metadata_core 0.0.2

Core interfaces and models for MetaVision.

metadata_core

A cohesive, platform-agnostic media metadata extraction, folder traversal, and data modeling architecture designed for production-grade Flutter applications.

metadata_core acts as the single source of truth for domain models and scanner interface contracts across the MetaVision suite. It provides essential structural definitions and fully immutable models (freezed) used for cross-platform media extraction.

🏗 System Architecture

The suite is segmented into three decoupled modules following clean architecture principles, allowing native recursive indexing on Desktop/Mobile platforms alongside native DOM API mapping for web environments.

                  ┌───────────────────────────────┐
                  │         metadata_core         │
                  │   (Models, Core Interfaces)   │
                  └───────────────┬───────────────┘
                                  │
         ┌────────────────────────┴────────────────────────┐
         ▼                                                 ▼
┌─────────────────────────────────┐       ┌─────────────────────────────────┐
│       metadata_extractor        │       │          metadata_web           │
│   (Native Recursive Scanner)    │       │    (DOM Bridge & Object URLs)   │
└─────────────────────────────────┘       └─────────────────────────────────┘

🧩 The MetaVision Suite

To fully utilize the suite based on your target platform, you can combine metadata_core with our specialized add-on packages:

• metadata_extractor: Targets native environments (macOS, Windows, Linux, Android, iOS). Uses dart:io recursive directory listings to traverse folders cleanly and extracts deep EXIF/video properties.
• metadata_web: Targets standard browser environments. Uses native DOM JS interop to recursively unravel folders dropped into the canvas and extracts EXIF data efficiently in the browser.

---

📦 Core Domain Models

MediaFile

Represents the structural file entity.

MediaFile({
  required String id,
  required String fileName,
  required String path, // Absolute native path OR Web blob: URL string
  required String relativePath,
  required int size,
  required String mimeType,
  required DateTime createdAt,
  required DateTime modifiedAt,
  String? hash,
})

MetadataResult

Encapsulates all extracted semantic insights mapped to a specific file.

MetadataResult({
  required String fileId,
  ImageMetadata? imageMetadata, // Dimensions, Exposure, F-number, ISO
  VideoMetadata? videoMetadata, // FPS, Bitrate, Duration, Codecs
  LocationData? location,       // Lat/Long EXIF geocoding
  DeviceData? device,           // Camera Hardware Make/Model
})

---

Getting started

Add metadata_core to your pubspec.yaml:

dependencies:
  metadata_core: ^0.0.1

Usage

Import the package:

import 'package:metadata_core/metadata_core.dart';

Extracting EXIF Metadata Manually

While the add-on packages handle extraction automatically during file scanning, you can manually extract EXIF metadata from image bytes using the bundled ExifMetadataExtractor:

import 'dart:io';
import 'package:metadata_core/metadata_core.dart';
import 'package:metadata_core/src/extractors/exif_extractor.dart';

void main() async {
  // Read image bytes
  final file = File('path/to/image.jpg');
  final bytes = await file.readAsBytes();

  // Extract EXIF data
  final result = await ExifMetadataExtractor.extractFromBytes('unique_file_id', bytes);

  // Access the extracted device data
  print('Device Brand: ${result.device?.brand}');
  print('Device Model: ${result.device?.model}');
  
  // Access location data
  if (result.location != null) {
    print('Latitude: ${result.location!.latitude}');
    print('Longitude: ${result.location!.longitude}');
  }
}

Building Custom Extractors

The package provides the IMetadataExtractor interface which you can implement to create custom metadata parsers for specialized file formats (e.g., RAW images, specialized video formats) while maintaining compatibility with the entire MetaVision suite.

Additional information

For more information, visit the repository.