"""
vTiger Sync Service for Time Tracking Module
=============================================

🚨 KRITISK: Denne service må KUN læse data fra vTiger.
Ingen opdateringer, ingen statusændringer, ingen skrivninger.

Formål:
- Hent ModComments (tidsregistreringer) fra vTiger
- Hent HelpDesk/ProjectTask (cases) fra vTiger  
- Hent Accounts (kunder) fra vTiger
- Gem alt i tmodule_* tabeller (isoleret)

Safety Flags:
- TIMETRACKING_VTIGER_READ_ONLY = True (default)
- TIMETRACKING_VTIGER_DRY_RUN = True (default)
"""

import logging
import hashlib
import json
import asyncio
from datetime import datetime
from typing import List, Dict, Optional, Any
from decimal import Decimal

import aiohttp
from fastapi import HTTPException

from app.core.config import settings
from app.core.database import execute_query, execute_insert, execute_update, execute_query_single
from app.timetracking.backend.models import TModuleSyncStats
from app.timetracking.backend.audit import audit

logger = logging.getLogger(__name__)


class TimeTrackingVTigerService:
    """
    vTiger integration for Time Tracking Module.
    
    🔒 READ-ONLY service - ingen skrivninger til vTiger tilladt.
    """
    
    def __init__(self):
        self.base_url = settings.VTIGER_URL
        self.username = settings.VTIGER_USERNAME
        self.api_key = settings.VTIGER_API_KEY
        self.rest_endpoint = f"{self.base_url}/restapi/v1/vtiger/default"
        
        # Safety flags
        self.read_only = settings.TIMETRACKING_VTIGER_READ_ONLY
        self.dry_run = settings.TIMETRACKING_VTIGER_DRY_RUN
        
        # Log safety status
        if self.read_only:
            logger.warning("🔒 TIMETRACKING vTiger READ-ONLY mode: Enabled")
        if self.dry_run:
            logger.warning("🏃 TIMETRACKING vTiger DRY-RUN mode: Enabled")
        
        if not self.read_only:
            logger.error("⚠️ WARNING: TIMETRACKING vTiger READ-ONLY disabled! This violates module isolation!")
    
    def _get_auth(self) -> aiohttp.BasicAuth:
        """Get HTTP Basic Auth"""
        # Prefer API key over password
        auth_value = self.api_key if self.api_key else self.password
        return aiohttp.BasicAuth(self.username, auth_value)
    
    def _calculate_hash(self, data: Dict[str, Any]) -> str:
        """Calculate SHA256 hash of data for change detection"""
        # Sort keys for consistent hashing
        json_str = json.dumps(data, sort_keys=True)
        return hashlib.sha256(json_str.encode()).hexdigest()
    
    async def _retrieve(self, record_id: str) -> Dict:
        """
        Retrieve full record details via vTiger REST API.
        This gets ALL fields including relationships that query doesn't return.
        
        🔍 READ-ONLY operation
        """
        try:
            async with aiohttp.ClientSession() as session:
                async with session.get(
                    f"{self.rest_endpoint}/retrieve",
                    params={"id": record_id},
                    auth=self._get_auth(),
                    timeout=aiohttp.ClientTimeout(total=30)
                ) as response:
                    text = await response.text()
                    
                    if response.status != 200:
                        logger.error(f"❌ vTiger retrieve failed: HTTP {response.status} for {record_id}")
                        return {}
                    
                    try:
                        data = json.loads(text)
                    except json.JSONDecodeError:
                        logger.error(f"❌ Invalid JSON in retrieve response for {record_id}")
                        return {}
                    
                    if not data.get('success'):
                        logger.error(f"❌ vTiger retrieve failed for {record_id}: {data.get('error', {})}")
                        return {}
                    
                    return data.get('result', {})
        
        except Exception as e:
            logger.error(f"❌ Error retrieving {record_id}: {e}")
            return {}
    
    async def _query(self, query_string: str) -> List[Dict]:
        """
        Execute SQL-like query against vTiger REST API.
        
