sigma-terminal 3.4.0__py3-none-any.whl → 3.5.0__py3-none-any.whl

sigma/tools/registry.py

@@ -0,0 +1,108 @@
+from typing import Any, Callable, Dict, List, Optional
+from pydantic import BaseModel
+import inspect
+
+class ToolDefinition(BaseModel):
+    name: str
+    description: str
+    input_schema: Dict[str, Any]
+    output_schema: Optional[Dict[str, Any]] = None
+    func: Callable
+    enabled: bool = True
+    provider: str = "internal" # yfinance, polygon, etc.
+
+class ToolRegistry:
+    def __init__(self):
+        self._tools: Dict[str, ToolDefinition] = {}
+
+    def register(self, name: str, description: str, provider: str = "internal"):
+        def decorator(func):
+            # Extract schema from type hints (simplified)
+            # In a real impl, utilize pydantic.TypeAdapter or similar if args are models.
+            # Here we assume simple args or we use a manual schema if provided?
+            # For now, let's keep it simple: no schema auto-extraction in this snippet 
+            # unless we implement a Schema generator.
+            # But the prompt says "typed tool registry... input_schema (pydantic)".
+            
+            # minimal schema generation
+            sig = inspect.signature(func)
+            params = {}
+            required = []
+            for pname, p in sig.parameters.items():
+                if pname == "self": continue
+                p_type = "string"
+                if p.annotation == int: p_type = "integer"
+                if p.annotation == float: p_type = "number"
+                if p.annotation == bool: p_type = "boolean"
+                
+                params[pname] = {"type": p_type}
+                if p.default == inspect.Parameter.empty:
+                    required.append(pname)
+            
+            schema = {
+                "type": "object",
+                "properties": params,
+                "required": required
+            }
+            
+            self._tools[name] = ToolDefinition(
+                name=name,
+                description=description,
+                input_schema=schema,
+                func=func,
+                provider=provider
+            )
+            return func
+        return decorator
+
+    def get_tool(self, name: str) -> Optional[ToolDefinition]:
+        return self._tools.get(name)
+
+    def list_tools(self) -> List[ToolDefinition]:
+        return [t for t in self._tools.values() if t.enabled]
+
+    def to_llm_format(self) -> List[Dict[str, Any]]:
+        tools_list = []
+        for t in self.list_tools():
+            # Clone schema to avoid mutating original
+            import copy
+            schema = copy.deepcopy(t.input_schema)
+            
+            # Inject thought parameter for Gemini/HackClub compatibility
+            # Some providers like Gemini via proxies require a "thought" field in tool calls
+            if "properties" in schema:
+                schema["properties"]["thought_signature"] = {
+                    "type": "string", 
+                    "description": "Internal reasoning for why this tool is being called. MUST be provided."
+                }
+                # Force model to use it
+                if "required" in schema and isinstance(schema["required"], list):
+                    schema["required"].append("thought_signature")
+                else:
+                    schema["required"] = ["thought_signature"]
+                
+            tools_list.append({
+                "type": "function",
+                "function": {
+                    "name": t.name,
+                    "description": t.description,
+                    "parameters": schema
+                }
+            })
+        return tools_list
+
+    async def execute(self, name: str, args: Dict[str, Any]) -> Any:
+        tool = self.get_tool(name)
+        if not tool:
+            raise ValueError(f"Tool {name} not found")
+        
+        # Remove provider-specific fields if present
+        args.pop("thought", None)
+        args.pop("thought_signature", None)
+        
+        if inspect.iscoroutinefunction(tool.func):
+            return await tool.func(**args)
+        else:
+            return tool.func(**args)
+
+TOOL_REGISTRY = ToolRegistry()

sigma/utils/extraction.py

@@ -0,0 +1,83 @@
+import re
+from datetime import datetime, timedelta
+from typing import List, Optional, Tuple
+
+COMMON_TICKERS = {
+    # Ticker map for normalization if needed
+    "APPLE": "AAPL",
+    "GOOGLE": "GOOGL",
+    "MICROSOFT": "MSFT",
+    "AMAZON": "AMZN",
+    "TESLA": "TSLA",
+    "NVIDIA": "NVDA",
+    "META": "META",
+    "FACEBOOK": "META",
+    "NETFLIX": "NFLX",
+    "S&P 500": "SPY",
+    "SP500": "SPY",
+    "S&P": "SPY",
+    "NASDAQ": "QQQ",
+    "DOW": "DIA",
+    "DOW JONES": "DIA",
+    "BITCOIN": "BTC-USD",
+    "ETHEREUM": "ETH-USD",
+}
+
+def extract_tickers(text: str) -> List[str]:
+    """Extract and normalize tickers from text."""
