Fancien
commented
1 year ago Possible models:
Session:
-
id: Unique session identifier
-
subject: The url towards the resource to evaluate (be careful of path if it is a local file sent by user)
-
status: The running status of the session (queued, preprocessing, running, postprocessing, finished, error).
-
list of tasks: The list of task objects associated with the session
-
Essential score: The FAIR evaluation score when only essential tasks are evaluated
-
Score: The FAIR evaluation score when all tasks are evaluated (including non essential ones)
-
Not applicable number: Number of not_applicable tasks
Task:
- id: Unique task identifier
- name: The task name (e.g. f1_metadata_unique_identifier). This would serve as a key for the list of possible tasks
- status: The running status of the specific task (queued, started, success, failed, warnings, error)
- priority: Defines how important the task is (essential, important, useful)
- session id: The unique session identifier for the session this task belongs to
- score: The FAIR evaluation score for this task (0, 0.5, 1 or None if task is not finished)
- children: A list of task that cannot start until this task has the success status (or is finished?)
TaskDescription: This describes the different tasks that are evaluated.
- name: The name of a task (e.g. f1_metadata_unique_identifier)
- priority: Defines how important the task is (essential, important, useful)
- short description: For example "Is metadata of model identified by a persistent identifier"
- long description: A text explaining the task in more details, what is evaluated, and how to get the evaluation to pass
The different classes the user interacts with needs to be defined as pydantic Models. This includes:
These models needs to be defined in such a way that the documentation created by FastAPI is correct and includes all relevant attributes and parameters. For example, these models cannot define @property methods, as these are not picked up by FastAPI.