solventum-image-metadata/backend/app/processors/template_manager.py

"""Metadata template manager with variable substitution."""

import json
from pathlib import Path
from typing import Dict, List, Optional
from datetime import datetime
from .utils import get_logger

logger = get_logger(__name__)


class TemplateManager:
    """Manage metadata templates with variable substitution."""

    # Available variables for substitution
    AVAILABLE_VARIABLES = {
        '{filename}': 'Original filename without extension',
        '{date}': 'Current date (YYYY-MM-DD)',
        '{datetime}': 'Current date and time',
        '{user}': 'Current username',
        '{year}': 'Current year',
        '{month}': 'Current month',
        '{day}': 'Current day'
    }

    def __init__(self, templates_path: Optional[str] = None):
        """
        Initialize template manager.

        Args:
            templates_path: Path to JSON file for storing templates
        """
        self.templates_path = templates_path or 'metadata_templates.json'

    def create_template(
        self,
        name: str,
        title_template: str,
        subject_template: str,
        keywords_template: str,
        description: str = ''
    ) -> Dict:
        """
        Create a new metadata template.

        Args:
            name: Template name
            title_template: Title template with variables (e.g., "{filename} - Product Guide")
            subject_template: Subject template with variables
            keywords_template: Keywords template with variables
            description: Optional description of template usage

        Returns:
            Template dictionary
        """
        template = {
            'name': name,
            'description': description,
            'title': title_template,
            'subject': subject_template,
            'keywords': keywords_template,
            'created_at': self._get_timestamp(),
            'updated_at': self._get_timestamp()
        }

        # Validate template
        validation = self.validate_template(template)
        if validation['invalid']:
            logger.warning(f"Template '{name}' has invalid variables: {validation['invalid']}")

        return template

    def save_template(self, template: Dict) -> bool:
        """
        Save template to storage.

        Args:
            template: Template dictionary

        Returns:
            True if successful
        """
        try:
            templates = self._load_templates()
            template['updated_at'] = self._get_timestamp()
            templates[template['name']] = template

            with open(self.templates_path, 'w', encoding='utf-8') as f:
                json.dump(templates, f, indent=2, ensure_ascii=False)

            logger.info(f"Saved template: {template['name']}")
            return True

        except Exception as e:
            logger.error(f"Failed to save template '{template['name']}': {e}")
            return False

    def load_template(self, name: str) -> Optional[Dict]:
        """
        Load template by name.

        Args:
            name: Template name

        Returns:
            Template dictionary or None if not found
        """
        templates = self._load_templates()
        template = templates.get(name)

        if template:
            logger.info(f"Loaded template: {name}")
        else:
            logger.warning(f"Template not found: {name}")

        return template

    def list_templates(self) -> List[Dict]:
        """
        List all available templates.

        Returns:
            List of template summaries
        """
        templates = self._load_templates()

        return [
            {
                'name': name,
                'description': data.get('description', ''),
                'created_at': data.get('created_at', ''),
                'updated_at': data.get('updated_at', ''),
                'variables_used': self._extract_variables(data)
            }
            for name, data in templates.items()
        ]

    def delete_template(self, name: str) -> bool:
        """
        Delete a template.

        Args:
            name: Template name

        Returns:
            True if deleted, False if not found
        """
        templates = self._load_templates()

        if name in templates:
            del templates[name]

            try:
                with open(self.templates_path, 'w', encoding='utf-8') as f:
                    json.dump(templates, f, indent=2, ensure_ascii=False)

                logger.info(f"Deleted template: {name}")
                return True
            except Exception as e:
                logger.error(f"Failed to delete template '{name}': {e}")
                return False

        logger.warning(f"Template not found: {name}")
        return False

    def apply_template(
        self,
        template: Dict,
        filename: str,
        user: str = 'Unknown',
        custom_vars: Optional[Dict[str, str]] = None
    ) -> Dict[str, str]:
        """
        Apply template to generate metadata for a file.

        Args:
            template: Template dictionary
            filename: Filename to process
            user: Username for {user} variable
            custom_vars: Additional custom variables (e.g., {'product_line': 'Dental'})

        Returns:
            Dictionary with title, subject, keywords
        """
        # Build variable substitution map
        variables = self._build_variable_map(filename, user, custom_vars)

        # Apply substitutions
        metadata = {
            'title': self._substitute_variables(template.get('title', ''), variables),
            'subject': self._substitute_variables(template.get('subject', ''), variables),
            'keywords': self._substitute_variables(template.get('keywords', ''), variables)
        }

        logger.info(f"Applied template '{template['name']}' to {filename}")
        return metadata

    def validate_template(self, template: Dict) -> Dict[str, List[str]]:
        """
        Validate template for correct variable usage.