+    found = []
+    
+    # Check known names first (case insensitive bounds)
+    upper_text = text.upper()
+    for name, ticker in COMMON_TICKERS.items():
+        # Simple word boundary check
+        if re.search(r'\b' + re.escape(name) + r'\b', upper_text):
+            found.append(ticker)
+            
+    # Regex for standard tickers (capitals, 1-5 chars)
+    # Exclude common words like I, A, AM, PM, IS, AT, VS, OR, AND...
+    # Strict mode: must be uppercase in original text? The prompt implies natural language which might be mixed.
+    # But usually users type tickers in caps or "Apple".
+    
+    # For simplicity, extract probable tickers
+    matches = re.findall(r'\b[A-Z]{2,5}\b', text)
+    stopwords = {"AND", "OR", "THE", "FOR", "GET", "SET", "NOT", "BUT", "BY", "OF", "AT", "IN", "ON", "TO", "FROM", "VS", "GDP", "CPI", "USD", "YTD", "CEO", "CFO", "SEC", "API", "LLM", "AI"}
+    
+    for m in matches:
+        if m not in stopwords and m not in found:
+            found.append(m)
+            
+    return list(set(found))
+
+def extract_timeframe(text: str) -> Tuple[str, Optional[str], Optional[str]]:
+    """Extract timeframe description, start date, end date."""
+    
+    today = datetime.now()
+    
+    # "5y", "10 years", "start of 2020"
+    
+    # Simple regex for periods
+    match_years = re.search(r'\b(\d+)\s*y(ears?)?\b', text)
+    if match_years:
+        years = int(match_years.group(1))
+        start_date = (today - timedelta(days=years*365)).strftime("%Y-%m-%d")
+        return f"{years}y", start_date, None
+        
+    match_months = re.search(r'\b(\d+)\s*m(onths?)?\b', text)
+    if match_months:
+        months = int(match_months.group(1))
+        start_date = (today - timedelta(days=months*30)).strftime("%Y-%m-%d")
+        return f"{months}m", start_date, None
+        
+    # "Since 2021"
+    match_since = re.search(r'\bsince\s+(\d{4})\b', text)
+    if match_since:
+        year = int(match_since.group(1))
+        return f"since {year}", f"{year}-01-01", None
+        
+    # "YTD"
+    if "YTD" in text.upper():
+        start_date = f"{today.year}-01-01"
+        return "YTD", start_date, None
+        
+    return "default", None, None

sigma_terminal-3.5.0.dist-info/METADATA

@@ -0,0 +1,184 @@
+Metadata-Version: 2.4
+Name: sigma-terminal
+Version: 3.5.0
+Summary: Sigma - Finance Research Agent
+Project-URL: Homepage, https://github.com/desenyon/sigma
+Project-URL: Documentation, https://github.com/desenyon/sigma/wiki
+Project-URL: Repository, https://github.com/desenyon/sigma
+Author: Sigma Team
+License: Proprietary
+License-File: LICENSE
+Keywords: ai,analytics,backtesting,finance,investment,portfolio,quantitative,research,stocks,terminal,trading
+Classifier: Development Status :: 4 - Beta
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Financial and Insurance Industry
+Classifier: License :: Other/Proprietary License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Office/Business :: Financial :: Investment
+Requires-Python: >=3.11
+Requires-Dist: aiohttp>=3.9.0
+Requires-Dist: anthropic>=0.18.0
+Requires-Dist: google-genai>=1.0.0
+Requires-Dist: groq>=0.4.0
+Requires-Dist: httpx>=0.26.0
