Import Flow and Metadata Mapping

This document describes the comprehensive flow of data import in OpenRegister, including the metadata mapping functionality that allows schema-based field mapping to standard metadata fields.

Overview

OpenRegister supports importing data from CSV and Excel files with automatic metadata extraction and mapping based on schema configuration. The system handles both single-schema and mixed-schema bulk operations with optimized performance paths.

Import Flow Diagram

The following diagram illustrates the complete import processing flow:

Import Processing Paths

1. Single-Schema Import (Optimized Path)

Used when all objects in the import belong to the same schema (most CSV imports).

Entry Point: SaveObjects::prepareSingleSchemaObjectsOptimized()

Features:

• Pre-calculated metadata values
• Single schema analysis cache
• Optimized metadata hydration
• Performance: ~8400 objects/second

2. Mixed-Schema Import (Standard Path)

Used when objects belong to different schemas (multi-sheet Excel imports).

Entry Point: SaveObjects::prepareObjectsForBulkSave()

Features:

• Per-object schema validation
• Dynamic schema cache building
• Individual metadata hydration
• Performance: Variable based on schema diversity

Metadata Mapping Configuration

Metadata mapping is configured in the schema's configuration field:

{
  "objectNameField": "naam",
  "objectSummaryField": "beschrijvingKort", 
  "objectDescriptionField": "beschrijvingLang",
  "objectImageField": "logo",
  "objectSlugField": "slug",
  "objectPublishedField": "publicatieDatum",
  "objectDepublishedField": "einddatum"
}

Supported Metadata Fields

Configuration Key	Target Field	Description
objectNameField	name	Primary display name
objectDescriptionField	description	Long description text
objectSummaryField	summary	Short summary text
objectImageField	image	Image URL or path
objectSlugField	slug	URL-friendly identifier
objectPublishedField	published	Publication date
objectDepublishedField	depublished	Depublication date

Twig-like Concatenation

The system supports Twig-like template syntax for combining multiple fields:

{
  "objectNameField": "{{ voornaam }} {{ tussenvoegsel }} {{ achternaam }}"
}

Examples:

• "{{ contact.voornaam }} {{ contact.achternaam }}" - Nested field access
• "{{ titel }}: {{ naam }}" - Mixed template and literal text
• "naam" - Simple field mapping

Implementation Details

Core Components

SaveObject::hydrateObjectMetadata()

Location: lib/Service/ObjectHandlers/SaveObject.php:461

Handles metadata extraction from object data based on schema configuration.

/**
 * Hydrates object metadata fields based on schema configuration.
 * 
 * @param ObjectEntity $entity The entity to hydrate
 * @param Schema       $schema The schema containing the configuration
 */
public function hydrateObjectMetadata(ObjectEntity $entity, Schema $schema): void

SaveObjects Bulk Processing

Location: lib/Service/ObjectHandlers/SaveObjects.php

Two main processing paths:

• prepareSingleSchemaObjectsOptimized() - Line 1112
• prepareObjectsForBulkSave() - Line 957

Both paths now include metadata hydration using temporary ObjectEntity instances.

Published Date Handling

DEPRECATED / REMOVED (2026-06-15). Object-level publish metadata was removed as a breaking change by the deprecate-published-metadata OpenSpec change: the published/depublished columns were dropped by migration Version1Date20260313130000, ImportService::addPublishedDateToObjects() was removed, and passing $publish=true to import now only logs a deprecation warning (locked by tests/Unit/Service/ImportServicePublishDeprecationTest.php). Publication/anonymous visibility is now governed by RBAC (a public-group read scope with a date-window match such as publicatiedatum <= $now), owned by the downstream consuming apps — not by import-time object metadata. The historical behaviour below no longer applies.

The legacy behaviour, when publish=1 was set during import (now a no-op + deprecation warning):

1. ImportService added a published date to all objects in @self.published
2. SaveObjects preserved published dates during processing
3. Database stored published dates in the published column

Performance Characteristics

Import Performance

• Single-schema bulk: ~8400 objects/second
• Mixed-schema bulk: ~2000-4000 objects/second (varies by schema diversity)
• Memory usage: Optimized for imports up to 200,000 objects

Metadata Processing Overhead

• Twig-like templates: ~0.1ms per object
• Simple field mapping: ~0.01ms per object
• Schema caching: Eliminates redundant schema lookups

API Usage

Import Endpoints

# CSV Import with metadata mapping
curl -X POST \
  -F 'file=@data.csv' \
  -F 'publish=1' \
  -F 'validation=1' \
  'http://localhost/index.php/apps/openregister/api/registers/{registerId}/import?schemaId={schemaId}'

# Excel Import with multiple schemas
curl -X POST \
  -F 'file=@data.xlsx' \
  -F 'publish=1' \
  'http://localhost/index.php/apps/openregister/api/registers/{registerId}/import'

Parameters

Parameter	Description	Default
file	CSV or Excel file	Required
schemaId	Target schema for CSV imports	Required for CSV
publish	Set published date on all objects	false
validation	Enable schema validation	false
includeObjects	Return full object data in response	false

Troubleshooting

Common Issues

1. Metadata not mapping: Check schema configuration format
2. Published dates null: Verify publish=1 parameter
3. Performance issues: Use single-schema imports when possible

Debug Information

Enable debug logging to trace metadata extraction:

$this->logger->debug('Metadata extraction', [
    'schemaConfig' => $schema->getConfiguration(),
    'extractedName' => $entity->getName(),
    'objectData' => $objectData
]);

Future Improvements

• Complete published date persistence fix
• Add validation for Twig-like syntax
• Optimize mixed-schema import performance
• Add support for conditional metadata mapping

Code References

See the following files for implementation details:

• Import flow: lib/Service/ImportService.php
• Bulk processing: lib/Service/ObjectHandlers/SaveObjects.php
• Metadata hydration: lib/Service/ObjectHandlers/SaveObject.php:461
• Schema configuration: website/docs/core/schema.md