        Args:
            template: Template dictionary

        Returns:
            Dictionary with 'valid' and 'invalid' variable lists
        """
        result = {
            'valid': [],
            'invalid': []
        }

        # Extract all variables from template
        all_text = (
            template.get('title', '') +
            template.get('subject', '') +
            template.get('keywords', '')
        )

        # Find all {variable} patterns
        import re
        variables = re.findall(r'\{[^}]+\}', all_text)

        for var in variables:
            if var in self.AVAILABLE_VARIABLES:
                if var not in result['valid']:
                    result['valid'].append(var)
            else:
                if var not in result['invalid']:
                    result['invalid'].append(var)

        return result

    def _load_templates(self) -> Dict:
        """Load all templates from file."""
        if Path(self.templates_path).exists():
            try:
                with open(self.templates_path, 'r', encoding='utf-8') as f:
                    return json.load(f)
            except Exception as e:
                logger.error(f"Failed to load templates: {e}")
                return {}
        return {}

    def _get_timestamp(self) -> str:
        """Get current timestamp as ISO format string."""
        return datetime.now().isoformat()

    def _build_variable_map(
        self,
        filename: str,
        user: str,
        custom_vars: Optional[Dict[str, str]]
    ) -> Dict[str, str]:
        """
        Build variable substitution map.

        Args:
            filename: Filename (with or without extension)
            user: Username
            custom_vars: Custom variables

        Returns:
            Dictionary mapping variable names to values
        """
        # Get filename without extension
        filename_stem = Path(filename).stem

        # Current date/time
        now = datetime.now()

        variables = {
            '{filename}': filename_stem,
            '{date}': now.strftime('%Y-%m-%d'),
            '{datetime}': now.strftime('%Y-%m-%d %H:%M:%S'),
            '{user}': user,
            '{year}': str(now.year),
            '{month}': now.strftime('%m'),
            '{day}': now.strftime('%d')
        }

        # Add custom variables
        if custom_vars:
            for key, value in custom_vars.items():
                # Ensure custom variables are wrapped in {}
                var_key = f'{{{key}}}' if not key.startswith('{') else key
                variables[var_key] = value

        return variables

    def _substitute_variables(self, template_text: str, variables: Dict[str, str]) -> str:
        """
        Substitute variables in template text.

        Args:
            template_text: Text with {variable} placeholders
            variables: Variable substitution map

        Returns:
            Text with variables replaced
        """
        result = template_text

        for var, value in variables.items():
            result = result.replace(var, value)

        return result

    def _extract_variables(self, template: Dict) -> List[str]:
        """
        Extract all variables used in a template.

        Args:
            template: Template dictionary

        Returns:
            List of variable names (e.g., ['{filename}', '{date}'])
        """
        import re
        all_text = (
            template.get('title', '') +
            template.get('subject', '') +
            template.get('keywords', '')
        )

        variables = re.findall(r'\{[^}]+\}', all_text)
        return list(set(variables))

    def get_available_variables(self) -> Dict[str, str]:
        """
        Get list of available variables with descriptions.

        Returns:
            Dictionary mapping variable names to descriptions
        """
        return self.AVAILABLE_VARIABLES.copy()

    def preview_template(
        self,
        template: Dict,
        sample_filename: str = 'example.pdf',
        user: str = 'User',
        custom_vars: Optional[Dict[str, str]] = None
    ) -> Dict[str, str]:
        """
        Preview template output with sample data.

        Args:
            template: Template dictionary
            sample_filename: Sample filename for preview
            user: Sample username
            custom_vars: Sample custom variables

        Returns:
            Preview metadata
        """
        return self.apply_template(template, sample_filename, user, custom_vars)

    def export_template(self, name: str, export_path: str) -> bool:
        """
        Export single template to JSON file.

        Args:
            name: Template name
            export_path: Path to save template

        Returns:
            True if successful
        """
        template = self.load_template(name)
        if not template:
            return False

        try:
            with open(export_path, 'w', encoding='utf-8') as f:
                json.dump(template, f, indent=2, ensure_ascii=False)

            logger.info(f"Exported template '{name}' to {export_path}")
            return True

        except Exception as e:
            logger.error(f"Failed to export template '{name}': {e}")
            return False

    def import_template(self, import_path: str) -> Optional[Dict]:
        """
        Import template from JSON file.

        Args:
            import_path: Path to template JSON file

        Returns:
            Imported template dictionary or None
        """
        try:
            with open(import_path, 'r', encoding='utf-8') as f:
                template = json.load(f)

            # Validate required fields
            required_fields = ['name', 'title', 'subject', 'keywords']
            if not all(field in template for field in required_fields):
                logger.error(f"Invalid template file: missing required fields")
                return None

            logger.info(f"Imported template from {import_path}")
            return template

        except Exception as e:
            logger.error(f"Failed to import template: {e}")
            return None