+Requires-Dist: kaleido>=0.2.1
+Requires-Dist: numpy>=1.26.0
+Requires-Dist: openai>=1.12.0
+Requires-Dist: pandas>=2.2.0
+Requires-Dist: pillow>=10.2.0
+Requires-Dist: plotext>=5.2.8
+Requires-Dist: plotly>=5.18.0
+Requires-Dist: pydantic-settings>=2.1.0
+Requires-Dist: pydantic>=2.6.0
+Requires-Dist: pyobjc-framework-cocoa>=10.0; sys_platform == 'darwin'
+Requires-Dist: python-dotenv>=1.0.0
+Requires-Dist: requests>=2.31.0
+Requires-Dist: rich>=13.7.0
+Requires-Dist: scipy>=1.12.0
+Requires-Dist: textual>=0.47.0
+Requires-Dist: yfinance>=0.2.36
+Provides-Extra: all
+Requires-Dist: black>=24.0.0; extra == 'all'
+Requires-Dist: lean>=1.0.0; extra == 'all'
+Requires-Dist: mypy>=1.8.0; extra == 'all'
+Requires-Dist: py2app>=0.28.0; extra == 'all'
+Requires-Dist: pytest-asyncio>=0.23.0; extra == 'all'
+Requires-Dist: pytest>=8.0.0; extra == 'all'
+Requires-Dist: ruff>=0.2.0; extra == 'all'
+Provides-Extra: dev
+Requires-Dist: black>=24.0.0; extra == 'dev'
+Requires-Dist: mypy>=1.8.0; extra == 'dev'
+Requires-Dist: py2app>=0.28.0; extra == 'dev'
+Requires-Dist: pytest-asyncio>=0.23.0; extra == 'dev'
+Requires-Dist: pytest>=8.0.0; extra == 'dev'
+Requires-Dist: ruff>=0.2.0; extra == 'dev'
+Provides-Extra: lean
+Requires-Dist: lean>=1.0.0; extra == 'lean'
+Description-Content-Type: text/markdown
+
+<h1 align="center">
+  <code>σ</code> SIGMA
+</h1>
+
+<p align="center">
+  <strong>The Terminal-Based Financial Research Agent</strong>
+</p>
+
+<p align="center">
+  <img src="https://img.shields.io/badge/version-3.5.0-blue.svg" alt="Version 3.5.0"/>
+  <img src="https://img.shields.io/badge/python-3.11+-green.svg" alt="Python 3.11+"/>
+  <img src="https://img.shields.io/badge/license-Proprietary-red.svg" alt="License"/>
+  <img src="https://img.shields.io/badge/UI-Textual-purple.svg" alt="UI Framework"/>
+</p>
+
+---
+
+## What is Sigma?
+
+Sigma is a financial analysis terminal powered by modern AI. It unifies natural language research, quantitative backtesting, and real-time market data into a single, high-performance CLI application.
+
+Unlike generic chat tools, Sigma is built for finance:
+
+- **Deterministic Tools**: Real APIs for quotes, financials, and news—no hallucinations.
+- **Local & Cloud AI**: Route queries to OpenAI, Anthropic, Gemini, or run locally with Ollama.
+- **Integrated Backtesting**: First-class support for LEAN engine to test strategies instantly.
+- **Privacy First**: Your API keys and strategies stay on your machine.
+
+---
+
+## Installation
+
+### Prerequisites
+
+- Python 3.11+
+- [Optional] Docker (for LEAN) or LEAN CLI
+- [Optional] Ollama (for local inference)
+
+### One-Command Setup
+
+Sigma includes an intelligent **Setup Agent** that handles the heavy lifting.
+
+```bash
+# Clone and install
+pip install sigma-terminal
+
+# Launch (triggers Setup Agent on first run)
+python -m sigma
+```
+
+The Setup Agent will:
+
+1. Detect your OS and Python environment.
+2. Install the LEAN backtesting engine (if missing).
+3. Install and configure Ollama (if missing).
+4. help you add API keys for data providers (Polygon, Alpha Vantage, etc.).
+
+---
+
+## Usage
+
+Sigma is designed for natural language. Just type what you need.