        🔍 READ-ONLY operation
        """
        try:
            async with aiohttp.ClientSession() as session:
                async with session.get(
                    f"{self.rest_endpoint}/query",
                    params={"query": query_string},
                    auth=self._get_auth(),
                    timeout=aiohttp.ClientTimeout(total=60)
                ) as response:
                    text = await response.text()
                    
                    if response.status != 200:
                        logger.error(f"❌ vTiger query failed: HTTP {response.status}")
                        logger.error(f"Query: {query_string}")
                        logger.error(f"Response body: {text[:1000]}")
                        logger.error(f"Response headers: {dict(response.headers)}")
                        raise HTTPException(
                            status_code=response.status,
                            detail=f"vTiger API error: {text[:200]}"
                        )
                    
                    # vTiger returns text/json, not application/json
                    try:
                        data = json.loads(text)
                    except json.JSONDecodeError as e:
                        logger.error(f"❌ Invalid JSON in query response: {text[:200]}")
                        raise HTTPException(status_code=500, detail="Invalid JSON from vTiger")
                    
                    # Check vTiger success flag
                    if not data.get('success'):
                        error_msg = data.get('error', {}).get('message', 'Unknown error')
                        logger.error(f"❌ vTiger query failed: {error_msg}")
                        logger.error(f"Query: {query_string}")
                        raise HTTPException(status_code=400, detail=f"vTiger error: {error_msg}")
                    
                    result = data.get('result', [])
                    logger.info(f"✅ Query returned {len(result)} records")
                    return result
        
        except aiohttp.ClientError as e:
            logger.error(f"❌ vTiger connection error: {e}")
            raise HTTPException(status_code=503, detail=f"Cannot connect to vTiger: {str(e)}")
    
    async def _retrieve(self, record_id: str) -> Dict:
        """
        Retrieve single record by ID.
        
        🔍 READ-ONLY operation
        """
        try:
            async with aiohttp.ClientSession() as session:
                async with session.get(
                    f"{self.rest_endpoint}/retrieve",
                    params={"id": record_id},
                    auth=self._get_auth(),
                    timeout=aiohttp.ClientTimeout(total=30)
                ) as response:
                    if response.status != 200:
                        error_text = await response.text()
                        logger.error(f"❌ vTiger retrieve failed: {response.status} - {error_text}")
                        return {}
                    
                    data = json.loads(await response.text())
                    
                    if not data.get('success', False):
                        return {}
                    
                    return data.get('result', {})
        
        except aiohttp.ClientError as e:
            logger.error(f"❌ vTiger retrieve error: {e}")
            return {}
    
    async def test_connection(self) -> bool:
        """Test vTiger connection"""
        try:
            logger.info("🔍 Testing vTiger connection...")
            
            async with aiohttp.ClientSession() as session:
                async with session.get(
                    f"{self.rest_endpoint}/me",
                    auth=self._get_auth(),
                    timeout=aiohttp.ClientTimeout(total=10)
                ) as response:
                    if response.status == 200:
                        logger.info("✅ vTiger connection successful")
                        return True
                    else:
                        logger.error(f"❌ vTiger connection failed: {response.status}")
                        return False
        
        except Exception as e:
            logger.error(f"❌ vTiger connection error: {e}")
            return False
    
    async def update_timelog_billed(self, vtiger_ids: List[str], hub_order_id: int) -> Dict[str, int]:
        """
        Update Timelog records in vTiger with Hub ordre ID (markér som faktureret).
        
        🚨 WRITE operation - respekterer safety flags
        
        Args:
            vtiger_ids: Liste af vTiger Timelog IDs (format: "43x1234")
            hub_order_id: Hub ordre ID til at sætte i billed_via_thehub_id felt
            
        Returns:
            Dict with success/error counts
        """
        if self.read_only:
            logger.warning(f"🔒 READ-ONLY mode: Would update {len(vtiger_ids)} timelogs with Hub order {hub_order_id}")
            return {"updated": 0, "errors": 0, "message": "READ-ONLY mode"}
        
        if self.dry_run:
            logger.warning(f"🏃 DRY-RUN mode: Would update {len(vtiger_ids)} timelogs with Hub order {hub_order_id}")
            return {"updated": 0, "errors": 0, "message": "DRY-RUN mode"}
        
        stats = {"updated": 0, "errors": 0}
        
        logger.info(f"📝 Updating {len(vtiger_ids)} timelogs in vTiger with Hub order {hub_order_id}...")
        
