Scanner Microservice

[Arthi-chaud]

Scanning for files and processing images requires a lot of CPU (open file, parse metadata, extract illustrations, compute blurhash, etc...)

There should be a microservice dedicated to this feature:

• [ ] We can start scan through an endpoint
  • [ ] By default, scanning considers all libraries
  • [ ] Scan can be library-specific
• [ ] We can start a cleaning task through an endpoint
  • [ ] By default, cleaning considers all libraries
  • [ ] Cleaning can be library-specific
• [ ] We can refresh files' metadata task through an endpoint
  • [ ] By default, refreshing is library specific
  • [ ] Refreshing can be album-specific
  • [ ] Refreshing can be release-specific
  • [ ] Refreshing can be song-specific
  • [ ] Refreshing can be track-specific
• [ ] Tasks cannot run in parallel
• [ ] We should be able to get the status (i.e. the currently running task) through an endpoint
• [x] A Swagger should be present

---

The Server/API should expose routes dedicated to the scanner:

• [ ] Delete a File
  • [ ] It should take a query parameter to trigger (or not) a housekeeping task
• [ ] Add a File
• [ ] Get all files
• [ ] Get a release's illustrations
• [ ] Post an illustration
• [ ] Post a thumbnail
• [ ] These routes should be protected, and only accessible by the scanner

---

• [x] Find the language to implement the microservice
• [x] Investigate whether ffprobe could be a good candidate to replace the music-metadata library

Roadmap

• [x] Setup Project
• [x] Connect to API
• [x] Read settings.json
• [x] Get Metadata from Path
• [x] Get Metadata from Embedded Data
• [ ] Know if a file has already been registered