_SYSTEM, max_tokens=8192
+        )
+
+        try:
+            result = json.loads(response)
+        except Exception:
+            result = {"summary": response, "claims": [], "key_metrics": [],
+                      "thesis": {}, "key_assumptions": [], "catalysts": [], "risks": []}
+
+        # Build Evidence Graph nodes
+        now = datetime.now(timezone.utc).isoformat()
+        node_ids = []
+        for i, c in enumerate(result.get("claims", [])):
+            if not c.get("claim"):
+                continue
+            node = EvidenceNode(
+                claim=c["claim"],
+                source_doc_id=doc_id,
+                page_number=max(int(c.get("page") or 1), 1),
+                paragraph_hash=f"{doc_id}_claim_{i}",
+                extraction_timestamp=now,
+                model_version="nemotron-3-super",
+                confidence=float(c.get("confidence") or 0.8),
+                ontology_tags=c.get("ontology_tags", [])
+            )
+            nid = await self._graph.add_node(node)
+            node_ids.append(nid)
+
+        result["evidence_node_ids"] = node_ids
+        result["ticker"]  = ticker
+        result["doc_id"]  = doc_id
+        result["status"]  = ValidationStatus.EVIDENCE_EXTRACTED.value
+        result["prior_thesis_count"] = len(history)
+
+        # Thesis Tracker: compare against history
+        try:
+            from ..memory.thesis_tracker import ThesisTracker
+            tracker = ThesisTracker(self._memory)
+            current_thesis = result.get("thesis", {}).get("base", result.get("summary", ""))
+            current_assumptions = result.get("key_assumptions", [])
+            comparison = await tracker.compare_against_history(ticker, current_thesis, current_assumptions)
+            result["thesis_change"] = {
+                "prior_count": comparison.prior_count,
+                "drift_detected": comparison.drift_detected,
+                "drift_description": comparison.drift_description,
+                "conviction_trend": comparison.conviction_trend,
+                "assumptions_changed": comparison.assumptions_changed,
+                "status": "FALSIFIED" if comparison.drift_detected and comparison.conviction_trend == "decreasing"
+                    else "WEAKENED" if comparison.conviction_trend == "decreasing"
+                    else "STRENGTHENED" if comparison.conviction_trend == "increasing"
+                    else "NEW" if comparison.prior_count == 0
+                    else "UNCHANGED",
+            }
+        except Exception:
+            result["thesis_change"] = {"status": "NEW", "prior_count": 0}
+
+        return result
+
+
+# ─────────────────────────────────────────────────────────
+#  MACRO AGENT
+# ─────────────────────────────────────────────────────────
+
+MACRO_SYSTEM = """You are the YUCLAW Macro Agent. Analyze macro events and their market impact.
+Output JSON only:
+{
+  "event": str,
+  "severity": "low|medium|high|critical",
+  "first_order_impacts": [
+    {"asset_class":str, "direction":"positive|negative|neutral", "magnitude":"small|medium|large", "reasoning":str}
+  ],
+  "second_order_impacts": [{"asset_class":str, "mechanism":str}],
+  "sectors": {"beneficiaries":[str], "casualties":[str]},
+  "portfolio_actions": [str],
+  "time_horizon": {"days_1_5":str, "weeks_1_4":str, "months_1_3":str}
+}
+JSON only. No preamble."""
+
+
+class MacroAgent:
+    def __init__(self):
+        self._router = get_router()
+
+    async def analyze_event(self, event: str) -> dict:
+        response = await self._router.complete(
+            prompt=f"Macro event to analyze: {event}",
+            system=MACRO_SYSTEM,
+            max_tokens=4096
+        )
+        try:
+            return json.loads(response)
+        except Exception:
+            return {"event": event, "raw_analysis": response, "error": "parse_failed"}
+
+
+# ─────────────────────────────────────────────────────────
+#  QUANT AGENT
+# ─────────────────────────────────────────────────────────
+
+QUANT_SYSTEM = """You are the YUCLAW Quant Agent. Generate quantitative strategy specifications.
+Output JSON only:
+{
+  "strategy_name": str,
+  "strategy_description": str,
+  "universe": str,
+  "signal_logic": str,
+  "holding_period": str,
+  "rebalance_frequency": str,
+  "risk_controls": [str],
+  "expected_calmar_ratio": float,
+  "python_pseudocode": str,
+  "known_failure_conditions": [str]
+}
+JSON only. No preamble."""
+
+
+class QuantAgent:
+    def __init__(self):
+        self._router = get_router()
+
+    async def generate_strategy(self, instruction: str) -> dict:
+        response = await self._router.complete(
+            prompt=f"Generate a quantitative strategy for: {instruction}",
+            system=QUANT_SYSTEM,
+            max_tokens=4096
+        )
+        try:
+            return json.loads(response)
+        except Exception:
+            return {"strategy_description": response, "error": "parse_failed"}