        try:
            async with aiohttp.ClientSession() as session:
                for vtiger_id in vtiger_ids:
                    try:
                        # Build update payload
                        update_url = f"{self.rest_endpoint}/update"
                        payload = {
                            "elementType": "Timelog",
                            "element": {
                                "id": vtiger_id,
                                "billed_via_thehub_id": str(hub_order_id),
                                "cf_timelog_invoiced": "1"  # Marker som faktureret
                            }
                        }
                        
                        async with session.post(
                            update_url,
                            json=payload,
                            auth=self._get_auth(),
                            timeout=aiohttp.ClientTimeout(total=30)
                        ) as response:
                            text = await response.text()
                            
                            if response.status == 200:
                                data = json.loads(text)
                                if data.get('success'):
                                    logger.debug(f"✅ Updated timelog {vtiger_id}")
                                    stats["updated"] += 1
                                else:
                                    error_msg = data.get('error', {}).get('message', 'Unknown error')
                                    logger.error(f"❌ vTiger error for {vtiger_id}: {error_msg}")
                                    stats["errors"] += 1
                            else:
                                logger.error(f"❌ HTTP {response.status} for {vtiger_id}: {text[:200]}")
                                stats["errors"] += 1
                    
                    except Exception as e:
                        logger.error(f"❌ Error updating timelog {vtiger_id}: {e}")
                        stats["errors"] += 1
                        
        except Exception as e:
            logger.error(f"❌ Batch update error: {e}")
            stats["errors"] += len(vtiger_ids)
        
        logger.info(f"✅ vTiger update complete: {stats['updated']} updated, {stats['errors']} errors")
        return stats
    
    async def _fetch_user_name(self, user_id: str) -> str:
        """
        Fetch user name from vTiger using retrieve API.
        
        Args:
            user_id: vTiger user ID (e.g., "19x1")
            
        Returns:
            User's full name or user_id if not found
        """
        try:
            user_data = await self._retrieve(user_id)
            
            if not user_data:
                return user_id
            
            # Build full name from first + last, fallback to username
            first_name = user_data.get('first_name', '').strip()
            last_name = user_data.get('last_name', '').strip()
            user_name = user_data.get('user_name', '').strip()
            
            if first_name and last_name:
                return f"{first_name} {last_name}"
            elif first_name:
                return first_name
            elif last_name:
                return last_name
            elif user_name:
                return user_name
            else:
                return user_id
        
        except Exception as e:
            logger.debug(f"Could not fetch user {user_id}: {e}")
            return user_id
            return False
    
    async def sync_customers(self, limit: int = 1000) -> Dict[str, int]:
        """
        Sync Accounts (customers) from vTiger to tmodule_customers.
        
        Uses ID-based pagination to fetch all accounts.
        
        Returns: {imported: X, updated: Y, skipped: Z}
        """
        logger.info("🔍 Syncing customers from vTiger...")
        
        stats = {"imported": 0, "updated": 0, "skipped": 0, "errors": 0}
        
        try:
            # Fetch ALL accounts using pagination (vTiger has 200 record limit)
            all_accounts = []
            last_id = None
            page = 1
            
            while True:
                if last_id:
                    query = f"SELECT * FROM Accounts WHERE id > '{last_id}' ORDER BY id LIMIT 200;"
                else:
                    query = "SELECT * FROM Accounts ORDER BY id LIMIT 200;"
                
                accounts = await self._query(query)
                
                if not accounts:
                    break
                
                all_accounts.extend(accounts)
                last_id = accounts[-1]['id']
                
                logger.info(f"📥 Fetched page {page}: {len(accounts)} accounts (last_id: {last_id})")
                
                # Safety: if we got less than 200, we're done
                if len(accounts) < 200:
                    break
                    
                page += 1
            
            logger.info(f"📥 Total fetched: {len(all_accounts)} accounts from vTiger")
            
            for account in all_accounts:
                try:
                    vtiger_id = account.get('id', '')
                    if not vtiger_id:
                        logger.warning("⚠️ Skipping account without ID")
                        stats["skipped"] += 1
                        continue
                    