+
+### Market Research
+
+> "Analyze AAPL and compare it with MSFT for the last 5 years"
+> "Get me the latest earnings report for NVDA and summarize risks"
+> "Show me a chart of SPY vs QQQ YTD"
+
+### Quantitative Backtesting
+
+> "Backtest a simple moving average crossover on BTC-USD from 2020"
+> "Run a momentum strategy on TSLA, weekly rebalance"
+
+### Tool & System Control
+
+> "Switch model to local llama3"
+> "List all available tools"
+> "/backtest AAPL -s sma_cross" (Shortcuts available)
+
+---
+
+## Architecture
+
+Sigma v3.5.0 is built on a modular, event-driven architecture:
+
+| Component                  | Description                                                                                           |
+| -------------------------- | ----------------------------------------------------------------------------------------------------- |
+| **Core Engine**      | Orchestrates intent parsing, tool routing, and result synthesis.                                      |
+| **LLM Router**       | Intelligent routing between OpenAI, Anthropic, Google, and Ollama. Handles rate limits and fallbacks. |
+| **Tool Registry**    | A typed system connecting the LLM to 30+ financial data functions.                                    |
+| **Backtest Service** | Wraps the LEAN engine to stream logs and results directly to the TUI.                                 |
+| **UI (Textual)**     | A multi-pane terminal interface with real-time streaming, trace logs, and plotting.                   |
+
+---
+
+## Configuration
+
+Configuration is stored in `~/.sigma/` and managed automatically. You can also edit it manually:
+
+`~/.sigma/config.json`:
+
+```json
+{
+  "default_model": "gpt-4o",
+  "ollama_url": "http://localhost:11434",
+  "data_providers": {
+    "polygon": "ENABLED",
+    "yfinance": "FALLBACK"
+  }
+}
+```
+
+### Supported Providers
+
+- **AI**: OpenAI, Anthropic (Claude), Google (Gemini), Ollama (Local)
+- **Data**: Polygon.io, Alpha Vantage, Financial Modeling Prep, YFinance (Default)
+
+---
+
+## License
+
+Proprietary / Closed Source.
+Copyright (c) 2026 Sigma Team. All Rights Reserved.

sigma_terminal-3.5.0.dist-info/RECORD

@@ -0,0 +1,46 @@
+sigma/__init__.py,sha256=dcYKxj4v5MyKiYVaZ4_P1mLHBN9xxJWzToBeTO6Y0To,3574
+sigma/__main__.py,sha256=JR33oXObSRLsirM_8pvzTIUBr16QXS294WkQRIFI95g,114
+sigma/app.py,sha256=waFgdUzD8s67ctM9tsmzUxatK4y7rGgqZ4xQwY47v30,14493
+sigma/charts.py,sha256=FMQ6kkX0OE1SMD_fcb2z-YGwX9Lcbc2bRpWpqLMzPGo,12784
+sigma/cli.py,sha256=GCSkeQ7g1puHrgpUl7ZkyuKIMpKEzLfVhVPg3UB9Bl8,16935
+sigma/comparison.py,sha256=0q-zWGVVuI8EMz7KsLVG5isORC2zRUTFCrht2Qjt8zk,21725
+sigma/config.py,sha256=Wq9NeUn1ImRC-ZPRVVUDlpq8NZzIkx46_uGvg390Cgo,17397
+sigma/monitoring.py,sha256=7iC9DaTX8dFIygwAyEgpFQFpLPQrmLg6UW1bBeB_n7M,22672
+sigma/portfolio.py,sha256=ySypECWntDfa4pRVrNgxRsNDyOCEEgTOCUhrQIiMgBY,24020
+sigma/reporting.py,sha256=3qzuG-uITeO2z9ENOst4QcJSY2z8FyJs_eGQVe7JGM4,19157
