merge: chore/cleanup-remove-bloat-and-secrets into main

app/routers/mbal_market.py

@@ -0,0 +1,161 @@
+#!/usr/bin/env python3
+"""#17 — MBAL 10M-Label Market Maker. Public searchable database of the MBAL dataset
+(10M labeled addresses). Free search, paid API. Kaggle: cryptorugmuncher."""
+
+import contextlib
+import os
+import sqlite3
+from pathlib import Path
+
+from fastapi import APIRouter, HTTPException, Query
+from pydantic import BaseModel
+
+router = APIRouter(prefix="/api/v1/mbal", tags=["mbal-market-maker"])
+
+MBAL_DB = os.environ.get("MBAL_DB", str(Path.home() / "rmi/backend/data/mbal.db"))
+
+
+def _get_db() -> sqlite3.Connection | None:
+    """Open MBAL database connection."""
+    db_path = Path(MBAL_DB)
+    if not db_path.exists():
+        return None
+    conn = sqlite3.connect(str(db_path))
+    conn.row_factory = sqlite3.Row
+    return conn
+
+
+@router.get("/search/{address}")
+async def search_mbal(address: str):
+    """Search MBAL database for an address. Returns label if found."""
+    conn = _get_db()
+    if not conn:
+        return {
+            "address": address,
+            "found": False,
+            "database": "unavailable",
+            "note": "MBAL database not found. Visit: https://www.kaggle.com/datasets/cryptorugmuncher/mbal-10m-labeled-addresses",
+        }
+
+    try:
+        cursor = conn.execute(
+            "SELECT address, label, category, confidence, source FROM addresses WHERE address = ? LIMIT 1",
+            (address.lower(),),
+        )
+        row = cursor.fetchone()
+        if row:
+            return {
+                "address": address,
+                "found": True,
+                "label": row["label"],
+                "category": row["category"],
+                "confidence": row["confidence"],
+                "source": row["source"],
+                "dataset": "MBAL (Multi-chain Blockchain Address Labels)",
+            }
+        return {"address": address, "found": False, "database": "available"}
+    except sqlite3.Error:
+        return {"address": address, "found": False, "database": "error"}
+    finally:
+        conn.close()
+
+
+@router.get("/stats")
+async def mbal_stats():
+    """MBAL dataset statistics."""
+    conn = _get_db()
+    stats = {
+        "dataset": "MBAL — Multi-chain Blockchain Address Labels",
+        "total_addresses": 0,
+        "categories": [],
+        "chains_covered": [],
+        "kaggle_url": "https://www.kaggle.com/datasets/cryptorugmuncher/mbal-10m-labeled-addresses",
+        "database_available": conn is not None,
+    }
+    if conn:
+        try:
+            stats["total_addresses"] = conn.execute("SELECT COUNT(*) FROM addresses").fetchone()[0]
+            stats["categories"] = [
+                r[0]
+                for r in conn.execute(
+                    "SELECT DISTINCT category FROM addresses WHERE category IS NOT NULL LIMIT 20"
+                ).fetchall()
+            ]
+            with contextlib.suppress(sqlite3.Error):
+                stats["chains_covered"] = [
+                    r[0]
+                    for r in conn.execute(
+                        "SELECT DISTINCT chain FROM addresses WHERE chain IS NOT NULL LIMIT 20"
+                    ).fetchall()
+                ]
+        except sqlite3.Error:
+            pass
+        finally:
+            conn.close()
+
+    return stats
+
+
+class BulkSearchRequest(BaseModel):
+    addresses: list[str]
+    max_results: int = 100
+
+
+@router.post("/bulk-search")
+async def bulk_search_mbal(request: BulkSearchRequest):
+    """Search multiple addresses in MBAL. Paid tier (x402)."""
+    conn = _get_db()
+    if not conn:
+        raise HTTPException(503, "MBAL database unavailable")
+
+    results: list[dict] = []
+    try:
+        for addr in request.addresses[: request.max_results]:
+            cursor = conn.execute(
+                "SELECT address, label, category, confidence FROM addresses WHERE address = ? LIMIT 1", (addr.lower(),)
+            )
+            row = cursor.fetchone()
+            if row:
+                results.append(
+                    {
+                        "address": addr,
+                        "found": True,
+                        "label": row["label"],
+                        "category": row["category"],
+                        "confidence": row["confidence"],
+                    }
+                )
+            else:
+                results.append({"address": addr, "found": False})
+    except sqlite3.Error:
+        raise HTTPException(500, "Database query error")
+    finally:
+        conn.close()
+
+    return {
+        "searched": len(request.addresses[: request.max_results]),
+        "found": sum(1 for r in results if r.get("found")),
+        "results": results,
+    }
+
+
+@router.get("/category/{category}")
+async def search_by_category(category: str, limit: int = Query(20, le=100)):
+    """Search MBAL addresses by category (e.g., 'exchange', 'scam', 'mixer')."""
+    conn = _get_db()
+    if not conn:
+        raise HTTPException(503, "MBAL database unavailable")
+
+    try:
+        rows = conn.execute(
+            "SELECT address, label, category, confidence FROM addresses WHERE category = ? LIMIT ?", (category, limit)
+        ).fetchall()
+        results = [
+            {"address": r["address"], "label": r["label"], "category": r["category"], "confidence": r["confidence"]}
+            for r in rows
+        ]
+        return {"category": category, "count": len(results), "addresses": results}
+    except sqlite3.Error:
+        raise HTTPException(500, "Database query error")
+    finally:
+        conn.close()