                    # Calculate hash for change detection
                    data_hash = self._calculate_hash(account)
                    
                    # Check if exists
                    existing = execute_query_single(
                        "SELECT id, sync_hash FROM tmodule_customers WHERE vtiger_id = %s",
                        (vtiger_id,))
                    
                    if existing:
                        # Check if data changed
                        if existing['sync_hash'] == data_hash:
                            logger.debug(f"⏭️ No changes for customer {vtiger_id}")
                            stats["skipped"] += 1
                            continue
                    if existing:
                        # Update existing
                        execute_update(
                            """UPDATE tmodule_customers 
                               SET name = %s, email = %s, economic_customer_number = %s,
                                   vtiger_data = %s::jsonb, sync_hash = %s, 
                                   last_synced_at = CURRENT_TIMESTAMP
                               WHERE vtiger_id = %s""",
                            (
                                account.get('accountname', 'Unknown'),
                                account.get('email1', None),
                                int(account.get('cf_854')) if account.get('cf_854') else None,
                                json.dumps(account),
                                data_hash,
                                vtiger_id
                            )
                        )
                        logger.debug(f"✏️ Updated customer {vtiger_id}")
                        stats["updated"] += 1
                    else:
                        # Insert new
                        execute_insert(
                            """INSERT INTO tmodule_customers 
                               (vtiger_id, name, email, economic_customer_number, 
                                vtiger_data, sync_hash, last_synced_at)
                               VALUES (%s, %s, %s, %s, %s::jsonb, %s, CURRENT_TIMESTAMP)""",
                            (
                                vtiger_id,
                                account.get('accountname', 'Unknown'),
                                account.get('email1', None),
                                int(account.get('cf_854')) if account.get('cf_854') else None,
                                json.dumps(account),
                                data_hash
                            )
                        )
                        logger.debug(f"➕ Imported customer {vtiger_id}")
                        stats["imported"] += 1
                
                except Exception as e:
                    logger.error(f"❌ Error processing account {account.get('id', 'unknown')}: {e}")
                    stats["errors"] += 1
            
            logger.info(f"✅ Customer sync complete: {stats}")
            return stats
        
        except Exception as e:
            logger.error(f"❌ Customer sync failed: {e}")
            raise
    
    async def sync_cases(self, limit: int = 5000, fetch_comments: bool = False) -> Dict[str, int]:
        """
        Sync HelpDesk tickets (cases) from vTiger to tmodule_cases.
        
        Args:
            limit: Maximum number of cases to sync
            fetch_comments: Whether to fetch ModComments for each case (slow - rate limited)
        
        Returns: {imported: X, updated: Y, skipped: Z}
        """
        if fetch_comments:
            logger.info(f"🔍 Syncing up to {limit} cases from vTiger WITH comments (slow)...")
        else:
            logger.info(f"🔍 Syncing up to {limit} cases from vTiger WITHOUT comments (fast)...")
        
        stats = {"imported": 0, "updated": 0, "skipped": 0, "errors": 0}
        
        try:
            # vTiger API doesn't support OFFSET - use id-based pagination instead
            all_tickets = []
            last_id = "0x0"  # Start from beginning
            batch_size = 100  # Conservative batch size to avoid timeouts
            max_batches = limit // batch_size + 1
            
            for batch_num in range(max_batches):
                # Use id > last_id for pagination (vTiger format: 39x1234)
                query = f"SELECT * FROM Cases WHERE id > '{last_id}' ORDER BY id LIMIT {batch_size};"
                batch = await self._query(query)
                
                if not batch:  # No more records
                    break
                
                all_tickets.extend(batch)
                last_id = batch[-1].get('id', last_id)  # Get last ID for next iteration
                
                logger.info(f"📥 Fetched {len(batch)} cases (total: {len(all_tickets)}, last_id: {last_id})")
                
                if len(batch) < batch_size:  # Last batch
                    break
                
                if len(all_tickets) >= limit:  # Reached limit
                    break
            
            tickets = all_tickets[:limit]  # Trim to requested limit
            logger.info(f"✅ Total fetched: {len(tickets)} HelpDesk tickets from vTiger")
            
            for ticket in tickets:
                try:
                    vtiger_id = ticket.get('id', '')
                    if not vtiger_id:
                        stats["skipped"] += 1
                        continue
                    
