#!/usr/bin/env python3
"""
🔓 LOCAL OUTLOOK HARVESTER - Ingen admin nødvendig!
Læser direkte fra din lokale Outlook installation via COM

Krav: Outlook installeret og logget ind med din TDC konto
"""
import win32com.client
import pythoncom
import json
import hashlib
from pathlib import Path
from datetime import datetime, timedelta
from neo4j import GraphDatabase

NEO4J_URI = "neo4j+s://054eff27.databases.neo4j.io"
NEO4J_USER = "neo4j"
NEO4J_PASSWORD = "Qrt37mkb0xBZ7_ts5tG1J70K2mVDGPMF2L7Njlm7cg8"

# Søgetermer for TDC intern data
SEARCH_KEYWORDS = [
    "strategi",
    "cyber",
    "NIS2",
    "SOC",
    "MDR",
    "cloud",
    "Azure",
    "AI",
    "Copilot",
    "Columbus",
    "ERP",
    "budget",
    "forecast",
    "kunde",
    "kontrakt",
    "rammeaftale",
    "SKI",
    "produkt",
    "CloudKey",
    "arkitektur",
    "roadmap"
]

class LocalOutlookHarvester:
    """Harvester der læser direkte fra lokal Outlook via COM"""
    
    def __init__(self):
        pythoncom.CoInitialize()
        self.outlook = None
        self.namespace = None
        self.emails = []
        self.stats = {
            "folders_scanned": 0,
            "emails_found": 0,
            "emails_matched": 0,
            "attachments": 0
        }
        
        # Neo4j
        self.neo4j = GraphDatabase.driver(NEO4J_URI, auth=(NEO4J_USER, NEO4J_PASSWORD))
        
        # Output
        self.output_dir = Path("data/outlook_local_harvest")
        self.output_dir.mkdir(parents=True, exist_ok=True)
    
    def connect(self) -> bool:
        """Forbind til lokal Outlook"""
        try:
            print("🔌 Forbinder til Outlook...")
            self.outlook = win32com.client.Dispatch("Outlook.Application")
            self.namespace = self.outlook.GetNamespace("MAPI")
            
            # Vis konti
            accounts = self.namespace.Accounts
            print(f"✅ Outlook forbundet!")
            print(f"   Konti fundet: {accounts.Count}")
            for i in range(1, accounts.Count + 1):
                acc = accounts.Item(i)
                print(f"   • {acc.DisplayName} ({acc.SmtpAddress})")
            
            return True
        except Exception as e:
            print(f"❌ Kunne ikke forbinde til Outlook: {e}")
            print("   Sørg for at Outlook er installeret og kører")
            return False
    
    def get_folders(self, parent=None, level=0):
        """Rekursivt hent alle mapper"""
        folders = []
        
        try:
            if parent is None:
                # Start fra root
                for store in self.namespace.Stores:
                    try:
                        root = store.GetRootFolder()
                        folders.append({
                            "name": store.DisplayName,
                            "path": store.DisplayName,
                            "folder": root,
                            "level": 0
                        })
                        folders.extend(self.get_folders(root, 1))
                    except:
                        continue
            else:
                for folder in parent.Folders:
                    try:
                        path = f"{parent.FolderPath}\\{folder.Name}"
                        folders.append({
                            "name": folder.Name,
                            "path": path,
                            "folder": folder,
                            "level": level
                        })
                        if level < 3:  # Max 3 niveauer dybt
                            folders.extend(self.get_folders(folder, level + 1))
                    except:
                        continue
        except:
            pass
        
        return folders
    
    def search_folder(self, folder_info: dict, keywords: list, days_back: int = 365) -> list:
        """Søg i en mappe efter keywords"""
        results = []
        folder = folder_info["folder"]
        
        try:
            items = folder.Items
            items.Sort("[ReceivedTime]", True)  # Nyeste først
            
            # Filtrer på dato
            cutoff = datetime.now() - timedelta(days=days_back)
            
            count = 0
            for item in items:
                try:
                    # Check om det er en mail
                    if item.Class != 43:  # 43 = MailItem
                        continue
                    
                    # Check dato
                    received = item.ReceivedTime
                    if hasattr(received, 'year') and datetime(received.year, received.month, received.day) < cutoff:
                        continue
                    
                    # Søg i subject og body
                    subject = str(item.Subject or "").lower()
                    body = str(item.Body or "")[:2000].lower()
                    sender = str(item.SenderEmailAddress or "")
                    
                    # Match keywords
                    matched_keywords = []
                    for kw in keywords:
                        if kw.lower() in subject or kw.lower() in body:
                            matched_keywords.append(kw)
                    
                    if matched_keywords:
                        # Hent attachments info
                        attachments = []
                        try:
                            for att in item.Attachments:
                                attachments.append({
                                    "name": att.FileName,
                                    "size": att.Size if hasattr(att, 'Size') else 0
                                })
                                self.stats["attachments"] += 1
                        except:
                            pass
                        
                        results.append({
                            "id": item.EntryID,
                            "subject": item.Subject,
                            "sender": sender,
                            "sender_name": str(item.SenderName or ""),
                            "received": received.strftime("%Y-%m-%d %H:%M") if hasattr(received, 'strftime') else str(received),
                            "body_preview": body[:500],
                            "folder": folder_info["path"],
                            "keywords": matched_keywords,
                            "has_attachments": len(attachments) > 0,
                            "attachments": attachments[:5],  # Max 5
                            "importance": item.Importance,
                            "categories": str(item.Categories or "")
                        })
                        self.stats["emails_matched"] += 1
                    
                    count += 1
                    self.stats["emails_found"] += 1
                    
                    # Progress
                    if count % 100 == 0:
                        print(f"      Scannet {count} emails...", end="\r")
                    
