WaveGuardClient 3.0.0__tar.gz → 3.3.0__tar.gz

{waveguardclient-3.0.0 → waveguardclient-3.3.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: WaveGuardClient
-Version: 3.0.0
+Version: 3.3.0
 Summary: Python SDK for WaveGuard — physics-based anomaly detection, fingerprinting & crypto risk API
 Author: Greg Partin
 License-Expression: MIT
@@ -36,7 +36,7 @@ Dynamic: license-file
 
 <p align="center">
   <img src="https://img.shields.io/pypi/v/WaveGuardClient?style=for-the-badge&color=blueviolet" alt="PyPI">
-  <img src="https://img.shields.io/badge/API-v2.2.0_stateless-brightgreen?style=for-the-badge" alt="v2.2.0">
+  <img src="https://img.shields.io/badge/API-v3.1.0_stateless-brightgreen?style=for-the-badge" alt="v3.1.0">
   <img src="https://img.shields.io/badge/GPU-CUDA_accelerated-76B900?style=for-the-badge&logo=nvidia" alt="CUDA">
   <img src="https://img.shields.io/badge/MCP-Claude_Desktop-orange?style=for-the-badge" alt="MCP">
   <a href="https://smithery.ai/servers/emergentphysicslab/waveguard"><img src="https://smithery.ai/badge/emergentphysicslab/waveguard" alt="Smithery"></a>
@@ -46,7 +46,7 @@ Dynamic: license-file
 
 <p align="center">
   <strong>Anomaly detection powered by wave physics. Not machine learning.</strong><br>
-  One API call. Fully stateless. Works on any data type. Zero false alarms.
+  One API call. Fully stateless. Works on any data type.
 </p>
 
 <p align="center">
@@ -125,6 +125,20 @@ Results saved to `benchmarks/benchmark_results.json`. Benchmarks use determinist
 
 </details>
 
+## Real-World Validation: Crypto Crash Detection
+
+WaveGuard powers [CryptoGuard](https://github.com/gpartin/CryptoGuard), a crypto risk scanner. Backtested against 7 historical crashes (LUNA, FTX, Celsius, 3AC, UST, SOL/FTX, TITAN):
+
+| Method | Recall | Avg Lead Time | False Positive Rate |
+|--------|--------|---------------|---------------------|
+| **WaveGuard** | **100% (7/7)** | **27.4 days** | **6.1%** |
+| Z-score baseline | 100% (7/7) | 28.4 days | 29.9% |
+| Rolling volatility | 86% (6/7) | 15.5 days | 4.0% |
+
+WaveGuard flagged FTT (FTX token) at CAUTION on October 16, 2022 — **23 days before the 94% crash** — while z-score analysis showed nothing unusual.
+
+5× fewer false alarms than statistical baselines with the same recall. Full results: [CryptoGuard backtest](https://github.com/gpartin/CryptoGuard/tree/main/backtest).
+
 ## Install
 
 ```bash

{waveguardclient-3.0.0 → waveguardclient-3.3.0}/README.md

@@ -1,6 +1,6 @@
 <p align="center">
   <img src="https://img.shields.io/pypi/v/WaveGuardClient?style=for-the-badge&color=blueviolet" alt="PyPI">
-  <img src="https://img.shields.io/badge/API-v2.2.0_stateless-brightgreen?style=for-the-badge" alt="v2.2.0">
+  <img src="https://img.shields.io/badge/API-v3.1.0_stateless-brightgreen?style=for-the-badge" alt="v3.1.0">
   <img src="https://img.shields.io/badge/GPU-CUDA_accelerated-76B900?style=for-the-badge&logo=nvidia" alt="CUDA">
   <img src="https://img.shields.io/badge/MCP-Claude_Desktop-orange?style=for-the-badge" alt="MCP">
   <a href="https://smithery.ai/servers/emergentphysicslab/waveguard"><img src="https://smithery.ai/badge/emergentphysicslab/waveguard" alt="Smithery"></a>
@@ -10,7 +10,7 @@
 
 <p align="center">
   <strong>Anomaly detection powered by wave physics. Not machine learning.</strong><br>
-  One API call. Fully stateless. Works on any data type. Zero false alarms.
+  One API call. Fully stateless. Works on any data type.
 </p>
 
 <p align="center">
@@ -89,6 +89,20 @@ Results saved to `benchmarks/benchmark_results.json`. Benchmarks use determinist
 
 </details>
 
+## Real-World Validation: Crypto Crash Detection
+
+WaveGuard powers [CryptoGuard](https://github.com/gpartin/CryptoGuard), a crypto risk scanner. Backtested against 7 historical crashes (LUNA, FTX, Celsius, 3AC, UST, SOL/FTX, TITAN):
+
+| Method | Recall | Avg Lead Time | False Positive Rate |
+|--------|--------|---------------|---------------------|
+| **WaveGuard** | **100% (7/7)** | **27.4 days** | **6.1%** |
+| Z-score baseline | 100% (7/7) | 28.4 days | 29.9% |
+| Rolling volatility | 86% (6/7) | 15.5 days | 4.0% |
+
+WaveGuard flagged FTT (FTX token) at CAUTION on October 16, 2022 — **23 days before the 94% crash** — while z-score analysis showed nothing unusual.
+
+5× fewer false alarms than statistical baselines with the same recall. Full results: [CryptoGuard backtest](https://github.com/gpartin/CryptoGuard/tree/main/backtest).
+
 ## Install
 