                    # Get related account (customer)
                    account_id = ticket.get('parent_id', '')
                    if not account_id:
                        logger.warning(f"⚠️ HelpDesk {vtiger_id} has no parent account")
                        stats["skipped"] += 1
                        continue
                    
                    # Find customer in our DB
                    customer = execute_query_single(
                        "SELECT id FROM tmodule_customers WHERE vtiger_id = %s",
                        (account_id,))
                    
                    if not customer:
                        logger.warning(f"⚠️ Customer {account_id} not found - sync customers first")
                        stats["skipped"] += 1
                        continue
                    
                    customer_id = customer['id']
                    
                    # Fetch internal comments for this case (with rate limiting) - ONLY if enabled
                    internal_comments = []
                    if fetch_comments:
                        internal_comments = await self._get_case_comments(vtiger_id)
                        # Small delay to avoid rate limiting (vTiger allows ~2-3 requests/sec)
                        await asyncio.sleep(0.4)  # 400ms between comment fetches
                    
                    # Merge comments into ticket data before storing
                    ticket_with_comments = ticket.copy()
                    if internal_comments:
                        ticket_with_comments['internal_comments'] = internal_comments
                    
                    # Calculate hash AFTER adding comments (so changes to comments trigger update)
                    data_hash = self._calculate_hash(ticket_with_comments)
                    
                    # Check if exists
                    existing = execute_query_single(
                        "SELECT id, sync_hash FROM tmodule_cases WHERE vtiger_id = %s",
                        (vtiger_id,))
                    
                    if existing:
                        if existing['sync_hash'] == data_hash:
                            stats["skipped"] += 1
                            continue
                        
                        # Update
                        execute_update(
                            """UPDATE tmodule_cases 
                               SET customer_id = %s, title = %s, description = %s,
                                   status = %s, priority = %s, module_type = %s,
                                   vtiger_data = %s::jsonb, sync_hash = %s,
                                   last_synced_at = CURRENT_TIMESTAMP
                               WHERE vtiger_id = %s""",
                            (
                                customer_id,
                                ticket.get('title') or ticket.get('ticket_title', 'No Title'),
                                ticket.get('description', None),
                                ticket.get('ticketstatus', None),
                                ticket.get('ticketpriorities', None),
                                'HelpDesk',
                                json.dumps(ticket_with_comments),
                                data_hash,
                                vtiger_id
                            )
                        )
                        stats["updated"] += 1
                    else:
                        # Insert
                        case_id = execute_insert(
                            """INSERT INTO tmodule_cases 
                               (vtiger_id, customer_id, title, description, status, 
                                priority, module_type, vtiger_data, sync_hash, last_synced_at)
                               VALUES (%s, %s, %s, %s, %s, %s, %s, %s::jsonb, %s, CURRENT_TIMESTAMP)""",
                            (
                                vtiger_id,
                                customer_id,
                                ticket.get('title') or ticket.get('ticket_title', 'No Title'),
                                ticket.get('description', None),
                                ticket.get('ticketstatus', None),
                                ticket.get('ticketpriorities', None),
                                'HelpDesk',
                                json.dumps(ticket_with_comments),
                                data_hash
                            )
                        )
                        stats["imported"] += 1
                
                except Exception as e:
                    logger.error(f"❌ Error processing case {ticket.get('id', 'unknown')}: {e}")
                    stats["errors"] += 1
            
            logger.info(f"✅ Case sync complete: {stats}")
            return stats
        
        except Exception as e:
            logger.error(f"❌ Case sync failed: {e}")
            raise
    
    async def _get_case_comments(self, case_id: str) -> List[Dict]:
        """
        Fetch all ModComments (internal comments) for a specific case from vTiger.
        
