Media Service

The media workflow project produces a package that gets loaded into the Orchestration Application to provide the OTT media stations, tiles on the homepage, and, potentially, some extensions. See the Mosaic developer documentation for a full description of how the micro-frontends are registered.

The OTT media workflow package implements workflows to manage movies, TV shows, seasons, episodes, and collections. Each of them registers a tile on the homepage that starts the workflow to view and manage the corresponding entities. In addition to those tiles, the template also adds tiles for the ingest, snapshot registry, and settings for the movie and TV show genres.

Domains are the central area to manage the business requirements for the OTT media management solution. It contains all the entity types and their specific logic about how they are made available in the GraphQL API, how they can be ingested, published, or any other related logic. If you want to adjust something for one specific entity, this is the place where to do it. Similarly, if you want to add a completely new entity type, this would be the place where to add it. Code that includes multiple entity types at once resides in the domains/common folder.

Common examples from the OTT Media Service solution to define one entity type:

Finally, the domains folder contains the code to collect all the GraphQL plugins and message handlers, so they can be registered in the startup code. And it contains the logic that maps the operation groups to permissions (e.g. to the admin permission, or the movies view or movies edit permission).

The backend uses PostGraphile to generate its GraphQL API. PostGraphile introspects the PostgreSQL database (tables, columns, relationships, etc.) to create a fast and flexible GraphQL API. You can find more information about PostGraphile and how to best use it in your solutions in our Mosaic developer documentation.

In the Mosaic Media Template, the graphql folder contains the code to set up the PostGraphile middleware. It allows you to fine-tune all the aspects of how PostGraphile generates the API and to register your plugins.

The graphql/plugins folder contains mostly plugin factories that provide common logic to be used for different domain entities.

For asynchronous message processing, the Mosaic Media Template uses the Mosaic message bus (with RabbitMQ). The Mosaic message bus uses the notions of events (something happened) and commands (some service should do something).

The main objective of the messaging folder is the setup and registration of the message handlers and middleware to consume the messages and the setup of publishers to send out the messages. In addition, there is a media message handler that protects message handlers by checking the message for specific permissions.

Code in this folder and in the models folder uses the media-messages library that contains the schema definitions for all events and commands.

The OTT Media Service offers ingestion capabilities to import entity metadata into the system. The process takes an ingestion file and will either create new entities or update the existing ones based on a provided unique identifier.

In addition to handling metadata, the system will also orchestrate the ingest of associated assets. For example, a movie is ingested that includes image references and references a main and a teaser video.

The following state diagram shows the ingest process starting from the GraphQL API which receives an ingest document until all items in the ingest were processed.

You can find more information in the How-To Guide Start an Ingest and in the Media Service - Ingest documentation.

The ingest folder contains a GraphQL plugins folder with plugins to start the ingest process and to expose the ingest process details as GraphQL query operations. The handlers folder contains all the general logic for the metadata ingest via message handlers and the orchestration of image and video imports into the manage Image and Video Service. The specific details for the entity metadata mappings are not in this folder but in the domain area.

Publishing is the process of aggregating all entity metadata and exporting it in a self-contained format. Most importantly, it will make the data of the OTT media service (with the related data from other micro-services, such as the image and video service) available in the catalog service.

Publishing is done in two steps:

The publishing process is available for single entities or as bulk operations for many entities at once. All the core logic to create the publishing GraphQL API endpoints and the background message handlers is located in the publishing folder.

To learn more about the publishing process, navigate to the Media Service - Publishing documentation.

All the setup and bootstrapping happen in the src/index.ts file. In there, the express web-server is started. Moreover, there the configuration, database pools, GraphQL, messaging, logging, monitoring, authentication, etc. are initialized as well.

The package.json file contains the most important scripts to start and test the application as well as other utility scripts. The entire setup and startup logic is described in the README.md file.