 ```bash

{waveguardclient-3.0.0 → waveguardclient-3.3.0}/WaveGuardClient.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: WaveGuardClient
-Version: 3.0.0
+Version: 3.3.0
 Summary: Python SDK for WaveGuard — physics-based anomaly detection, fingerprinting & crypto risk API
 Author: Greg Partin
 License-Expression: MIT
@@ -36,7 +36,7 @@ Dynamic: license-file
 
 <p align="center">
   <img src="https://img.shields.io/pypi/v/WaveGuardClient?style=for-the-badge&color=blueviolet" alt="PyPI">
-  <img src="https://img.shields.io/badge/API-v2.2.0_stateless-brightgreen?style=for-the-badge" alt="v2.2.0">
+  <img src="https://img.shields.io/badge/API-v3.1.0_stateless-brightgreen?style=for-the-badge" alt="v3.1.0">
   <img src="https://img.shields.io/badge/GPU-CUDA_accelerated-76B900?style=for-the-badge&logo=nvidia" alt="CUDA">
   <img src="https://img.shields.io/badge/MCP-Claude_Desktop-orange?style=for-the-badge" alt="MCP">
   <a href="https://smithery.ai/servers/emergentphysicslab/waveguard"><img src="https://smithery.ai/badge/emergentphysicslab/waveguard" alt="Smithery"></a>
@@ -46,7 +46,7 @@ Dynamic: license-file
 
 <p align="center">
   <strong>Anomaly detection powered by wave physics. Not machine learning.</strong><br>
-  One API call. Fully stateless. Works on any data type. Zero false alarms.
+  One API call. Fully stateless. Works on any data type.
 </p>
 
 <p align="center">
@@ -125,6 +125,20 @@ Results saved to `benchmarks/benchmark_results.json`. Benchmarks use determinist
 