        Args:
            case_id: vTiger case ID (format: "32x1234")
            
        Returns:
            List of comment dicts with structure: {text, author, date, created_at}
            Sorted by creation date (newest first)
        """
        try:
            # Query ModComments where related_to = case_id
            query = f"SELECT * FROM ModComments WHERE related_to = '{case_id}' ORDER BY createdtime DESC;"
            comments = await self._query(query)
            
            if not comments:
                return []
            
            # Transform vTiger format to internal format
            formatted_comments = []
            for comment in comments:
                formatted_comments.append({
                    "text": comment.get("commentcontent", ""),
                    "author": comment.get("assigned_user_id", "Unknown"),  # Will be user ID - could enhance with name lookup
                    "date": comment.get("createdtime", "")[:10],  # Format: YYYY-MM-DD from YYYY-MM-DD HH:MM:SS
                    "created_at": comment.get("createdtime", "")
                })
            
            logger.info(f"📝 Fetched {len(formatted_comments)} comments for case {case_id}")
            return formatted_comments
            
        except HTTPException as e:
            # Rate limit or API error - log but don't fail sync
            if "429" in str(e.detail) or "TOO_MANY_REQUESTS" in str(e.detail):
                logger.warning(f"⚠️ Rate limited fetching comments for case {case_id} - skipping")
            else:
                logger.error(f"❌ API error fetching comments for case {case_id}: {e.detail}")
            return []
        except Exception as e:
            logger.error(f"❌ Failed to fetch comments for case {case_id}: {e}")
            return []  # Return empty list on error - don't fail entire sync
    
    async def sync_case_comments(self, case_vtiger_id: str) -> Dict[str, Any]:
        """
        Sync comments for a specific case (for on-demand updates).
        
        Args:
            case_vtiger_id: vTiger case ID (format: "39x1234")
            
        Returns:
            Dict with success status and comment count
        """
        try:
            # Fetch comments
            comments = await self._get_case_comments(case_vtiger_id)
            
            if not comments:
                return {"success": True, "comments": 0, "message": "No comments found"}
            
            # Update case in database
            execute_update(
                """UPDATE tmodule_cases 
                   SET vtiger_data = jsonb_set(
                       COALESCE(vtiger_data, '{}'::jsonb),
                       '{internal_comments}',
                       %s::jsonb
                   ),
                   last_synced_at = CURRENT_TIMESTAMP
                   WHERE vtiger_id = %s""",
                (json.dumps(comments), case_vtiger_id)
            )
            
            logger.info(f"✅ Synced {len(comments)} comments for case {case_vtiger_id}")
            return {"success": True, "comments": len(comments), "message": f"Synced {len(comments)} comments"}
            
        except Exception as e:
            logger.error(f"❌ Failed to sync comments for case {case_vtiger_id}: {e}")
            return {"success": False, "comments": 0, "error": str(e)}
    
    async def sync_time_entries(self, limit: int = 3000) -> Dict[str, int]:
        """
        Sync time entries from vTiger Timelog module to tmodule_times.
        
        vTiger's Timelog module contains detailed time entries with:
        - timelognumber: Unique ID (TL1234)
        - duration: Time in seconds
        - relatedto: Reference to Case/Account
        - is_billable: '1' = yes, '0' = no
        - cf_timelog_invoiced: '1' = has been invoiced
        
        We only sync entries where:
        - relatedto is not empty (linked to a Case or Account)
        - Has valid duration > 0
        
        NOTE: is_billable and cf_timelog_invoiced fields are not reliably populated in vTiger,
        so we sync all timelogs and let the approval workflow decide what to bill.
        """
        logger.info(f"🔍 Syncing all timelogs from vTiger with valid relatedto...")
        
        stats = {"imported": 0, "updated": 0, "skipped": 0, "errors": 0}
        
        try:
            # Cache for user names (avoid fetching same user multiple times)
            user_name_cache = {}
            
