<aside> 🧑‍💻 This documentation is part of the Technical Kick-start Collection. If you’re a developer interested in contributing to Thea, or just looking to understand how Thea works on a deeper level, then I’d highly recommend reading through this collection.

</aside>

Overview

The ingest service is a fundamental part of Thea, as it handles the importing and hydration of media. The Download & Import documentation covers this on a high-level, however this service is the implementation behind only automated ingestion of media. See Download Service for manual downloads**.**

On a high level, this service:

• Detects new content to auto-ingest via a directory watcher on the import directory.
• Creates internal state to keep track of this ingest.
• Gathers as much metadata as possible about the content by looking at the title, file metadata, and the output of ffprobe. See Metadata Scraping.
• Hydrates the internal state using TMDB Searching.
• Performs Database Finalization.
• Facilitates any relevant transcoding workflows by emitting success events back to Thea core via the event bus.

Troubles

Due to the complexity of this service, a number of troubles can occur during the ingestion of media:

• TODO

Technical Implementation

The ingest service is responsible for a lot of heavy lifting in Thea, and it performs a lot of file and network I/O to do it.

External Interface

Ingest requests are initiated by two main mechanisms.

Firstly, there’s content that is simply being ‘imported’ - this is handled by watching a specific directory for new files and waiting for the file size to stop changing for a configurable amount of time. This path is internal as the ingest service watches the import directory.