 </details>
 
+## Real-World Validation: Crypto Crash Detection
+
+WaveGuard powers [CryptoGuard](https://github.com/gpartin/CryptoGuard), a crypto risk scanner. Backtested against 7 historical crashes (LUNA, FTX, Celsius, 3AC, UST, SOL/FTX, TITAN):
+
+| Method | Recall | Avg Lead Time | False Positive Rate |
+|--------|--------|---------------|---------------------|
+| **WaveGuard** | **100% (7/7)** | **27.4 days** | **6.1%** |
+| Z-score baseline | 100% (7/7) | 28.4 days | 29.9% |
+| Rolling volatility | 86% (6/7) | 15.5 days | 4.0% |
+
+WaveGuard flagged FTT (FTX token) at CAUTION on October 16, 2022 — **23 days before the 94% crash** — while z-score analysis showed nothing unusual.
+
+5× fewer false alarms than statistical baselines with the same recall. Full results: [CryptoGuard backtest](https://github.com/gpartin/CryptoGuard/tree/main/backtest).
+
 ## Install
 
 ```bash

{waveguardclient-3.0.0 → waveguardclient-3.3.0}/mcp_server/server.py

@@ -371,7 +371,7 @@ class MCPStdioServer:
     def __init__(self) -> None:
         self.server_info = {
             "name": "waveguard",
-            "version": "2.3.0",
+            "version": "3.1.0",
         }
 
     def handle_message(self, msg: dict) -> Optional[dict]:
@@ -403,6 +403,20 @@ class MCPStdioServer:
                 "result": {"tools": TOOLS},
             }
 
+        elif method == "resources/list":
+            return {
+                "jsonrpc": "2.0",
+                "id": msg_id,
+                "result": {"resources": []},
+            }
+
+        elif method == "prompts/list":
+            return {
+                "jsonrpc": "2.0",
+                "id": msg_id,
+                "result": {"prompts": []},
+            }
+
         elif method == "tools/call":
             tool_name = params.get("name", "")
             arguments = params.get("arguments", {})
@@ -435,7 +449,7 @@ class MCPStdioServer:
     def run_stdio(self) -> None:
         """Run the MCP server on stdin/stdout."""
         sys.stderr.write(
-            f"WaveGuard MCP server v2.3.0 started (API: {API_URL})\n"
+            f"WaveGuard MCP server v3.1.0 started (API: {API_URL})\n"
         )
         sys.stderr.flush()
 
@@ -470,7 +484,7 @@ def run_http_server(port: int = 3001) -> None:
         print("HTTP transport requires: pip install fastapi uvicorn")
         sys.exit(1)
 
-    mcp_app = FA(title="WaveGuard MCP Server", version="2.3.0")
+    mcp_app = FA(title="WaveGuard MCP Server", version="3.1.0")
     server = MCPStdioServer()
 
     @mcp_app.post("/mcp")
@@ -481,7 +495,7 @@ def run_http_server(port: int = 3001) -> None:
     async def mcp_tools() -> dict:  # type: ignore[type-arg]
         return {"tools": TOOLS}
 
-    print(f"WaveGuard MCP HTTP server v2.3.0 on port {port}")
+    print(f"WaveGuard MCP HTTP server v3.1.0 on port {port}")
     uvicorn.run(mcp_app, host="0.0.0.0", port=port)
 
 
@@ -492,7 +506,7 @@ def run_http_server(port: int = 3001) -> None:
 def main():
     """Entry point for `waveguard-mcp` console script."""
     parser = argparse.ArgumentParser(
-        description="WaveGuard MCP Server v2.3.0"
+        description="WaveGuard MCP Server v3.0.0"
     )
     parser.add_argument(
         "--http",

{waveguardclient-3.0.0 → waveguardclient-3.3.0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "WaveGuardClient"
-version = "3.0.0"
+version = "3.3.0"
 description = "Python SDK for WaveGuard — physics-based anomaly detection, fingerprinting & crypto risk API"
 readme = "README.md"
 license = "MIT"

{waveguardclient-3.0.0 → waveguardclient-3.3.0}/waveguard/client.py

@@ -42,7 +42,7 @@ from .exceptions import (
     ServerError,
 )
 
-__version__ = "3.0.0"
+__version__ = "3.2.0"
 
 logger = logging.getLogger("waveguard")
 
@@ -145,15 +145,17 @@ class FingerprintResult:
     Attributes
     ----------
     fingerprint : list[float]
-        52-dimensional physics embedding vector.
+        Physics embedding vector (52-dim at L0, 62-dim at L1).
     dimensions : int