            # vTiger API doesn't support OFFSET - use id-based pagination instead
            all_timelogs = []
            last_id = "0x0"  # Start from beginning
            batch_size = 100  # Conservative batch size
            max_batches = limit // batch_size + 1
            
            for batch_num in range(max_batches):
                # Use id > last_id for pagination (vTiger format: 43x1234)
                # NOTE: vTiger query API ignores WHERE on custom fields, so we fetch all and filter later
                query = f"SELECT * FROM Timelog WHERE id > '{last_id}' ORDER BY id LIMIT {batch_size};"
                batch = await self._query(query)
                
                if not batch:  # No more records
                    break
                
                all_timelogs.extend(batch)
                last_id = batch[-1].get('id', last_id)  # Get last ID for next iteration
                
                logger.info(f"📥 Fetched {len(batch)} timelogs (total: {len(all_timelogs)}, last_id: {last_id})")
                
                if len(batch) < batch_size:  # Last batch
                    break
                
                if len(all_timelogs) >= limit:  # Reached limit
                    break
            
            logger.info(f"✅ Total fetched: {len(all_timelogs)} Timelog entries from vTiger")
            
            # We don't filter here - the existing code already filters by:
            # 1. duration > 0
            # 2. relatedto not empty
            # These filters happen in the processing loop below
            
            timelogs = all_timelogs[:limit]  # Trim to requested limit
            logger.info(f"📊 Processing {len(timelogs)} timelogs...")
            
            # NOTE: retrieve API is too slow for batch operations (1500+ individual calls)
            # We'll work with query data and accept that relatedto might be empty for some
            
            for timelog in timelogs:
                try:
                    vtiger_id = timelog.get('id', '')
                    if not vtiger_id:
                        stats["skipped"] += 1
                        continue
                    
                    # Get duration in hours (stored as seconds in vTiger)
                    duration_seconds = float(timelog.get('duration', 0) or 0)
                    if duration_seconds <= 0:
                        logger.debug(f"⏭️ Skipping timelog {vtiger_id} - no duration")
                        stats["skipped"] += 1
                        continue
                    
                    hours = Decimal(str(duration_seconds / 3600.0))  # Convert seconds to hours
                    
                    # Get related entity (Case or Account)
                    related_to = timelog.get('relatedto', '')
                    case_id = None
                    customer_id = None
                    
                    if related_to:
                        # Try to find case first, then account
                        case = execute_query_single(
                            "SELECT id, customer_id FROM tmodule_cases WHERE vtiger_id = %s",
                            (related_to,))
                        
                        if case:
                            case_id = case['id']
                            customer_id = case['customer_id']
                        else:
                            # Try to find customer directly
                            customer = execute_query_single(
                                "SELECT id FROM tmodule_customers WHERE vtiger_id = %s",
                                (related_to,))
                            
                            if customer:
                                customer_id = customer['id']
                                case_id = None  # No specific case, just customer
                            else:
                                logger.debug(f"⏭️ Related entity {related_to} not found in our database - will skip")
                                stats["skipped"] += 1
                                continue
                    
                    # If no customer found at all, skip this timelog
                    if not customer_id:
                        logger.warning(f"⚠️ Timelog {vtiger_id} has no valid customer reference - skipping")
                        stats["skipped"] += 1
                        continue
                    
                    # Get user name with caching
                    assigned_user_id = timelog.get('assigned_user_id', '')
                    if assigned_user_id and assigned_user_id not in user_name_cache:
                        user_name_cache[assigned_user_id] = await self._fetch_user_name(assigned_user_id)
                    user_name = user_name_cache.get(assigned_user_id, assigned_user_id)
                    
                    data_hash = self._calculate_hash(timelog)
                    
                    # Check if exists
                    existing = execute_query_single(
                        "SELECT id, sync_hash FROM tmodule_times WHERE vtiger_id = %s",
                        (vtiger_id,))
                    
                    if existing:
                        if existing['sync_hash'] == data_hash:
                            stats["skipped"] += 1
                            continue
                        
                        # Update only if NOT yet approved AND NOT yet billed
                        result = execute_update(
                            """UPDATE tmodule_times 
                               SET description = %s, original_hours = %s, worked_date = %s,
                                   user_name = %s, billable = %s, vtiger_data = %s::jsonb, 
                                   sync_hash = %s, last_synced_at = CURRENT_TIMESTAMP
                               WHERE vtiger_id = %s 
                               AND status = 'pending' 
                               AND billed_via_thehub_id IS NULL""",
                            (
                                timelog.get('name', ''),
                                hours,
                                timelog.get('startedon', None),
                                user_name,
                                timelog.get('isbillable', '0') == '1',
                                json.dumps(timelog),
                                data_hash,
                                vtiger_id
                            )
                        )
                        
