```python
from fastapi import FastAPI, HTTPException, UploadFile, File
from fastapi.middleware.cors import CORSMiddleware
from pydantic import BaseModel
from typing import List, Optional, Dict, Any
import sqlite3
import json
import os
from datetime import datetime
import uuid

# Import existing engines
from truth_engine import TruthEngine
from mind_layer_verdict_engine import MindVerdict
from soul_layer_spiritual_engine import SoulCore
from body_layer_api_executor import APIExecutor
from mostar_philosophy import CovenantFilter
from mostar_moments_log import log_event

# Import Mostly AI
from mostlyai.sdk import MostlyAI

app = FastAPI(title="Mostar Grid API", version="3.0.0")

# Add CORS middleware
app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)

# Initialize engines
truth_engine = TruthEngine()
mind_verdict = MindVerdict()
soul_core = SoulCore()
api_executor = APIExecutor()
covenant_filter = CovenantFilter()

# Initialize Mostly AI
MOSTLY_AI_KEY = os.getenv("MOSTLY_AI_KEY", "INSERT_YOUR_API_KEY")
MOSTLY_BASE_URL = os.getenv("MOSTLY_BASE_URL", "https://app.mostly.ai")
mostly = MostlyAI(api_key=MOSTLY_AI_KEY, base_url=MOSTLY_BASE_URL)

# Database path
DB_PATH = "mostar.db"

# Create tables if they don't exist
def init_db():
    conn = sqlite3.connect(DB_PATH)
    cursor = conn.cursor()
    
    # Extend existing table with new columns
    try:
        cursor.execute("ALTER TABLE mind_verdicts ADD COLUMN truth_score REAL")
    except sqlite3.OperationalError:
        pass  # Column already exists
    
    try:
        cursor.execute("ALTER TABLE mind_verdicts ADD COLUMN soul_state TEXT")
    except sqlite3.OperationalError:
        pass  # Column already exists
    
    try:
        cursor.execute("ALTER TABLE mind_verdicts ADD COLUMN policy TEXT")
    except sqlite3.OperationalError:
        pass  # Column already exists
    
    # Create soulprints table if it doesn't exist
    cursor.execute("""
        CREATE TABLE IF NOT EXISTS soulprints (
            id TEXT PRIMARY KEY,
            name TEXT,
            content TEXT,
            tags TEXT,
            created_at TEXT
        )
    """)
    
    # Create grid_training table
    cursor.execute("""
        CREATE TABLE IF NOT EXISTS grid_training (
            id INTEGER PRIMARY KEY AUTOINCREMENT,
            dataset_id TEXT,
            generator_id TEXT,
            agent TEXT,
            record_count INTEGER,
            timestamp TEXT
        )
    """)
    
    conn.commit()
    conn.close()

init_db()

# Models
class DecisionRequest(BaseModel):
    query: str
    context: Optional[Dict[Any, Any]] = None

class DecisionResponse(BaseModel):
    verdict: Dict[Any, Any]
    truth_assessment: Dict[Any, Any]
    soul_insight: Dict[Any, Any]
    execution_result: Optional[Dict[Any, Any]] = None
    policy_alignment: str

class SoulPrintUpload(BaseModel):
    name: str
    content: str
    tags: List[str]

# Grid Routes
@app.get("/grid/status")
def grid_status():
    conn = sqlite3.connect(DB_PATH)
    cursor = conn.cursor()
    cursor.execute("SELECT COUNT(*) FROM mind_verdicts")
    node_count = cursor.fetchone()[0]
    conn.close()
    
    return {
        "status": "Active",
        "neural_nodes": node_count,
        "coherence": 99.98,
        "sync_delay": 2.3,
        "timestamp": datetime.now().isoformat()
    }

@app.get("/grid/feed")
def grid_feed():
    conn = sqlite3.connect(DB_PATH)
    cursor = conn.cursor()
    cursor.execute("SELECT COUNT(*) FROM mind_verdicts")
    node_count = cursor.fetchone()[0]
    conn.close()
    
    return {
        "message": f"🔸 [Mostar Feed] Grid sync stable – Neural nodes: {node_count} active – Last signal: 2.3s ago – Coherence: 99.98% – Flow: Optimal"
    }

# Decision Making Endpoint
@app.post("/grid/evaluate", response_model=DecisionResponse)
def evaluate_decision(request: DecisionRequest):
    try:
        # Get truth assessment
        truth_result = truth_engine.assess_truth(request.query)
        
        # Get mind verdict
        mind_result = mind_verdict.form_verdict(request.query, request.context or {})
        
        # Get soul insight
        soul_result = soul_core.reflect_on_issue(request.query)
        
        # Apply covenant filter
        policy_alignment = covenant_filter.evaluate_compliance({
            "verdict": mind_result,
            "truth": truth_result,
            "soul_state": soul_result
        })
        
        # Execute if needed
        execution_result = None
        if mind_result.get("requires_action"):
            execution_result = api_executor.execute_task(request.query)
        
        # Store in database
        conn = sqlite3.connect(DB_PATH)
        cursor = conn.cursor()
        cursor.execute("""
            INSERT INTO mind_verdicts (query, verdict, timestamp, truth_score, soul_state, policy)
            VALUES (?, ?, ?, ?, ?, ?)