-        Number of dimensions (typically 52).
+        Number of dimensions (52 at Level 0, 62 at Level 1).
     labels : list[str]
         Human-readable label for each dimension.
     encoder_type : str
         Encoder used to map input data onto the lattice.
     latency_ms : float
         Server-side processing time.
+    field_level : int
+        Field level used (0=real scalar, 1=complex).
     raw : dict
         Full JSON response.
     """
@@ -163,6 +165,7 @@ class FingerprintResult:
     labels: List[str]
     encoder_type: str
     latency_ms: float
+    field_level: int = 0
     raw: Dict[str, Any] = field(default_factory=dict)
 
 
@@ -177,9 +180,9 @@ class CompareResult:
     distance : float
         Euclidean distance between fingerprints.
     fingerprint_a : list[float]
-        52-dim embedding of data_a.
-    fingerprint_b : list[float]
-        52-dim embedding of data_b.
+        Physics embedding of data_a.
+    fingerprint_b: List[float]
+        Physics embedding of data_b.
     dimensions : int
         Number of dimensions.
     encoder_type : str
@@ -266,6 +269,7 @@ class WaveGuard:
         test: List[Any],
         encoder_type: Optional[str] = None,
         sensitivity: Optional[float] = None,
+        field_level: int = 0,
     ) -> ScanResult:
         """Scan test data for anomalies against a training baseline.
 
@@ -282,11 +286,15 @@ class WaveGuard:
             1+ samples to check for anomalies.
         encoder_type : str, optional
             Force a specific encoder: ``"json"``, ``"numeric"``,
-            ``"text"``, ``"timeseries"``, ``"tabular"``.
+            ``"text"``, ``"timeseries"``, ``"tabular"``,
+            ``"complex_numeric"``.
             Leave *None* for auto-detection (recommended).
         sensitivity : float, optional
             Detection sensitivity in the range 0.5–3.0.
             Lower values are more sensitive (flag more anomalies).
+        field_level : int
+            Physics field complexity. 0 = real scalar (default).
+            1 = complex field (phase-aware, 62-dim fingerprints).
 
         Returns
         -------
@@ -302,6 +310,8 @@ class WaveGuard:
             body["encoder_type"] = encoder_type
         if sensitivity is not None:
             body["sensitivity"] = sensitivity
+        if field_level:
+            body["field_level"] = field_level
 
         resp = self._post("/v1/scan", body)
         return self._parse_scan(resp, len(training), len(test))
@@ -310,8 +320,9 @@ class WaveGuard:
         self,
         data: Any,
         encoder_type: Optional[str] = None,
+        field_level: int = 0,
     ) -> FingerprintResult:
-        """Get a 52-dimensional physics embedding of any data item.
+        """Get a physics embedding of any data item.
 
         Parameters
         ----------
@@ -319,16 +330,21 @@ class WaveGuard:
             A single data item to fingerprint (JSON object, list, string, etc.).
         encoder_type : str, optional
             Force a specific encoder.  Leave *None* for auto-detection.
+        field_level : int
+            0 = real scalar (52-dim, default).
+            1 = complex field (62-dim, includes phase statistics).
 
         Returns
         -------
         FingerprintResult
-            ``.fingerprint`` is the 52-dim vector.
+            ``.fingerprint`` is the embedding vector (52 or 62 dims).
             ``.labels`` names each dimension.
         """
         body: Dict[str, Any] = {"data": data}
         if encoder_type is not None:
             body["encoder_type"] = encoder_type
+        if field_level:
+            body["field_level"] = field_level
 
         resp = self._post("/v1/fingerprint", body)
         return FingerprintResult(
@@ -337,6 +353,7 @@ class WaveGuard:
             labels=resp.get("labels", []),
             encoder_type=resp.get("encoder_type", "auto"),
             latency_ms=resp.get("latency_ms", 0.0),
+            field_level=resp.get("field_level", 0),
             raw=resp,
         )