+sigma/robustness.py,sha256=rvy-LQx9b7IZRYTI1srnL_yrVWeNAwcUFra0YSD6ZrE,23502
+sigma/setup_agent.py,sha256=W9HfaNQTLEDnDeUvzZnyzw1AQE9A67RJ-nYf_IBVEPU,7686
+sigma/strategy.py,sha256=1IucLv3vcWyQgAsMhav6Q1qf-i3yk3obYEdaOkRA4Nw,28881
+sigma/visualization.py,sha256=w5amseNd-sO8oCUpE_jnZFQD2xW-Twx6W6kk6xSb-ZA,24749
+sigma/analytics/__init__.py,sha256=WkyZgfvDSxbXKox-VV7mvnBhMC7l7y8qAv4Tfy0lYcQ,23191
+sigma/backtest/__init__.py,sha256=Wko8jkF76S_WZSpIIfFK_GDaFnLZvsnH7HAVJPB4Y_M,112
+sigma/backtest/service.py,sha256=_UzEOnmMV8sntAt9gYgUnpvn853Ne8EudoJf1Z0Zshg,4448
+sigma/backtest/simple_engine.py,sha256=7hImn0A4eO6gokWVLKERWqapN1N-T9ohc8Vof1XJo-0,13258
+sigma/core/__init__.py,sha256=vpMwe22zgDRze9LuGumTNvyRS2rzaeAgNd9_4dkaFrk,136
+sigma/core/command_router.py,sha256=sJ4TCdAV99jBn6X4gBQh03PAt0BgeCdlnizj3SvaMUk,3442
+sigma/core/engine.py,sha256=u8Z8ecvrPzaqyijHDjI4rDfdTpNjbAhxqx-5ahR8-0A,17566
+sigma/core/intent.py,sha256=w3VOUIH7fQ0nSrG2Ch9l9CVfSU7dp0wqz5FCSu3e3HQ,23519
+sigma/core/models.py,sha256=IkpvIK28rvGsjqV9cvC3NhbdiHaRmNzEJUevJkpLGyA,15745
+sigma/data/__init__.py,sha256=nqFtjuuTGhEO7bs-eVAzesKIf0w7U_38qgkGi5O53Ek,23896
+sigma/data/models.py,sha256=Zoh4s90ihPPVpUw3Nh5SqInNzf7l-jB8_I1de1OUDNQ,3694
+sigma/llm/__init__.py,sha256=_ntoRdQKScfMA8aClrcuM4PT5WxJtXAZMcmxSCmBrWI,109
+sigma/llm/rate_limit.py,sha256=vDHhQg142Wd8LrPLn0vsmO6LBsiuao-NuvHX_zE0Kh4,1537
+sigma/llm/registry.py,sha256=OXcqUaqEMkRQ7e9RaFXPyvstX9X36NH6rG1bBk1qOPU,2866
+sigma/llm/router.py,sha256=GlYwVO2lqfivRcEE-x8zOeYCZzPssLS1UTKdyws4rko,4605
+sigma/llm/providers/anthropic_provider.py,sha256=eLE-z4R_tqNGQetPqr9Nl-Vh5G4Te4CxpHnH5_8k9d0,7663
+sigma/llm/providers/base.py,sha256=OFWaCbIxWgY6bMxN5qnxydS-vWgvyLFvUacIzGaiLGM,866
+sigma/llm/providers/google_provider.py,sha256=xryjOItRLhSWcSITSSnc2OkdfNiL3Y-fh9Asiyxy6cY,8381
+sigma/llm/providers/ollama_provider.py,sha256=yqpejtZLtTjddaS99vkcUFq_8vDVHKMeMWn83fvln9E,6607
+sigma/llm/providers/openai_provider.py,sha256=UFfkGBMGVBA-Zcprc-0XrGNWxuFl0aTghYPprtbRDVY,6903
+sigma/llm/providers/sigma_cloud_provider.py,sha256=mCT445D0LSwE7Th79QTT-6Y9QRzoSmZfXBmzADrSp9M,1838
+sigma/tools/__init__.py,sha256=kD6zjXorRuxBuWUfTZXGC1LPzFdg7FMUH8Uwa-g9Bdg,710
+sigma/tools/adapter.py,sha256=2ChjA4I7wUVA26NBeGGddSOFeRp4_C-hikb4OpNzFF8,1238
+sigma/tools/backtest.py,sha256=xYunp_dm46uGBw0pxNcWsSh4UuLmE1xuaHvWzzmcGv4,53869
+sigma/tools/library.py,sha256=Yn1I9ro8Ifp1x4mtVp1uXGf3qYtQQjR4kUlv1bDFEPc,70933
+sigma/tools/registry.py,sha256=SuFT22TsaVXgboiAalCVXHOdg2FV_Vy-v5H3FFdLdm4,4072
+sigma/utils/extraction.py,sha256=vmSnDzrP_rUrYM_iUpDTFOGAtp6I5Ff_Z6RYR_SH9Og,2806
+sigma_terminal-3.5.0.dist-info/METADATA,sha256=Q0E5P7IBXYvcAJX-f1wWC2Je6Ww8g7kHCAA1ZAWPrec,6227
+sigma_terminal-3.5.0.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+sigma_terminal-3.5.0.dist-info/entry_points.txt,sha256=lFpcHMWvr4jUIO-MRrBwKhxWp-AIsRJDaWMU4dOQzlw,121
+sigma_terminal-3.5.0.dist-info/licenses/LICENSE,sha256=YGf6UbSNXayxJnonOYohaDbbV1yZQfQ1wyStkTuHS6w,1099
+sigma_terminal-3.5.0.dist-info/RECORD,,