#!/usr/bin/env python3
"""
Knowledge Graph Review Marker

Marks incomplete research sessions for manual review without auto-fixing.
Creates review queue entries and annotations in the KG database.
"""

import json
import time
from datetime import datetime
from pathlib import Path
from typing import List, Dict


class IncompleteResearchReviewMarker:
    """Mark incomplete research for manual review."""

    REVIEW_QUEUE_DIR = Path('/home/admin/conductor/review')
    REVIEW_TAG = 'needs_human_review'

    def __init__(self):
        """Initialize the review marker."""
        self.REVIEW_QUEUE_DIR.mkdir(parents=True, exist_ok=True)

    def mark_for_review(self, entity_id: str, entity_name: str, reason: str,
                       severity: str = 'medium', pattern: str = 'unknown') -> Dict:
        """
        Mark a single KG entity for review.

        Args:
            entity_id: KG entity ID
            entity_name: Human-readable entity name
            reason: Why this needs review (pattern type)
            severity: 'high' | 'medium' | 'low'
            pattern: Pattern type that triggered marking

        Returns:
            Dict with review_id, file_path, status
        """
        review_id = f"{entity_id}_review_{int(time.time())}"
        review_file = self.REVIEW_QUEUE_DIR / f"{review_id}.json"

        review_data = {
            'review_id': review_id,
            'entity_id': entity_id,
            'entity_name': entity_name,
            'pattern': pattern,
            'severity': severity,
            'reason': reason,
            'marked_at': datetime.now().isoformat(),
            'marked_by': 'kg_health_checker',
            'status': 'pending_review',
            'action_required': self._get_action_for_pattern(pattern),
            'notes': ''
        }

        review_file.write_text(json.dumps(review_data, indent=2))

        return {
            'review_id': review_id,
            'entity_id': entity_id,
            'file_path': str(review_file),
            'status': 'marked'
        }

    def mark_findings_batch(self, findings: List[Dict]) -> Dict:
        """
        Mark multiple findings for review.

        Args:
            findings: List of finding dicts from KGPatternDetector

        Returns:
            Dict with:
                - 'marked_count': Number of sessions marked
                - 'review_files': List of created files
                - 'summary': Breakdown by severity
        """
        marked = []
        review_files = []
        summary = {'high': 0, 'medium': 0, 'low': 0}

        for finding in findings:
            # Only mark KG database findings (not files)
            if finding['source'] != 'kg_database':
                continue

            result = self.mark_for_review(
                entity_id=finding['id'],
                entity_name=finding.get('name', 'unknown'),
                reason=f"Pattern: {finding['pattern']}",
                severity=finding.get('severity', 'medium'),
                pattern=finding['pattern']
            )

            marked.append(result)
            review_files.append(result['file_path'])
            summary[finding.get('severity', 'medium')] += 1

        return {
            'marked_count': len(marked),
            'review_files': review_files,
            'summary': summary,
            'review_queue_path': str(self.REVIEW_QUEUE_DIR),
            'timestamp': time.time()
        }

    def create_review_queue(self, findings: List[Dict]) -> Dict:
        """
        Create review queue from findings (alias for mark_findings_batch).

        Args:
            findings: List of findings to create review queue for

        Returns:
            Review queue creation result
        """
        return self.mark_findings_batch(findings)

    def get_pending_reviews(self) -> List[Dict]:
        """Get list of all pending review items."""
        reviews = []

        if not self.REVIEW_QUEUE_DIR.exists():
            return reviews

        for review_file in self.REVIEW_QUEUE_DIR.glob('*_review_*.json'):
            try:
                data = json.loads(review_file.read_text())
                if data.get('status') == 'pending_review':
                    reviews.append(data)
            except Exception:
                pass

        return sorted(reviews, key=lambda x: x.get('severity') == 'high', reverse=True)

    def mark_review_complete(self, review_id: str, resolution: str = '') -> Dict:
        """
        Mark a review as complete.

        Args:
            review_id: The review ID to complete
            resolution: What was done to resolve the issue

        Returns:
            Updated review data
        """
        review_file = self.REVIEW_QUEUE_DIR / f"{review_id}.json"

        if not review_file.exists():
            return {'status': 'error', 'message': f'Review {review_id} not found'}

        data = json.loads(review_file.read_text())
        data['status'] = 'completed'
        data['completed_at'] = datetime.now().isoformat()
        data['resolution'] = resolution

        review_file.write_text(json.dumps(data, indent=2))

        return {
            'status': 'completed',
            'review_id': review_id,
            'resolution': resolution
        }

    def get_review_stats(self) -> Dict:
        """Get statistics about review queue."""
        reviews = []

        if not self.REVIEW_QUEUE_DIR.exists():
            return {
                'pending': 0,
                'completed': 0,
                'by_severity': {},
                'by_pattern': {}
            }

        pending = 0
        completed = 0
        by_severity = {}
        by_pattern = {}

        for review_file in self.REVIEW_QUEUE_DIR.glob('*_review_*.json'):
            try:
                data = json.loads(review_file.read_text())
                severity = data.get('severity', 'unknown')
                pattern = data.get('pattern', 'unknown')

                if data.get('status') == 'pending_review':
                    pending += 1
                elif data.get('status') == 'completed':
                    completed += 1

                by_severity[severity] = by_severity.get(severity, 0) + 1
                by_pattern[pattern] = by_pattern.get(pattern, 0) + 1
            except Exception:
                pass

        return {
            'pending': pending,
            'completed': completed,
            'total': pending + completed,
            'by_severity': by_severity,
            'by_pattern': by_pattern,
            'review_queue_path': str(self.REVIEW_QUEUE_DIR)
        }

    def _get_action_for_pattern(self, pattern: str) -> str:
        """Get recommended action for a pattern."""
        actions = {
            'unresolved_question': 'Resume research session with user feedback; complete final analysis',
            'incomplete_duration': 'Investigate why session ended early; may need deeper research',
            'claude_no_conclusion': 'Add missing conclusion/summary section to analysis',
            'unknown': 'Review and complete research session'
        }
        return actions.get(pattern, actions['unknown'])


if __name__ == '__main__':
    marker = IncompleteResearchReviewMarker()

    # Test: create sample review
    result = marker.mark_for_review(
        entity_id='research_12345',
        entity_name='Structured Repositories & Trusted Data Sources for AI Agents',
        reason='Session ended with unresolved user question',
        severity='high',
        pattern='unresolved_question'
    )

    print(f"Created review: {result['review_id']}")
    print(f"Review file: {result['file_path']}")

    # Show stats
    stats = marker.get_review_stats()
    print(f"\nReview queue stats:")
    print(f"  Pending: {stats['pending']}")
    print(f"  Completed: {stats['completed']}")
    print(f"  By severity: {stats['by_severity']}")