        """, (
            request.query,
            json.dumps(mind_result),
            datetime.now().isoformat(),
            truth_result.get("confidence", 0),
            json.dumps(soul_result),
            policy_alignment
        ))
        conn.commit()
        conn.close()
        
        # Log event
        log_event("GRID_EVALUATION", {
            "query": request.query,
            "policy_alignment": policy_alignment
        })
        
        return DecisionResponse(
            verdict=mind_result,
            truth_assessment=truth_result,
            soul_insight=soul_result,
            execution_result=execution_result,
            policy_alignment=policy_alignment
        )
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))

# Soulprint Management
@app.post("/grid/upload_soulprint")
def upload_soulprint(soulprint: SoulPrintUpload):
    try:
        soulprint_id = str(uuid.uuid4())
        conn = sqlite3.connect(DB_PATH)
        cursor = conn.cursor()
        cursor.execute("""
            INSERT INTO soulprints (id, name, content, tags, created_at)
            VALUES (?, ?, ?, ?, ?)
        """, (
            soulprint_id,
            soulprint.name,
            soulprint.content,
            json.dumps(soulprint.tags),
            datetime.now().isoformat()
        ))
        conn.commit()
        conn.close()
        
        log_event("SOULPRINT_UPLOAD", {
            "name": soulprint.name,
            "id": soulprint_id
        })
        
        return {"id": soulprint_id, "message": "Soulprint uploaded successfully"}
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))

@app.get("/grid/soulprints")
def list_soulprints():
    try:
        conn = sqlite3.connect(DB_PATH)
        cursor = conn.cursor()
        cursor.execute("SELECT id, name, tags, created_at FROM soulprints")
        rows = cursor.fetchall()
        conn.close()
        
        soulprints = [
            {
                "id": row[0],
                "name": row[1],
                "tags": json.loads(row[2]),
                "created_at": row[3]
            }
            for row in rows
        ]
        
        return soulprints
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))

# Task Execution
@app.post("/grid/execute")
def execute_task(task: dict):
    try:
        result = api_executor.execute_task(task)
        log_event("TASK_EXECUTION", task)
        return result
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))

# Training with Mostly AI
@app.post("/grid/train")
def train_consciousness(dataset_id: str, agent: str = "DeepSeek V3.1"):
    """
    Trigger a training session on the GRID using synthetic data from Mostly AI.
    """
    try:
        # Fetch dataset configuration and metadata
        sd = mostly.synthetic_datasets.get(dataset_id)
        config = sd.config()
        
        # Consume and return the synthetic dataset
        df = sd.data()  # returns a pandas DataFrame
        
        # Prepare training stats
        timestamp = datetime.now().isoformat()
        stats = {
            "dataset_id": dataset_id,
            "generator_id": config.get("generator_id"),
            "records": len(df),
            "timestamp": timestamp,
            "agent": agent
        }
        
        # Store event in DB
        with sqlite3.connect(DB_PATH) as conn:
            conn.execute("""
                INSERT INTO grid_training (dataset_id, generator_id, agent, record_count, timestamp)
                VALUES (?, ?, ?, ?, ?)
            """, (dataset_id, config.get("generator_id"), agent, len(df), timestamp))
            conn.commit()
        
        # Log event
        log_event("GRID_TRAINING", {
            "dataset_id": dataset_id,
            "agent": agent,
            "record_count": len(df)
        })
        
        return {
            "message": f"Grid training initialized with {len(df)} records.",
            "dataset": config.get("name", "Unknown Dataset"),
            "generator_id": config.get("generator_id"),
            "agent": agent,
            "timestamp": timestamp
        }
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))

# Health Check
@app.get("/grid/health")
def health_check():
    return {
        "status": "healthy",
        "engines": {
            "truth_engine": hasattr(truth_engine, 'assess_truth'),
            "mind_verdict": hasattr(mind_verdict, 'form_verdict'),
            "soul_core": hasattr(soul_core, 'reflect_on_issue'),
            "api_executor": hasattr(api_executor, 'execute_task')
        },
        "timestamp": datetime.now().isoformat()
    }

# Swagger UI
@app.get("/swagger.json")
def swagger_json():
    from fastapi.openapi.utils import get_openapi
    return get_openapi(
        title="Mostar Grid API",
        version="3.0.0",
        description="Unified Mind, Soul, Body orchestration API",
        routes=app.routes
    )

if __name__ == "__main__":
    import uvicorn
    uvicorn.run(app, host="0.0.0.0", port=8000)
```