                        if result > 0:
                            stats["updated"] += 1
                        else:
                            logger.debug(f"⏭️ Time entry {vtiger_id} already approved or billed")
                            stats["skipped"] += 1
                    else:
                        # Insert new
                        execute_insert(
                            """INSERT INTO tmodule_times 
                               (vtiger_id, case_id, customer_id, description, original_hours,
                                worked_date, user_name, billable, vtiger_data, sync_hash, 
                                status, last_synced_at)
                               VALUES (%s, %s, %s, %s, %s, %s, %s, %s, %s::jsonb, %s, 'pending', CURRENT_TIMESTAMP)""",
                            (
                                vtiger_id,
                                case_id,
                                customer_id,
                                timelog.get('name', ''),
                                hours,
                                timelog.get('startedon', None),
                                user_name,
                                timelog.get('isbillable', '0') == '1',
                                json.dumps(timelog),
                                data_hash
                            )
                        )
                        stats["imported"] += 1
                
                except Exception as e:
                    logger.error(f"❌ Error processing timelog {timelog.get('id', 'unknown')}: {e}")
                    stats["errors"] += 1
            
            logger.info(f"✅ Time entry sync complete: {stats}")
            return stats
        
        except Exception as e:
            logger.error(f"❌ Time entry sync failed: {e}")
            raise
    
    async def full_sync(
        self,
        user_id: Optional[int] = None,
        fetch_comments: bool = False
    ) -> TModuleSyncStats:
        """
        Perform full sync of all data from vTiger.
        
        Order: Customers -> Cases -> Time Entries (dependencies)
        
        Args:
            user_id: User performing the sync
            fetch_comments: Whether to fetch ModComments (slow - adds ~0.4s per case)
        """
        if fetch_comments:
            logger.info("🚀 Starting FULL vTiger sync WITH comments (this will be slow)...")
        else:
            logger.info("🚀 Starting FULL vTiger sync WITHOUT comments (fast mode)...")
        
        start_time = datetime.now()
        
        # Log sync started
        audit.log_sync_started(user_id=user_id)
        
        try:
            # Test connection first
            if not await self.test_connection():
                raise HTTPException(
                    status_code=503,
                    detail="Cannot connect to vTiger - check credentials"
                )
            
            # Sync in order of dependencies
            customer_stats = await self.sync_customers()
            case_stats = await self.sync_cases(fetch_comments=fetch_comments)
            time_stats = await self.sync_time_entries()
            
            end_time = datetime.now()
            duration = (end_time - start_time).total_seconds()
            
            # Build result
            result = TModuleSyncStats(
                customers_imported=customer_stats["imported"],
                customers_updated=customer_stats["updated"],
                customers_skipped=customer_stats["skipped"],
                cases_imported=case_stats["imported"],
                cases_updated=case_stats["updated"],
                cases_skipped=case_stats["skipped"],
                times_imported=time_stats["imported"],
                times_updated=time_stats["updated"],
                times_skipped=time_stats["skipped"],
                errors=(
                    customer_stats["errors"] + 
                    case_stats["errors"] + 
                    time_stats["errors"]
                ),
                duration_seconds=duration,
                started_at=start_time,
                completed_at=end_time
            )
            
            # Log completion
            audit.log_sync_completed(
                stats=result.model_dump(),
                user_id=user_id
            )
            
            logger.info(f"✅ Full sync completed in {duration:.2f}s")
            logger.info(f"📊 Customers: {customer_stats['imported']} new, {customer_stats['updated']} updated")
            logger.info(f"📊 Cases: {case_stats['imported']} new, {case_stats['updated']} updated")
            logger.info(f"📊 Times: {time_stats['imported']} new, {time_stats['updated']} updated")
            
            return result
        
        except Exception as e:
            logger.error(f"❌ Full sync failed: {e}")
            audit.log_sync_failed(error=str(e), user_id=user_id)
            raise


# Singleton instance
vtiger_service = TimeTrackingVTigerService()