Added simple filter queries for collections in the server API. Such
queries should cover simpler cases, while more complex filters should
use the JSON syntax.
- Added simple filters for collection endpoints in API (e.g.
`/tasks?project_id=42&user=myusername`)
- Removed duplicating collection endpoints in API in favor of their full
versions:
- `/projects/{id}/tasks` -> `/tasks?project_id={id}`
- `/tasks/{id}/jobs` -> `/jobs?task_id={id}`
- `/jobs/{id}/issues` -> `/issues?job_id={id}`
- `/issue/{id}/comments` -> `/comments?issue_id={id}`
- Corresponding owning objects now return a link to the collection:
- `/projects/{id}`: `.tasks`
- `/tasks/{id}`: `.jobs`
- `/jobs/{id}`: `.issues`
- `/issue/{id}`: `.comments`
- Fixed errors in the generated server API schema:
- Input and output types in all the basic model methods (`Create`,
`Update`, `Retrieve`, `Delete`)
- Removed invalid separation for the project list operation response
type
- File structure for projects in the dataset cache is updated to store
the new project representation. Old caches will be invalidated.
- Added tests
This will let users to run their PyTorch code without network access,
provided that they have already cached the data.
### How has this been tested?
<!-- Please describe in detail how you tested your changes.
Include details of your testing environment, and the tests you ran to
see how your change affects other areas of the code, etc. -->
Unit tests.
Maxim Zhiltsov
Roman Donchenko