                    # Limit per mappe
                    if count >= 1000:
                        break
                        
                except Exception as e:
                    continue
                    
        except Exception as e:
            pass
        
        return results
    
    def save_to_neo4j(self, email: dict):
        """Gem email i Neo4j"""
        content_hash = hashlib.md5(f"{email['subject']}:{email['id']}".encode()).hexdigest()
        
        with self.neo4j.session() as session:
            # Opret email node
            session.run("""
                MERGE (e:OutlookEmail {contentHash: $hash})
                ON CREATE SET
                    e.entryId = $id,
                    e.subject = $subject,
                    e.sender = $sender,
                    e.senderName = $senderName,
                    e.received = $received,
                    e.bodyPreview = $body,
                    e.folder = $folder,
                    e.keywords = $keywords,
                    e.hasAttachments = $hasAtt,
                    e.importance = $importance,
                    e.harvestedAt = datetime()
                ON MATCH SET
                    e.lastSeen = datetime()
                
                MERGE (ds:DataSource {name: 'TDC_Outlook_Local'})
                ON CREATE SET ds.type = 'local_exchange'
                MERGE (e)-[:HARVESTED_FROM]->(ds)
            """,
                hash=content_hash,
                id=email["id"],
                subject=email["subject"][:500],
                sender=email["sender"],
                senderName=email["sender_name"],
                received=email["received"],
                body=email["body_preview"],
                folder=email["folder"],
                keywords=email["keywords"],
                hasAtt=email["has_attachments"],
                importance=email["importance"]
            )
            
            # Opret keyword relationships
            for kw in email["keywords"]:
                session.run("""
                    MERGE (k:SearchKeyword {name: $kw})
                    WITH k
                    MATCH (e:OutlookEmail {contentHash: $hash})
                    MERGE (e)-[:MATCHES_KEYWORD]->(k)
                """, kw=kw, hash=content_hash)
    
    def run(self, days_back: int = 365):
        """Kør fuld harvest"""
        print("\n" + "=" * 60)
        print("🔓 LOCAL OUTLOOK HARVESTER")
        print("   Ingen admin nødvendig!")
        print("=" * 60)
        
        if not self.connect():
            return
        
        # Hent alle mapper
        print("\n📁 Scanner mapper...")
        folders = self.get_folders()
        print(f"   Fundet {len(folders)} mapper")
        
        # Filtrer til relevante mapper
        relevant_folders = []
        skip_names = ["Junk", "Spam", "Deleted", "Drafts", "Outbox", "Slettet", "Kladder", "Udbakke"]
        
        for f in folders:
            if not any(skip in f["name"] for skip in skip_names):
                relevant_folders.append(f)
        
        print(f"   Relevante mapper: {len(relevant_folders)}")
        
        # Søg i hver mappe
        print(f"\n🔍 Søger efter {len(SEARCH_KEYWORDS)} keywords...")
        print(f"   Periode: Sidste {days_back} dage")
        
        all_results = []
        
        for i, folder_info in enumerate(relevant_folders):
            indent = "  " * folder_info["level"]
            print(f"\n{indent}📂 {folder_info['name']}")
            
            results = self.search_folder(folder_info, SEARCH_KEYWORDS, days_back)
            self.stats["folders_scanned"] += 1
            
            if results:
                print(f"{indent}   ✅ {len(results)} matches!")
                for email in results:
                    # Gem i Neo4j
                    self.save_to_neo4j(email)
                    all_results.append(email)
        
        # Gem lokal JSON
        output_file = self.output_dir / f"outlook_harvest_{datetime.now().strftime('%Y%m%d_%H%M%S')}.json"
        with open(output_file, 'w', encoding='utf-8') as f:
            json.dump({
                "timestamp": datetime.now().isoformat(),
                "stats": self.stats,
                "keywords": SEARCH_KEYWORDS,
                "emails": all_results
            }, f, indent=2, ensure_ascii=False)
        
        # Summary
        print("\n" + "=" * 60)
        print("📊 HARVEST COMPLETE")
        print("=" * 60)
        print(f"   📁 Mapper scannet:    {self.stats['folders_scanned']}")
        print(f"   📧 Emails gennemgået: {self.stats['emails_found']}")
        print(f"   ✅ Emails matched:    {self.stats['emails_matched']}")
        print(f"   📎 Attachments:       {self.stats['attachments']}")
        print(f"\n   💾 Saved: {output_file}")
        print("=" * 60)
        
        # Top keywords
        if all_results:
            print("\n🏷️ TOP KEYWORDS:")
            keyword_counts = {}
            for email in all_results:
                for kw in email["keywords"]:
                    keyword_counts[kw] = keyword_counts.get(kw, 0) + 1
            
            for kw, count in sorted(keyword_counts.items(), key=lambda x: -x[1])[:10]:
                print(f"   • {kw}: {count} emails")
        
        self.neo4j.close()
        pythoncom.CoUninitialize()


if __name__ == "__main__":
    import argparse
    parser = argparse.ArgumentParser(description="Local Outlook Harvester")
    parser.add_argument("--days", type=int, default=365, help="Dage tilbage at søge (default: 365)")
    args = parser.parse_args()
    
    harvester = LocalOutlookHarvester()
    harvester.run(args.days)